1998 Oxford University Press
Nucleic Acids Research, 1998, Vol. 26, No. 7
1801–1806
Correct usage of multiple transcription initiation sites and C/EBP-dependent transcription activation of the rat XDH/XO TATA-less promoter requires downstream elements located in the coding region of the gene Melissa P. Clark1,3, Chi-Wing Chow2, Jean E. Rinaldo1,2,3 and Roger Chalkley2,* 1Department
of Medicine and 2Department of Molecular Physiology and Biophysics, School of Medicine, Vanderbilt University and 3Nashville VA Medical Center, Nashville, TN 37232, USA Received October 31, 1997; Revised and Accepted February 12, 1998
ABSTRACT In the present study, we have shown that a downstream element located in the coding region of the TATA-less rat xanthine dehydrogenase/oxidase (XDH/XO) gene (–7 to +42) plays an important role in transcription initiation and C/EBP transcriptional activation. Previous work from our laboratory has shown that the promoter is organized with multiple initiator elements (Inr 1, 2, 3 and 4) which are important for transcription initiation. Additionally, we had identified two C/EBP binding sites upstream of this promoter. Deletional and mutational studies revealed that C/EBP binding was not essential for the basal level of transcriptional initation. However when XO-luciferase constructs include downstream sequence extending to +42 there is development of C/EBP sensitivity as well as a shift in the initiator usage. In the absence of the downstream element, primer extension analyses reveals Inr 3 and 4 to be the major start sites but in the presence of this additional sequence the usage is shifted to Inr 1 and 2. This shift in Inr usage more closely resembles that seen in intact macrophages or liver cells. Gel mobility shift assays indicate the presence of several binding factors located in this downstream region, one of which has been identified as YY-1. We postulate that YY-1 allows DNA bending which permits the upstream C/EBP elements to exhibit a transcriptional activation which is not seen when the downstream element is absent. This study presents a potential model for regulation of the XDH/XO promoter. INTRODUCTION The xanthine dehydrogenase/xanthine oxidase enzyme system (XDH/XO) plays an important role in purine metabolism, iron uptake and transport as well as in the defense against microbial agents. XDH/XO catalyzes the oxidation of hypoxanthine to xanthine, and subsequently to uric acid (1). XDH is converted to
xanthine oxidase (XO) by oxidation and by proteolytic modification (1–3). XO utilizes oxygen instead of NAD as the electron acceptor, and superoxide radicals are produced as a byproduct (4). Under certain pathophysiological conditions, overproduction of XO can lead to excessive superoxide production which leads to the formation of the hydroxyl radical, which is highly reactive and can cause severe tissue damage (4–8). Regulatory studies have shown that several cytokines, including TNF, IL-6 and IFN, can up-regulate XDH/XO expression (9–11). In addition, studies by Hausson et al. (12) have indicated that low oxygen may also stimulate XDH/XO gene expression. Our previous studies have shown that the promoter of the rat XDH/XO gene is TATA-less (13). Four functional initiators (Inrs) have been described just upstream of the translational start (ATG) (14). Using in vitro footprinting we have identified six footprints within the proximal XDH/XO promoter. Transcription factors including YY-1, C/EBP, NF-1, USF-like factor, Oct-1 related factor and Myc have been identified among these footprints. We postulated that the USF-like factor is likely to be TFII-I. Interestingly, YY-1 and a USF-like factor, which have been shown to bind known initiators, map to Inr 4 and Inr 3, respectively (15,16). Identification of C/EBP binding sites upstream of the XDH/XO promoter is intriguing as C/EBP factors have been postulated to play an important role in acute phase response (17–19). However, previous deletional and mutational analyses of the rat XDH/XO promoter have indicated that the XDH/XO transcriptional activity is independent of the upstream C/EBP binding sites in the constructs which include all of the proximal promoter to –7 bp (14). In addition, subsequent primer extension analyses have shown that the usage of multiple initiators in these constructs has been altered so that Inr 3 and 4 are the major start sites. In this report, we describe a downstream element, which is located in the coding region of the XDH/XO gene between +1 and +42, and which plays an important role in the transcription initiation. The presence of this downstream element allows correct usage of the multiple initiators in in vitro transcription analyses. In addition, C/EBP dependent activation of the promoter is now observed in constructs containing this downstream element.
*To whom correspondence should be addressed at: Department of Molecular Physiology and Biophysics, 741 Light Hall, Vanderbilt University School of Medicine, Nashville, TN 37232, USA. Tel: +1 615 323 4611; Fax: +1 615 322 7236; Email:
[email protected]
1802 Nucleic Acids Research, 1998, Vol. 26, No. 7 MATERIALS AND METHODS Gel mobility shift assay Procotols for preparing nuclear extracts and for performing gel shift assays were as described previously (20–22). Antibodies for all of the gel shift assays were purchased from Santa Cruz Biotechnology Inc. Oligonucleotides were synthesized on a Milligen/Biosearch 7500 DNA synthesizer. Oligonucleotides used as competitors were administered at 100-fold molar excess and most sequences have been described previously (14), for additional sequence information see below. For gel shift assays with antibodies, protein–DNA complexes were allowed to form in the presence of binding buffer and poly-(dI.dC) before addition of specific antibody. Sequences for the oligonucleotides not described previously are: Inr 2, 5′-GGCAGTGATATCTACAACTACTTCTCAAGAGCTCA-3′ (from –53 to –19); Inr 1, 5′-CAAGAGCTCAGTGACTCCAGCAGCCACGATG-3′ (from –28 to +3); ATG, 5′-GCAGCCACGATGACTGCGGATGAGTTGGTCT-3′ (from –9 to +22); +42, 5′-CAACCAGAAGAAACACTTACCGTTTTTCCATTCGTCCCCAGACGA-3′ (from +16 to +60); mInr 4, 5′-CCTCCCAGCAAATTGAAAGAAAGTTACCC-3′. In vitro transcription XO-luciferase reporter plasmid (1.5 µg) was used as a DNA template in in vitro transcription. A cocktail mix that will adjust the final concentration of the reaction sample to 600 µM NTPs, 7.5 mM MgCl2 and 75 mM NaCl was added. About 400 µg of HeLa nuclear extract (a kind gift from Tony Weil) was added and the sample was incubated at 30C for an hour. The reaction was stopped by adding 50 µl of stop solution (100 mM NaOAc, 0.4% SDS and 1 µg/µl tRNA). In vitro transcribed mRNA was isolated by two phenol-CHCl3 extractions followed by ethanol precipitation. Isolated mRNA was used for primer extension as described previously (14). Deletional and mutational analysis Desired fragments for subcloning were generated from PCR reactions using appropriate oligonucleotides. The PCR products were subcloned into a luciferase reporter construct as described previously (13). Positive clones were identified by colony hybridization. All constructs were amplified by liquid culture and sequenced to verify identity before use. Transient transfection assays were performed in HeLa cells and NIH/3T3 cells using the calcium phosphate precipitation method as described previously (13). DNase I footprinting A 236 bp fragment (XDH sequence from –116 to +119) was end labelled with [γ-32P]dATP using polynucleotide kinase. About 3–5 fmol of labelled probe was mixed with increasing concentrations of rat liver nuclear extract (5–15 µg) under the binding conditions described above for gel mobility shift assays. After incubating the mixture at room temperature for 30 min, 50 U/ml of DNase I was added and the mixture incubated for 1 min at 23C. Digestion was stopped by adding 4 vol of DNase I stop solution (1% SDS, 10 mM EDTA, 10 µg/ml proteinase K and 10 µg/ml denatured salmon sperm DNA) and incubation at 65C for 20 min. After phenol-CHCl3 extracion and ethanol precipitation, the products were resolved in an 8% polyacrylamide gel under denaturing conditions and visualized by audioradiography.
Figure 1. An analysis of in vitro and in vivo transcription start sites using two different XO-Luc reporter constructs. XO-Luc constructs encompassing sequence either from –102 to +42 bp or from –102 to –7 bp were used to direct in vitro transcription of luciferase using HeLa cell nuclei extracts. The transcription initiation sites were detected by primer extension from within the luciferase transcript. As a comparison for initiator usage in vivo, primer extension was performed on total RNA isolated from rat bone marrow macrophage (Mø) as well as from liver using a xanthine oxidase mRNA primer.
RESULTS Correct choice of Inr start sites requires sequences within the coding region We have previously shown that the XDH promoter contains at least four Inrs (14). In the constructs we had made to analyze XDH promoter activity, we found that qualitatively the usage of the Inrs was recapitulated faithfully, but that the relative quantitative usage frequency was unphysiologic. In addition the C/EBP sites located in the upstream part of the promoter showed no regulatory activity. A resolution to these concerns came serendipitiously when we attempted to understand the role of factors binding to the Inrs by studying transcription in vitro in the presence of competitor oligonucleotides. In a series of control experiments we employed a construct which included sequences out to +42 bp which had been made for experiments involved in identifying the start site of transcription of this gene. The presence of the extra sequence led to a more physiological usage of the various Inrs. Thus, as shown in Figure 1, RNA made from constructs which include sequence out to +42, shows primer extended products which resemble those generated from RNA made from intact cells such as macrophages or liver cells in which we see that transcription inititation is primarily from Inrs 1 and 2. In contrast the RNA products from the shorter constructs (–7 bp) show that most of the initiation was directed from Inrs 3 and 4. Similar observations were obtained by RNase protection (13,14). Effects of extra sequences on basal promoter and C/EBP activity This then suggested to us that the additional sequences from –7 to +42 may play an important role in the regulation of this gene. Initially we assayed the effect of the extra sequences on the basal promoter activity of the Inrs themselves. These results are shown
1803 Nucleic Acids Acids Research, Research,1994, 1998,Vol. Vol.22, 26,No. No.17 Nucleic
1803
Figure 2. 5′ Deletional analysis with various basal XO-Luc constructs containing increasing amounts of Inr sequences all of which extended to +42 bp. The indicated sense and antisense reporter constructs were transiently transfected into HeLa cells with RSV-β galactosidase reporter plasmid as a control. At least three independent assays were performed with the activity from the promoterless construct normalized as 1.0.
Figure 3. Effect of upstream C/EBP binding sites on XO promoter activity when the constructs contain sequences out to +42 bp. The various constructs indicated in the figure, and appropriate mutations, were transiently transfected into HeLa cells as described in Materials and Methods. At least three independent assays were performed with the activity from the promoterless construct normalized as 1.0. In the chart a filled block represents a C/EBP binding site. A cross through a filled box represents a mutated C/EBP binding site.
in Figure 2. As expected, the non-TATA initiated transcription is polar and requires that the Inrs are in the native orientation. Second, the inclusion of Inr1 to the –7 to +42 bp sequence provides the greatest stimulation over the promoterless control, whereas the other Inrs provide less additional response. This
result is consistent with the transcription results from intact cells which indicate that in vivo Inr 1 is the major initiation site in this promoter. Overall in the transfection experiments, the presence of all of the Inrs increases promoter activity by 9-fold.
1804 Nucleic Acids Research, 1998, Vol. 26, No. 7
Figure 4. The transcriptional effect of the various C/EBP constructs in NIH/3T3 cells. The same constructs as described in Figure 3 were utilized. At least three independent transient transfection assays in NIH/3T3 cells were performed, with the activity from the promoterless construct normalized as 1.0.
The results of an analysis of additional XDH promoter fragments including the two C/EBP sites is shown in Figure 3. Inclusion of sequences beyond the basic promoter described above, out to –116, except that the C/EBP site is mutated, provided no additional increase in activity. However, in contrast to our previous observations for constructs lacking the +42 bp sequences, if the C/EBP site is maintained intact, then a dramatic increase to 50-fold activation over the control is seen. This is also seen for the second C/EBP site at –160. Thus, the presence of the additional 49 bp of sequence beyond Inr 1 imparts a more physiological response to the cloned XDH promoter, in terms of not only usage frequency of Inrs but also in terms of response to a well documented transcriptional activator. Similar results were obtained when we utilized NIH/3T3 cells (Fig. 4). This cell line is of interest in that it is somewhat unusual as it has maintained the capacity to transcribe the XDH gene in tissue culture. We note that the direction of the various changes in activity parallels that seen in HeLa cells, though the magnitude of the change is decreased. Analysis of factors binding to the –7 to +42 bp region Since sequence elements in the promoter between –7 and +42 appear to play a major role in recapitulating the properties of this gene seen in vivo, we have asked what factors may be involved in binding to this region of the XDH promoter. As a first step we performed a DNase I footprint analysis with the results shown in Figure 5. The factors involved in interactions with this region form clearly defined, though not especially strong footprints,
when nuclear extracts were exposed to labelled DNA from this region. One unusual aspect of the footprinting over these sequences is the appearance of some new hypersensitive sites. One of these sites at +10 is not cut by the nuclease at all in naked DNA, suggesting that a new, more enzymatically sensitive conformation may have been produced at this site. In the region –7 to +42, footprints can be detected from –4 to +7 and from +32 to +40 bp. Inspection of the sequence of the presumptive binding domains indicated that these sites are just outside (and downstream of) the reported transcription initiation zone. Oligonucleotides were synthesized which encompass the binding regions. They were then utilized in mobility shift analyses to determine the nature of the factor(s) which might be involved in binding to these regions. Incubation of a nuclear extract with the ‘ATG’ oligonucleotide generated one major and two minor, more slowly migrating, complexes. The major complex was identified as YY-1 based upon overall mobility, competition with known YY-1 binding sequences and disruption of the complex with YY-1 antibodies, as indicated in Figure 6 (16,23–28). There is a YY-1 site on the antisense strand that is a very close match to the canonical site GTCATN(n)TG. One of the major hypersensitive sites induced by protein binding was immediately adjacent to the presumptive site for YY-1 binding (16,27). YY-1 is known to generate major changes in DNA secondary structure as it binds to the minor groove and bends the DNA through some 80 (29). The faster moving of the minor complexes was sensitive to USF antibody and the slower moving complex was competed effectively by the EFII oligonucleotide despite the competition with EFII there is no evidence for C/EBP interaction with the ATG oligonucleotide.
1805 Nucleic Acids Acids Research, Research,1994, 1998,Vol. Vol.22, 26,No. No.17 Nucleic
Figure 5. In vitro DNase I footprinting of the XO promoter over the additional sequences out to +42. A fragment of the XO promoter from –116 to +119 bp (236 mer) was end labelled using T4 polynucleotide kinase. The fragment was then allowed to bind in control buffer or in increasing concentratations of a rat liver nuclear extract and digested with 50 U/ml of DNase I for 1 min at 23C. After purification of DNA the material was then electrophoresed on an 8% denaturing polyacrylamide gel and visualized by autoradiography. A Maxam and Gilbert reaction on adenine and guanosine nucleotides of the probe was used a size marker. The open boxes represent several weakly identified footprints (FP). Lane 1, partial digestion of the probe in the absence of nuclear extract. Lanes 2 and 3, partial digestion of the probe in the presence of increasing amounts of unheated rat liver nuclear extract.
Incubation of a nuclear extract with the ‘+42’ oligonucleotide containing the protected sequences from +32 to +40 generates a pattern which somewhat resembles that seen for the ATG oligonucleotide in as much as there is a single major product which moves close to that seen for the ATG oligonucleotide. However, there is no evidence for the presence of YY-1 as judged by competition assays and the composition of this band is unknown at the present time. DISCUSSION Sequences from –7 to +42 bp in the XDH promoter play a role in two quite distinct biological functions. They are needed for the correct frequency usage of the multiple Inrs from which transcription is initiated in this gene in vivo. In addition the upstream C/EBP binding domains do not exert a transcriptional stimulatory effect in the absence of these sequences. We have shown that the –7 to + 42 sequence is the site of binding of YY-1 (from –7 to +7) and of an as yet unidentified factor (from +30 to + 42). One of the major new sites of DNase I hypersensitivity in
1805
Figure 6. Gel shift analyses of factors binding to footprinted regions between Inr 1 and +42. An oligonucleotide was constructed encompassing nucleotides from –9 to +22 bp (designated AUG oligonucleotide) and used as a probe in EMSA. The binding of rat liver nuclear extracts was studied in the presence of various competitor oligonucleotides and antibodies as indicated in the figure itself. M identifies the major complex which is competed by 100-fold excess of YY-1 oligonucleotide (lane 2).
DNase I footprints is found at +10 bp, immediately adjacent to the YY-1 site. This may well reflect a profound disturbance in the DNA secondary structure due to the DNA bending induced by YY-1. If so then it is curious that such a disturbance is only found in the DNA on one side of the YY-1 complex. We have determined that there is also a YY-1 site immediately upstream of the transcription initiation region. In other words the four Inrs are flanked by YY-1 sites. This suggests that the possible attendant DNA bending might bring sequences upstream and downstream of the Inr region into proximity, thereby permitting the upstream C/EBP cis elements to exhibit a transcriptional activation not seen previously when the downstream YY-1 site as well as the factor binding domain from +30 to +42 were absent from the XDH reporter constructs. In this regard it is interesting to note that Bauknecht et al. recently reported that YY-1 and C/EBP β are able to form a stable complex (30). The effects of the various cis acting sequences on initiation and on C/EBP activity are sumarized in Figures 3 and 4. In the absence of Inr 1 and the +42 cis sequences we see insensitivity to C/EBP and excessive firing of Inrs 3 and 4, relative to Inr 2 (14). If Inr 1 sequences are added to this construct then we see a diminution of overall transcription, which is, however, still insensitive to the presence of the upstream C/EBP site (14). Evidently the information added in the Inr 1 sequence is sufficient to repress the transcriptional activity of Inrs 3 and 4. In this same
1806 Nucleic Acids Research, 1998, Vol. 26, No. 7 ACKNOWLEDGEMENTS This work was supported by a SCOR grant from the NIH (#HL19153) and by grants from the Department of Veterans Affairs (J.R. and M.P.C.). REFERENCES
Figure 7. Cartoon model illustrating potential regulation of XDH promoter. (A) C/EBP binding at –110 bp (represented by the oval) has no effect on the initiator activity at Inrs 2, 3 or 4 in the absence of sequences to +42 bp. In the absence of Inr 1 there is uncharacteristically high activity of Inrs 3 and 4 which is indicated by the height of the arrow. (B) The addition of sequence to –7 bp includes Inr 1. The result is an inactivation of the high activity from Inrs 3 and 4. (C) The presence of a downstream element at +42 appears to exert a positive effect through the addition of extra factors such as YY-1 at +2 and a factor binding at +42. The addition of this downstream element permits the upstream C/EBP element to exert a transcriptional activation which is not seen when the downstream element is absent.
construct Inr 1 is still, however, relatively inactive in contrast to its activity in vivo. If the additional sequence to +42 is included in the construct we now see the development of sensitivity to C/EBP as well as increased Inr 1 activity more closely resembling that seen in the cell (Figs 1 and 4). Since C/EBP does not activate Inr 1 unless the +42 sequences are present we surmise that C/EBP and the +42 cis elements are acting jointly in this regard as shown in the model in Figure 7. In this figure we see that C/EBP binding at –110 bp has no effect on the intiator activity at Inrs 2, 3 or 4 in the absence of the sequences to +42 bp. Interestingly, the absence of Inr 1 leads to an unphysiologically high activity of Inrs 3 and 4. The addition of sequence to –7 bp includes Inr 1 which serves to inactivate the high response from Inrs 3 and 4, but does not lead to a dependence on C/EBP. Addition of sequences to +42 bp brings in at least two extra factors including YY-1 at +2 to +10 as well as a factor binding towards the +42 position. It is tempting to imagine that YY-1 may permit DNA bending so that the additional, as yet unidentified factor can exert its synergism in concert with C/EBP so that now this latter factor is seen to play a role in the regulation of XDH/XO.
1 Parks, D. A. and Granger, N. D. (1986) Acta Physiol. Scand. Suppl., 548, 87–99. 2 Stirpe, F. and Della Corte, E. (1969) J. Biol. Chem., 244, 3855–3863. 3 Waud, W. and Rajagopalan, K. (1976) Arch. Biochem. Biophys., 172, 365–379. 4 Mc Cord, J. M. (1985) New Engl. J. Med., 312, 159–163. 5 McCord, J. M., Roy, R. S. and Schaffer, S. W. (1985) Adv. Myocardio., 5, 183–189. 6 McCord, J. M. (1987) Fed. Proc., 46, 2402–2406. 7 Granger, D. N. (1988) Am. J. Physio., 255, H1269–H1275. 8 Brass, C. A., Narciso, J. and Gollan, J. L. (1991) J. Clin. Invest., 87, 424–431. 9 Dupont, G. P., Huecksteadt, T. P., Marshall, B. C., Ryan, U. S., Michael, J. R. and Hoidal, J. R. (1992) J. Clin. Invest., 89,197–202. 10 Pfeffer, K., Huecksteadt, T. and Hoidal, J. (1994) J. Immunol., 153, 1789–1797. 11 Falciani, F., Ghezzi, P., Terao, M., Cassaniga, G. and Garattini, E. (1992) Biochem. J., 285, 1001–1008. 12 Hassoun, P., Yu, F., Shedd, A., Zulueta, J., Thannickal, V., Lanzillo, J. and Fanburg, B. (1994) Am. J. Physio., 266, L163–L171. 13 Chow, C., Clark, M., Rinaldo, J. and Chalkley, R. (1994) Nucleic Acids Res., 22, 1846–1854. 14 Chow, C., Clark, M. P., Rinaldo, J. and Chalkley, R. (1995) Nucleic Acids Res., 23, 3132–3140. 15 Roy, A., Meisterernst, M., Pognonec, P. and Roeder, R. (1991) Nature, 354, 245–248. 16 Seto, E., Shi, Y. and Shenk, T. (1991) Nature, 354, 241–245. 17 Alam, T., An, M., Mifflin, R., Hsieh, C., Ge, X. and Papaconstantinou, J. (1993) J. Biol. Chem., 268, 15681–15688. 18 Baumann, H., Morella, K., Campos, S., Cao, Z. and Jahreis, G. (1992) J. Biol. Chem., 267, 19744–19751. 19 Isshiki, H., Akira, S., Sugita, T., Nishio, Y., Hashimoto, S., Pawlowski, T., Suematsu, S. and Kishimoto, T. (1991) New Biol., 3, 63–70. 20 Gorski, K., Carneiro, M. and Schibler, U. (1986) Cell, 47, 767–776. 21 Sears, R. and Sealy, L. (1992) J. Virol., 66, 6338–6352. 22 Cheyette, T., Ip, T., Faber, S., Matsui, Y. and Chalkley, R. (1992) Nucleic Acids Res., 20, 3427–3433. 23 Seto, E., Lewis, B. and Shenk, T. (1993) Nature, 365, 462–464. 24 Shi, Y., Seto, E., Chang, L. S. and Shenk, T. (1991) Cell, 67, 377–378. 25 Hariharan, N., Kelley, D. E. and Perry, R. P. (1991) Proc. Natl. Acad. Sci. USA, 88, 8799–9803. 26 Shrivastava, A. Saleque, S., Kalpana, G. V., Artandi, S., Goff, S. P. and Calame, K. (1993) Science, 262, 1889–1892. 27 Shrivastava, A. and Calame, K. (1994) Nucleic Acids Res., 22, 5151–5155. 28 Lee, J. S., Galvin, K. M. and Shi, Y. (1993) Proc. Natl. Acad. Sci. USA, 90, 6145–6149. 29 Natesan, S. and Gilman, M. (1993) Genes Dev., 7, 2497–2509. 30 Bauknecht, T., See, R. and Shi Y. (1996) J. Virol., 70, 7695–7705.
