Nuclear Respiratory Factor 1 Mediates the Transcription Initiation of Insulin-Degrading Enzyme in a TATA Box-Binding Protein-Independent Manner

CpG island promoters often lack canonical core promoter elements such as the TATA box, and have dispersed transcription initiation sites. Despite the prevalence of CpG islands associated with mammalian genes, the mechanism of transcription initiation from CpG island promoters remains to be clarified. Here we investigate the mechanism of transcription initiation of the CpG island-associated gene, insulin-degrading enzyme (IDE). IDE is ubiquitously expressed, and has dispersed transcription initiation sites. The IDE core promoter locates within a 32-bp region, which contains three CGGCG repeats and a nuclear respiratory factor 1 (NRF-1) binding motif. Sequential mutation analysis indicates that the NRF-1 binding motif is critical for IDE transcription initiation. The NRF-1 binding motif is functional, because NRF-1 binds to this motif in vivo and this motif is required for the regulation of IDE promoter activity by NRF-1. Furthermore, the NRF-1 binding site in the IDE promoter is conserved among different species, and dominant negative NRF-1 represses endogenous IDE expression. Finally, TATA-box binding protein (TBP) is not associated with the IDE promoter, and inactivation of TBP does not abolish IDE transcription, suggesting that TBP is not essential for IDE transcription initiation. Our studies indicate that NRF-1 mediates IDE transcription initiation in a TBP-independent manner, and provide insights into the potential mechanism of transcription initiation for other CpG island-associated genes.


Introduction
DNA methylation, an epigenetic modification that regulates chromatin structure and gene expression [1,2], occurs predominantly at cytosines of CpG dinucleotides in vertebrates [3,4]. There are three striking features of CpG sites in mammalian genomes. First, the majority (60%-90%) of CpG sites are methylated [5]. Second, mammalian genomes are depleted of CpG sites, a phenomenon called CG suppression. The observed rate of CpG sites is approximately one-fifth of the rate expected based on the GC content [6]. This rarity of CpG sites arises from the spontaneous mutation of methylated cytosines to thymidines by deamination, which converts CpG to TpG dinucleotides [7,8]. Thirdly, unmethylated CpG sites cluster to form CpG islands [9]. CpG islands, which have a high GC content and a high observedto-expected CpG ratio relative to the bulk genome, are often associated with mammalian gene promoters, including all housekeeping genes and some tissue-specific genes [10]. Actually, CpG islands serve as an important criterion for gene prediction [11,12,13]. In human, it was estimated that approximately 72% of genes are associated with CpG islands [14].
The core promoter is the minimal promoter region necessary to direct transcription initiation. Several core promoter elements, including the TATA box, the Inr (initiator) element, the downstream core promoter element (DPE), the motif ten element (MTE) and TFIIB recognition elements (BREs), have been identified for RNA polymerase II-transcribed genes, and they function individually or cooperatively in recruiting the transcription apparatus to the promoter region [15]. Promoters with these core promoter elements initiate transcription at a focused start site, which varies by only a few nucleotides. However, CpG island promoters often lack these canonical core promoter elements and have dispersed transcription initiation sites over 50-150 nucleotides [16]. Although genes with focused promoters are well studied, the mechanism for transcription apparatus recruitment to CpG island promoters remains to be clarified.
Insulin-degrading enzyme (IDE) is a ubiquitously expressed zinc metalloprotease that degrades several substrates [17], including insulin and b-amyloid (Ab) [18,19]. IDE is closely related to diabetes mellitus (DM) and Alzheimer's disease (AD), because insulin and Ab play critical roles in the pathogenesis of DM and AD, respectively. GK rats, which are a well-characterized model for type II DM, have IDE missense mutations, and exhibit impaired insulin and Ab degradation [20]. IDE knockout mice also show accumulation of endogenous Ab, hyperinsulinemia and glucose intolerance [21]. However, the exact role of IDE in DM pathogenesis seems a little elusive. A recent study indicates that the diabetic phenotype of IDE knockout mice is an emergent compensatory response to chronic hyperinsulinemia, but not a direct consequence of IDE deficiency [22]. In addition, human genetic studies suggest that IDE polymorphisms are associated with the pathogenesis of both type II DM [23] and AD [24]. These results imply that IDE dysfunction may cause DM or AD and underline the importance of characterizing the transcriptional regulation of IDE.
In this article, we focus on the mechanism of IDE transcription initiation. The mouse IDE promoter contains a CpG island and has dispersed transcription initiation sites. Promoter deletion analysis indicates that the IDE core promoter contains a nuclear respiratory factor 1 (NRF-1) binding motif which is essential for transcription initiation. We further demonstrated that NRF-1 binds to the IDE promoter and dominant negative NRF-1 represses IDE transcription. Finally, we showed that TBP is not essential for IDE transcription initiation.

Mapping the Core Promoter Region of Mouse IDE
Because the sequence of the mouse IDE promoter was incomplete in the NCBI genomic database ( Figure 1A), we cloned and sequenced the unknown region [Genbank: JN038396]. The mouse IDE promoter has a high CpG frequency and contains a CpG island approximately 1300 bp long ( Figure 1A). The transcription initiation sites of mouse IDE were determined using the 59-rapid amplification of cDNA ends (59-RACE) method. The IDE promoter possesses dispersed transcription initiation sites located within a 62-bp region ( Figure 1 A and B). IDE transcription starts with a purine at position +1 at a frequency of 91% and with a pyrimidine-purine dinucleotide at position 21,+1 at a frequency of 67%, which is consistent with previous studies showing that RNA polymerase II-mediated transcription is preferentially initiated at a pyrimidine-purine dinucleotide at position 21,+1 in mammals [16].
Searching the mouse IDE promoter revealed no TATA box, Inr or other canonical core promoter elements around the transcription initiation sites. To determine the region that is essential for IDE transcription initiation, we constructed luciferase reporter plasmids of the mouse IDE promoter with different truncations at the 59 or 39 terminus. For convenience, we set the first transcription initiation site that we identified ( Figure 1B) as position +1. Deleting the 59 terminus of the mouse IDE promoter from 22774 to 223 retained its transcriptional activity, while further deletion to +7 abolished its transcriptional activity ( Figure 2A). On the other hand, deleting the 39 terminus of the mouse IDE promoter from +139 to +9 retained its transcriptional activity, while further deletion to 26 abolished its transcriptional activity ( Figure 2B). These results indicate that the region between 223 and +9 of the mouse IDE promoter is essential for transcription initiation and behaves as the core promoter. Indeed, this region was capable of driving transcription initiation in both NIH-3T3 and HeLa cells ( Figure 2B).

The IDE Core Promoter Contains a Functional NRF-1 Binding Site that is Essential for Transcription Initiation
Analysis of the region between 223 and +9 of the mouse IDE promoter revealed three CGGCG repeats and a NRF-1 binding motif ( Figure 3). To identify which element is essential for IDE transcription initiation, we sequentially mutated this region and detected changes in its transcriptional activity. Mutations within the NRF-1 binding motif abolished transcription initiation, while mutations in other regions resulted in little or no decrease in the transcriptional activity (Figure 3), suggesting that the NRF-1 binding motif is critical for IDE transcription initiation. Next, we aimed to demonstrate whether this NRF-1 binding motif is functional. Mammalian cells are not sensitive to NRF-1 stimulation [25,26]. To determine the effect of NRF-1 on IDE transcription, we employed a dominant negative form of NRF-1which lacks the trans-activating domain [27]. In both NIH-3T3 and HeLa cells, dominant negative NRF-1 repressed IDE promoter activity dramatically ( Figure 4 A-C). When the NRF-1 binding motif was mutated, the effect of dominant negative NRF-1 on IDE transcription was abolished ( Figure 4 A-C). We subsequently tested whether NRF-1 binds to the predicted NRF-1 binding motif in vivo by chromatin immunoprecipitation assays (ChIP). Indeed, NRF-1 binds to this site and the promoter of a positive control gene, cytochrome c, but not to the promoter of a negative control gene, GAPDH ( Figure 4D). PCR products were cloned and sequenced to verify the specificity of the primers. Therefore, the NRF-1 binding motif in the mouse IDE core promoter is functional.

The NRF-1 Binding Site in the IDE Promoter is Conserved between Mouse and Human
The NRF-1 binding site in the IDE promoter is conserved among different species ( Figure 5A). The human IDE promoter contains a NRF-1 binding motif, which locates between 23 and +9 relative to the first transcription initiation site ( Figure 5B). Deleting the 59 terminus of the human IDE promoter from 2484 to 23 retained its transcriptional activity, while further deletion to +10 abolished its transcriptional activity ( Figure 5B). On the other hand, deleting the 39 terminus of the human IDE promoter from +173 to +9 retained its transcriptional activity, while further deletion to 24 abolished its transcriptional activity ( Figure 5B). Furthermore, mutation of the NRF-1 binding motif reduced the human IDE promoter activity by approximately 70% ( Figure 6B). These results indicate that the NRF-1 binding motif is critical for human IDE transcription initiation. ChIP experiments showed that NRF-1 binds to the human IDE promoter in HeLa cells ( Figure 5C). Furthermore, dominant negative NRF-1 repressed the human IDE promoter activity dramatically, which was abolished by the mutation of the NRF-1 binding motif ( Figure 5D). Therefore, the NRF-1 binding motif in the human IDE promoter is functional.
To test whether NRF-1 regulates endogenous IDE expression, HeLa cells were transiently transfected with wild-type or dominant negative NRF-1 expression plasmids. Both the mRNA and protein levels of IDE were not sensitive to wild-type NRF-1 stimulation ( Figure 6 A and B). However, enforced expression of dominant negative NRF-1 decreased IDE mRNA levels by 25%, and also reduced IDE protein levels, as compared with the vector control ( Figure 6 A and B).

TBP is not Essential for IDE Transcription Initiation
The IDE promoter is CpG-rich and contains no TATA box. Therefore, we wondered whether TBP is required for IDE transcription initiation. ChIP assays showed that TBP is not associated with the IDE promoter in both NIH-3T3 ( Figure 7A) and HeLa ( Figure 7B) cells. By contrast, TBP is recruited to the promoters of GAPDH and elongation factor 1 a 1 (EF1a1), which contain the TATA box. These results suggest that TBP is not essential for IDE transcription initiation.
To further address this question, we carried out in vitro transcription assays. The CMV promoter, which contains the TATA box, was used as a control. In vitro transcription occurred from all three DNA templates of the CMV, mouse IDE and human IDE promoters, which was abolished by a-amanitin ( Figure 7C). When the DNA template or the nuclear extract was omitted, no transcripts were detected ( Figure 7C). TBP was reported to be specifically inhibited by being heated at 47uC for 15 minutes, which was rescued by adding back of TBP [28]. In vitro transcription assays showed that transcription from the CMV promoter was inhibited by heat-inactivation of TBP, while transcription from the mouse or human IDE promoters was activated ( Figure 7D), indicating that TBP is not required for IDE transcription initiation.

Discussion
Despite the prevalence of CpG islands associated with mammalian promoters, the mechanism of transcription initiation from this type of promoters is not well understood. CpG island promoters often lack the TATA box or other common core promoter elements. It has been proposed that Sp1 may participate in the transcription initiation of CpG island promoters, because Sp1 binds to a consensus sequence of GGGCGG which is present in many CpG islands due to their high GC content [29]. NRF-1 is a transcription factor that activates nuclear genes required for respiration as well as genes regulating heme biosynthesis and mitochondrial DNA transcription and replication. NRF-1 has been reported to play an important role in the transcriptional regulation of several genes [30,31,32,33]. Interestingly, NRF-1 binding motifs are frequently found around the transcription initiation sites of genes [27,34], typically those without the TATA box [35], suggesting that NRF-1 may function in regulating the transcription initiation of TATA box-less genes.
Here we studied the mechanism of transcription initiation of the ubiquitously expressed gene, IDE, which has a CpG island promoter and dispersed transcription initiation sites. We mapped the core promoter region of IDE, and showed that the IDE core promoter contains a NRF-1 binding motif that is essential for transcription initiation. We further demonstrated that NRF-1 binds to this motif and dominant negative NRF-1 represses IDE transcription. Our studies suggest that, besides Sp1, NRF-1 may be another important transcription factor that regulates the transcription initiation from CpG island promoters. These results are suggestive for the mechanism of transcription initiation from other CpG island promoters. Since CpG island promoters often lack canonical core promoter elements, some transcription factors, such as Sp1 and NRF-1, may function in recruiting the transcription apparatus to the promoter region without the assistance of other core promoter elements. In this hypothesis, the transcription apparatus is not ''fixed'' to the promoter through its own interactions to the promoter, which may explain the fact that CpG island promoters often have dispersed transcription initiation sites.
TBP was originally regarded as a universal factor that participated in the transcription initiation of genes transcribed by all three RNA polymerases in eukaryotes [36]. However, emerging studies have indicated the existence of TBP-independent transcription by RNA polymerase II [37,38]. In addition, A TBP-free TAF II -containing complex was reported to initiate

NRF-1 Regulates IDE Transcription Initiation
PLoS ONE | www.plosone.org transcription from both TATA box-free and TATA boxcontaining promoters [39]. The IDE promoter contains no TATA box, and is free of TBP binding. Furthermore, inactivation of TBP does not block IDE transcription. These results suggest that TBP is not required for IDE transcription initiation. It is fascinating to demonstrate whether TBP is essential for the transcription initiation from other CpG island promoters which do not contain the TATA box.
IDE is a zinc metalloprotease that typically degrades insulin and Ab, and is associated with both type II DM and AD. However, the mechanism of the transcriptional regulation of IDE has not been studied in detail. Here, we showed that NRF-1 acts as a central transcription factor that regulates IDE transcription. Thus, NRF-1 dysfunction may result in a down-regulation of IDE expression and DM or AD pathogenesis. Indeed, human genetic studies suggest that NRF-1 polymorphisms are associated with the pathogenesis of type II DM [40,41]. The transcriptional regulation of IDE by NRF-1 reported here provides a mechanism that links NRF-1 to type II DM. Although IDE is predominantly distributed in the cytosol, an isoform of IDE generated by alternative translation initiation is targeted to mitochondria [42]. NRF-1 is an important transcription factor that mediates nuclear-mitochondrial interactions. The transcriptional regulation of IDE by NRF-1 may be of significance to the mitochondrial function.
In conclusion, our studies indicate that NRF-1 mediates IDE transcription initiation in a TBP-independent manner, and provide insights into the potential mechanism of the transcription initiation from other CpG island promoters. NRF-1 has been reported to regulate the transcription of a variety of genes. It is fascinating to investigate whether NRF-1 is essential for the transcription initiation of these genes. In addition, NRF-1 interacts with the peroxisome proliferator-activated receptor c coactivator 1 (PGC-1) family of proteins, including PGC-1a [43], PGC-1b [44] and PGC-1-related coactivator (PRC) [45]. It remains to be clarified whether NRF-1 regulates IDE transcription initiation through these coactivators.

Plasmid Constructs
Genome DNA from C57/BL6 mice or HeLa cells was used as the template to construct the promoter reporter plasmids. Different truncations of the mouse and human IDE promoter were cloned into the Xho I and Hind III restriction sites of the pGL3-basic vector (Promega, Madison, WI). The primers used during the plasmid construction are shown in Table S1 and S2. The NRF-1 binding sites in the mouse and human IDE promoter were mutated from TGGGCATGCGCA and AGAG-CATGCGCA to TGGGCAGCGACA and AGATAGGTAGCA, respectively. The expression plasmids for wild-type NRF-1and dominant negative NRF-1 which contains 1-304 amino acids of full-length NRF-1 [27] were constructed by cloning the corresponding sequences into the Bam HI and Xba I restriction sites of the pcDNA 3.1 myc-His A vector (Invitrogen, Carlsbad, CA).

Cell Culture and Cell Transfection
NIH-3T3 and HeLa cells were maintained in Dulbecco's modified Eagle's medium (Thermo Scientific HyClone, Logan, UT) containing 10% fetal bovine serum (Thermo Scientific HyClone). HeLa cells were seeded into 6-well plates at a density of 4610 5 cells per well in antibiotic-free medium the day before transfection. Each well of cells were transiently transfected with 4 mg of pcDNA 3.1, wild-type or dominant negative NRF-1 Figure 7. TBP is not required for IDE transcription initiation. (A) and (B) TBP is not associated with the IDE promoter. Binding of TBP and RNA polymerase II to the IDE promoter in NIH-3T3 (A) and HeLa (B) cells was tested by ChIP assays. Promoters of GAPDH and EF1a1 were used as positive controls for TBP and RNA polymerase II binding. Negative controls were also included. Data are represented as relative enrichment to the input. (C) In vitro transcription assays. DNA template for the CMV, mouse IDE or human IDE promoter was incubated with HeLa nuclear extracts and ribonucleotides. The resulting transcripts were purified, digested with DNase I to eliminate the DNA template and detected by RT-PCR. An internal control RNA was included to indicate the purification efficiency of different samples. Complete elimination of the DNA template was confirmed by PCR. Transcription from all the three DNA templates proceeded only when both the DNA template and HeLa nuclear extracts existed, and was inhibited by 1 mg/mL of a-amanitin. (D) Heat-inactivation of TBP does not block IDE transcription. HeLa nuclear extracts were heated at 47uC for 15 min before in vitro transcription assays. doi:10.1371/journal.pone.0042035.g007 expression plasmids using Lipofectamine TM 2000 (Invitrogen). Twenty-four hours later, cells were collected, and analyzed by quantitative real-time RT-PCR or Western immunoblotting.

5'-RACE
59-RACE was carried out using the SMART TM RACE cDNA Amplification Kit (Clontech, Shiga, Japan) according to the manufacturer's instructions, to determine the transcription initiation sites of mouse IDE.59-RACE-ready cDNA was synthesized from total RNA from NIH-3T3 cells. The 59-flanking region of IDE was amplified using the universal primer A mix (Clontech) and an IDEspecific primer: 59-TCTTCAACGTGCAATAACCCT-39. The resulting PCR products were used as a template for a nested PCR reaction using the nested universal primer A (Clontech) and a nested IDE-specific primer: 59-TGTGGTGGGATCGCTGATGA-GAAG-39. The nested PCR products were then cloned into the pMD19-T vector (Takara Bio Inc., Shiga, Japan), and a total of 33 positive clones were sequenced.

Dual-luciferase Reporter Assays
NIH-3T3 and HeLa cells were seeded into 24-well plates at a density of 8610 4 cells per well in antibiotic-free culture medium the day before transfection. Each well of cells were transiently cotransfected with 0.8 mg of IDE reporter plasmids and 8 ng of Renilla reporter plasmid (pCMV-RL, Promega) as an internal control using Lipofectamine TM 2000. Alternatively, cells were cotransfected with 0.4 mg of IDE reporter plasmids, 4 ng of pCMV-RL plasmid and 0.4 mg of pcDNA 3.1 or NRF-1 expressing plasmids. Twenty-four hours later, cells were lysed and the luciferase acitvity was detected using the Dual-Luciferase Reporter Assay System (Promega). Firefly luminescence signal was normalized based on the Renilla luminescence signal.

RNA Isolation and Quantitative Real-time RT-PCR
RNA was isolated using Trizol reagent (Invitrogen), and treated with DNase I (Promega) before cDNA synthesis. cDNA was synthesized with 2 mg of total RNA with anchored oligo (dT) 20 primers using the Transcript First-Strand cDNA Synthesis Kit (Transgene Biotechnology Inc., Beijing, China). Quantitative realtime PCR was performed using the UltraSYBR mixture (Cwbiotech, Beijing, China) in the Stratagene Mx3000P TM Real-Time PCR System (Agilent Technologies, La Jolla, CA). PCR primers for human IDE were forward: 59-AAAGACAAAGAGAGGCCACGGGG-39 and reverse: 59-TGGCAACCCGGACATTTTCTGGTC-39. PCR primers for human GAPDH were forward: 59-AC-CGTCAAGGCTGAGAACGGGA- 39 and reverse: 59-CCTGCAAATGAGCCCCAGCCTT-39. Both the IDE and GAPDH primers are intron-spanning. Melting curves were performed on the products, which confirmed the specificity of the primers. The thermal cycling conditions were as follows: initial denaturation at 95uC for 10 min, followed by 40 cycles of denaturing at 95uC for 15 s, annealing at 60uC for 30 s and extension at 72uC for 30 s.

Western Immunoblotting
Cells were collected and lysed with the cell lysis buffer (Beyotime, Shanghai, China). Protein concentrations were determined using the BCA method. Whole cell extracts were boiled for 4 min, resolved by 10% SDS-PAGE and electrophoretically transferred to polyvinylidene difluoride membranes (Millipore, Schwalbach, Germany). The membranes were blocked with 5% non-fat dry milk in Tris-buffered saline containing 0.05% Tween 20 at room temperature for 1 h, and then incubated with mouse monoclonal anti-IDE (1:1000, Abcam), mouse monoclonal anti-His tag (1:2000, Cwbiotech) or rabbit polyclonal anti-GAPDH (1:1000, Cwbiotech) antibodies at 4uC overnight, followed by an incubation with the appropriate horseradish peroxidase-conjugated secondary antibodies (1:10000, Cwbiotech) at room temperature for 1 h. The chemiluminescence reaction was performed using ECL reagent (Thermo Scientific).

ChIP
ChIP was performed using the Magnetic Chromatin Immunoprecipitation Kit (Active Motif, Carlsbad, CA) as described in the manufacturer's protocol. Briefly, NIH-3T3 or HeLa cells were washed once with phosphate buffered saline, and fixed in culture medium containing 1% formaldehyde at room temperature for 10 min. The cells were then collected and lysed to release the nucleus. The nucleus was digested with the enzyme mix at 37uC for 10 min to shear the chromatin into small segments. The sheared chromatin was then immunoprecipitated with 2 mg of NRF-1 antibodies (Abcam), 2 mg of TBP antibodies (Abcam), 2 mg of RNA polymerase II antibodies (Active Motif), or a negative control IgG at 4uC for 4 h. The pulled-down chromatin was washed, reverse-crosslinked and purified. Semi-quantitative or quantitative real-time PCR was performed using the primers shown in Table S3.

In vitro Transcription
In vitro transcription was performed using the HeLaScribeH Nuclear Extract in vitro Transcription System (Promega) with some modifications. DNA template of the mouse IDE promoter was amplified from mouse genomes using primers forward: 59-TGCCGAGACGACGACCCACC-39 and reverse: 59-TGA-GAAGCGGTTTCCCACGAC-39. DNA template of the human IDE promoter was amplified from human genomes using primers forward: 59-CTCACAGTCAGACACACGTCGCCACC -39 and reverse: 59-CCACACGGTCCTGGAAACTCAGTGCC-39. DNA template of the CMV promoter was amplified from a mouse peroxisome proliferator-activated receptor a expression plasmid drived by the CMV promoter using primers forward: 59-TGTTGGAGGTCGCTGAGTAGTGC-39 and reverse: 59-AGGCGGGTTGTTGCTGGTCT-39. The DNA template (100 ng) was incubated with HeLa nuclear extracts and four types of ribonucleotides at 37uC for 30 minutes. The resulting transcripts were purified, digested with DNase I (Promega) at 37uC for 1 h to eliminate the DNA template, and then used as RT-PCR template. An internal control RNA was included to indicate the purification efficiency of different samples. RT-PCR was performed using the PrimeScriptH One Step RT-PCR Kit Ver.2 (Takara), using the following primers-forward: 59-GCACCTTGCGCTCCATCCTCG- 39 and reverse: 59-CCGCTGCACTTGTGGCGTTCTC-39 for mouse IDE, forward: 59-CCAGCACCTTCCGCTCAGTCC-39 and reverse: 59-GGTGACTCTGTCCGCCCCTGC-39 for human IDE, and forward: 59-GACCTGGAAAGTCCCTTATCT-39 and reverse: 59-AGCCCTTACAGCCTTCACAT-39 for CMV. Elimination of DNA template was confirmed by PCR. To confirm that IDE is transcribed by RNA polymerase II, 1 mg/mL of a-amanitin was included in the in vitro transcription system. To determine the role of TBP in transcription initiation, HeLa nuclear extracts were heated at 47uC for 15 min before the in vitro transcription assay.

Statistical Analysis
Results are represented as means 6 S.E.M from three independent experiments. Statistical significance was determined by one-way ANOVA, followed by the post-hoc Tukey multiple comparison test.