Novel Functions of Protein Arginine Methyltransferase 1 in Thyroid Hormone Receptor-Mediated Transcription and in the Regulation of Metamorphic Rate in Xenopus laevis

Abstract

Protein arginine methyltransferase 1 (PRMT1) acts as a transcription coactivator for nuclear receptors through histone H4 R3 methylation. The in vivo function of PRMT1 is largely unknown. Here we investigated the role of PRMT1 in thyroid hormone (T3) receptor (TR)-mediated transcription in vivo during vertebrate development. By using intestinal remodeling during T3-dependent Xenopus laevis metamorphosis for in vivo molecular analysis, we first showed that PRMT1 expression was upregulated during metamorphosis when both TR and T3 were present. We then demonstrated a role for PRMT1 in TR-mediated transcription by showing that PRMT1 enhanced transcriptional activation by liganded TR in the frog oocyte transcription system and was recruited to the T3 response element (TRE) of the target promoter in the oocyte, as well as to endogenous TREs during frog metamorphosis. Surprisingly, we found that PRMT1 was only transiently recruited to the TREs in the target during metamorphosis and observed no PRMT1 recruitment to TREs at the climax of intestinal remodeling when both PRMT1 and T3 were at peak levels. Mechanistically, we showed that overexpression of PRMT1 enhanced TR binding to TREs both in the frog oocyte model system and during metamorphosis. More importantly, transgenic overexpression of PRMT1 enhanced gene activation in vivo and accelerated both natural and T3-induced metamorphosis. These results thus indicate that PRMT1 functions transiently as a coactivator in TR-mediated transcription by enhancing TR-TRE binding and further suggest that PRMT1 has tissue-specific roles in regulating the rate of metamorphosis.

Thyroid hormone (T3) is essential for normal development in vertebrates, including humans (4, 29, 40, 68, 74, 87). High levels of T3 present during late embryonic and neonatal development, during the last few months of fetal development, and after birth are critical for brain development and the growth and maturation of other organs in humans, and T3 deficiency causes a number of developmental abnormalities, including cretinism, which is characterized by extremely short stature and severe mental retardation. Unfortunately, the difficulty in manipulating uterus-enclosed mammalian embryos has severely limited molecular and functional studies of T3 action during the critical late embryonic developmental period.

An anuran amphibian undergoes metamorphosis during late development, a period developmentally equivalent to the late embryonic and neonatal periods in humans (4, 74). During metamorphosis, an anuran changes from an aquatic herbivorous larval tadpole to a terrestrial carnivorous frog. This process involves three major types of transformations (23, 68). The tadpole-specific organs such as the tail are completely resorbed while the frog-specific ones such as the limb develop de novo. The majority of the organs are present in both tadpoles and frogs but undergo drastic remodeling. Interestingly, all of these changes are controlled by T3 (4, 68, 74). This metamorphic effect of T3 is mediated through gene regulation by the T3 receptor (TR) (13, 15, 17, 51, 67). TRs form heterodimers with 9-cis retinoic acid receptors (RXRs), and these dimers bind to the T3 response element (TRE) in or around the promoters of target genes (40, 47, 78, 87). In the absence of T3, TR/RXR functions as a repressor, while in the presence of T3, TR/RXR functions as an activator. In both transcriptional activation and repression, different cofactor complexes are recruited by TR to TREs to affect transcription (18, 30, 35-37, 60, 61, 87, 88). Previously, we and others have shown that the p160 family coactivator SRC3 (steroid receptor coactivator 3) and the histone acetyltransferase p300 are recruited to the TREs of endogenous target genes during frog metamorphosis and that the SRC/p300 coactivator complexes are required for gene regulation by TR and metamorphosis (28, 56-58).

The p160 coactivator proteins (SRC1 to -3) and p300 are known to form complexes with protein arginine methyltransferase 1 (PRMT1) and coactivator-associated arginine methyltransferase 1 (CARM1 or PRMT4) (5, 19, 38, 41, 45, 75, 81). PRMT1 is an arginine methyltransferase that belongs to the expanding PRMT family broadly classified as type I and II enzymes in vertebrates. Type I enzymes (PRMT1, -3, -4, -6, and -8) catalyze the formation of N^G-monomethylarginine and asymmetric N^G,N^G-dimethylarginine, whereas type II enzymes (PRMT5, -7, and -9) form symmetric dimethylarginine via a monomethylarginine as the intermediate (8, 39, 54). In vitro and in cell cultures, PRMT1 can function as a coactivator in transcriptional regulation by nuclear receptors, including TRs, through histone H4 R3 methylation (38, 73, 81). In addition, PRMT1 can also methylate other proteins, with more than 20 substrates identified so far (8, 9, 42, 53). PRMT1 has been implicated in many biological events, including RNA processing (12, 21, 46), DNA repair (10), signal transduction (1, 50), and transcription (2, 63, 89). Furthermore, PRMT1-null embryonic stem (ES) cells retain only 15% of their total methyltransferase activity and 46% of their asymmetric methylation, suggesting that PRMT1 is the major PRMT in ES cells (59). Most of the research on PRMT1 has so far been performed in vitro or with cell cultures, leaving the in vivo function of PRMT1 largely unknown. In vivo studies of PRMT1 function are further hindered by the fact that PRMT1 knockout or knockdown is embryonically lethal in mice (59) and Xenopus (6).

Here we investigated the function of Xenopus laevis PRMT1 in TR-mediated transcription during metamorphosis. For this purpose, we have chosen the animal intestine as a model organ for detailed molecular analysis. The intestine is remodeled from a larval to an adult form functionally and morphologically during metamorphosis (70). This process involves both degeneration of the larval epithelium, the predominant tissue of the tadpole intestine, and de novo development of the adult epithelium. The other major tissues of the intestine, the connective tissue and muscles, also undergo extensive growth and differentiation. Like other metamorphic processes, all of these changes are completely controlled by T3. The molecular transformations during intestinal remodeling in X. laevis have been extensively characterized and documented (15, 32, 71), making the Xenopus intestine a powerful model to study the molecular mechanism of TR-mediated transcription in vivo. Thus, we analyzed the expression of PRMT1 mRNA and protein by quantitative real-time reverse transcription-PCR (qRT-PCR) and Western blotting, respectively. The results showed that the expression of PRMT1 was upregulated during intestinal metamorphosis. Next we investigated PRMT1 function in reconstituted Xenopus oocytes, where we can study transcriptional regulation by TR in the context of chromatin (84, 85). The results revealed that PRMT1 activated TR-mediated transcription by enhancing the binding of TR to TREs of target genes and histone modifications. The involvement of PRMT1 in TR-mediated transcription in vivo was then demonstrated by chromatin immunoprecipitation (ChIP) assays, which demonstrated PRMT1 recruitment to endogenous target genes during metamorphosis. Surprisingly, we observed a transient, stage-dependent recruitment of PRMT1 despite the continuous presence of high levels of liganded TR and PRMT1, suggesting tissue-specific functions for PRMT1 during metamorphosis. Finally, by transgenesis, we showed that overexpression of PRMT1 enhanced gene activation by TR through increased TR binding to target genes, resulting in accelerated natural, as well as T3-induced, metamorphosis.

MATERIALS AND METHODS

Experimental animals.

Wild-type tadpoles of X. laevis were reared in the laboratory or obtained from NASCO, and developmental stages were determined according to reference 52. Oocytes were prepared from adult female frogs purchased from NASCO. Stage 54 tadpoles were treated with 10 nM T3 for the indicated number of days at 18°C.

qRT-PCR.

qRT-PCR with a TaqMan probe was carried out to quantify gene expression levels on an ABI 7000 (Applied Biosciences) as described previously (26). For detection of PRMT1, forward primer CGTAACTCAATGTTTCACAACAGACA, reverse primer TCCGGCCTTTGCTGCAA, and 6-carboxyfluorescein-labeled TaqMan probe CCACTCCCAACATCC were used. For detection of TRβ and TH/bZIP, a primer-probe set was used for each (66). A primer-probe set for ribosomal protein rpL8 (26) was used as an internal control for each sample, and the expression levels of PRMT1, TRβ, and TH/bZIP in each sample were normalized to that of rpL8.

Western blot analysis of intestinal proteins.

Tadpole intestine was homogenized in nuclear extract buffer (0.5% Triton X-100, 10 mM Tris-HCl [pH 7.5], 3 mM CaCl₂, 0.25 M sucrose, a proteinase inhibitor tablet [Roche], 0.2 mM phenylmethylsulfonyl fluoride). The nuclei were isolated through a 100-μm nylon cell strainer (BD Falcon). After centrifugation, the pellet was mixed into immunoprecipitation buffer (20 mM HEPES [pH 7.5], 5 mM KCl, 1.5 mM MgCl₂, 1 mM EGTA, 10 mM β-glycerophosphate, 0.2 mM phenylmethylsulfonyl fluoride, protease inhibitor cocktail [Roche]) and the solution was sonicated. After centrifugation, the supernatant was quantified and used for Western blotting with anti-human PRMT1 (Upstate) and anti-histone H4 antibodies (Upstate).

Cloning and expression constructs.

A T7TS-6Myc vector was first constructed. Six copies of the Myc tag were added to the T7TS vector (76). Forward (catcataccggtCCAAAGAAGAAGCGTAAGGT) and reverse (atgatgggtaaccgagctcactagtTTCTAGAGGCTCGAGAGGCC) primers were used to clone six copies of the Myc tag from the pCS2+NLSMT vector, a gift from David L. Turner, University of Michigan (79). The PCR product was cloned into the T7TS vector.

Total RNA was extracted from the intestines of tadpoles at stage 56 by TRIZOL (Invitrogen), and a cDNA library was constructed by using the SMART RACE cDNA amplification kit (BD Biosciences) to clone PRMT1. Forward (GACAGACTACTCGAGATGGCCGAAGCGACCACCTGCAAC) and reverse (GTCGTATCTACTAGTTCAACGCATTCTGTAGTCTGTTG) primers were used to amplify the full-length PRMT1 sequence. The PCR product was cloned into the T7TS-6Myc vector. Several clones were isolated and sequenced for confirmation. One clone, T7TS-6Myc-PRMT1v1, was used in this study. To generate PRMT1 without the tag, the six-Myc tag was removed from T7TS-6Myc-PRMT1v1 by digestion with AgeI and XhoI and the vector was religated to produce T7TS-NT-PRMT1v1.

For transgenesis, the p(I-SceI)DPHG vector was first constructed. A 3-kb DNA fragment containing I-SceI recognition sites was amplified with primers 5′-AATGTCGACTAGGGATAACAGGGTAATGGTATCGATAAGCTTCGAGAAAGCTCG-3′ and 5′-AATACGCGTTAGGGATAACAGGGTAATTATGACCATGATTACGCCAAGCGCGCG-3′ and pDPHG, a modified double-promoter vector derived from pCGHG (25), as the template. The PCR product was double digested with SalI and MluI and inserted into pDPHG predigested with SalI and BssHII. The constructed was confirmed by sequencing.

To make the PRMT1 transgenic vector p(I-SceI)DPHG-3F-PRMT1v1, the T7TS-3F vector (48) and T7TS-6Myc-PRMT1v1 were digested with both XhoI and SpeI. The PRMT1 fragment isolated from T7TS-6Myc-PRMT1v1 was subcloned into the digested T7TS-3F vector to produce T7TS-3F-PRMT1, which had three copies of the FLAG tag fused to the N terminus of PRMT1. The p(I-SceI)DPHG vector and T7TS-3F-PRMT1 were digested with EcoRI and SpeI, respectively, filled with Klenow, and then digested again with AgeI. The 3F-PRMT1 fragment thus isolated from T7TS-3F-PRMT1 was then subcloned into digested p(I-SceI)DPHG to produce p(I-SceI)DPHG-3F-PRMT1v1.

Assay of transcription in Xenopus oocytes and Western blot analysis.

Transcription was assayed in Xenopus oocytes as previously described (77). To express proteins in the oocytes, expression plasmids were used to make the corresponding mRNAs with T7 or SP6 in vitro transcription kits (Ambion). The mRNAs (1.15 or 5.75 ng/oocyte for FLAG-tagged TR and RXR, 1.15 or 4.6 ng/oocyte for Myc-tagged PRMT1, and 1.15 ng/oocyte for PRMT1) were injected into the cytoplasm. After mRNA injection, the firefly luciferase reporter plasmid TRE-Luc (0.33 or 0.99 ng/oocyte), containing the T3-regulated promoter of the gene for Xenopus TRβA, and the control Renilla luciferase vector phRG-TK (0.03 or 0.09 ng/oocyte) were coinjected into the nucleus. After overnight incubation at 18°C with or without 100 nM T3, oocyte lysates were prepared for luciferase assay with the dual-luciferase assay kit (Promega) to determine the relative levels of expression from the T3-regulated promoter and the control promoter. Nine injected oocytes for each sample were divided into three groups, and a luciferase assay was performed for each group. A portion of the lysates was used for Western blot analysis with anti-FLAG M2 (Sigma), anti-Myc (Invitrogen), or anti-PRMT1 (Upstate) antibodies.

ChIP assay.

A ChIP assay on Xenopus oocytes was done as previously described (48). For FLAG ChIP, anti-FLAG-M2-agarose beads (Sigma) were used. Anti-PRMT1 (Upstate), anti-SRC3 (57), anti-p300 (55), anti-CARM1 (Upstate), anti-acetylated lysine 9 of histone H3 [ac(K9)H3; Upstate], anti-acetylated histone H4 (acH4; Upstate), and anti-methylated arginine 17 of histone H3 [me(R17)H3] antibodies (Upstate) were used for ChIP assays of PRMT1, ac(K9)H3, acH4, and me(R17)H3, respectively. Salmon sperm DNA-protein A-agarose (Upstate) was used for ChIP of PRMT1, ac(K9)H3, acH4, and me(R17)H3. After reversal of the DNA-protein cross-links, purification of the immunoprecipitated DNA was carried out with a PCR purification kit (Qiagen). After elution, the DNA was analyzed by qPCR with gene-specific primers and probes for the TRβ promoter (17) and the ampicillin resistance gene (48).

ChIP assay on the tadpole intestine was done as described previously (16). For ChIP on TR, PRMT1, and Id14, anti-TR(PB) (84), anti-PRMT1 (Upstate), and anti-Id14 (14) antibodies were used. qPCR was carried out on the immunoprecipitated DNA with gene-specific primers and probes for the TRβ promoter, the TH/bZIP promoter, and exon 5 of TRβ (16).

In vitro methylation assay.

One microgram of recombinant human PRMT1 (U.S. Biological), X. laevis histone H4 (U.S. Biological), and/or Xenopus TRβ (PROTEIN ONE) was incubated in the presence of 33.5 pmol of S-adenosyl-l-[methyl-³H]methionine ([³H]AdoMet; 81.9 Ci/mmol; Perkin-Elmer Life Sciences) in 20 μl of 25 mM Tris-HCl (pH 7.5) at 37°C for 60 min (3). After the reaction, samples were subjected to sodium dodecyl sulfate-polyacrylamide gel electrophoresis and the ³H signal was detected with a Storage Phospho Screen (GE Healthcare) and visualized with a Storm 840 (Amersham). A portion of the samples was used for Western blot analysis with anti-PRMT1 (Upstate), anti-histone H4 (Upstate), or anti-TR (new PB) (82) antibodies.

Histology.

Tadpole intestine fragments were fixed in 4% paraformaldehyde for 2 h at 4°C. After washing in phosphate-buffered saline, the tissues were immersed in 30% sucrose in phosphate-buffered saline at 4°C for cryoprotection, embedded, and frozen in optimal cutting temperature compound (Fisher Scientific, Pittsburgh, PA). Tissue sections were prepared at 8 μm. For histology, tissue sections were stained with methyl green-pyronine Y (Muto, Tokyo, Japan) for 5 min as described previously (33).

Transgenesis.

Transgenesis was carried out as described previously (25), by using the double-promoter construct p(I-SceI)DPHG-3F-PRMT1v1 to produce F₀ founder animals. F₁ generation animals were produced by mating an F₀ transgenic animal with a wild-type animal. Transgenic F₁ animals were identified by the green fluorescent protein (GFP) expression in the lens, and wild-type siblings (animals without GFP expression in the lens) were used as controls. For consistency, both wild-type and transgenic animals were reared together in the same container. To study the effect of the transgene, stage 53/54 tadpoles were heat shocked at 33 to 34°C for 30 min twice daily for 3 days, starting at day zero, as described previously (26). After this first period of heat shock treatment, tadpoles were heat shocked at 33 to 34°C for 30 min twice again and treated with or without 2 or 10 nM T3 for the indicated periods of time. For morphological analysis, the animals were subsequently heat shocked at 33 to 34°C for 30 min once daily during the T3 treatment period. The rearing water was changed every day.

To study natural metamorphosis, we heat shocked stage 58 tadpoles at 33 to 34°C for 30 min once daily (7). The morphology of the animals was examined every day.

RESULTS

The expression of PRMT1 mRNA and protein is upregulated in the intestine during natural and T3-induced metamorphosis.

To investigate the possible developmental role of PRMT1, we first analyzed PRMT1 expression in the intestine during Xenopus metamorphosis. Total RNA was isolated from the intestines of tadpoles from premetamorphic stage 54 to the end of metamorphosis (stage 66). Analysis of PRMT1 mRNA expression by qRT-PCR showed that PRMT1 mRNA was upregulated by stage 58 (Fig. 1A), the onset of intestinal remodeling (70). The expression increased further at stage 60, when adult stem cells appear as cell islets and the highest number of apoptotic larval epithelial cells are observed in the remodeling intestine. The PRMT1 mRNA levels peaked at stage 62, when maximal levels of T3 are detectable in tadpole plasma and the larval epithelium is almost completely replaced by proliferating adult epithelial cells in the intestine. By the end of metamorphosis at stage 66, when intestinal metamorphosis is complete, PRMT1 expression was dramatically reduced (Fig. 1A).

FIG. 1.

PRMT1 expression is upregulated in the intestine during metamorphosis. The expression of PRMT1 was analyzed during natural (A and C) and T3-induced (B and D) metamorphosis by qRT-PCR (A and B) and Western blotting (C and D). Note that the levels of PRMT1 mRNA and protein were gradually upregulated during natural development and reached a peak at stage 62 (A and C). The expression of both mRNA and protein was then reduced by stage 66, the end of metamorphosis (A and C). In T3-induced metamorphosis, the expression of both mRNA and protein was upregulated after 4 to 8 days of T3 treatment (B and D). A Western blot assay with anti-histone H4 antibody was used as an internal control for protein loading (C and D). α, anti.

To investigate PRMT1 expression during T3-induced metamorphosis, we treated premetamorphic tadpoles at stage 54 with 10 nM T3, a concentration close to the peak levels of T3 during metamorphosis (43). Total RNA was isolated from the intestine and analyzed by qRT-PCR. The results showed that PRMT1 expression was gradually upregulated in the intestine during the 8 days of T3 treatment (Fig. 1B). Furthermore, Western blot analysis with an anti-human PRMT1 antibody, which also recognizes its Xenopus counterpart (see below), showed that PRMT1 protein levels in the intestine were regulated similarly to mRNA levels during both natural and T3-induced metamorphosis (Fig. 1C and D).

Endogenous PRMT1 is recruited to the TRE of the TRβ promoter by liganded TR in vivo.

To investigate whether PRMT1 participates in metamorphosis by affecting gene regulation by liganded TR, we next studied whether PRMT1 can function as a coactivator of Xenopus TR in vivo by using the reconstituted frog oocyte as a model, which allows the study of gene regulation in the context of chromatin in vivo.

PRMT1 mRNA is known to be expressed in fruit fly, mouse, bovine, and Xenopus oocytes (3, 11, 80), although there have been no reports of the presence of PRMT1 protein in oocytes. Therefore, we first determined if PRMT1 protein was present in Xenopus oocytes. We microinjected frog oocytes with the mRNA encoding Myc-tagged Xenopus PRMT1. After overnight incubation, we isolated protein extracts from the injected and control oocytes and performed a Western blot analysis with an antibody against human PRMT1. As shown in Fig. 2A, the antibody recognized an endogenous band in the uninjected control oocytes of the size expected for Xenopus PRMT1. In the injected oocytes, another band corresponding to Myc-PRMT1 was also detected. Furthermore, when oocytes were injected with mRNA encoding untagged Xenopus PRMT1, a single band with higher intensity but the same size as in the control oocytes was detected by the antibody. These results indicate that the antibody recognizes Xenopus PRMT1 and that endogenous PRMT1 is present in the frog oocyte.

FIG. 2.

Xenopus PRMT1 is recruited to the TRE of a target promoter by liganded TR. (A) An anti (α)-human PRMT1 antibody recognizes Xenopus PRMT1, which is highly homologous to the human protein (data not shown). Oocytes were injected into the cytoplasm with mRNAs encoding Myc-tagged Xenopus PRMT1 (1.25 ng/oocyte). After overnight incubation, the protein extract from the samples was analyzed by Western blotting with an anti-human PRMT1 antibody. Note that an endogenous band was detected which corresponded to endogenous PRMT1 (see panel B) and was smaller than Myc-PRMT1 because of the Myc tag on the latter. (B) Endogenous PRMT1 is expressed in Xenopus oocytes. Extracts from uninjected oocytes or oocytes injected with PRMT1 mRNA (1.15 ng/oocyte) were analyzed by Western blotting with the anti-PRMT1 antibody or anti-histone H4 antibody as a loading control. Note that the injected oocytes had the same PRMT1 band (40.4 kDa) as the uninjected oocytes but with higher intensity due to the overexpression caused by the injection. (C) TR regulates transcription in a ligand-dependent manner in the oocyte. Oocytes were injected into the cytoplasm with mRNAs (5.75 ng/oocyte for FLAG-tagged TR and RXR). Subsequently, oocytes were injected into nuclei with reporter DNAs (T3-dependent reporter Xenopus TRβ promoter TRE-Luc at 0.99 ng/oocyte and control plasmid phRG-TK at 0.09 ng/oocyte). The oocytes were then incubated with or without T3 before harvesting for luciferase assays (top). The protein extract from the samples was also analyzed by Western blotting with anti-FLAG (for TR) or anti-PRMT1 antibody (bottom). (D) Endogenous PRMT1 is recruited to the TRE of the reporter TRβ promoter in the presence of T3. Oocytes were injected and incubated as described for panel C. A ChIP assay was performed with the anti-PRMT1 antibody to analyze PRMT1 recruitment. The presence of the TRE region of the TRβ promoter or the ampicillin resistance gene, which was in the same TRE-Luc reporter plasmid but far away from the TRE region, in the immunoprecipitated DNA was determined by qPCR. *, P < 0.05.

To study the effect of PRMT1 on gene regulation by TR, we microinjected mRNAs for FLAG-tagged TR and untagged RXR together with reporter plasmids (TRE-Luc, with the firefly luciferase reporter under the control of the T3-regulated promoter of Xenopus TRβA gene, and the control Renilla luciferase vector phRG-TK). The oocytes were incubated overnight with or without T3 before being harvested for luciferase assay. As we reported previously (48, 84), TR/RXR heterodimers repressed and activated transcription from the T3-inducible promoter in the absence or presence of T3, respectively (Fig. 2C). Under the conditions used, endogenous PRMT1 was present at a constant level in the oocytes (Fig. 2C). PRMT1 binds directly to SRCs and thus is expected to be recruited to target genes as multicomponent complexes with SRCs. Indeed, a ChIP assay of the same oocytes with the anti-PRMT1 antibody demonstrated T3-dependent recruitment of the endogenous PRMT1 to the TRE of the promoter when TR and RXR were overexpressed (Fig. 2D), suggesting that endogenous PRMT1 participates in gene activation by T3-bound TR in frog oocytes in vivo.

PRMT1 enhances TR-mediated transcription through increased TR binding to TRE and histone modification.

The effect of PRMT1 on gene activation by TR was analyzed by overexpressing PRMT1 in Xenopus oocytes. As shown in Fig. 3, microinjection of mRNA for Myc-PRMT1 had little effect on luciferase activity with or without TR/RXR in the absence of T3. On the other hand, in the presence of T3, overexpression of Myc-PRMT1 enhanced luciferase activity in a dose-dependent manner in the presence of both T3 and TR/RXR (Fig. 3, compare lanes 8 and 9 to lane 7), suggesting that PRMT1 functions as a coactivator for liganded TR.

FIG. 3.

Xenopus PRMT1 enhances TR-mediated transcription. Oocytes were injected with reporter DNA (0.33 ng/oocyte for TRE-Luc and 0.03 ng/oocyte for phRG-TK) and the indicated mRNAs (1.15 ng/oocyte for FLAG-tagged TR and RXR, 1.15 or 4.6 ng/oocyte for Myc-tagged PRMT1). After overnight incubation in the presence or absence of T3, the oocytes were isolated and subjected to a luciferase assay (top) or Western blot analysis with anti-Myc antibody for the expression of Myc-tagged PRMT1 (bottom). Note that PRMT1 overexpression enhanced transcription in the presence of both T3 and TR/RXR but not in the absence of T3 and/or TR/RXR. *, P < 0.05.

To investigate how PRMT1 enhanced promoter activity, a ChIP assay was performed. A ChIP assay with the anti-PRMT1 antibody showed that overexpression of PRMT1, as expected, led to increased recruitment of PRMT1 to the TRE region of the reporter (Fig. 4, compare lanes 6 and 5 in the PRMT1 panel). Interestingly, analysis of the FLAG-tagged TR with the anti-FLAG antibody showed increased TR binding to the TRE when PRMT1 was overexpressed with TR/RXR in the presence, but not in the absence, of T3 (Fig. 4, compare lanes 6 and 5 in the FLAG panel). Furthermore, ChIP assay with anti-SRC3, -p300, and -CARM1 antibodies showed that recruitment of these coactivators by liganded TR to the TRE was also increased when PRMT1 was overexpressed (Fig. 4, compare lanes 6 and 5 in the SRC3, p300, and CARM1 panels).

FIG. 4.

PRMT1 enhances TR binding and histone modifications at the TRE in the presence of T3. Oocytes were injected with reporter DNA and the indicated mRNAs (5.75 ng/oocyte each for FLAG-tagged TR, RXR, and Myc-tagged PRMT1). After overnight incubation, a ChIP assay of the oocytes was performed with antibodies to PRMT1, FLAG (for TR binding), SRC3, p300, CARM1, acetylated histone H4 (AcH4), acetylated lysine 9 of histone H3 [Ac(K9)H3], and methylated arginine 17 of histone H3 [Me(R17)H3]. Note that overexpression of PRMT1 led to increased binding of TR and cofactors to the TRE, as well as enhanced histone modifications. *, P < 0.05 for comparisons between samples with or without PRMT1 overexpression in the presence of T3. α, anti.

PRMT1 is a histone methyltransferase that targets histone H4 arginine 3 (H4 R3) for methylation. Its methyltransferase activity activates transcription and induces histone H4 and H3 acetylation by CBP/p300 and histone H3 methylation by CARM1 (PRMT4) (2, 22, 31, 81). Unfortunately, ChIP assay with a commercial antibody against mammalian R3-methylated (meR3) histone H4 failed to detect any changes in H4 methylation at the promoter in the presence or absence of T3 with or without PRMT1 overexpression (data not shown). This is likely due to the inability of the commercial antibody to recognize Xenopus meR3 H4, as a similar failure was reported with gene activation by the fusion protein of the DNA-binding domain of yeast transcription factor Gal4 and the ligand-binding domain of TR in the oocyte (44). On the other hand, overexpression of PRMT1 led to increased acetylation of histones H3 and H4, as well as the methylation of histone H3 R17 (Fig. 4), which was likely due to the increased recruitment, as shown above, of histone acetyltransferases such as SRC3 and p300 and the histone arginine methyltransferase CARM1, respectively. In all of our ChIP assays, only background signals were detected for the coding region of the ampicillin resistance gene on the same plasmid but far away from the TRE and no effect of PRMT1 overexpression was observed (Fig. 4, bottom panels). These results suggest that PRMT1 enhances the binding of TR and the recruitment of coactivators to the TRE in chromatin, leading to elevated local histone modification and transcriptional activation, at least in Xenopus oocytes.

PRMT1 fails to methylate TR in vitro.

PRMT1 has been reported to enhance DNA binding by HNF4, an orphan nuclear receptor, by methylating HNF4 (5). To investigate whether a similar mechanism exists for TR, we carried out in vitro methylation studies with recombinant proteins and [³H]AdoMet. As shown in Fig. 5, while PRMT1 methylated Xenopus histone H4 as expected (lane 1), no radioactive signal was detected for TR (lane 3), suggesting that PRMT1 does not methylate TR under the in vitro conditions used. This result is also consistent with the fact that Xenopus TRα, TRβ, and RXRs have no conserved arginine residue at the corresponding position in the D box of the DNA-binding domain of HNF4 that is methylated by PRMT1 (5; data not shown).

FIG. 5.

PRMT1 fails to methylate TR in vitro. One microgram of each of the recombinant proteins was mixed as indicated and incubated with [³H]AdoMet at 37°C for 1 h. The reaction products were subjected to sodium dodecyl sulfate-polyacrylamide gel electrophoresis for analysis by autoradiography (top) or a Western blot assay with anti (α)-PRMT1, anti-TR, or anti-histone H4 antibodies (bottom). Note that all of the reaction mixtures had the same amount of total proteins and that H4, but not TR, was methylated, as shown by the ³H label.

PRMT1 is recruited to the TREs of endogenous T3-inducible genes during metamorphosis.

PRMT1 is expressed throughout metamorphosis, with the highest levels at the climax of metamorphosis in the intestine (Fig. 1). Earlier studies have shown that the TR-binding coactivator SRC3 is recruited by liganded TR to endogenous promoters during metamorphosis (28, 56, 57). PRMT1 is thus likely recruited to endogenous promoters by TR since it binds to SRCs to form a multicomponent complex. To investigate this possibility, we analyzed the association of PRMT1 with the TREs of two T3-inducible genes, those for TRβ and TH/bZIP. Both genes are direct target genes (27, 62), and TR has been shown to bind to their TREs and recruit cofactors to the TREs in the intestine in vivo during metamorphosis (16, 28, 56, 57, 64, 65, 77). Thus, we isolated intestines from tadpoles at different stages of metamorphosis (Fig. 6A) and carried out ChIP assays with three different antibodies, anti-TR, anti-PRMT1, and anti-Id14 (an extracellular protein that served as a negative control) (14). The immunoprecipitated DNA was analyzed by qPCR with three different primer-probe sets for detecting TRE regions of the TRβ and TH/bZIP genes and exon 5 of the TRβ gene (another negative control).

FIG. 6.

PRMT1 is recruited to the TREs of T3-inducible genes in the intestine during natural metamorphosis. (A) Cross sections of the intestine at different stages, stained with methyl green/pyronin Y. (B) A ChIP assay was performed on intestines isolated from tadpoles at the indicated stages during metamorphosis with anti (α)-TR, anti-PRMT1, or anti-Id14 (extracellular protein, used as a negative control) antibody to analyze the recruitment of PRMT1 into the TREs of the endogenous TRβ and TH/bZIP promoters. The TRE regions of the two genes, as well as the region of exon 5 of TRβ, used as a negative control, in the immunoprecipitated DNA were analyzed by qPCR. *, P < 0.05.

A ChIP assay with the anti-TR antibody showed that TR-TRE binding at both genes increased during metamorphosis and reached a peak at stage 60 (Fig. 6B), similar to that reported previously (16). A ChIP assay of the same samples with anti-PRMT1 antibody showed that at stage 54, a low background signal was detected at the TREs of both the TRβ and TH/bZIP genes (Fig. 6B). As the endogenous T3 levels rose at stages 58 and 60, the PRMT1 signal at the TREs of both the TRβ and TH/bZIP genes increased dramatically. Surprisingly, the PRMT1 signals were reduced to premetamorphic levels at stage 62, the climax of intestinal remodeling, when the levels of plasma T3 and PRMT1 expression were at a peak (Fig. 1) (43). (Note that TR binding to the TRE is constitutive, although higher levels of binding were observed during metamorphosis. This was supported by the fact the TR signals at the TREs at stage 54 or 66 were much higher than the background levels detected with the anti-Id14 antibody while the PRMT1 signals at the TREs at stage 54 were close to the background levels with the anti-Id14 antibody [Fig. 6B].) The PRMT1 signal remained low at stage 66, the end of metamorphosis. In all of our ChIP assays, the signals with the anti-Id14 antibody or at exon 5 of the TRβ gene were very low and changed little during metamorphosis. Thus, PRMT1 was transiently recruited to the TREs of both target genes during intestinal metamorphosis.

The dramatic increase in PRMT1 recruitment to endogenous TR target genes correlated with the rise in the endogenous T3 level during early metamorphosis, suggesting that the recruitment was due to T3 binding to the chromatin-bound TR. To directly demonstrate this, premetamorphic tadpoles at stage 54 were treated with 10 nM T3 for up to 8 days. Their intestines were isolated and subjected to a ChIP assay as described above. Similar to natural metamorphosis, TR binding to the TRE of the TRβ gene was increased by T3 after 2 to 4 days of T3 treatment and then returned to pretreatment levels (Fig. 7). Likewise, TR binding to the TRE of the TH/bZIP gene was also elevated to peak levels after 2 days, although it remained above pretreatment levels throughout the treatment period. In contrast to TR binding but similar to that during natural metamorphosis, the association of PRMT1 with the TREs was elevated only after 2 days of T3 treatment and subsequently the binding returned pretreatment background levels (comparable to anti-Id14 signals or the signals at exon 5 of the TRβ gene) (Fig. 7). Again, the signals with the anti-Id14 antibody or at exon 5 of the TRβ gene were very low and changed little during T3 treatment. These results indicate that PRMT1 was transiently recruited to the TREs of both target genes by liganded TR during intestinal metamorphosis, and consistent with observations from ChIP assays with oocytes (Fig. 4), the maximal binding of TR to the TREs of the TRβ and TH/bZIP promoters coincides with the increased recruitment of PRMT1 to the promoters.

FIG. 7.

PRMT1 is transiently recruited to the TREs of T3-inducible genes in the intestine during T3-induced metamorphosis. Premetamorphic tadpoles at stage 54 were treated with T3 for the indicated numbers of days. The intestine was isolated and subjected to a ChIP assay with anti (α)-TR, anti-PRMT1, or anti-Id14 antibody. The TRE regions of the endogenous TRβ and TH/bZIP promoters, as well as the region of exon 5 of TRβ, used as a negative control, in the immunoprecipitated DNA were analyzed by qPCR. Note that enhanced PRMT1 binding was observed only after 2 days of T3 treatment, although various levels of TR binding were observed throughout the treatment period. *, P < 0.05.

Transgenic overexpression of PRMT1 enhances TR-TRE binding and target gene expression during metamorphosis.

The results described above strongly suggest that PRMT1 plays a role in gene activation by TR during metamorphosis. Currently, it is impossible to carry out gene knockout in Xenopus, and no successful knockdown of endogenous genes with small interfering RNAs and a transgenic expression plasmid in metamorphosing tadpoles has been reported (note that microinjected small interfering RNAs or antisense RNA oligonucleotides are not stable enough to survive the few weeks needed for metamorphic studies). Thus, we chose to overexpress PRMT1 to investigate whether PRMT1 can affect TR function in vivo and frog metamorphosis.

To overexpress PRMT1 in vivo, we used a double-promoter construct (25) in which the coding sequence for FLAG-tagged Xenopus PRMT1 was placed under the control of a heat shock-inducible promoter and the marker GFP was under the control of the second, eye-specific γ-crystallin promoter in the same plasmid (Fig. 8A). Transgenic founder animals were generated. Mature transgenic animals were used to produce F₁ generation tadpoles, with the transgenic F₁ tadpoles easily identified because of the GFP expression in their lenses (Fig. 8B). When premetamorphic transgenic tadpoles and wild-type siblings at stage 54 were subjected to heat shock treatment, the expression of PRMT1 was strongly upregulated in the intestines of transgenic but not wild-type animals (Fig. 8C), demonstrating the heat shock induction of the transgenic promoter.

FIG. 8.

Transgenic overexpression of PRMT1 enhances gene regulation by T3 through increased TR binding at target genes. (A) Schematic representation of the double-promoter construct used to generate transgenic animals. The Xenopus heat shock protein 70 promoter drives FLAG-tagged PRMT1 (F-PRMT1) expression. The lens-specific γ-crystallin promoter drives GFP expression in the animal eyes. (B) Two representative heat shock-treated sibling tadpoles during natural metamorphosis. The transgenic animal (Tg), but not the wild-type (Wt) animal, had green eyes (arrowheads) because of the expression of GFP from the construct encoding the transgene. (C) Heat shock (Hs) treatment of transgenic, but not wild-type, animals leads to increased PRMT1 expression. Wild-type and transgenic animals at stage 54 were treated with or without heat shock for 3 days. Total RNA was isolated from the intestine and subjected to qRT-PCR analysis. (D) Transgenic overexpression of PRMT1 enhances gene regulation by T3. Wild-type and transgenic animals at stage 54 were treated with heat shock for 3 days. They were then treated with or without 10 nM T3. Total RNA was isolated from the intestine after 0, 12, 24, 48, and 144 h of T3 treatment and subjected to qRT-PCR analysis for the expression of PRMT1 and two T3-iducible genes, those for TRβ and TH/bZIP. (E) Transgenic overexpression of PRMT1 leads to increased binding of TR and enhanced recruitment of PRMT1 to the TREs of the target genes for TRβ and TH/bZIP. Wild-type and transgenic animals at stage 54 were treated with heat shock for 3 days. They were then treated with or without 10 nM T3. The intestine was isolated from the animals after 0, 12, 24, 48, and 144 h of T3 treatment and subjected to a ChIP assay with anti (α)-TR, anti-PRMT1, and anti-Id14 (negative control) antibodies. The TRE regions of the two genes, as well as the region of exon 5 of TRβ, used as a negative control, in the immunoprecipitated DNA was analyzed by qPCR. *, P < 0.05 compared to wild-type animals.

To study the effect of PRMT1 overexpression on gene regulation by TR, the transgenic and wild-type sibling tadpoles at stage 54 were heat shocked for 3 days and then treated with or without 10 nM T3 and daily heat shocks. Total RNA was isolated from the intestines after 0, 12, 24, 48, and 144 h of T3 treatment and analyzed by qRT-PCR. As expected, PRMT1 expression was higher in the transgenic animals than in the wild-type animals throughout the treatment period (Fig. 8D). In the wild-type animals, PRMT1 expression gradually increased because of the induction of the endogenous PRMT1 gene by the T3 treatment (Fig. 8D). To study the effect of the overexpression of PRMT1 on TR function, we analyzed the expression of the two endogenous TR target genes, those for TRβ and TH/bZIP (Fig. 8D). The T3 induction was enhanced for both genes in the transgenic animals at 12, 24, and 144 h of T3 treatment (Fig. 8D). At 48 h of T3 treatment, there was a reduction in TRβ expression while no effect was observed on the TH/bZIP expression by the PRMT1 transgene (Fig. 8D). These findings demonstrate that PRMT1 enhances gene activation by liganded TR in vivo, at least during the early period of T3 induction. It should be pointed out that during natural metamorphosis or T3 treatment of premetamorphic tadpoles, the expression of both the TRβ and TH/bZIP genes first increases dramatically and then decreases toward the end of metamorphosis or after prolonged T3 treatment (34, 69, 72, 83). For example, during T3 treatment of wild-type tadpoles, their expression peaked at around 48 h of T3 treatment but remained much higher than in untreated animals even after 144 h (Fig. 8D). Thus, it may not be surprising that no enhancement for either gene due to PRMT1 overexpression was observed after 48 h of T3 treatment. Furthermore, the slight reduction in TRβ expression in the transgenic animals after 48 h suggests that TRβ expression reached the peak level earlier in the transgenic animals because of the acceleration of T3 induction by the PRMT1 transgene.

To investigate how PRMT1 overexpression enhanced TR-mediated transcription in the transgenic animals, a ChIP assay was performed as described above on the intestines of animals heat shocked for 3 days and then treated with 10 nM T3 for 0, 12, 24, 48, and 144 h. Again, a ChIP assay with the anti-TR antibody showed that TR binding to the TREs of both the TRβ and TH/bZIP genes was elevated by T3 treatment (Fig. 8E). TR binding to the TREs was enhanced in transgenic animals compared to that in their wild-type siblings at 12, 24, and 144 h, but not 48 h, of T3 treatment, and the recruitment of PRMT1 to the TREs followed a similar pattern (Fig. 8E). In addition, in both wild-type and transgenic animals, PRMT1 recruitment first rose and then decreased during T3 treatment, although higher levels of PRMT1 were found in the transgenic animals than in their wild-type siblings, even at 144 h of treatment. These findings are in agreement with the effect of the PRMT1 transgene on TRβ and TH/bZIP expression (Fig. 8D). The lack of any enhancement by PRMT1 overexpression at 48 h of T3 treatment on either target gene expression or the binding of TR or PRMT1 to the TREs of the target genes was likely because both genes normally reach their peak levels of transcription at around 48 h of T3 treatment in wild-type animals. PRMT1 overexpression likely altered the kinetics of their T3 induction, with the TRβ gene reaching peak levels of expression and TR binding to its TRE peaking at around 24 h of T3 treatment while TH/bZIP had higher levels of expression and TR binding to its TRE in the first 24 h of T3 treatment. In addition, consistent with a role for PRMT1 in TR function in the presence but not in the absence of T3, overexpression of PRMT1 had no effect on TR binding to the TREs in the absence of T3 (Fig. 8E, 0 h). Again, the negative controls with the anti-Id14 antibody or at exon 5 of the TRβ gene had very low, constant background signals during T3 treatment.

Transgenic overexpression of PRMT1 accelerates metamorphosis.

As gene activation by TR is sufficient for metamorphosis (17), the enhanced activation of TR target genes by transgenic PRMT1 suggests that PRMT1 overexpression would enhance T3-induced metamorphosis. To investigate this possibility, we subjected transgenic and wild-type sibling animals at stage 58, the onset of metamorphic climax, to daily heat shock and examined their external morphology daily. As shown in Fig. 9A, after 6 days of heat shock treatment, the wild-type animal reached stage 60 while the transgenic one was at stage 61. Statistical analysis showed that the transgenic animals reached stages 60 and 62 significantly faster than their wild-type siblings, requiring only 5.20 ± 0.05 days (to stage 60) and 8.69 ± 0.07 days (to stage 62), compared to 6.50 ± 0.12 and 10.08 ± 0.21 days, respectively, for the wild-type siblings (Fig. 9B). Thus, transgenic overexpression of PRMT1 accelerates natural metamorphosis.

FIG. 9.

Transgenic overexpression of PRMT1 accelerates metamorphosis. (A and B) PRMT1 accelerates natural metamorphosis. Twelve wild-type (Wt) and 26 transgenic (Tg) tadpoles were heat shocked for 30 min daily, starting at stage 58. Each animal was individually analyzed morphologically after daily heat shock treatment until it reached stage 62. Representative photographs (A) show that after 6 days, the wild-type animal reached stage 60 while the transgenic animal was at about stage 61. The average time for the transgenic animals to develop from the starting point of stage 58 to stage 60 or 62 was shorter than that for the wild-type animals (B). *, P < 0.05 compared to wild-type animals. (C) PRMT1 transgenic animals develop to more advanced metamorphic stages when treated with T3. Sixteen wild-type and 17 transgenic animals were heat shocked for 30 min twice daily, starting at stage 56. Three days after the first heat shock treatment, all of the animals were treated with 2 nM T3 and continued to be subjected to heat shock treatment daily. Ten to 12 tadpoles (5 or 6 wild-type and 5 or 6 transgenic animals) were kept in the same container and analyzed morphologically every 3 days throughout the 15-day experiment. The average developmental stages were plotted. *, P < 0.05 compared to wild-type animals.

To study the effect of PRMT1 overexpression on T3-induced metamorphosis, we subjected animals at stage 56, when endogenous T3 levels were low, to daily heat shock treatment. Three days later, 2 nM T3 was added to the rearing water. The animals were continuously heat shocked daily, and the rearing water was changed every day. As shown in Fig. 9C, the transgenic animals overexpressing PRMT1 developed into statistically significantly more advanced stages within 3 days of T3 treatment and remained more advanced throughout the 12 days of T3 treatment. Thus, PRMT1 overexpression not only enhanced gene regulation by liganded TR but also accelerated tadpole metamorphosis, consistent with the causative role of gene activation by TR in metamorphosis.

DISCUSSION

Anuran metamorphosis is one of the most dramatic events mediated by TR-dependent gene regulation and serves as an excellent model to understand the function and molecular mechanism of transcriptional regulation by TR in vivo during postembryonic development in vertebrates. TR has been shown to be both necessary and sufficient to mediate the metamorphic effects of T3 (13, 15, 17, 51, 67). We have further demonstrated that both SRC3 and p300 function as coactivators in TR-mediated transcription during metamorphosis and that SRC/p300 coactivator complexes are necessary for gene regulation by TR and metamorphosis (55-58). Our studies reported here have demonstrated that the SRC/p300-interacting histone arginine methyltransferase PRMT1 participates in gene regulation by TR during frog metamorphosis. We have shown for the first time that overexpression of a TR coactivator accelerates metamorphosis, suggesting that the levels of cofactor expression influence the metamorphic rate. Furthermore, we have discovered that, surprisingly, liganded TR recruits PRMT1 only transiently during intestinal metamorphosis, suggesting that PRMT1 has tissue-specific and developmental stage-dependent functions during metamorphosis.

PRMT1 enhances transcription by increasing the binding of TR to TREs in vivo.

It is generally believed that PRMT1 facilitates the formation of active chromatin through histone modifications, including directly methylating residue R3 of histone H4 (73, 81) and indirectly altering histone H3 and H4 acetylation and histone H3 R17 methylation (2, 22, 31). In the absence of a good antibody against R3-methylated Xenopus H4 for a ChIP assay, we and others have not been able to show whether H4 R3 methylation is induced by liganded TR (data not shown; 44). On the other hand, in the oocyte, PRMT1 enhanced acetylation of histone H3 K9 and histone H4 and arginine methylation of histone H3 R17 in the presence of TR and T3, which is likely due to the observed enhanced recruitment of other coactivators such as CARM1 (for H3 R17 methylation). In addition, Xenopus PRMT1 is capable of methylating histone H4 in the in vitro histone methyltransferase assay (6). Together, these findings suggest that PRMT1 recruited to the TRE facilitates local H4 R3 methylation, which in turn enhances other histone modifications associated with active chromatin. More importantly, we discovered a novel function for PRMT1 in gene regulation by TR, that is, enhancing the binding of TR to TREs in chromatin in the presence of T3. Furthermore, the recruitment of other coactivators such as p300 and CARM1 is enhanced by the overexpression of PRMT1. It is quite possible that the enhanced binding of liganded TR to the TREs in the presence of PRMT1 is responsible for more recruitment of other cofactors to further modify local histones and/or other proteins for transcriptional activation.

How does PRMT1 enhance TR binding to TREs? It has been reported that PRMT1 enhances DNA binding by HNF4, an orphan nuclear receptor, in chromatin, because of the methylation of HNF4 by PRMT1 (5). In vitro, the binding of TR/RXR heterodimers to the TREs of both the TRβ and TH/bZIP genes is independent of T3 (16). On the other hand, a ChIP assay has shown that TR-TRE binding in chromatin in the frog oocyte is higher in the presence than in the absence of T3 (Fig. 4) (48). The Xenopus TRα, TRβ, and RXRs have no conserved arginine residue at the corresponding position in the D box of the DNA-binding domain of HNF4 that is methylated by PRMT1 (5; data not shown). There has also been no report of arginine-methylated sites in any TR or RXR. Consistently, PRMT1 failed to methylate Xenopus TRβ under conditions in which it was able to methylate Xenopus histone H4. On the other hand, there are arginine residues in TRs and RXRs, including the DNA-binding domains. Thus, we cannot completely exclude the possibility that PRMT1 regulates TR-TRE binding by methylating TRs and RXRs under in vivo conditions. In addition, it is also possible that PRMT1 enhances TR binding to TRE in chromatin through histone modifications at or around the TRE. PRMT1 is a histone methyltransferase that targets H4 R3 for methylation. While we have been unable to detect any changes in H4 methylation at the promoter (data not shown), likely because of the inability of the commercial antibody to recognize Xenopus meR3 H4, PRMT1 may affect TR binding to TRE in chromatin through histone methylation. Finally, PRMT1 may simply function as an integral part of a large coactivator complex through its interactions with SRC and p300 and the association of this coactivator complex may stabilize the TR-TRE interaction. Thus, as the PRMT1 level rises, more coactivator complexes are available, leading to increased TR-TRE association.

PRMT1 may also indirectly enhance TR-TRE binding by methylating other proteins. More than 20 substrates have been identified for PRMT1 (9, 53). PRMT1 may regulate the interactions of these substrates with nucleic acids or other proteins. One of them, PGC-1, a known coactivator for nuclear receptors, including TRs (86), is methylated by PRMT1, which enhances transcriptional activation by estrogen receptor and TR (75). In addition, arginine methylation is reduced by over 85% in PRMT1-null ES cells (59), suggesting the existence of many additional PRMT1 substrates. Some of them may encode enzymes that can modify TR to enhance its DNA binding. Interestingly, PRMT1 enhances TR binding to TREs only in the presence of T3, suggesting that PRMT1 regulates TR-TRE in the context of a larger complex associated with liganded TR. Clearly, further analysis is needed to determine the exact mechanism governing this enhanced binding to TREs by TR in vivo.

Transient recruitment of PRMT1 to TREs by liganded TR during metamorphosis.

We have shown the PRMT1 is recruited by TR in the presence of T3. PRMT1 binds to and forms a multicomponent complex with SRCs (5, 19, 38, 41, 45, 75, 81). We and others have shown previously that SRC3 is recruited by liganded TR in the frog oocyte and during frog metamorphosis and, more importantly, that SRC-p300 complexes are required for gene activation by TR and frog metamorphosis (28, 55-57). Thus, it is likely that PRMT1 is recruited to TREs as a complex with SRCs.

A surprising finding from our in vivo analyses is that PRMT1 is only transiently associated with the TREs of endogenous target genes during either natural or T3-induced metamorphosis. The genes for both TRβ and TH/bZIP are direct T3 response genes. During both natural and T3-induced metamorphosis, PRMT1 is recruited to the TREs of both the TRβ and TH/bZIP genes. However, PRMT1 is subsequently released from the TREs of both genes even though their expression remains at high levels for much longer period during both natural and T3-induced metamorphosis (Fig. 8D) (34, 69, 72, 83). In addition, this transient nature of PRMT1 recruitment was also observed in transgenic animals overexpressing PRMT1. It is unclear how PRMT1 is released from the TRE after its recruitment by liganded TR, as TR remains bound to the TREs (a small reduction was observed during natural metamorphosis when PRMT1 binding to the TREs were reduced to the background level [Fig. 6], which is consistent with the enhancement of TR-TRE binding by PRMT1) and the levels of T3 and PRMT1 are high. One possibility is that PRMT1 may methylate a component of the TR coactivator complex, which then facilitates the dissociation of PRMT1. In addition, it has been reported that histone acetylation inhibits histone methylation by PRMT1 (81). We have shown previously that histone H4 acetylation is upregulated at the TREs during metamorphosis (56, 65, 77). It is possible that this upregulation of histone H4 acetylation somehow results in the dissociation of PRMT1 from the TREs. It is difficult to determine the exact mechanism responsible for the release of PRMT1 from the promoters.

Tissue- and stage-dependent roles of PRMT1 during metamorphosis.

Our studies with frog oocytes have clearly demonstrated the ability of PRMT1 to function as a coactivator for TR in the presence of T3. More importantly, we have shown that PRMT1 is recruited to the TREs of TR target genes during both natural and T3-induced metamorphosis. Furthermore, transgenic overexpression of PRMT1 not only enhances gene activation by liganded TR through increased TR binding to the TREs but also accelerates natural, as well as T3-induced, metamorphosis, a process mediated by TR. These results strongly support a role for PRMT1 in facilitating gene activation by liganded TR to mediate the metamorphic effects of T3 during development.

We have also observed that PRMT1 is only transiently recruited to the TREs in vivo, while target gene activation lasts much longer. In addition, PRMT1 expression remains high even after PRMT1 recruitment is reduced. These findings suggest two types of functions for PRMT1 during metamorphosis, at least in the animal intestine. During early metamorphosis, from stage 58 to stage 60, PRMT1 functions as a coactivator in TR-mediated transcription to facilitate the activation of target genes such as those for TRβ and TH/bZIP. This explains the recruitment of PRMT1 to the TREs during this period and the ability of PRMT1 overexpression to accelerate metamorphosis. After stage 60, when drastic histological changes occur in the intestine, PRMT1 recruitment to the TREs drops while its expression continues to rise and the expression of TR target genes such as the genes for TRβ and TH/bZIP is further induced by T3. This suggests that PRMT1 is involved in the regulation of other genes during this period of intestinal remodeling, possibly independently of TR. In addition, PRMT1 might regulate TR targets other than TRβ and TH/bZIP subsequent to initiation of the remodeling process. Clearly, it will be interesting to identify these other PRMT1-regulated genes and determine whether their regulation by PRMT1 is dependent on TR.

The metamorphosis of the intestine involves two major changes, the degeneration of the larval epithelium through apoptosis and de novo development of the adult epithelium (70). Adult cells are first microscopically visible around stage 60 and nearly completely replace the larval epithelial cells by stage 62, when PRMT1 recruitment to target genes is reduced to background levels. During this period, the cells in the other two major tissues, connective tissue and muscles, undergo extensive proliferation. After stage 62, the cells in all three major tissues differentiate to form the adult intestine. On the other hand, we have previously shown that the genes for both TRβ and TH/bZIP are strongly activated in the larval epithelium but not or weakly in the other tissues before stage 60, although the expression of both genes is subsequently activated in all cell types of the intestine (34, 72). Thus, it is reasonable to speculate that PRMT1 mainly functions as a coactivator in larval epithelial cells during the early metamorphic period up to stage 60.

In adult epithelial cells, which are present at high levels at stage 62 when little PRMT1 recruitment is found at the TREs, PRMT1 is likely not involved in the activation of these TR target genes. Instead, PRMT1 may be involved in the proliferation and/or differentiation of adult epithelial cells during this late metamorphic period. A potential role for PRMT1 in cell proliferation has been suggested by earlier studies. For example, PRMT1 is recruited to the cyclin E1 promoter in G₁/S phase (24). It has also been reported that PRMT1 is involved in oncogenesis (20). It will be of interest to further investigate this interesting role of PRMT1 in cell proliferation in the future. In addition, although not microscopically identifiable, adult epithelial cells are first formed prior to stage 60 (by stage 60, adult cells are present as multicell islets [49, 70]). During this early phase of adult cell development, PRMT1 potentially functions as a TR coactivator in the activation of TR target genes important for this process because of strong recruitment of PRMT1 to the TREs at stage 60 and earlier. It will be important in the future to separate larval epithelial cells from proliferating adult epithelial cells to determine whether PRMT1 has similar or different effects on TR target genes in larval versus adult epithelial cells.

In conclusion, we have shown here that PRMT1 is capable of functioning as a coactivator for liganded TR in the context of chromatin. Its involvement in gene activation by TR during metamorphosis is supported by (i) its recruitment to TREs of endogenous target genes during natural and T3-induced metamorphosis when TR and T3 are present and (ii) the fact that overexpression of PRMT1 enhances TR binding to TREs in vivo and gene activation by TR and, more importantly, accelerates metamorphosis. Furthermore, the surprising finding of the transient recruitment of PRMT1 to the TREs suggest that PRMT1 has tissue- and developmental stage-dependent roles during metamorphosis through both TR-dependent and -independent mechanisms.