Interaction of PIMT with transcriptional coactivators CBP, p300, and PBP differential role in transcriptional regulation.

PIMT (PRIP-interacting protein with methyltransferase domain), an RNA-binding protein with a methyltransferase domain capable of binding S-adenosylmethionine, has been shown previously to interact with nuclear receptor coactivator PRIP (peroxisome proliferator-activated receptor (PPAR)-interacting protein) and enhance its coactivator function. We now report that PIMT strongly interacts with transcriptional coactivators, CBP, p300, and PBP but not with SRC-1 and PGC-1alpha under in vitro and in vivo conditions. The PIMT binding sites on CBP and p300 are located in the cysteine-histidine-rich C/H1 and C/H3 domains, and the PIMT binding site on PBP is in the region encompassing amino acids 1101-1560. The N-terminal of PIMT (residues 1-369) containing the RNA binding domain interacts with both C/H1 and C/H3 domains of CBP and p300 and with the C-terminal portion of PBP that encompasses amino acids 1371-1560. The C-terminal of PIMT (residues 611-852), which binds S-adenosyl-l-methionine, interacts respectively with the C/H3 domain of CBP/p300 and with a region encompassing amino acids 1101-1370 of PBP. Immunoprecipitation data showed that PIMT forms a complex in vivo with CBP, p300, PBP, and PRIP. PIMT appeared to be co-localized in the nucleus with CBP, p300, and PBP. PIMT enhanced PBP-mediated transcriptional activity of the PPARgamma, as it did for PRIP, indicating synergism between PIMT and PBP. In contrast, PIMT functioned as a repressor of CBP/p300-mediated transactivation of PPARgamma. Based on these observations, we suggest that PIMT bridges the CBP/p300-anchored coactivator complex with the PBP-anchored coactivator complex but differentially modulates coactivator function such that inhibition of the CBP/p300 effect may be designed to enhance the activity of PBP and PRIP.

The next step in the multistep transcriptional activation process involves participation of the TRAP/DRIP/ARC mediator complex anchored by the coactivator PBP (4,5,26). While CBP/p300 and the p160 family of cofactors that form the initial multiprotein complex function by exhibiting HAT and arginine methyltransferase activities, there is limited functional information about the coactivator PBP and other proteins that form the TRAP/DRIP/ARC complex. This complex facilitates interaction with RNA polymerase II complexes of the basal transcriptional machinery, but how this is accomplished remains unclear (4,5,16,26,34,35). PBP lacks HAT activity, and it is uncertain if the other members of the TRAP/DRIP/ARC complex in mammals possess enzymatic activity of any kind. Disruption of the PBP and CBP/p300 genes in the mouse results in embryonic lethality around E11.5 days, indicating that disruption of these pivotal anchoring coactivators affects the function of many nuclear receptors and possibly other transcription factors (27, 36 -40). The recently identified coactivator designated PRIP/ASC2/RAP250/NRC/TRBP has been shown to interact with several nuclear receptors and also with CBP/p300 and TRAP130 of the TRAP/DRIP/ARC complex (18 -22). Thus, PRIP appears to serve as a bridge between the initial histone acetyltransferase-histone methyltransferase complex of CBP/ p300 and p160 coactivators and the downstream TRAP/DRIP/ ARC complex. Disruption of this PRIP gene in the mouse leads to embryonic lethality around E13.5 days, implying that PRIP (like CBP/p300 and PBP) is also critical for embryonic development and survival. 2 We have previously isolated a PRIPinteracting protein, designated PIMT, using the yeast twohybrid approach with PRIP as bait. We found that PIMT enhances the nuclear receptor coactivator function of PRIP (41). PIMT, which has an invariant GXXGXXI segment found in K-homology motifs and many RNA-binding proteins, binds RNA (41). PIMT also has a nine amino acid VVDAFCGVG methyltransferase motif I and binds S-adenosyl-L-methionine, the methyl donor for the methyltransfer reaction, suggesting that it may be a putative RNA methyltransferase (41). We report here that PIMT interacts with CBP, p300, and PBP but not with SRC-1 and PGC-1␣, suggesting that PIMT functions as a link between the two major multiprotein complexes an-chored by CBP/p300 and PBP, respectively. Of considerable interest is that while PIMT enhances the PRIP and PBPmediated nuclear receptor transcriptional activity, it inhibits the CBP/p300-mediated nuclear receptor function.
Construction of Adenovirus PIMT-To construct recombinant adenovirus-PIMT (Ad/PIMT), the full-length PIMT coding region was cloned by PCR from plasmid pCMV-PIMT (41) and inserted into the BamHI/ XhoI site of pFastBac HTc (Invitrogen). The entire coding region of PIMT with the hexahistidine affinity tag was cut with RsrII/XhoI and transferred to the SalI site of pShuttle vector (Quantum Biotechnologies, Inc.) by blunting the ends. The recombinant Ad/PIMT was generated as described previously (42). The adenoviral construct of Ad/LacZ was the generous gift of Dr. W. El-Deiry (University of Pennsylvania, Philadelphia) and has been described previously (43).
GST Pull-down Assays-GST fusions or GST alone was expressed in Escherichia coli BL21 (DE3) bound to glutathione-Sepharose 4B beads (Amersham Biosciences) and incubated with [ 35 S]methionine-labeled coactivator/coactivator-binding protein expressed by in vitro translation by using the TNT-coupled transcription-translation system (Promega). Briefly, the binding assays were carried out by incubating 10 l of [ 35 S]methionine-labeled coactivator/coactivator-binding protein for 2 h at 4°C with the immobilized GST fusion protein in GST binding buffer (180 mM KCl, 20 mM Tris-HCl, pH 7.9, 1 mM EDTA, 0.05% Nonidet P-40, 1 mM dithiothreitol, 1 mM phenylmethylsulfonyl fluoride). Beads were washed four times with 1 ml of binding buffer containing 0.1% Nonidet P-40, and bound protein was eluted by boiling for 2 min in 20 l of SDS sample buffer, was analyzed by SDS-PAGE, and was subjected to autoradiography.
Immunoprecipitation and Western Blotting-COS-7 cells were cultured in Dulbecco's modified Eagle's medium containing 10% fetal bovine serum and transfected with pCMV-PIMT-FLAG plasmid encoding FLAG-tagged PIMT along with one of the coactivator expression plasmids. After 72 h cells were harvested and lysed at 4°C by vortexing in lysis buffer (100 mM Tris-HCl, pH 8.0, 250 mM NaCl, 1% Nonidet P-40, 1 mM EDTA) containing protease inhibitors (1 mM phenylmethylsulfonyl fluoride, 1 g/ml leupeptin, 1 g/ml pepstatin, 1 g/ml aprotinin). Lysates were clarified by centrifugation at 18,000 ϫ g for 20 min and immunoprecipitated at 4°C for 2 h with anti-CBP, anti-p300, anti-PBP (Santa Cruz Biotechnology), or anti-PRIP, respectively, followed by precipitation with 50 l of protein A/G-Sepharose (Sigma) overnight at 4°C. After four washes with lysis buffer (described above), the immunoprecipitates were eluted by boiling for 5 min in Laemmli sample buffer. Resulting immunoprecipitates were electrophoresed in 10% SDS-PAGE, transferred onto a nitrocellulose membrane, immunoblotted with anti-FLAG monoclonal antibody M2 (Sigma), and detected using ECL chemiluminescence (Amersham Biosciences).
Alternatively, PIMT-binding proteins were immunoprecipitated from HEK293 cells that were infected with Ad/PIMT or Ad/LacZ. HEK293 cells were cultured to 95% confluence in 150-cm plates. The cells were at first incubated with 10 ml of Dulbecco's modified Eagle's medium containing 10% fetal bovine serum with Ad/PIMT or Ad/LacZ respectively at a multiplicity of infection of 100. After 1.5 h, 35 ml of Dulbecco's modified Eagle's medium containing 10% fetal bovine serum was added. Infected HEK293 cells were harvested 24 h after infection. The cell lysates were immunoprecipitated with anti-His tag (Santa Cruz Biotechnology) followed by precipitation with protein A/G-Sepharose as described above and immunoblotted with anti-CBP, anti-p300, anti-PRIP, and anti-PBP. The antigen-antibody complexes were then detected using the alkaline phosphatase-conjugated substrate kit (Bio-Rad). 2 C. Qi and J. K. Reddy, unpublished observations. Immunofluorescence-COS-1 cells were transfected with pCMV-PIMT-FLAG along with one of the coactivator expression plasmids, namely CBP, p300, and PBP using Polyfect (Qiagen). After 48 h transfection, the cells were washed and fixed in 1% formaldehyde and washed twice with phosphate-buffered saline, pH 7.4, after which autofluorescence was quenched with 50 mM ammonium chloride in phosphate-buffered saline. Cells were washed with phosphate-buffered saline, permeabilized with 0.1% Triton X-100, and then blocked with 0.2% fish skin gelatin. Cells were incubated with the primary antibody anti-FLAG (Sigma) along with one of the antibodies (anti-CBP, anti-p300, or anti-PBP) followed by incubation with secondary antibodies. Fluorescence microscopy and digital image collection were performed by using an Olympus (New Hyde Park, NY) microscope and a photometrix cooled charge-coupled device camera driven by DELTAVISION software from Applied Precision (Seattle).
Transactivation Assay-CV1 cells (for CBP transactivation) and HEK293 cells (for p300 and PBP transactivation) were plated in Dulbecco's modified Eagle's medium with 10% fetal bovine serum without antibiotics in six well plates and cultured for 24 h before transfection. Transfections were carried out with LipofectAMINE Plus reagent (Invitrogen) according to the manufacturer's instructions for 5 h. Each transfection contains 20 g of PPAR expression vector, 1.5 g of luciferase reporter plasmid DNA, 0.1 g of ␤-galactosidase expression vector, and 1 g of appropriate expression vector. Cell extracts were analyzed 24 h after transfection for luciferase and ␤-galactosidase activities (Tropix).
Having confirmed that PGC-1␣ and SRC-1 failed to interact with truncated PIMT (PIMTC), we then tested the interaction of GST-PIMTA (aa 1-334) and GST-PIMTB (aa 326 -852) with in vitro translated [ 35 S]methionine-labeled full-length SRC-1 and PGC-1␣ (Fig. 1C). We failed to detect any binding of SRC-1 and PGC-1␣ with these PIMT fragments. We also used fulllength in vitro translated PIMT and investigated its in vitro interaction with various SRC-1-and PGC-1␣-truncated proteins expressed as GST fusions. In these assays full-length PIMT did not interact with these two coactivators (data not shown).
Mapping of the CBP and PIMT Domains Required for Interaction-In order to define the interacting domain(s) of CBP on PIMT, we prepared GST fusion proteins containing various peptides of CBP and determined the ability of each of these to interact with [ 35 S]methionine-labeled in vitro translated fulllength PIMT. GST and GST-CBP fusion proteins (shown schematically in Fig. 2A) were bacterially expressed, purified, coupled to glutathione-Sepharose beads, and used for in vitro binding assays. CBP2 peptide (aa 301-600) containing the cysteine-histidine-rich 1 (C/H1) region and CBP7 peptide (aa 1801-2100) containing the cysteine-histidine-rich 3 (C/H3) region interacted significantly with full-length PIMT (Fig. 2B). The extent of interaction of PIMT with CBP2 peptide is decreased when a pull-down assay was done at the higher salt concentration of 250 mM KCl, indicating weak interaction of PIMT with peptide containing the C/H1 domain compared with that of peptide containing the C/H3 domain (data not shown). Using further truncated CBP2 and CBP7 fragments, we narrowed down the PIMT binding sites on CBP to two regions encompassing amino acids 301-400 (Fig. 2B, CBP2A) and 1801-1900 (Fig. 2B, CBP7A), respectively.
We [ 35 S]methionine, and equal portions of these peptides were used to assess their binding to GST alone, GST-CBP2A, and GST-CBP7A in the in vitro pull-down assays (Fig. 2C). Both PIMTN (aa 1-369) and PIMTC (aa 611-852) interacted with CBP7A (aa 1801-1900), whereas only PIMTN interacted with CBP2A (aa 301-400) (Fig. 2C), indicating two independent domains of interaction of PIMTN on CBP. PIMTM did not bind these GST-CBP fusion peptides. We can conclude that PIMT has two independent binding sites on CBP, one in the region of aa 301-400 and the other in the region of aa 1801-1900. These CBP regions also interact with adenoviral E1A oncoprotein (44).
Mapping of the p300 and PIMT Domains Required for Interaction-We also determined the interacting domains of PIMT and p300 using GST fusion proteins containing various peptides of p300 (Fig. 3). Of the GST-p300 fusion proteins (depicted schematically in Fig. 3, A and B), p300/2 (aa 301-597) and p300/5 (aa 1501-1800) interacted with full-length in vitro translated PIMT (Fig. 3B). Further analysis revealed that p300/2A and p300/5C interacted strongly with PIMT, indicating that the PIMT binding domain in p300 is in the region encompassing aa 301-920 and aa 1701-1800 (Fig. 3B). The results of the binding of PIMT peptides (PIMTN, PIMTM, and PIMTC) to p300 clearly established that the PIMT N and C domains interact(s) with p300 (Fig. 3C).

Mapping of the PBP and PIMT Domains Required for Interaction-In vitro GST pull-down analyses were performed with recombinant GST fusion proteins of PBP and [ 35 S]methioninelabeled in vitro translated PIMT proteins. GST-PBP fragments
were generated to map the regions required for interaction with PIMT (Fig. 4, A and B). Three overlapping fragments in the C-terminal of PBP (PBP5, PBP6, and PBP7 covering aa 740 -1130, 981-1370, and 1360 -1560, respectively) interacted with PIMT, but PIMT failed to interact with fragments containing the LXXLL region (PBP2, PBP3, and PBP4) and GST alone (Fig. 4B). To determine the domain(s) of PIMT involved in interaction with PBP, we used different truncated PIMT peptides in GST pull-down assays (Fig. 4C). The results show that PIMTN (aa 1-369) interacts with the extreme C-terminal of PBP in the region encompassing aa 1371-1560 (Fig. 4C,  PBP11), whereas PIMTC (aa 611-852) interacts with PBP in the region encompassing aa 1101-1370 (Fig. 4C, PBP10), indicating two different interactive sites of PIMT for PBP.
Interaction of PIMT with CBP, p300, PBP, and PRIP In Vivo (PIMT Complex)-Given that PIMT interacts with the nuclear receptor coactivators CBP, p300, and PBP in vitro, we wanted to determine whether PIMT interacts with CBP, p300, and PBP in the context of intact cells. We used two different approaches. First, a vector encoding human PIMT with the Cterminal FLAG epitope was cotransfected separately along with one of the coactivators (namely CBP, p300, PBP, and PRIP) (used as positive control) into COS-1 cells derived from African Green Monkey Kidney (American Type Culture Collection, CRL1651). The potential complex between PIMT and each coactivator was immunoprecipitated separately using respective antibodies (i.e. anti-CBP, anti-p300, anti-PBP, and anti-PRIP), and the products were analyzed by immunoblot using anti-FLAG to demonstrate the presence of PIMT in the precipitates (Fig. 5A). The results showed that all antibodies precipitated PIMT, demonstrating that PIMT interacts in vivo with these nuclear receptor coactivators (Fig. 5A) and suggesting the existence of a PIMT-coactivator complex in vivo. Second, to pull down such a complex, we infected HEK293 cells with Ad/PIMT or Ad/LacZ and immunoprecipitated recombinant PIMT using anti-His antibodies. The immunoprecipitate obtained in this fashion contained CBP, p300, PBP, PRIP, as well as PIMT (Fig. 5B), indicating the existence of a complex of at least five proteins anchored by or containing PIMT as a component.
PIMT Colocalizes with CBP, p300, and PBP in the Nucleus-We have shown that PIMT interacts with CBP, p300, PBP, and PRIP both in vitro and in vivo. Previously, we demonstrated by immunofluorescence microscopy that PIMT and PRIP colocalize within the nucleus. To evaluate whether PIMT colocalizes with its newly identified binding partners, a plasmid containing three FLAG epitopes linked to the C-terminal portion of the PIMT protein was separately cotransfected into COS-1 cells along with one of the coactivator plasmids (i.e. CBP, p300, or PBP). Immunofluorescence (green) with anti-FLAG revealed that the expressed PIMT protein is localized in the nucleus (Fig. 6, A, D, and G). Localizations of CBP, p300, and PBP using anti-CBP, anti-p300, and anti-PBP, respectively, revealed that CBP, p300, and PBP are also localized in the nucleus (red fluorescence, Fig. 6, B, E, and H). When localized images of PIMT (green) were separately merged with localized images of CBP, p300, and PBP (red), an appreciable degree of overlapping localization (yellow) of PIMT with CBP ( Fig. 6C), p300 (Fig. 6F), and PBP (Fig. 6I) have been noted. These results demonstrated colocalization of PIMT with coactivator CBP, p300, and PBP in the nucleus similar to the PIMT and PRIP colocalization described before (41).
Differential Modulation of Transcriptional Activities of Coactivator CBP, p300, and PBP by PIMT-To determine the functional relevance of the interaction of PIMT with CBP, p300, and PBP, we transiently overexpressed PIMT with CBP in CV1 cells and PBP and p300 in HEK293 cells along with PPAR␥. We monitored the transcriptional activity of PPAR␥ with the expression of the thymidine kinase promoter-driven PPRE-linked reporter luciferase gene. When transfected individually, CBP, p300, and PBP consistently increased the transcription of the PPAR␥-mediated luciferase gene by approximately 2-fold in the case of CBP (Fig. 7A), 5-fold in the case of p300 (Fig. 7B), and 3-fold in the case of PBP (Fig. 7C) in the presence of the PPAR␥ ligand BRL49653. Cotransfection of PIMT and CBP or p300 in this assay resulted in a repression of ligand-dependent reporter gene expression (Fig. 7, A and B), a phenomena that is well established in the case of adenoviral E1A oncoprotein-mediated transcriptional repression (14). Cotransfection of PIMT and PBP resulted in further enhancement (Fig. 7C) indicating a synergism of PIMT and PBP activities. Earlier, we showed that within the context of transient transfection PIMT also synergized PRIP action (41). These results demonstrate that PIMT differentially modulates the coactivator-mediated transcriptional activity, acting as a transcriptional repressor of CBP and p300 while functioning as an enhancer of PBP and PRIP coactivator function. DISCUSSION Coactivators play a central role in mediating nuclear receptor transactivation by functioning as at least two large multiprotein complexes, one anchored by CBP/p300 and the other by PBP (4,5). The multistep model of transcription proposes that the acetylation-methylation functions of the initial CBP/p300mediated complex leads to a transition from CBP/p300-dependent to a mediator-dependent stage of transcription involving the TRAP/DRIP/ARC complex of coactivators anchored by PBP (4,33). The central importance of CBP/p300 and PBP in transcription is underscored by embryonic lethality observed in gene knockout studies in mice (36 -40). One other coactivator of equal importance appears to be the recently cloned PRIP/ ASC-2/RAP250/TRBP/NRC (18 -22), which has not been identified as part of either the CBP/p300 or the PBP-anchored multiprotein coactivator complexes. However, disruption of the PRIP gene appears to result in embryonic lethality. 2 The embryonic lethality of null mutation as well as PRIP's ability to interact with a variety of nuclear receptors and CBP (18 -22) suggested that PRIP is as indispensable as CBP, p300, and PBP in mediating the transcriptional activity of nuclear receptors and other transcription factors (41). To elucidate the functional role of PRIP and to identify coactivator-binding protein(s), we used PRIP as bait in a yeast two-hybrid screen and identified PIMT as a PRIP-interacting protein capable of enhancing PRIP's coactivator function (41). Interestingly, PIMT binds to S-adenosyl-L-methionine and RNA, implying that PIMT may function as a putative methyltransferase (41). Truncated PIMT, without the methyltransferase domain, still showed its ability to enhance the PRIP-mediated transcriptional activity of PPAR␥ and RXR␣, suggesting that putative enzyme activity of PIMT may not be crucial to transcription under transient transcription conditions (41).
In this study, we have examined the ability of PIMT to interact with other coactivators such as CBP, p300, PBP, SRC-1, and PGC-1␣ in an attempt to further explore the role of PIMT in nuclear receptor signal transduction. Our results show that PIMT interacts with CBP, p300, and PBP under in vitro and in vivo conditions, but it failed to interact with SRC-1 and PGC-1␣, implying selectivity of association. Using GST pull-down assays, we mapped the PIMT binding site(s) on CBP and p300 to their C/H1 and C/H3 domains. Further analysis revealed that the N-terminal of PIMT containing the RNA binding domain interacts with these two sites, whereas the C-terminal S-adenosyl-L-methionine binding domain of PIMT appears to interact with only the C/H3 domain. These two regions, namely C/H1 and C/H3, of CBP/p300 participate in the binding and activity of a variety of transcription factors (26,45). In particular, a small domain of C/H3 binds diverse proteins including adenoviral E1A oncoprotein and the coactivator TIF-2/SRC-2 (45). The PIMT binding site on PBP was localized to region aa 1101-1560. Since this region is devoid of the LXXLL motif, it appears that binding of PIMT to PBP is not contingent upon the presence of this motif. Nevertheless, the PIMT binding site on PRIP was in the region (aa 773-927) of PRIP that has an LXXLL that is considered necessary for PRIP-nuclear receptor interaction (18,41).
The in vivo interaction of PIMT with CBP, p300, and PBP was ascertained by immunoprecipitation using antibodies against specific coactivators and then immunoblotting with anti-FLAG to detect the presence of this epitope-tagged PIMT (Fig. 5). Alternatively, the presence of CBP, p300, PBP, and PRIP were detected by immunoblot analysis of immunoprecipitates obtained using anti-His antibodies to precipitate adenovirally expressed PIMT with hexahistidine. These approaches clearly established the presence in vivo of PIMT-coactivator complex containing at least four of what appear to be general coactivators. Furthermore, immunofluorescence studies with anti-FLAG revealed the presence of expressed epitope-tagged PIMT protein in a speckled pattern in nuclei. A merging of the distribution patterns of CBP, p300, or PBP with that of PIMT revealed overlapping colocalization of PIMT with these coactivators. From these studies it is reasonable to assume that the PIMT complex includes, in addition to PRIP, some components of HAT-methyltransferase containing the CBP/p300-coactivator complex as well as PBP, the major component of the TRAP/ DRIP/ARC complex. Thus, PIMT and PRIP appear to serve as linkers between two major coactivator complexes involved in the multistep model of transcription (4,5). However, it is plausible to consider the existence of one major coactivator complex (Fig. 8) and that different signals pass selectively through different components of this megacomplex to accomplish transcription factor-specific gene transcription instead of the sequential recruitment of two complexes. It is important to recall that SRC-1 with its histone acetyltransferase activity was found to be dispensable in generating the PPAR␣-dependent signal transduction in SRC-1-null mice. This implies a redundancy of the p160 family of coactivators for PPAR␣ signaling despite the fact that SRC-1 is an essential component of the CBP/p300-coactivator complex (46). In contrast, partial blunting of steroid hormone responses was found in SRC-1-null mice (47), further attesting to the use of different coactivators for different circuits necessary for receptor-mediated signal transduction.
We found that PIMT functions as a strong enhancer of PBP coactivator function when the PPAR␥-mediated transcription function was assayed using transient transfection in the presence of PPAR␥ ligand BRL49653. In contrast, our results also show that PIMT represses the CBP/p300 coactivator function. The integrator protein p300 was demonstrated to interact with PPAR␣ and enhance its transcriptional activity (48). The repressor effect of PIMT on the CBP/p300 coactivator function is opposite of what we observed in this study with PBP and previously with PRIP (41) where PIMT augmented the coactivator functions of both PRIP and PBP. The mechanism by which PIMT exerts this coactivator-dependent differential transcriptional activity is not clear. It is possible that PIMT by virtue of its ability to bind CBP/p300 may be disrupting the initial coactivator complex formation in a manner analogous to that of E1A oncoprotein (44). E1A binding to CBP/p300 interferes with normal cellular transcription and cell cycle progression (14). However, the C/H3 domain where E1A binds is not required for the E1A inhibition of transcription. This is due to the E1A inhibition of the assembly of CBP-nuclear receptorcoactivator complex formation (44). PIMT also binds to CBP/ p300 at this C/H3 site, and overabundance/overexpression of PIMT or lower concentrations of CBP/p300 may lead to disassembly or nonassembly of the CBP/p300-coactivator complex. Another possibility is that the RNA binding property of PIMT together with its putative methyltransferase activity, as suggested by its ability to bind S-adenosyl-L-methionine, may influence transcription or RNA processing machinery. PIMT reveals the presence of S-adenosyl-L-methionine binding site VVDAFCGVG, which is similar to the highly conserved methyltransferase motif I with a consensus hh(D/S)(L/P)FXGXG (where h represents a hydrophobic residue and X represents any amino acid) (41). The FCGVGGN present in PIMT corresponds with a single methylating consensus sequence (FG-GRGGF) motif found in many methyl-accepting substrates including RNA and DNA methyltransferases (49,50). One of the substrates happens to be the NPL3 gene product in yeast that is implicated in a wide variety of cellular processes including nuclear import, rRNA processing, association with and export of poly(A ϩ ) messenger RNA from the nucleus, and mRNA splicing (51)(52)(53). It would be important to examine if PIMT serves as a substrate for CARM1 and participates in mRNA processing during transcription. Further work is essential to characterize the biochemical functions of PIMT and also to ascertain its in vivo function by generating mice with a disrupted PIMT gene.
In summary, recent studies have highlighted the critical importance of histone acetylation, histone methylation, and coactivator methylation in regulating gene transcription (2, 4, 5, 31-33). Our results strongly implicate the involvement of RNA-binding proteins such as PIMT and RNA methyltransferases in transcription.