Oct-1 cooperates with the TATA binding initiation complex to control rapid transcription of human iNOS

Abstract

Expression of the human inducible nitric oxide synthase (hiNOS) is generally undetectable in resting cells, but stimulation by a variety of signals including cytokines induces transcription in most cell types. The tight transcriptional regulation of the enzyme is a complex mechanism many aspects of which remain unknown. Here, we describe an octamer (Oct) element in hiNOS proximal promoter, located close to the TATA box. This site constitutively binds Oct-1 and its deletion abrogates cytokine-induced transcription, showing that it is indispensable though not sufficient for transcription. Increasing the distance between Oct and the TATA box by inserting inert DNA sequence inhibits transcription, and footprinting of this region shows no other protein binding in resting cells, suggesting an interaction between the two complexes. Chromatin immunoprecipitation assays detect the presence of Oct-1, RNA polymerase II and trimethyl K4 histone H3 on the proximal promoter in resting cells, confirming that the gene is primed for transcription before stimulation. RT-PCR of various fragments along the hiNOS gene shows that transcription is initiated in resting cells and this is inhibited by interference with Oct-1 binding to the proximal site of the promoter. We propose that, through interaction with the initiation complex, Oct-1 regulates hiNOS transcription by priming the gene for the rapid response required in an immune response.

Introduction

The high levels of nitric oxide (NO) produced by the inducible nitric oxide synthase (iNOS, NOSII, NOS2) have a variety of effects on cell homeostasis and pathology. The potentially harmful, cytotoxic effects of high levels of NO make a tight regulation of iNOS essential, but the paramount role of iNOS and NO in the defence mechanisms against infection implies that the enzyme must be rapidly available in response to immune signalling.

The iNOS gene is typically not expressed in resting cells, but stimulation with infectious agents, microbial products such as lipopolysaccharides and immune mediators such as various cytokines, induces expression of the enzyme in a variety of cells. It is widely thought that, although regulated at every step of its synthesis, the main control of iNOS expression is transcriptional and a combination of cytokines including Il1-β, TNF-α and IFN-γ is required for maximal induction. However, with a promoter spanning at least 16 kb in the human enzyme (hiNOS) [1], a detailed mechanism of transcriptional control is difficult to unravel. Most of the cis-acting DNA elements responsive to the main transcription factors implicated in hiNOS transcriptional regulation have been localised to the region between −5 and −6 kb of the promoter. This region harbours at least three NFκB [2], two STAT-1 [3] and two AP-1 [4] sites that have been shown to play an important role in hiNOS induction.

The mouse iNOS, on the other hand, has a much simpler promoter of only 1.7 kb that has helped to understand iNOS regulation to a certain extent [5]. Although similar DNA elements and the corresponding transcription factors such as NFκB [5, 6] have been described, a marked difference in inducibility of the murine and human enzymes is notorious. This transcriptional species distinction [7] has been at least partly explained by dramatic differences in the DNA methylation patterns of the promoters [8], highlighting the need for a thorough understanding of hiNOS-specific regulation.

One of the transcription factors identified in mouse iNOS is Oct-1 [9], which is permanently bound to a proximal Oct site and is essential but not sufficient for transcriptional activity [10]. Oct-1 was shown to contribute to a DNA–protein complex with the p50 subunit of NFκB to modulate iNOS transcription [11]. However, the only report of Oct participation in the regulation of human iNOS describes an Oct site in the far upstream promoter (10 kb) that is bound by Oct-1 and plays a role in modulating transcriptional activation by NFκB [12].

Though transcriptional mechanisms are paramount for the control of gene expression, it has become apparent in recent years that all the steps of transcription are subject to regulation. Thus, the dominant view has been that the assembly of the pre-initiation complex (PIC), leading to the recruitment of RNA polymerase II (Pol II) is the rate-limiting step of transcription, and therefore the principal step at which synthesis of a complete mRNA transcript is regulated. However, genome-wide studies in a variety of species from yeast to man indicated that preassembled PICs can be found on many inactive genes [13]. Furthermore, since the determination of the chromatin structure and the unravelling of its regulatory functions, its fundamental role in transcription has become increasingly clear. Chromatin has a basic repeating unit, the nucleosome, composed of 147 bp of DNA wrapped 1.65 times around an octamer of core histone proteins comprising two molecules each of H2A, H2B, H3 and H4. DNA methylation and posttranslational modifications of histones determine the state of accessibility of the DNA to the transcription machinery. A multitude of modifications have been described that have a variety of functions, some of which are associated with heterochromatin while others characterise permissive euchromatin. Thus, for example, acetylation of histone 3 on lysine 9 and 14 (H3K9,4Ac) is associated with active transcription, while trimethylation of lysine 27 of histone 3 is generally correlated to transcriptional repression [14, 15]. Of these modifications, the trimethylation of lysine 4 in histone 3 (H3K4me3) is present exclusively at active genes [16] and has become a widely accepted predictor of transcription [17]. Mapping histone modifications also showed epigenetic marks associated with transcriptional activation on inactive genes [17]. These studies imply that recruitment of Pol II is not always the rate-limiting step of transcriptional regulation and introduce the elongation step as an important checkpoint of this process. Indeed, this type of regulation is thought to govern as much as 30% of the genome, particularly genes of rapid response such as immediate early genes, heat shock, DNA repair proteins, immune response, etc. [13, 17–19].

The ubiquitous distribution of iNOS and its paramount function in many physiological processes as well as in disease make it vital to understand the precise mechanism of expressional control of the human enzyme. This is particularly important in the efforts to modulate the NO system for disease therapy. We aimed at defining in more detail the transcriptional mechanism that controls expression of the human iNOS. Here, we describe for the first time a functional Oct binding site in the proximal hiNOS promoter that is indispensable for transcription. Located close to the TATA box, it is constitutively bound by Oct-1 which participates in the formation of the initiation complex. This complex initiates transcription of the gene, which is completed upon binding of inducible transcription factors such as NFκB, thus providing a highly controlled rapid expression of hiNOS.

Materials and methods

Cell culture and cytokine stimulation

HCT-8R and HCT-116 cells were maintained in D-MEM supplemented with 10% foetal calf serum and 2% glutamine. A mix of recombinant human cytokines (CM): IL-1β (3 ng/ml) (Sanofi, Labège, France), IFN-γ (200 U/ml) (Roussel-Uclaf, Romainville, France) and TNF-α (75 ng/ml) (Dainippon, Osaka, Japan) were added to the same medium for the appropriate period of time.

Plasmids

A pXP2 plasmid containing 7 kb of the hNOSII promoter regulating luciferase as a reporter gene [1] was used to examine transcriptional activity. A pbluescript plasmid containing human hepatocyte NOSII cDNA was used for northern blotting. The expression vector of the non-phosphorylatable IkB-α was obtained from S. Chouaib [20]. NFκB-driven transcription was studied using a (pNFκB-luc) reporter gene driven by four tandem copies of the k enhancer (B4) in a pUC vector [21].

RNA extraction northern blot and RT-PCR

Cells were stimulated with CM for the appropriate time and RNA was extracted using the Bioline mini kit (London, UK). Northern blot was performed as described previously [22]. A specific 600-bp fragment of the human NOSII cDNA (46–644) was used for probing of NOSII mRNA. RT was performed with 2 μg of RNA and a MuLV reverse transcriptase, followed by 25–40 cycles of PCR with specific primers.

Cell transfection and reporter gene assay

Cells were plated on six-well plates at 1 x 10⁶ cells per cm². After 24 h, the medium was removed and cells were transfected with the SuperFect transfection reagent (Qiagen, Courtaboeuf, France). Transfection efficiency was normalised by co-transfection with a TK-renilla luciferase plasmid. Cells were transfected overnight and stimulated with CM for the indicated time. The dual luciferase reporter assay kit from Promega (Madison, WI, USA) was used to measure transcriptional activity in a luminometer Lumat LB9507 (EG&G Berthold, Bad Wildbad, Germany).

Electrophoretic mobility shift assays (EMSA)

Assays were performed as described previously [23]. The oligonucleotides corresponding to the different Octamer elements found in the NOSII promoter (Table 1) were hybridised with the corresponding anti-sense oligonucleotides, labelled, and purified on Sephadex 50 columns (Boehringer Mannheim). The consensus Oct and NFκB sites used for EMSAs were obtained from Promega, and the specific antibodies to NFκB subunits and Oct-1 were purchased from Santa Cruz Biotechnology (CA, USA).

Table 1.

Oct-like sites in the hiNOS promoter

Octamer-like sequences in the hiNOS promoter
Sequence	Location		Name
ATGAAAAT	−7,132	−7,125	U-Oct
ATAGCAAAT	−7,005 −1,852	−6,997 −1,844	D-Oct
ATGCATAT	−1,088	−1,081	M-Oct
ATGAAAT	−1,046 −726 −721	−1,040 −720 −715	R-Oct
ATGCAAAA	−64	−57	P-Oct

Site-directed mutagenesis

Mutagenesis of the hiNOS promoter was performed using the QuikChange Site-directed Mutagenesis kit (Stratagene, La Jolla, CA, USA). The resulting plasmids were verified by enzymatic digestion and confirmed by sequencing. The oligonucleotides used were: for deletion of the P-Oct site:

5′-GAGACCTTTATGAATGCACTCTCTGG-3′ and the corresponding anti-sense.

For insertion of DNA sequences (in bold) between P-Oct and TATA box:

5 bp inserted at −48: F 5′-AAGCTTATAAATACTTCTTGG-3′ R 5′-CTCACCCCATGCCATCC-3′.

10 bp inserted at −30: F 5′-GGGAGCTCCCGGATGGCATGGGG-3′.

R 5′-AGAGAGTTGTTTTTGCATAAAGG-3′.

In vivo footprinting

The genomic footprinting analysis was performed by the DMS methylation/ligation-mediated PCR method as previously described [9]. The primers used for the footprinting of the hiNOS gene are as follows:

1S

+33 5′-GGAGCCTCAGTTTTCGA-3′ +17.

2S

+14 5′-CGCTACAAAGTTATGAACACACTGGC-3′−12.

3S

+14 5′-CGCTACAAAGTTATGAACACACTGGCAGCCA-3′ −17.

1AS

−276 5′-TCTTAGCAGCCACCCT-3′ −261.

2AS

−231 5′-CCAGAAAGAGGTGGGTTGGG-3′ −209.

3AS

−2315′-CCAGAAAGAGGTGGGTTGGGTGAAGAGGC-3′ −203.

Double-stranded linker molecule:

5′-GCGGTGACCCGGGAGATCTGAATTC-3′ 3′-CTAGACTTAAG-5′.

Nuclear run-on

Assessment of the rate of transcription of the hiNOS gene was carried out using the procedure described by Srivastava [24]. Cells were cultured with or without cytokines for 2, 4 or 24 h. Then, 5 μg of each probe and of actin cDNA were immobilized onto zeta probe membranes (Biorad). Nuclei from non-stimulated or stimulated HCT-8R cells were collected and in vitro transcription was performed. The transcripts were hybridised onto probed membranes, washed extensively and exposed to Kodak Biomax films overnight.

The primers used to generate the probes for these experiments are indicated below:

Probe 1: F 5′GCAGAGAACTCAGCCTCATTC3′, R 5′GGTAAGGACAGTCAAACCAG3′.

Probe 2: F 5′AATGTGGAGAAAGCCCC3′, R 5′CATCTGGAGGGGTAGGC3′.

Probe 3: F 5′GAGGAAGTGGCAGGAGAAT3′, R 5′CAGCATACAGGCAAAGAGCA3′.

Chromatin immunoprecipitation (ChIP)

Chromatin immunoprecipitation experiments (ChIP) were performed as described in [25] with minor modifications. Briefly, after crosslinking, the cells were resuspended in 1 ml of SDS lysis buffer (Millipore) and sonicated under conditions that reduced the length of the DNA to 200–1,000 bp. Chromatin extracts were diluted 10× in ChIP Dilution buffer (Millipore) and pre-cleared for 1–2 h with protein G Agarose coated with Salmon sperm DNA (Millipore). Extracts were incubated overnight with 10 μg of either anti-IgG (Rabbit control IgG-ChIP grade, ab 46540; Abcam) or specific antibodies (anti-Pol II sc899X N20, anti-Oct1 sc232X; Santa Cruz Biotechnology, or anti-trimethyl Histone H3 (Lys4); Millipore, CS200580). Immune complexes were captured for 2 h with protein G Agarose (Millipore) and washed once with low salt, high salt and LiCl wash buffers (Millipore) and twice with TE. Immune complexes were then eluted and crosslinking reversed overnight. DNA was recovered by phenol extraction and ethanol precipitation. Transcription factor binding was analysed by PCR of the specific fragments for 35 cycles resolved in 1.5% agarose gels. The primers used were as follows:

P1F 5′-AACAGCAAGATCAGGTCACCCAC-3′ P1R 5′-ATGAACACACTGGCAGCCAAGAAG-3′.

P2F 5′-AAGGCACAGGTCTCTTCCTGGTTT-3′ P2R 5′-AATGAAGGCAACTCACCTTGCAGC-3′.

P6F 5′-TCTGTAGGAAGTGGGCAGGAGAAT-3′ P6R 5′ACCTGGAAACTATGGAGCACAGCA3′.

Results

The hiNOS promoter contains a proximal Oct site that binds Oct-1

We investigated the participation of Oct-1 in the regulation of hiNOS transcription in the human colon cancer cell line HCT-8R. We have shown previously that this cell line expresses iNOS in response to stimulation with a cytokine mix (CM) containing IL1-β, IFN-γ and TNF-α [22]. Expression of Oct-1 was examined, using a consensus oct oligonucleotide for electrophoretic mobility shift assays (EMSA). Figure 1a shows Oct-1 DNA binding activity as identified by a specific antibody. The binding of Oct-1 was verified in another human colon cancer cell line HCT-116. Interestingly, Oct-1 binding is present in resting un-stimulated cells and only a slight increase is observed when HCT-8R cells are stimulated with CM.

Fig. 1.

Constitutive presence of Oct-1 in human colon cancer cell lines. a EMSA of nuclear proteins (2 μg) from control (CT) or CM-stimulated (2 h) cells (CM), using a consensus Oct probe. Unlabelled probe (UP) was added in 50× excess as competitor and an anti-Oct-1 antibody (Ab) was used to identify the specific complex. b The Oct sites found in the promoter, described in Table 1, were used as probes in EMSA with nuclear extracts (2 μg) from CM-stimulated (+) and non-stimulated (−) cells. c Labelled proximal Oct site (P-Oct) was incubated with 2 μg nuclear extracts from control (CT) or stimulated (CM) cells. Specificity was confirmed with 50-fold excess of the un-labelled (UP) probe and the protein complex was identified using a specific anti-Oct-1 antibody (Ab). d Chromatin immunoprecipitation with specific anti-Oct-1 antibodies was performed on control and stimulated (CM) cells. Primers specific for the proximal promoter (P1) were used for the PCR which was run on 1.5% agarose gels

Once an active Oct-1 was demonstrated, the proximal 7 kb of hiNOS promoter were analysed searching for potential Oct binding sequences (TRANSFAC [26]). Several putative Oct sites were detected all along the promoter which are summarised in Table 1. Double-stranded oligonucleotides corresponding to these sequences were synthesised and probed in EMSA to investigate the binding of transcription factors (Fig. 1b). Most of these sequences are inactive showing no protein binding. A distal Oct element (D-Oct in Table 1) forms two protein complexes, but their fast migration indicates that they are smaller polypeptides possibly corresponding to smaller members of the POU family such as Pit-1 or Brn, which have been reported to bind the hiNOS promoter [27]. The proximal Oct element (P-Oct), located at −57 to −64 is the only sequence recognized by a complex of nuclear proteins migrating in a similar pattern to that obtained with the consensus Oct probe. Closer analysis of this sequence revealed specific binding of Oct-1 in both cell lines tested (Fig. 1c). The binding of Oct-1 to the hiNOS proximal promoter in vivo was confirmed by chromatin immunoprecipitation (ChIP) using a specific antibody (Fig. 1d).

The proximal Oct element in the hiNOS promoter is essential but not sufficient for transcription

To test the role of this new Oct site, we used a luciferase reporter gene containing 7 kb of the hiNOS promoter [1]. Deletion of the P-Oct site from this construct abrogates transcription in both cell lines (Fig. 2a), demonstrating a fundamental role of this Oct site for hiNOS transcription.

Fig. 2.

Role of P-Oct in NOSII transcription. a The proximal Oct element (P-Oct) was deleted from the 7 kb hiNOS promoter and cells were transiently transfected with the full length (P7) or the deleted (OD) plasmids. Luciferase activity was measured in control cells (CT) or after 6 h CM stimulation (CM) and activity is reported as fold induction of non-stimulated cells for each plasmid. b Northern blot showing hiNOS mRNA accumulation in control (CT) cells and after 6 h cytokine stimulation (CM)

However, Fig. 1c shows binding of Oct-1 to the P-Oct site in the absence of cytokine stimulation when iNOS is not transcribed (Fig. 2a). The lack of iNOS mRNA accumulation in the absence of stimulation was confirmed by a northern blot (Fig. 2b). This demonstrates that Oct-1 binds to the hiNOS promoter without inducing transcriptional activity and therefore is not sufficient to drive expression of the enzyme.

Cytokines trigger intracellular signals resulting in the activation of transcription factors that control gene expression. A major pathway consists in the phosphorylation and degradation of IκB, which allows nuclear translocation of the NFκB complexes to activate iNOS transcription [28]. Thus, in contrast to Oct-1, which is constitutively bound to the promoter, NFκB is only detectable upon stimulation as shown by EMSA (Fig. 3a). NFκB binding leads to transcriptional activity as demonstrated by a luciferase reporter gene driven by four tandem NFκB sites [21] (Fig. 3b). Several NFκB sites have been described in the hiNOS promoter which are involved in transcriptional regulation [2, 3]. So, in order to study their contribution to hiNOS regulation, the 7-kb hiNOS reporter gene was co-transfected with an expression vector containing a dominant negative non-phosphorylatable mutant of IκB (IκBM). As this protein is resistant to phosphorylation, it prevents the nuclear translocation of NFκB. Expression of IκBM abolishes transcriptional activity of the promoter (Fig. 3c) to the same degree as the deletion of P-Oct (Fig. 2a). These experiments show that both Oct-1 and NFκB are required to induce hiNOS transcription and, while NFκB activity is induced by extracellular stimuli, Oct-1 sits constitutively on the promoter.

Fig. 3.

NFκB activation by CM. a EMSA of nuclear proteins (2 μg) from control cells (CT) or cells stimulated with CM for 2 h, using a consensus NFκB probe. Components of the induced complexes are identified using specific antibodies as p50/p65 (P1) and p50/p50 (P2) by the formation of supershifts (S). NS non-specific band. b Transcriptional activity of NFκB was examined using a luciferase construct containing four tandem NFκB sites. Luciferase activity in transiently transfected control (CT) or 6 h stimulated (CM) cells is reported as fold induction. c Cells were transiently transfected with the 7-kb hiNOS reporter gene together with an empty vector (vector) or an expression vector for a non-phosphorylatable IκB (IκBM) and activity was determined in control (CT) or 6 h stimulated (CM) cells

Oct-1 binding to P-Oct is implicated in the transcription initiation complex

Using the luciferase reporter gene containing 7 kb of the hiNOS promoter described above, we showed in previous work that two major deletions between the distant enhancer and P-Oct, spanning 3 and 4.6 kb, have no effect on the transcriptional activity of the hiNOS promoter in response to cytokines [22]. The result of these modifications is that the distance between the enhancer (where the NFκB sites described in the literature are located) and the P-Oct is decreased from 5 to 2 kb and to as little as 600 bp. This implies that the distance between the enhancer and the P-Oct is not relevant for transcriptional regulation.

So, the question emerged whether the distance between P-Oct and the transcription start site (TSS) is important. Examining the iNOS proximal promoter in various species revealed that the sequence between Oct and the TATA box is identical in human, chimpanzee and gorilla, and that the distance between these sites is very conserved throughout species such as mouse, rat and cow (Table 2). Thus, we hypothesised that, due to its constitutive binding and proximity to the TATA box, P-Oct might be implicated in the initiation complex. In order to examine this possibility, several constructs were designed introducing small inert DNA sequences between the TATA box and P-Oct, to increase the separation between them by half a turn of DNA (5 bp) at a time (Fig. 4a). Since all the deleted constructs described above have equivalent transcriptional activity, the smaller H3 construct was used for this strategy (activity of the full length promoter P7 is shown for comparison). Increasing the P-Oct-TATA distance by 5 bp already disrupts transcriptional activity, indicating that Oct-1 is important for the initiation complex formation. This is not an orientation effect because introducing 10 bp between the sites, which preserves the binding positions on the DNA shows no activity either (Fig. 4a).

Table 2.

Sequences of the Oct-TATA promoter regions throughout species

Species	Oct-TATA box
Human	ATGCAAAAACAACTCTCTGGATGGCATGGGGTGAGTATAAA
Gorilla	ATGCAAAAACAACTCTCTGGATGGCATGGGGTGAGTATAAA
Mouse	ATGCAAAATAGCTCTGCAGAGCCTGGAGGGGTATAAA
Rat	ATGCAAAACAGCTCTGCAGAGCGTGGATGGGTATAAA
Cow	ATGCAAAACGGCTCCCTGGATGCCACAGGTTGGGTATAAA

Fig. 4.

Interaction between P-Oct and the TATA box. a Schematic representation of the constructs introducing short sequences between the two sites. Transcriptional activity of the H3 hiNOS reporter gene and the modified constructs with 5 bp (H3 + 5) or 10 bp (H3 + 10) insertions, in response to CM or LPS stimulation for 6 h is shown. The activity of the full length P7 hiNOS reporter gene after 6 h CM stimulation is shown for reference. b In vivo footprinting of the hiNOS proximal promoter in non-stimulated cells. Genomic samples were obtained as described in “Materials and methods”. In vitro and in vivo samples were treated for ligation-mediated PCR

This region was then examined by footprinting to detect in vivo protein binding (Fig. 4b). Ligation-mediated PCR footprinting enables the visualisation of changes in the sensitivity to methylation of guanine residues within or adjacent to DNA recognition sites for transcription factors in living cells. Genomic DNA methylated in vitro shows the control guanine methylation profile of this region. The in vivo genomic DNA samples isolated from untreated cells show hypermethylations at the sequence corresponding to the P-Oct site and hypomethylations at the TATA box as compared to the in vitro control. This indicates the constitutive presence of proteins on the sequences corresponding to the TATA box and P-Oct and, importantly, shows no other proteins binding in this region.

Oct-1 binding to P-Oct primes hiNOS for transcription

To further investigate the role of the constitutive binding of Oct-1 to the promoter, we performed chromatin immunoprecipitation (ChIP) experiments to assess the state of the gene (Fig. 5). As expected, examination of the proximal region of the promoter where the Oct site and the TATA box are located (P1 primers) showed the presence of Oct-1 in non-stimulated cells. These experiments also demonstrated the presence of RNA polymerase II (Pol II) on this region as well as an enrichment for the well-known mark of active transcription tri-methylated Lys4 of Histone 3 (H3K4me3). These results were confirmed with primers specific for exon 1 (P2). When primers directed to a downstream region in the gene (exon 11) were used (P6), no enrichment was found for Pol II nor Oct-1. However, after stimulation with cytokines, all three sets of primers showed the presence of Pol II confirming active transcription. These results suggest the formation of the PIC on the hiNOS promoter with the presence of Oct-1, and enrichment for H3K4me3 confirms that the gene is primed for transcription.

Fig. 5.

Binding of Oct-1 and Pol II to hiNOS chromatin. Schematic representation of the localisation of the primer pairs on the hiNOS gene. ChIP assays were performed with specific antibodies to RNA Pol II, Oct-1 and H3K4me3 (MH3) as well as non-specific IgG on control (CT) and 2 h cytokine-stimulated (CM) cells. PCR with primers for the regions depicted was performed on the precipitated chromatin and on the input DNA (IN) and run on 1.5% agarose gels

To test whether the iNOS gene is transcriptionally active, we designed sets of primers to probe various regions along the hiNOS gene (Fig. 6a). These were used to perform RT-PCR on nuclear RNA of control and CM-treated cells. Figure 6a shows amplification of probe 1 (corresponding to exon 1) in resting cells, whereas no signal is detected from probes 2 (exons 2–5) and 3 (exons 11–14). Upon CM stimulation, all fragments are amplified within 1 h, indicating active elongation and processivity. Similar results were obtained with nuclear run-on experiments (not shown) that allow an estimation of the rate of de novo transcription. The activation index calculated from these experiments by normalisation to the actin signal (Fig. 6b) shows that once the cells are stimulated with CM the rate of transcription is higher at the 3′ end of the gene, confirming an increase in the processivity of transcription (Fig. 6b).

Fig. 6.

Transcription processivity of the hiNOS gene. a Schematic representation of the localization of the amplified fragments on the iNOS gene. RT-PCR of nuclear RNA for each fragment after the indicated CM stimulation was blotted and hybridised with labelled hiNOS probe. b Activation index obtained from nuclear run-on experiments and calculated for each probe after normalisation to the actin signal

In order to assess the involvement of Oct-1 in the initiation of transcription before stimulation, oligonucleotides corresponding to the P-Oct DNA site were synthesised in both orientations and hybridised to form a double-stranded binding element. The U-Oct sequence that shows no transcription factor binding (Fig. 1b) was used as a control. These double-stranded fragments were transfected into cells as competitors for Oct-1 binding, thus preventing binding to the iNOS promoter. iNOS transcription was then assessed by RT-PCR using primers for exon 1 (5′) and exons 11–17 (3′). As shown in Fig. 7, only transcription of exon 1 (5′) is observed in non-stimulated cells, which is abrogated by transfection of the P-Oct DNA element but not the U-Oct element, suggesting that Oct-1 binding to the iNOS promoter controls initiation of transcription.

Fig. 7.

Oct-1 control of transcription initiation. Non-stimulated cells were transiently transfected with double-stranded oligonucleotides harbouring the P-Oct and U-Oct sites identified in Table 1. Total mRNA was extracted and RT-PCR performed with primers to the proximal region of the hiNOS gene (5′) corresponding to exon 1 and distal region (3′) corresponding to exons 11–17

Taking these results together, we propose that rapid expression of hiNOS is regulated by Oct-1 through the formation of an initiation complex that primes transcription. Elongation of these transcripts is triggered by activation of transient transcription factors such as NFκB to rapidly produce full length iNOS mRNA as needed.

Discussion

Oct-1 is a ubiquitous transcription factor of the POU family, characterised by a bipartite POU domain that has intrinsic conformational flexibility, and specifically binds to an Oct sequence of DNA. The N terminus sub-domain of the protein (POU specific domain) contacts the ATGC sequence of the Oct element, while the C terminus or POU homeodomain rests in the AAAT site major groove [29]. It has been classified as a proximal activator that usually functions from a position close to the TATA box, typically activating transcription in response to a remote enhancer [27]. However, active Oct sites in upstream enhancers have also been described, such as the one recently identified [12] in the far upstream hiNOS promoter (10 kb).

We found that the only Oct-like sequence in the proximal 7 kb of the hiNOS promoter recognised by Oct-1 is located 27 bp upstream of the TATA box. This sequence is indispensable for transcription and even the slightest disruption of the close proximity to the TATA box abrogates promoter activity. In agreement with these findings, Bertolino and Singh [31], established that the proximity between the Oct and TATA sequences is crucial for transcriptional activity. This is due to the formation of a preassembled complex on the promoter which facilitates activation by a distant enhancer [30]. The importance of this structure is highlighted by its conservation throughout several species. Furthermore, no additional proteins are required and disruption of the helical phasing between the Oct site and the TATA box does not affect this function. The complex containing the POU domain and TBP then recruits Pol II to the promoter [32]. We also find in vivo binding of Oct-1, TBP and Pol II on the hiNOS promoter, indicating that the mechanism of hiNOS regulation by Oct-1 involves the formation of an initiation complex with the TATA binding transcription machinery. Indeed, a direct interaction between Oct-1 and TBP in vitro has been described involving the POU homeodomain of the protein [31, 32]. One of the mechanisms of Oct-1-regulated transcription is the recruitment of the basal transcription complex to the DNA [33, 34]. At the same time, TFIIB can increase the rate of binding and decrease the rate of dissociation of Oct-1 [35].

This mechanism is consistent with our observation that Oct-1 is present on the proximal promoter together with Pol II in resting cells. It would also explain how Oct, while being essential for transcription, is incapable of driving full-length mRNA synthesis until binding of transient transcription factors such as NFκB to the upstream enhancer is activated. Indeed, it has been shown that cytokine-induced transcription factors such as NFκB and STAT3 can activate transcription elongation by directly interacting with the positive transcription elongation factor b P-TEFb [37]. In this mechanism, the RelA subunit of NFκB is phosphorylated, which allows it to form a complex with P-TEFb. This complex harbours a CDK9 activity capable of phosphorylating Ser 2 of Pol II, a modification known to signal the transition into the processive mode [38, 39].

Further, we show that the assembly of the PIC allows transcription to start in readiness for the cytokine-stimulated signal to complete elongation of the transcripts. Transcriptional elongation is increasingly being acknowledged as an important step in the transcriptional regulation of eukaryotic genes [36], and many of the factors that cooperate with the general transcription machinery in this mechanism have been described [40, 41]. These mechanisms have been described for a handful of viral and mammalian genes [42], indicating that post-recruitment regulation occurs much more often than previously recognised. Recent genome-wide analyses in mammalian cells [13, 17, 19] renewed interest in these observations showing preassembled PICs on many inactive genes. The presence of the PICs is accompanied by marks of active transcription that characterise permissive euchromatin, such as H3K4me3, as we identified on the proximal hiNOS promoter. In these cases, transcription is truly initiated but elongation is not completed, indicating that the recruitment of Pol II is not the rate-limiting step for transcriptional regulation of many genes. This type of regulation occurs in functionally related groups of genes that are among the most rapidly induced, generally involved in growth, reprogramming of cellular metabolism, proliferation, and particularly immediate regulators of the inflammatory response [43].

Some of the few eukaryotic genes reported to be controlled by transcription elongation and processivity are important signalling molecules of the immune system [43] such as Igκ [44] and TNFα [44]. Here, we describe the regulation of hiNOS by this mechanism, which ensures a rapid response that in cases of infection and inflammation is critical for cell survival. Understanding this process at the molecular level may lead to potential new therapeutics that can be used to modify inflammation, immune responses and cell behaviour.