|Home | About | Journals | Submit | Contact Us | Français|
Sirtuins, homologs of the yeast SIR2 family, are protein deacetylases that require nicotinamide adenosine dinucleotide as cofactor. To determine whether the sirtuin family of deacetylases is involved in progesterone receptor (PR)-mediated transcription, the effect of sirtuin inhibitor, nicotinamide (NAM), was monitored in T47D breast cancer cells. NAM suppressed hormone-dependent activation of PR-regulated genes in a dose-dependent manner. Surprisingly, NAM-mediated inhibition of PR-mediated transcription occurs independently of SIRT1 and PARP1. Chromatin immunoprecipitation experiments did not show that PR binding nor that of the coactivators CBP and SRC3 was compromised. Consistent with the recruitment of the BRG1 chromatin remodeling complex, promoter chromatin remodeling still occurs despite NAM inhibition of PR transactivation. Rather, we show that this inhibition of transcription is due to dramatic loss of recruitment of the basal transcriptional machinery to the promoter. These results show that NAM uncouples promoter chromatin remodeling from transcription preinitiation complex assembly and suggest the existence of vital NAM-regulated steps required for promoter chromatin remodeling and basal transcription complex communication.
Steroid hormone receptors such as the progesterone receptor (PR) and the glucocorticoid receptor (GR) are part of a large nuclear receptor (NR) family of eukaryotic transcription factors (41, 59). NRs play essential roles in numerous biological processes such as growth and development, reproduction, homeostasis, and metabolism by eliciting a transcriptional output from target genes in response to their cognate ligands which include steroids, retinoids, thyroid hormone, and vitamin D3, among others (12, 41, 59). Studies of NR action have not only provided insight into their physiological roles but have been vital in the overall understanding of the mechanism of transcription by transcription activators (34, 42, 49, 55).
The process of ligand-dependent transcription initiation by steroid hormone receptors (SHRs) such as PR and GR involve ligand binding, followed by receptor binding to the hormone responsive elements at the promoter DNA of target genes as dimers. The promoter-bound SHR leads to recruitment of a large number of coactivators that work in sequence and/or in combination to ultimately facilitate the recruitment of RNA polymerase II (RNAP II) and the transcription machinery to elicit a transcriptional response (34, 49, 59). These coactivators include the p160/SRC family of proteins (SRC1, -2, and -3) that directly interact with NRs through consensus LXXLL motifs (NR boxes) (25, 49). The p160/SRC family of proteins is also able to associate with histone-modifying enzymes such as histone acetyltransferase p300/CBP and histone methyltransferase CARM1, thereby playing a role in bridging the recruitment of chromatin modifiers to SHR target promoters (34, 49, 56). In addition to the p160/SRC family of proteins, PR and GR both also interact with and recruit the ATP-dependent chromatin remodeling complex, SWI/SNF to target promoters to render the promoter chromatin more accessible, allowing additional transcription factors, coactivators, and the general transcription factors access the promoter DNA (2, 6, 21, 34).
The SHRs also recruit the multisubunit Mediator complex (also known as TRAP/DRIP/ARC/CRSP/SMCC complexes) (5). The Mediator complex has been purified by various biochemical methods and have been found to interact directly with various NRs, including thyroid hormone receptor (19), vitamin D receptor (51, 52), estrogen receptor (ER) (30), and androgen receptor (AR) (63), as well as GR (26). These contacts are made through the NR box motif (LXXLL) containing subunit MED1 (also known as TRAP220, ARC/DRIP205, PBP, and CRSP200 but referred to here by the MED acronym nomenclature set forth by Bourbon et al. ), as well as MED14 (also known as TRAP170, ARC/DRIP150, and p110) in the case of GR (19, 26, 66). The Mediator complex is thought to aid in the recruitment RNAPII and the formation of the transcription preinitiation complex (PIC) machinery to ligand-activated promoters. Many of the RNAPII regulated genes appear to require the Mediator complex for gene expression, with specific subunits of the complex playing distinct roles in regulating target genes through their interactions with various transcription activators and the RNAPII transcription machinery (15, 35, 40).
Adding to the complexity of the combinatorial recruitment of large numbers of coactivators for gene activation is the regulation of transcription factor and coactivator activities themselves by posttranslational modifications. Some examples include SHR, which undergoes several types of modifications such as acetylation, phosphorylation, ubiquitylation, and sumoylation that affect receptor activity and stability (18). p300 acetylates promoter chromatin, as well as autoacetylates itself, which leads to p300 dissociation and enhancement of TFIID binding (8). In addition, NR coactivator ACTR (SRC3) is acetylated by p300/CBP, which leads to disruption of the ACTR (SRC3) interaction with ER and coincides with cessation of transcription (13).
In the present study, we were interested in identifying novel factors that may influence PR-mediated transcription. In particular, we focused on sirtuins, homologs of the yeast SIR2 family, which are protein deacetylases that require nicotinamide adenosine dinucleotide (NAD+) as a cosubstrate (9, 16, 27). The closest mammalian structural ortholog of the yeast Sir2 protein, SIRT1, exerts its effects on a wide range of cellular metabolism by acting as a deacetylase of histones, as well as nonhistone substrates such as PCAF, p300, p53, PGC-1α, AR, and possibly ERα (22, 32, 65). The effect of the sirtuin inhibitor, nicotinamide (NAM) (4, 17), on PR-mediated transcription was monitored in T47D human breast cancer cells to determine whether the sirtuin family of deacetylase is involved in PR-mediated transcription. While NAM is a known sirtuin inhibitor, it is also known to inhibit the poly-ADP-ribosylation (PAR) activities of poly-ADP-ribose polymerase (PARP) family of proteins (37). PARP1, the best characterized member of the PARP family, has been closely linked to transcription through various mechanisms such as influencing chromatin structure, ribosylation of transcription factors, and altering Mediator activities, in addition to its role in DNA repair (29, 31, 33, 48). Our data show that NAM suppresses hormone-dependent activation of PR-regulated genes in a dose-dependent manner. However, unexpectedly, small interfering RNA (siRNA) knockdown experiments demonstrate that NAM inhibition of PR mediated transcription occurs independently of SIRT1 and PARP1. The inhibition of PR-mediated transcription is due to a dramatic loss of recruitment of the basal transcriptional machinery (RNAPII, TBP, Mediator) as determined by chromatin immunoprecipitation (ChIP) assays, while hormone-dependent association of PR and various coactivators at the promoter is not compromised in the presence of NAM. Interestingly, restriction enzyme hypersensitivity assay shows that promoter chromatin remodeling still occurs despite NAM inhibition of PR transactivation. These results suggest that NAM inhibits the coordination of basal transcription machinery assembly after chromatin remodeling of the promoter. Chromatin remodeling therefore must be followed by distinct critical steps in which the remodeling event is efficiently communicated with the transcription PIC formation event.
T47D/2963.1 (2963.1) cells were derived from human T47D breast cancer cells by stable transfection of the chimeric bovine papillomavirus-based construct pJ83d carrying the mouse mammary tumor virus (MMTV) long terminal repeat (LTR) attached to the bacterial chloramphenicol acetyltransferase (CAT) gene (45). T47D/A1-2 (A1-2) cells were derived from human T47D breast cancer cells by stable transfection of the GR expression plasmid pGRneo and the plasmid pHHLuc that contains the MMTV-LTR sequences attached to the luciferase gene (47). Cells were grown at 37°C with 5% CO2 in Dulbecco modified Eagle medium (Invitrogen, Carlsbad, CA) containing 10% fetal bovine serum (HyClone, Logan, UT) supplemented with 10 mM HEPES and 2 mM glutamine (Invitrogen).
2963.1 and A1-2 cells grown in six-well plates were treated with NAM or vehicle control (H2O) for 30 min, followed by treatment with hormone or ethanol (EtOH) for 4 h as indicated in the figure legends. Total RNA was isolated by using TRIzol reagent (Invitrogen) according to the manufacturer's protocol. 2 μg of total RNA was used to perform the RT reaction according to the First-Strand synthesis protocols (Invitrogen). PCR analysis was performed by real-time PCR with the following primers: MMTV-Nucleosome A (5′-AGT CCT AAC ATT CAC CTC TTG TGT GT-3′ and 5′-ACC CTC TGG AAA GTG AAG GAT AAG T-3′), GAPDH (5′-TCG GAG TCA ACG GAT TTG G-3′ and 5′-GGC AAC AAT ATC CAC TTT ACC AGA GT-3′), CEBP/β (5′-CGT GCC CGC TGC AGT T-′3 and 5′-CTC GCA GTT TAG TGG TGG TAA GTC-3′), SGK (5′-GAC CCC GAG TTT ACC GAA GAG-3′ and 5′-GGA AAG CCT CGG CAG CTT-3′), and EZF (5′-CGC TCC ATT ACC AAG AGC TCA T-3′ and 5′-CGA TCG TCT TCC CCT CTT TG-3′). Real-time PCRs were performed by using the Stratagene SYBR green QPCR master mix and Stratagene Mx3000p instrument (La Jolla, CA). After the gene transcript levels were normalized by that of GAPDH, the level of transcription in the absence of any treatment (−siRNA, −NAM, −hormone) was set to 1 as described in the figures. The data presented are an average of three independent experiments with the standard mean error indicated.
2963.1 and A1-2 cells were seeded in six-well plates (105/well) and grown overnight. 2963.1 cells were transfected with 100 pmol of siRNA against nontargeting scrambled sequence (Dharmacon, Lafayette, CO), SIRT1 (Dharmacon), or PARP1 (Santa Cruz Biotechnology, Santa Cruz, CA) or 50 pmol each of SIRT1 and PARP1 siRNA, while A1-2 cells were transfected with 200 pmol of nontargeting scrambled sequence siRNA or 100 pmol each of MED1 and MED14 siRNA (Dharmacon) per well, using Lipofectamine 2000 reagent (Invitrogen) as according to the manufacturer's protocol for 48 h.
2963.1 and A1-2 cells were lysed in buffer X (100 mM Tris-Cl [pH 8.5], 250 mM NaCl, 1% [vol/vol] Nonidet P-40, and 1 mM EDTA) with protease inhibitor cocktail (Sigma, St. Louis, MO). Proteins were electrophoresed on 6% or 4 to 12% sodium dodecyl sulfate (SDS)-polyacrylamide gels and transferred to polyvinylidene difluoride membranes (Invitrogen). Membranes were probed with antibodies to SIRT1 (H-300), PARP1 (F-2), RNAP II (N-20), TBP (SI-1), and MED1/TRAP220 (M-255) (all from Santa Cruz Biotechnology), MED14/DRIP150 (gift from M. Garabedian), and α-tubulin (Sigma).
2963.1 cells (1.2 × 107) were seeded in 15-cm diameter tissue culture plates and treated the next day with 50 mM NAM or vehicle (H2O) for 30 min, followed by treatment with 10 nM R5020 or vehicle (EtOH) for 1 h. The cells were then fixed with 1% formaldehyde at 37°C for 10 min. Cells were collected by centrifugation in phosphate-buffered saline containing protease inhibitors. Nuclei were isolated as previously described (36), lysed in SDS lysis buffer, followed by the ChIP assay as described by Upstate Biotechnology. Immunoprecipitation was performed overnight with antibodies to PR (H-190), NF-1 (H-300), BRG1 (H-88), CBP (A-22), SRC3/NCoA-3 (M-397), RNAP II (N-20), and TBP (SI-1) (all from Santa Cruz Biotechnology) and MED1/TRAP220 and MED14/DRIP150 (gifts from M. Garabedian). For acetylated histone ChIP assays, 2963.1 cells (2 × 106) were seeded in 10-cm-diameter tissue culture plates and treated the next day with 50 mM NAM or vehicle (H2O) for 15 min, followed by treatment with 10 nM R5020 or vehicle (EtOH) for 30 min. The cells were then fixed with 1% formaldehyde at 37°C for 10 min. Cells were collected by centrifugation in PBS containing protease inhibitors and lysed in SDS lysis buffer. After immunoprecipitation, 60 μl of salmon sperm DNA-protein A-agarose was added for 1 h at 4°C to capture the immune complexes. The agarose beads were washed, chromatin extracted, and protein-DNA cross-link reversed, and the proteins were digested by proteinase K as indicated in the Upstate ChIP assay protocol. DNA was purified by QIAquick PCR purification kit (Qiagen, Valencia, CA) and analyzed by real-time PCR analysis using Stratagene SYBR green QPCR master mix and Stratagene Mx3000p instrument with the following primers: Nucleosome B (5′-GGT TAC AAA CTG TTC TTA AAA CGA GGA T-3′ and 5′-CAG AGC TCA GAT CAG AAC CTT TG-3′) or Nucleosome A (sequences listed above). The data presented are the average of three independent experiments with standard mean of error as indicated.
The sirtuins, SIRT1 in particular, have been known to participate in various endocrine signaling pathways, including those mediated by PPARγ, PGC1α, ERα, and AR (22, 32, 65). We wanted to examine the possibility of sirtuins being involved in PR-mediated transcription by using an inhibitor of sirtuins, NAM. NAM is a reaction intermediate of SIRT1 deacetylation activity that feeds back to inhibit the deacetylation activity of SIRT1 (4, 17). The effect of NAM on PR-mediated transcription was determined by using T47D/2963.1 (2963.1) human breast cancer cells that endogenously express PR (45) that were treated with or without NAM for 30 min, followed by treatment with synthetic progesterone R5020 or vehicle (EtOH) for 4 h. RT-PCRs were performed from isolated RNA, and the expression levels of known PR-regulated genes CEBP/β, EZF and SGK (53, 62) were analyzed by real-time PCR. The RT-PCR analysis revealed that NAM inhibits hormone-dependent activation of PR-regulated genes in a dose-dependent manner (Fig. (Fig.1A).1A). In addition, to determine whether this is a phenomenon that is particular to PR activity, the effect of NAM on the activity of a related receptor GR was tested in the same T47D human breast cancer cell line background that expresses rat GR [T47D/A1-2 (A1-2) cells] (47). RT-PCR analysis of GR- and PR-regulated genes in the A1-2 cell line demonstrated that NAM also inhibits glucocorticoid dexamethasone (Dex)-induced GR-mediated transcription in addition to PR-mediated transcription (Fig. (Fig.1B1B).
The inhibition of PR-mediated transcription by NAM suggested that perhaps SIRT1 and PARP1 are involved in this event. Both SIRT1 and PARP1 have been closely linked to NR-mediated transcription (9, 33, 65). To test this possibility, SIRT1 and PARP1 levels individually or together were reduced by siRNA treatment in 2963.1 cells, followed by analysis of PR-mediated transcription in the presence or absence of NAM. Efficient reduction of both SIRT1 and PARP1 protein levels after 48 h of siRNA treatment compared to treatment with scrambled nontargeted siRNA control were achieved as demonstrated on the Western blot of prepared whole-cell extracts in both NAM- and R5020-treated and untreated cells (Fig. (Fig.2A).2A). After siRNA treatment, the cells were treated with or without NAM for 30 min, followed by 4 h of R5020 treatment or vehicle control. The RT-PCR results show that knockdown of SIRT1 and PARP1 individually or together leads to a slight decrease in PR-mediated expression of EZF and SGK. However, these changes are modest and, most importantly, all of the PR-regulated genes analyzed showed a marked decrease in expression in the presence of NAM in spite of a significant loss of cellular SIRT1 and PARP1 protein (Fig. 2B and C). This siRNA experiment demonstrates that SIRT1 and PARP1 may play some role in the transcription of a subset of the genes but that the NAM inhibition of PR-mediated transcription of the three genes tested occurs independently of SIRT1 and PARP1.
Since the siRNA experiment (Fig. (Fig.2)2) excluded SIRT1 and PARP1 as the targets of NAM-mediated inhibition of PR-mediated transcription, a series of ChIP experiments were conducted to determine the mechanism by which NAM inhibits transcription. ChIP experiments were performed to monitor the recruitment of various cofactors and transcription factors to the well-characterized progesterone responsive MMTV promoter. The MMTV promoter has been used extensively to study the mechanism of GR- and PR-mediated transcription among others, and many of the important factors involved in the activation of this promoter are known. The MMTV promoter adopts a well-characterized chromatin architecture with six rotationally phased nucleosomes termed A to F when stably integrated into the host genome. Nucleosome A harbors the transcription start site, while PR binding sites reside in the nucleosome B region and the site of coactivator recruitment and SWI/SNF-induced nucleosome remodeling (2). The 2963.1 cell line has the full-length MMTV promoter integrated in the genome with the CAT reporter that allows of analysis of transcription driven by the MMTV promoter (45). Treatment of 2963.1 cells with NAM leads to the loss of PR-mediated transcription of MMTV much like that of endogenous genes (Fig. (Fig.3A).3A). In addition, the effect of NAM on PR-mediated transcription activation of MMTV also occurs independently of SIRT1 and PARP1, as shown by the siRNA knockdown-RT-PCR experiment performed and analyzed identically to those of the endogenous PR-regulated genes (Fig. (Fig.3B3B).
After we established that the MMTV transcription in response to NAM is comparable to those of the endogenous genes, we performed ChIP experiments to monitor the promoter association of the PR and a transcription factor nuclear factor 1 (NF1) and coactivators BRG1, CBP, and SRC3, all known to be important for efficient hormone-induced activation of MMTV (2, 38). The ChIP experiments were performed by treating 2963.1 cells with or without 30 min of NAM, followed by 1 h with or without R5020, and the results of the DNA precipitated with the indicated antibodies were analyzed by real-time PCR using MMTV promoter primers (Fig. (Fig.3C)3C) and GAPDH gene primers as a background negative control (data not shown). The results of the ChIP experiments show that the ability of the PR to associate with the MMTV promoter in a hormone-dependent manner is not compromised in the presence of NAM. Coactivators BRG1 of the SWI/SNF complex, the histone acetyltransferase CBP, p160/SRC family of coactivator SRC3, and the transcription factor NF1 were also efficiently recruited to the promoter in a hormone-dependent manner in the presence of NAM (Fig. (Fig.3C3C).
One of the hallmarks of MMTV activation is the remodeling of chromatin at the nucleosome B (promoter) region after recruitment of the SWI/SNF complex (21, 58). It is possible, however, that while the recruitment of BRG1, the ATPase catalytic subunit of SWI/SNF, and other coactivators is not compromised by NAM treatment (Fig. (Fig.3C),3C), the efficiency of the hormone-dependent increase in the accessibility of promoter chromatin region diminishes in the presence of NAM. To evaluate the extent of the chromatin accessibility of the MMTV promoter, a restriction enzyme hypersensitivity assay was performed using the SstI enzyme. The SstI enzyme cleavage site resides within the nucleosome B region of the MMTV promoter (Fig. (Fig.4A),4A), and the accessibility of this cleavage site was monitored upon treatment with NAM and R5020. Isolated nuclei from 2963.1 cells treated with or without NAM for 30 min, followed by 1 h with or without R5020, were digested with the SstI enzyme, followed by digestion of the isolated DNA with the HaeIII enzyme to completion for quantitation control. The extent of SstI cleavage was monitored by reiterative PCR analysis. Consistent with previous work, 2963.1 cells demonstrate constitutive hypersensitivity at the SstI site in the absence of R5020 (Fig. (Fig.4B,4B, lane 1) (45), which remains accessible in the presence of NAM (Fig. (Fig.4B,4B, lane 3). Upon hormone treatment, nucleosome B hypersensitivity increases slightly, although the effect is modest due to the constitutively open nature of nucleosome B in the absence of hormone in this cell line (Fig. (Fig.4B,4B, compare lanes 1 and 2). This hormone-dependent increase in hypersensitivity is maintained in the presence of NAM after R5020 treatment (Fig. (Fig.4B,4B, compare lanes 3 and 4). The analysis of MMTV nucleosome B remodeling was also extended to the A1-2 cell line, which also has MMTV promoter stably integrated in the genome with a luciferase reporter (47). The MMTV in the A1-2 cell line is highly responsive to treatment with Dex. Much like the endogenous GR-responsive genes, the induction of GR-dependent MMTV transcription is inhibited in the presence of NAM (Fig. (Fig.1B1B and data not shown). Unlike the 2963.1 cell line, the nucleosome B region in the A1-2 cell line is not constitutively hypersensitive and remains relatively inaccessible to SstI digestion, as expected (Fig. (Fig.4C,4C, lane 1). Since the MMTV promoter in this particular cell line is not efficiently induced by PR, changes in chromatin accessibility were assessed after Dex treatment (3). Upon Dex treatment, the accessibility at the SstI increases significantly (Fig. (Fig.4C,4C, lane 2). This increase in nucleosome B accessibility upon hormone treatment occurs just as efficiently in A1-2 cells treated with NAM as in those that were not exposed to NAM (Fig. (Fig.4C,4C, compare lanes 2 and 4).
In addition, ChIP experiments were performed using acetylated histone H3 and H4 antibodies to determine whether the hormone-dependent chromatin modification such as histone acetylation that is thought to contribute to the increased accessibility of chromatin is altered by NAM. ChIP experiments were performed after treatment of 2963.1 cells with or without NAM for 30 min, followed by 15 min of R5020, when the histone acetylation levels were determined to be maximal in our previous work or EtOH control (1). The ChIP results demonstrate that the hormone-dependent increases in the promoter histone acetylation levels are not altered upon NAM treatment (Fig. (Fig.4D).4D). These restriction enzyme hypersensitivity and ChIP assays demonstrate that NAM does not inhibit transcription by inhibiting the remodeling of the MMTV promoter upon treatment with hormone.
The ChIP and restriction enzyme hypersensitivity assays performed thus far indicate that NAM does not suppress PR-mediated transcription by inhibiting the recruitment of the receptor and coactivators or chromatin remodeling of the promoter. In order to determine whether NAM inhibits the recruitment of the RNAPII machinery, ChIP assays were performed to monitor RNAPII and TATA-binding protein (TBP) of the TFIID complex recruitment after treatment of 2963.1 cells with or without NAM, followed by R5020 exposure. The ChIP results show that, unlike the various coactivators that are recruited to the promoter, both RNAPII and TBP are inhibited from associating with the transcription start site region (nucleosome A) (Fig. (Fig.5A).5A). This raises the intriguing possibility that NAM inhibits the communication between the activator (PR) and the RNAPII transcription machinery. We postulated that perhaps NAM inhibits the recruitment of the Mediator complex by PR which is required for activation of many RNAPII-regulated genes (5, 40, 57). To test this idea, ChIP experiments were performed with antibodies to two of the Mediator complex subunits MED1 and MED14. ChIP experiments performed with or without NAM in the presence or absence of hormone R5020 demonstrates that the hormone-dependent recruitment of the Mediator complex, as assessed through MED1 and MED14 subunits, is inhibited by NAM (Fig. (Fig.5B).5B). The exclusion of RNAPII, TBP, and Mediator from the MMTV promoter is not due to changes in expression levels upon NAM treatment, as determined by Western blot analysis of whole-cell extracts (Fig. (Fig.5C).5C). These results show that NAM inhibits PR-mediated transcription by preventing the transcription PIC to form, possibly due to the loss of recruitment of the Mediator complex.
The Mediator complex is thought to play a critical role in bridging the interaction between the transcription activator and the RNAPII machinery and for the transcription of many of the RNAPII transcribed genes (5, 15, 35, 57). However, there are some genes that do not require Mediator complex for transcription activation. For example, different GR-responsive genes have been shown to require different subunits of the Mediator complex for activation. For example, the GILZ gene in the osteosarcoma cell line U2OS was shown by siRNA knockdown experiments to require neither MED1 nor MED14 subunits of the Mediator complex (14). We postulated that if NAM affects transcription at the Mediator recruitment step, genes that do not require the Mediator complex for activation will not be affected by NAM. A screen of additional PR- and GR-inducible genes identified the HEF1 gene as Dex and R5020 inducible but unaffected by the NAM treatment (Fig. (Fig.6A6A).
We predicted that the NAM-resistant HEF1 gene does not require the Mediator complex for transcription activation. To test this idea, both MED1 and MED14 were knocked down simultaneously in the A1-2 cell line by the use of siRNA to circumvent the possibility that one of the subunits may compensate for the other during hormone-induced transcription. The level of knockdown achieved for both MED1 and MED14 protein levels after 48 h of siRNA treatment compared to treatment with scrambled nontargeted siRNA control were achieved as demonstrated on the Western blot of prepared whole-cell extracts in R5020 treated and untreated cells (Fig. (Fig.6B).6B). After siRNA transfection, the cells were treated with R5020 or vehicle (EtOH). RT-PCR analysis of the isolated RNA shows that the R5020-induced transcription of CEBP/β, SGK, and EZF genes is decreased after MED1 and MED14 knockdown (Fig. (Fig.6C).6C). In contrast, the HEF1 gene transcription is completely unaffected by the decreased expression of MED1 and MED14, as predicted (Fig. (Fig.6C).6C). These siRNA experiments demonstrate that the hormone-dependent activation of the HEF1 gene does not require the Mediator complex and is therefore unaffected by NAM treatment, suggesting that NAM inhibition of PR-mediated transcription occurs at the Mediator recruitment step.
Transcription is a complex process involving multiple enzymes and signaling pathways that intersect to regulate gene expression. In recent studies we have described the dynamic changes in histone acetylation and deacetylation that accompany progesterone-induced transcription. These changes were linked to the occupancy of canonical histone deacetylases HDAC1 and HADC3 at the promoter (1). In the present study we focused on the sirtuin family of proteins and the roles they may play in PR-mediated transcription by analyzing the impact of NAM on PR-mediated transcription in T47D human breast cancer cells (2963.1 and A1-2 cells) (Fig. (Fig.1).1). There are seven homologs of sirtuins, SIRT1 to SIRT7, with SIRT1 being by far the most studied and the most closely linked to gene regulation (23). SIRT1 deacetylates not only histones (H1, H3, and H4) (9, 27, 60) but also coactivators involved in NR-mediated transcription such as p300 (10) and PGC1α (46, 54). In addition, SIRT1 has been shown to interact with PPARγ and downregulate its activity (50). SIRT1 has also been shown to deacetylate AR, and treatment with SIRT1 inhibitor, NAM, induced ligand-dependent AR-mediated transcription (22). Further, while NAM is a known sirtuin inhibitor, it is also known to inhibit the PAR activities of PARP family of proteins (4, 17, 37). PARP1, the best-characterized member of the PARP family of proteins, has been closely linked to transcription via its ability to alter chromatin structure, as well as to regulation of Mediator complex activity (33, 48).
Our results demonstrate that just half an hour of treatment with NAM prior to hormone treatment suppressed hormone-dependent activation of PR- and GR-regulated genes in a dose-dependent manner. However, the siRNA knockdown of SIRT1 and PARP1 demonstrated that NAM inhibition of transcription was occurring through neither of these two factors (Fig. (Fig.2B).2B). This was surprising given the known roles of SIRT1 and PARP1 in the regulation of transcription and suggests the presence of a novel mechanism by which NAM (a commonly used SIRT1 and PARP1 inhibitor) inhibits transcription. The novel transcription regulating factor that is inhibited by NAM may involve other members of the sirtuin proteins. SIRT2, SIRT6, and SIRT7 can be found in the nucleus and play a role in the deacetylation of histones during mitosis, base excision repair, and RNA polymerase I transcription, respectively (20, 23, 44, 61). It is possible that in addition to their known roles, they are part of the PR-mediated transcription process that is disrupted by NAM. Although the functions of the other members of the sirtuins that are found outside of the nucleus, such as the mitochondria, are not well known (9, 43), as their roles in biology are discovered one could envision how they may also affect transcription through processes such as posttranslational modification of transcription factors as they shuttle in and out of the nucleus. The PARP family of proteins, as well as other NAD+ metabolizing factors such as cyclic-ADP-ribosyl cyclases and members of the NAD+ salvage pathway, are also possible targets of NAM, and the identification of such a factor is under investigation (7, 39). While SIRT1 and PARP1 may not be the target of NAM inhibition of PR-mediated transcription under our conditions, it still remains to be seen whether they have a role in regulating the gene expression of some genes such as EZF and SGK which show decreased expression upon SIRT1 and PARP1 knockdown (Fig. (Fig.2B2B).
To further dissect the mechanism by which NAM inhibits PR-mediated transcription, ChIP experiments were performed using the MMTV as a model PR-activated promoter. Because the NR target promoters are regulated in the context of chromatin, the receptor must recruit various coactivators and ATP-dependent chromatin remodeling complexes to render the promoter DNA more accessible (24, 34, 55). The ChIP experiments we performed demonstrate that the hormone-dependent association of PR is not compromised (Fig. (Fig.3C).3C). PR also retains the ability to recruit various coactivators and allow the promoter chromatin to be more accessible as determined by ChIP and restriction enzyme hypersensitivity assays in the presence of NAM (Fig. (Fig.3C3C and and4).4). Interestingly, our ChIP results show that inhibition of PR-mediated transcription occurs at the step of PIC assembly. This suggests that NAM inhibits the assembly of basal transcription machinery after chromatin remodeling of the promoter and that the communication between PR-mediated chromatin remodeling events and the assembly of basal transcription machinery is disrupted. In particular, NAM inhibition of PIC assembly appears to be occurring at the Mediator complex recruitment step. The Mediator complex has been shown to facilitate the recruitment of TFIID by binding cooperatively with it on promoters and to facilitate the recruitment of RNAPII through the interaction with CTD of RNAPII and transcription activators, thereby acting as a “bridge” to connect the transcription factor activities with the PIC machinery (5, 28, 35). The observations from the siRNA knockdown experiments of subunits of Mediator (MED1 and MED14) are consistent with a clear role for the Mediator complex in vital steps in the assembly of basal transcription machinery, and NAM most likely interferes with this step, thereby leading to the inhibition of PR-mediated transcription (Fig. (Fig.6).6). NAM could potentially directly or through other factors affect either or both the PR or the Mediator complex itself by altering posttranslational modifications or inducing conformational changes, which leads to the inhibition of Mediator complex recruitment to target genes. In considering these possibilities, it is also important to determine the fate of NAM and whether this leads to changes in nuclear NAD+ concentration within our experimental parameters since the enzymes involved in the NAD+ consumption (by proteins such as SIRT1 and PARP1) and regeneration by the NAD+ salvage pathway are present in the nucleus (39, 64). This is an important factor to consider when we further characterize the mechanism by which NAM inhibits the recruitment of the Mediator complex. NAM will most probably affect transcription by other steroid hormone receptors such as ER and AR in a similar manner as we have demonstrated for PR, given the similarities in the mechanism of transcription activation involving the Mediator complex. Depending on the exact nature of NAM inhibition of Mediator recruitment and whether this step takes place in the activation of genes by other families of transcription activators, it remains to be seen whether NAM has a similar effect on a wide range of transcription pathways.
We demonstrate here that transcription factor recruitment and promoter chromatin remodeling events can be uncoupled from transcription activation. This NAM-mediated inhibition is independent of SIRT1 and PARP1. This indicates that NAM blocks the regulation of an as-yet-unknown but critical mechanism or pathway required for the chromatin remodeling events to be translated into PIC formation and the activation of gene expression.
We thank Paul Wade, Xioling Li, Pratibha Hebbar, and Harriet Kinyamu (NIEHS) for critical reviews for the manuscript; Micheal Garabedian (NYU School of Medicine) for the MED1 and MED14 antibodies; and Grace Kissling (NIEHS) for the statistical analyses.
This research was supported by the Intramural Research Program of the NIH and NIEHS.
Published ahead of print on 22 October 2007.