Interferon regulatory factors (IRFs) repress transcription of the chicken ovalbumin gene

2.1. Oviduct tubular gland cell culture and transfection

Sexually immature White Leghorn chicks were treated with estrogen and primary oviduct tubular gland cells isolated as previously described in detail (Sanders and McKnight, 1988). The cells were transfected using calcium phosphate coprecipitation as described (Sanders and McKnight, 1988). DNA concentrations were held constant by the addition of appropriate amounts of pcDNA-3.1 (Invitrogen, Carlsbad, CA USA). Cells were cultured for 24 hrs in F12/DMEM medium (1:1) with insulin (50 ng/ml) in the presence or absence of 17β-estradiol (1 × 10^-7 M) and corticosterone (1 × 10^-6 M) in 60 mm dishes. CAT assays were performed as previously described (Sanders and McKnight, 1985).

2.2. Nuclear protein extracts and gel mobility shift assays (GMSAs)

Nuclear protein extracts were prepared essentially as described (Gorski et al., 1986) using 3 g of frozen tissue. Final extracts were assayed for protein concentration, aliquoted into microfuge tubes, and stored at -80° C. All GMSAs were performed as previously described using 5 μg of nuclear protein and 50,000 cpm of Klenow P³²-labeled double-stranded DNA oligomers (Nordstrom et al., 1993). For the competition experiments, unlabeled oligomers were added at the same time as the radiolabeled oligomers and before the addition of the nuclear protein. For antibody supershift experiments, nuclear protein extracts were incubated with the antibody at 4° C overnight, with rocking, before the DNA binding experiment was performed. The top strand of the oligomer sequences used for the data in Fig. 3A are shown in Fig. 3B, and the top strand of the oligomers used for Fig. 3C are below. For all oligomers, the non-wild type sequences are shown as small letters.

• Wild type CAR oligomer: 5′-agctTATTGAAACTAAAATCTACCAATCCCATTa-3′

• Mutated CAR oligomer: 5′-agctTATTGgggCTAgggTCTACCAATCCCATTa-3′

• Wild Type Consensus ISRE oligomer: 5′-agcttGGAAGCGAAAATGAAATTGACa-3′

• Mutated Consensus ISRE oligomer: 5′-agcttGGAAGCGAggATGAggTTGACa-3′

2.3. RNA isolation and reverse transcriptase polymerase chain reaction (RT-PCR)

The tissues used in Fig. 5 were isolated from exsanguinated chickens to reduce the possibility that the IRFs measured represented contamination by hematopoetic cells. RNA was isolated using Trizol reagent (Invitrogen), and RT reactions were performed using 2 μg of total RNA and MMLV reverse transcriptase, both according to the manufacturer’s directions (Invitrogen). All RT-PCR reactions used 2 μl of the appropriate RT reaction as a template and recombinant Taq polymerase (Invitrogen) according to the manufacturer’s instructions. The primers and annealing temperatures are listed in Table 1. PCR cycles: 1 cycle at 95° C for 5 min., 25 cycles at 95° C for 30 sec., followed by annealing for 1 min. and an extension at 72° C for 2 min and the final cycle at 72° C for 10 min. All PCR reactions were run on a 2% metaphor gel, stained with ethidium bromide. The RT-PCR data were quantified using the Gel Logic 100 Imaging System with Kodak 1D Image Analysis software. The analysis was done on underexposed versions of the gels actually put into the paper to increase accuracy and prevent saturation. The analysis program was tested for the dynamic range with standards, and background was subtracted for all samples.

2.4. Site-directed mutagenesis

Primer pairs were manually designed, based upon the QuikChange® Site Directed Mutagenesis kit (Stratagene, La Jolla, CA USA), to introduce mutations into the sequences in the 5′-flanking region of the Ov gene. The PCR reactions were performed with the primers below according to the manufacturer’s (Stratagene) protocol except that the extension times were increased to 2 min per kb and the cycle number was increased from 12 to 18 cycles. Mutations were verified by DNA sequencing.

-126/-125 5′-CCATTCAATCCAAAAATGGACCTATTccAACTAAAATCTAACCCAATCCC-3′

-122/-121 5′-CCATTCAATCCAAAAATGGACCTATTAGAAtcAAAATCTAACCCAATCCC-3′

OTE 5′-GCAGAACAGGCAGtcgACACATTGTTACCCAGATTAAAAACTAATATG-3′

2.5. Plasmids

The IRF-4 cDNA was a gift from Dr. Henry Bose (University of Texas). IRF-8 was a gift from Dr. Ben-Zion Levi (Technion-Israel Institute of Technology). The IRF-10 cDNA was cloned from a cDNA library derived from estrogen-stimulated chicken oviduct. These IRF cDNAs were subcloned into the pcDNA-3.1 (Invitrogen) expression vector.

The pOvCAT-900 reporter vector was constructed using Ov sequences from -900 to +9 (GenBank Accession no. J00895), which encompass all of the elements necessary for induction of the gene by steroid hormones (Sanders and McKnight, 1988). Mutations introduced into the CAR and OTE sites are described under site-directed mutagenesis.

3.1. The OTE and CAR sites bind similar protein complexes

The repressive CAR element (-130 to -100) in the NRE prevents expression of Ov in the oviduct in the absence of steroid hormones (Sensenbaugh and Sanders, 1999) and at all times in other tissues (Sanders and McKnight, 1988; Monroe and Sanders, 2000). CAR shares identity with the OTE site (-198 to -170), which acts as an activator of Ov transcription, over two short sequence elements of 4 and 5 nucleotides separated by 10 bp (underlined, Fig. 1, panel B). To being to investigate whether these two elements bind the same protein(s), GMSAs were performed using nuclear protein isolated from laying hen oviduct tissue. A single major shifted band of identical migration was observed with both radiolabeled oligomers (Fig. 2. panels A and B, lane 2). Binding to the CAR oligomer is effectively competed off by 50X molar excess of self-oligomer (panel A, lane 6) or by the OTE oligomer (panel A, lane 4), but not by non-specific oligomer (panel A, lane 8). A minor upper band is competed by self-oligomer (panel A, lane 6) but not by the OTE oligomer (panel A, lane 4), suggesting that at least one protein binds to CAR that does not bind to the OTE oligomer. Conversely, both self-oligomer (panel B, lane 4) and CAR oligomer (panel B, lane 6) compete off binding to the OTE. These data raise the possibility that the CAR and OTE sites bind the same or a similar protein(s).

3.2. Protein binding to the CAR oligomer requires an intact ISRE

In order to determine which nucleotides are necessary for protein binding to the CAR site, 2 bp mutations were introduced into the CAR oligomer beginning at -129 and ending at -118. These mutated oligomers were tested in GMSAs for binding (data not shown). Interestingly, oligomers containing mutations in a potential interferon-stimulated response element (ISRE) from -126 to -121 were unable to compete effectively for binding. IRFs bind a core consensus sequence of GAAANN (Paun and Pitha, 2007), although other nucleotides may be important, depending on which IRF complex is binding. Subsequently, combinations of these 2 bp mutations were combined and tested in GMSAs using the labeled CAR probe (Fig. 3A). The protein complex binding to the CAR oligomer, designated by the arrow, is specific because cold self-competitor competes for the complex at 25X molar excess (Fig. 3A, Lane 3). None of the oligomers except the one containing the combined mutations at -122/-121 and -119/-118 are able to compete effectively with the wt (wild type) CAR oligomer at this low molar excess (Fig. 3A, Lane 7). The fact that this oligomer can compete is not too surprising as the ISRE can apparently tolerate any nucleotides in positions occupied by -122/-121 (GAAANN). The oligomer used for the competition in lane 8 only retains two of the six core ISRE nucleotides (at positions -124 and -123) and is ineffective at competing for the binding protein(s). Thus, these data provide strong support for the contention that the protein(s) binding to the CAR element is binding in the putative ISRE.

The data thus far suggest that the protein binding to the CAR element is doing so via nucleotides found in the ISRE. To investigate whether this putative ISRE is in fact binding a protein that recognizes a consensus ISRE, additional GMSAs were done using nuclear protein extracts from sexually immature chicken oviduct (Fig. 3C), which is why the banding pattern appears somewhat different from the other GMSA. The CAR probe was labeled and, as expected, an unlabeled wt CAR oligomer was able to compete off binding (Fig. 3C, lane 14) whereas a mutated CAR oligomer was not (lane 15) nor was an unrelated oligomer (lane 16). However, a wt consensus ISRE was able to compete (lane 17) and its mutated counterpart was not (lane 18). This supports the idea the CAR element harbors a bone fide ISRE.

3.3. The protein complex that binds to the CAR site cross reacts with an IRF antibody

To ascertain whether the protein binding to the CAR oligomer is in fact an IRF, GMSAs were done, shifted bands were excised (Fig. 4, panel A, areas 1, 2, and 3), and the eluted proteins were analyzed by western blotting with an IRF or a control antibody. No bands were detected with the control antibody nor with the control area 3 (panel B). However, western blotting with the IRF antibody revealed that the protein in shifted Band 2 cross-reacted and migrated at approximately 50 kDa, the expected size of an IRF family member (Takaoka et al., 2008). This suggests that the binding protein is an IRF, although we cannot rule out the possibility that an IRF protein comigrated in the gel with the shifted complex. Excision and western blot analysis of shifted Band 1 did not yield a protein that cross reacts with the IRF antibody, although the mass was so much lower that the assay likely would not detect IRF protein. Although this antibody was supposedly selective for IRF-4, it also recognizes IRF-8 and -10 transfected into cells where they are not normally expressed (data not shown). Because few antibodies are available that are selective for chicken IRFs, we were unable to definitively determine which IRF might be binding. Nonetheless, these data strongly suggest that an IRF family member binds the CAR oligomer in vitro and raise the possibility that an IRF may also bind to the OTE.

3.4. Tissue distribution of IRFs-4, -8 and -10

The mammalian IRF family consists of 9 members, IRFs-1 through -9 (Takaoka et al., 2008). While IRFs-1, -2, -5, and -7 are fairly ubiquitous, expression of IRFs-4 and -8 is often considered to be unique to hematopoietic cells (Takaoka et al., 2008). However, in our attempts to identify the IRF(s) binding to Ov, we cloned a novel IRF, now called IRF-10, out of an estrogen-stimulated oviduct library (Park and Sanders, unpublished observations). This was not published because another group reported the cloning of IRF-10 out of a chicken splenic cDNA library (Nehyba et al., 2002) while our manuscript was in preparation. Nonetheless, it led us to speculate that other IRFs might be expressed in oviduct. Because we cloned IRF-10 out of chick oviduct and because IRFs-4, -8 and -10 are in the same IRF family subgroup (the hematopoietic IRFs, Takaoka et al., 2008) and are highly homologous in the DNA binding domains, we hypothesized that one or more of these IRFs may be regulating Ov. This seemed logical as IRF-4 and IRF-8 are present in at least some non-immune tissues (Li et al., 1999; Dadoune et al., 2005; Eguchi et al., 2008).

In hopes of identifying which IRF(s) might regulate Ov, we examined the tissue distribution of IRF-4, -8 and -10 in chicken by RT-PCR (Fig. 5). Interestingly, IRF-4 mRNA was detected in estrogen-withdrawn oviduct, lung, kidney, thymus, and faintly in liver (Fig. 5A). Much less IRF-4 mRNA was observed in estrogen-stimulated animals, with the highest amount being in kidney. While IRF-4 mRNA is high in thymus from estrogen-withdrawn chicks, it is exceptionally low in estrogen-treated thymus. Although this might be ascribed to the atrophic effect that estrogen has on thymus (Wang et al., 2008 and references therein), IRF-8 mRNA was not similarly affected. IRF-4 mRNA is also very low in estrogen-stimulated oviduct, suggesting that estrogen selectively represses transcription of Irf4.

IRF-8 mRNA is relatively abundant in all tissues analyzed regardless of estrogen treatment with the exception of the oviduct (Fig. 5B). In estrogen-stimulated oviduct, the amount of IRF-8 mRNA is about 50% less than that observed in estrogen-withdrawn oviduct, suggesting estrogen directly or indirectly represses Irf8 in this tissue (Fig. 5B, histogram). IRF-10 mRNA, like IRF-8 mRNA, was found in all tissues analyzed, although, at lower levels (Figure 5C). As with IRFs-4 and -8, IRF-10 mRNA is lower in estrogen-stimulated oviduct. These data indicate that at least these three IRFs are expressed in oviduct tissue and that their expression is repressed by estrogen. Furthermore, they suggest that IRFs-4, -8, and/or -10 may play a role in repressing the Ov gene in absence of estrogen in oviduct tissue.

3.5. IRF-4 and IRF-10 repress Ov

To ascertain whether IRFs-4, -8 and -10 are capable of repressing Ov transcription, increasing amounts of these IRFs were cotransfected with pOvCAT-900 (Ov sequences -900 to +9), which contains all of the Ov sequences required for induction by steroid hormones (Sanders and McKnight, 1988). The amount of expression vector DNA was held constant at 4 pmol by using the empty expression vector. The cells were then cultured for 24 hrs without or with 17β-estradiol and corticosterone (+S). The results show that both IRF-4 and IRF-10 repress the steroid hormone-dependent expression of pOvCAT-900 in a dose-dependent manner (Fig. 6), with maximal repression being about 70%. IRF-10 also represses basal expression of pOvCAT-900 by about the same amount at the highest dose (panel B). Although IRF-8 was tested, we later found that several amino acids were missing from the cDNA, so those data are invalid. Nonetheless, these functional studies reveal for the first time that IRFs repress expression of Ov.

Having identified nucleotides within the CAR element ISRE that are critical for protein binding (Fig. 3), the functional significance of these nucleotides was analyzed in transfection experiments (Fig. 7). Site-directed mutagenesis was utilized to introduce mutations into the CAR site at -126 and -125 or at -122 and -121, and the constructs were transfected into primary oviduct cells, which were subsequently cultured with or without estrogen and corticosterone. Because the CAR element as a whole appears to be a repressive site, we anticipated that IRF binding to the ISRE would repress Ov and that mutation of those critical nucleotides in the ISRE (Paun and Pitha, 2007) would relieve repression. Surprisingly, mutation of nucleotides -126 and -125 abolished steroid-dependent induction of the Ov promoter (Fig. 7, first panel), so we were unable to assess whether those two nucleotides are required for repression of Ov by the IRFs, as predicted. Although transcription of the construct with mutations in Ov at -122 and -121 appears to be repressed by the IRFs, this decrease was not statistically significant (Fig 7, second panel). This is not surprising as those two nucleotides do not affect the ability of the CAR oligomer to compete for binding in GMSAs (Fig 3, lane 7) and as any nucleotide can be in those two positions in a consensus ISRE. Because the OTE has some sequence identity to CAR (Fig. 1), because the OTE oligomer competes for proteins bound by the CAR oligomer in GMSA (Fig. 2), and because the OTE region plays a role in the activation of Ov, this element may be confounding the overexpression experiments.

To investigate the role of the OTE, three nucleotides in the ISRE in the OTE element were mutated, and the resultant construct was transfected into primary oviduct cells. Expression of this construct did not differ from wt, and the IRFs were able to repress both steroid-stimulated and basal expression by about 50% (Figure 7, third panel). A construct was made in which both the CAR and OTE ISREs were mutated, but this proved uninformative because its expression was at background levels (data not shown). These data indicate that nucleotides -126 and -125 are required for Ov expression and suggest that only one functional ISRE is necessary for partial repression of Ov by IRFs. More extensive repression is seen when both ISREs are present (Fig. 6).