Hypoxia increases rate of transcription and stability of tyrosine hydroxylase mRNA in pheochromocytoma (PC12) cells.

Reduced arterial oxygen tension (i.e. hypoxia) is a powerful physiological stimulus that induces synthesis and release of dopamine from O2-sensitive (type I) cells in the mammalian carotid bodies. We reported recently that hypoxia stimulates gene expression for tyrosine hydroxylase (TH), the rate-limiting enzyme in catecholamine synthesis in type I cells of the carotid body. Efforts to identify the mechanisms regulating TH gene expression in O2-sensitive cells during hypoxia have been hampered by the lack of an appropriate model cell culture system. Here we report that TH gene expression in the rat pheochromocytoma cell line (PC12) is regulated during hypoxia in a manner similar to that measured in carotid body type I cells. PC12 cells might therefore be useful as an experimental model for identifying the molecular mechanisms that regulate TH gene expression during hypoxia. Nuclear runoff assays revealed that transcription of the wild type TH gene was enhanced during exposures to hypoxia lasting 12 h. Chloramphenicol acetyltransferase assays with constructs that contained different fragments of TH promoter revealed that the regulatory sequences that mediate the hypoxia-induced increase in transcription are located between bases -272 and +27 of the TH gene. Findings from experiments in which transcription was inhibited either with actinomycin D or 5,6-dichloro-1-D-ribofuranosylbenzimidazole, as well as pulse-chase experiments using 4-thiouridine showed that the half-life of TH mRNA was substantially increased during hypoxia. Thus, in the present paper we show that TH gene expression in PC12 cells during hypoxia is regulated by increases in both the rate of TH gene transcription and TH mRNA stability.

forts to identify the mechanisms regulating TH gene expression in 02-sensitive cells during hypoxia have been hampered by the lack of an appropriate model cell culture system. Here we report that TH gene expression in the rat pheochromocytoma cell line (PC12) is regulated during hypoxia in a manner similar to that measured in carotid body type I cells. PC12 cells might therefore be useful as an experimental model for identifying the molecular mechanisms that regulate TH gene expression during hypoxia. Nuclear runoff assays revealed that transcription of the wild type TH gene was enhanced during exposures to hypoxia lasting 12 h. Chloramphenicol acetyltrasnferase assays with constructs that contained different fragments of TH promoter revealed that the regulatory sequences that mediate the hypoxia-induced increase in transcription are located between bases -272 and +27 of the TH gene. Findings from experiments in which transcription was inhibited either with actinomycin D or 6,6-dichloro-l-~-ribofuranosylbenzimidazole, as well as pulse-chase experiments using 4-thiouridine showed that the half-life of TH mRNA was substantially increased during hypoxia. Thus, in the present paper we show that TH gene expression in PC12 cells during hypoxia is regulated by increases in both the rate of TH gene transcription and TH mRNA stability.
The primary mechanisms by which mammals adapt to reduced oxygen tension (hypoxia) are hyperventilation and polycythemia. Hyperventilation occurs within minutes and polycytemia within days of exposure to hypoxia. These physiological responses to hypoxia serve to increase the delivery of oxygen to tissues by increasing arterial oxygen tension and the 02-canying capacity of blood, respectively. Polycythemia results from increased production of red blood cells which is mediated by * The present study was supported by Grants HL 33831, HL 34919, and HD 28948 from the National Institutes of Health, Grant-in-aid 91-009960 from the American Heart Association, and a research grant from the American Lung Association. The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked "advertisement" in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.  (3). Findings from these studies revealed that the increase in Epo gene expression during hypoxia involves increases in both Epo gene transcription and Epo mRNA stability (2).
The hyperventilation that occurs during hypoxia is mediated by the carotid body chemoreceptors, which are located bilaterally at the bifurcation of the common carotid artery (4). It is now generally accepted that the type I (glomus) cells are the 02-sensitive cells in the carotid body and that they transmit information concerning arterial O2 tension to primary sensory afferent terminals by release of a neurotransmitter (5)(6)(7)(8)(9)(10)(11)(12)(13). It has been demonstrated that dopamine is released from type I cells during hypoxia (6)(7)(8). In addition, the activity (V,,,,) of tyrosine hydroxylase (TH), the rate-limiting enzyme in dopamine synthesis, is enhanced by hypoxia in type I cells (9)(10)(11).
We reported recently that environmental hypoxia stimulates TH gene expression in type I cells, which could account for the increased TH activity in the carotid body during hypoxia (12). Identification of the molecular mechanisms that regulate TH gene expression in the carotid body during hypoxia has been hampered by the paucity of type I cells (104/carotid body; Ref. 13) and the lack of an appropriate model cell culture line. Here we report that reduced oxygen tension induces TH mRNA in the rat pheochromocytoma PC12 cell line in a manner similar in time course and magnitude to that observed in type I cells.
We have therefore used PC12 cells as a model system to investigate further the mechanisms that regulate TH gene expression during hypoxia. We found that reduced O2 tension mediates an increase in TH mRNA in PC12 cells by a dual mechanism that involves increases in both the rate of TH gene transcription and TH mRNA stability ( i e . decreased rate of degradation).
Northern BZot Analysis-Total cellular RNA was isolated by lysing cells that had been exposed to various durations of either 21% or 5% 0, with 4 M guanidinium thiocyanate and ultracentrifugation of cellular lysates on 5.7 M CsCl, 0.01 M EDTA for 17 h at 34,000 rpm and 24 "C. Aliquots of RNA (10 pg) were dried and denatured in a mixture containing dimethyl sulfoxide (Me,SO), glyoxal, 0.2 M phosphate buffer (pH 6.8), and water (5:2:1:2) for 1 h at 50 "C (14) prior to electrophoresis on 1% agarose gels in 10 m phosphate buffer. The RNA was blotted on a synthetic transfer filter (Genescreen Plus, DuPont) overnight. The filters were then dried and prehybridized for 2 h at 42 "C in buffer containing 50% formamide, 50 m phosphate buffer, pH 6.5, 5 x SSC, 5 x Denhardt's, 250 p g / d sonicated salmon sperm, and 0.5% SDS. Next, the filters were hybridized with 10 x lo6 cpm of TH probe that was complementary to TH mRNA. Hybridization was performed overnight at 42 "C using essentially the same buffer but supplemented with 10% dextran sulfate. Blots were washed in 2 x SSC and 0.1% SDS twice for 15 min at 42 "C.
An oligodeoxyribonucleotide probe (30-mer) that was complementary to bases 1,230-1,259 of rat TH cDNA (15) was used. Analysis of the probe sequence using GenEMBL data base did not show homology with other cloned c D N h or genes. The band obtained with this probe in the Northern blots corresponds in size to rat TH mRNA (1,780 bases) (15). The probe was labeled at the 5' end with [Y-~~P~ATP using T4 polynucleotide kinase. Analysis of the ethidium bromide-stained 18 and 28 S ribosomal RNA bands was used to veri^ that equal amounts of RNA were loaded onto each lane. Quantitation of TH mRNA levels was accomplished by optical density measurements (ImagePro I1 Plus).
Nuclear Dunscription Runoff Assays-The runoff assays were performed as described by Marzluff and Huang (16). All procedures involving extraction of nuclei were performed at 4 "C. Following exposure to either 21% or 5% O,, PC12 cells were resuspended in buffer containing 0.32 M sucrose, 3 m CaCl,, 2 m magnesium acetate, 0.1 m EDTA, 0.1% Triton X-100, 1 m dithiothreitol (DTT), 10 m Tris-HC1, pH 8, and then homogenized with 20 strokes in a Dounce homogenizer. Suspensions were checked with a microscope for completeness of cell lysis. Nuclei were diluted twice with buffer containing 2 M sucrose, 5 m magnesium acetate, 0.1 m EDTA, 1 m Dm, 10 m Tris-HC1, pH 8, and layered onto a cushion of the same buffer representing one third of the volume of the suspension of nuclei before centrifugation at 30,000 x g for 45 min at 4 "C. Nuclei were counted, resuspended at 2-10 x 107/ml in a storage buffer (containing 25% glycerol, 5 m magnesium acetate, 0.1 m EDTA, 5 m DTT, 50 m Tris, pH 8), frozen in liquid nitrogen, and stored at -80 "C. In vitro transcription was performed in the presence of 0.5 m each ofATP, CTP, and GTP, 50 pCi of [C~-~~P]UTP, 0.05 m S-adenosylmethionine, 120 m KCl, 2.5 m magnesium acetate for 30 min at 30 "C. The reaction was stopped by addition of 25 units of RNase-free DNase for 30 min at 37 "C. The nuclei were then lysed by addition of 10 volumes of solution containing 1% SDS and 10 m EDTA, and the reaction mix was treated with proteinase K at 42 "C for 30 min. The radiolabeled RNA was extracted with equal volumes of phenolchloroform (21) at 55 "C, ethanol-precipitated, and purified on G-50 RNase-free columns. Incorporation of radioactivity was in the range of 0.5-2.0 cpdnucleus. Addition of the transcription blocker warnanitin to the reaction blocked incorporation of radioactivity into nascent RNA.
For each experimental condition, equal amounts of 3zP-labeled RNA was hybridized to nitrocellulose membranes that contained 3 pg each of the following linearized and denatured plasmids: 3.1-kilobase pair frag-ment of rat genomic TH DNA (exons 3-11) in SPf35 plasmid, the N1length cDNA for 0-actin, and SP65 plasmid DNA. Filters were prehybridized and hybridized overnight at 42 "C in buffer containing 50% formamide, 50 m phosphate buffer, 5 x SSC, 5 x Denhardt's, 250 pg/d sonicated salmon sperm, 0.5% SDS, and 10% dextran sulfate. Following hybridization filters were washed twice in 2 x SSC for 1 h at 65 "C, followed by treatment with RNase A (5 mg/ml) at 37 "C for 30 min and a wash in 2 x SSC at 37 "C for 1 h. The dried filters were exposed to x-ray film at -80 "C for 24-48 h. Optical density measurements were performed (ImagePro I1 Plus), and the level of TH RNA transcription was normalized at each time point to transcription of p-actin and plotted in arbitrary units (mean 2 S.E.).
Dansfections and CAT Assay-PC12 cells were transfected with plasmid constructs containing fragments of the TH promoter (-272/+27, -773/+27, and -4,800/+27) that had been cloned in front of the chloramphenicol acetyltransferase (CAT) bacterial reporter gene. Cells were also transfected with the promoterless pCAT plasmid, which served as a negative control, and with a Rous sarcoma virus promoter-CAT construct, which served as a positive control. PC12 cells were cotransfected with a 0-galactosidase expression plasmid, which was under control of the SV40 enhancer/promoter that controlled for efficiency of transfection. PC12 cells (2 x 106/100-mm2 dish) were transfected by liposomal fusion (17). Plasmid DNA (10 pg of test plasmid + 10 pg of pgalactosidase plasmid) was combined with Lipofectin (30 pg) in a polystyrene tube containing 1 ml of DMEM/F-12 medium for 10 min at room temperature and applied to PC12 cells. Following a 3-h incubation, 10 ml of DMEM/F-12 medium containing 10% fetal calf serum was added to each plate and cells were allowed to recover overnight. Fresh DMEM/F-12 with fetal calf serum was added prior to exposure of transfected cells to 21% or 5% 0, for 48 h. During these exposures the medium was changed aRer 24 h of exposure. In several experiments, the COS-7 cells were transfected in a manner identical to that for PC12 cells. CAT activity was measured by the method of Gorman et al. ( 18). The amount of cellular lysate used for measurement of CAT activity in each experimental condition was normalized to 0-galactosidase activity, which accounted for differences in the efficiency of transfection (19). Cellular lysates were incubated in the presence of [14Clchloramphenicol and acetyl-coA. 14C radioactivity representing both nonacetylated and monoacetylated chloramphenicol was measured using an Ambis radioanalytical system. Results were plotted as percent acetylation of total [14Clchloramphenicol.
Measurements of TH mRNA Stability-The half-life of TH mRNA was measured using two different methods. In the first method, transcription was blocked pharmacologically and the time course for the decay of TH mRNA was measured in PC12 cells exposed to either normoxia (21% 0,) or hypoxia (5% 0,) for various durations. In these experiments, PC12 cells were treated with fresh medium and with either one of two transcription blockers: actinomycin D (3 pg/ml; Ref. 20) or DRB (100 w; Ref. 21). Actinomycin D is a nonspecific blocker of all RNApolymerases (20), and DRB inhibits initiation of RNAsynthesis (21). The drugs were added 15 min prior to exposure to either 21% or 5% 0, for 6, 12, 18, 24, or 30 h. At the end of each exposure period, cells were collected and total cellular RNA was extracted and processed for Northern blot analysis as described above.
We also measured the rate of TH mRNA degradation with a pulsechase method using 4-thiouridine (22,23). PC12 cells were grown in the DMEM/F-12 medium with dialyzed serum for 24 h and then pulsed with 0.2 m 4-thiouridine for 1.5 h. The chase was then initiated by adding 10 m cytidine and 10 m uridine to fresh media. Cells were harvested at 0, 6, 12, and 24 h from the end of the pulse. In control experiments (n = 3), cells were kept in normoxia for the duration of the experiment. The exposure to hypoxia started in some experiments (n = 4) 12 h before the pulse and was continued for the duration of the pulse and chase periods. In other experiments (n = 2), the exposure to hypoxia was started at the onset of the pulse and continued throughout the chase period.
Total cellular RNA was isolated using TRI ReagentTM (RNA/DNA/ protein isolation kit) based on method described by Chomczynski and Sacchi (24). Thiolabeled RNA was isolated from the total cellular RNA using an organomercury affinity chromatography on an ARi-Gel 501 (23). Equal amounts of thiolabeled RNA were glycosylated and processed for Northern blot hybridization as described above.

RESULTS
Exposure of PC12 cells to 5% 0 2 caused an increase in TH mRNA over the control level that was apparent within the first hour of exposure and reached a peak (approximately 3-fold) after 6 h of exposure (Fig. 1, top). The TH mRNA remained elevated above the control level for the entire exposure to hypoxia (24 h). In contrast, mRNA encoding for p-actin was not augmented by hypoxia (Fig. 1, bottom 1. These results are representative of five experiments in which this paradigm was used and indicate that hypoxia causes a specific induction of TH mRNA in PC12 cells. This finding served as the basis for using PC12 cells as a model system to identify the molecular mechanism that regulates TH gene expression during hypoxia.

Regulation of TH Gene Expression by Hypoxia in PC12 Cells
The increase in the TH mRNA during hypoxia can result from an increase in the rate of TH gene transcription, decreased rate of TH mRNA degradation (Le. increased stability) or both mechanisms occumng simultaneously. To investigate the possibility that the rate of TH gene transcription is enhanced during hypoxia, nuclear runoff assays were performed on nuclei isolated from PC12 cells that had been exposed to 5% O2 for 1-12 h. Fig. 2 shows that the rate of TH gene transcription was increased above the control (normoxia) level throughout the entire hypoxia exposure period. The maximum rate of transcription (4.5-fold) occurred 6 h after the onset of hypoxia, which corresponds to the maximum increase in TH mRNA. After 6 h of exposure, there was a slight decline in transcription; however, i t remained elevated above the control level for the entire duration of hypoxia. In contrast, transcription of the p-actin gene was not increased by hypoxia (Fig. 2). In addition, we did not measure any nonspecific hybridization of a nascent transcript to SP65 plasmid DNA.
In order to determine if cis elements on the 5"flanking region of TH gene regulate transcription during hypoxia, PC12 cells were transfected with plasmid constructs that contained fragments of the TH gene (from -4,800, -773, and -272 to +27 bases) fused to a CAT reporter gene. The highest level of expression of the TH-CAT fusion gene during hypoxia was measured in cells transfected with the CAT recombinant containing the -272 to +27 fragment of TH gene (Fig. 3, A and B). The hypoxia-induced increase in CAT expression was repressed in cells transfected with constructs that contained longer fragments of the TH promoter (Fig. 3, A and B ) . This result indicates that the cis elements that mediate the increase in TH gene transcription during hypoxia are located between bases -272 and +27 of the TH gene and that repressor elements are located upstream from base -272. In another series of experiments, COS-7 cells were transfected with the -272 to +27 TH-CAT construct and expression of the TH-CAT gene was measured during normoxia and hypoxia (Fig. 3C). In contrast to the finding in PC12 cells, there was no increase in CAT expression in COS-7 cells during hypoxia (Fig. 3C), which indicates that the response is specific for PC12 cells.
To determine if hypoxia might affect TH mRNA stability in PC12 cells, the half-life of TH mRNA was measured during hypoxia and normoxia (Fig. 4). In one series of experiments. transcription was blocked prior to exposure to 2 1 9 (dashed lines) or 5 9 O2 (solid lines) with either actinomycin D ( 3 pg/ ml,) (Fig. 4, A and B, filled squares), or (Fig. 4B, filled circles ). In the second series, pulse-chase experiments were performed by labeling RNA with 4-thiouridine and then chasing the thiolabeled RNA with high concentrations of uridine and cytidine. The decrease in the amount of thiolabeled RNA over time al-  = 4)). During hypoxia "4 mRNA was markedly stabilized, which resulted in an increase in TH mRNA half-life in cells exposed to 5% 0,. Averaged results were calculated from optical density measurements of TH mRNA on Northern blots.
lows for estimations of the half-life of TH mRNA during both normoxia and hypoxia (Fig. 4C, dashed and solid lines, respectively). In two pulse-chase experiments, the hypoxic stimulus was started simultaneously with the pulse of thiouridine (Fig.  4C, filled triangles). In four other cases, hypoxia was started 12 h before the pulse and continued throughout the duration of the chase (Fig. 4C, filled diamonds). The important finding from these studies was that both methods, i.e. inhibition of transcription and pulse-chase, showed that the half-life of TH mRNA is increased in cells exposed to hypoxia (Fig. 4, compare  panels B and C ) . The level of TH mRNA in PC12 cells during normoxia dropped to 50% of the orignal value in -10 h. In contrast, during the 24-30-h exposure to hypoxia, TH mRNA did not decline to 50% of the initial value. Thus, the increased level of TH mRNA in PC12 cells during hypoxia is also due to an increase in TH mRNA stability.

DISCUSSION
In the present study, we have demonstrated that reduced oxygen tension (hypoxia) causes enhanced expression of tyrosine hydroxylase gene in the PC12 cells. We showed that hypoxia stimulates transcription of the wild type TH gene in PC12 cells. In addition, we showed that the increase in transcription of the TH-CAT fusion gene is mediated by a fragment of the 5' flanking region that extends from base -272 to base +27 relative to transcription start site. We also showed that hypoxia increases the stability of TH mRNA in PC12 cells.
Thus, hypoxia enhances TH gene expression in PC12 cells by a dual mechanism involving both increases in the rate of TH gene transcription and TH mRNA stability.
In this study, we used the PC12 cell line (26,27) as a model system to study regulation of TH gene expression during hypoxia. The observed increase in TH mRNA during hypoxia in PC12 cells is similar in time course and magnitude to that observed in the carotid body type I cells in the intact rat preparation (12). The carotid body is believed to have specific sensory and signal transducing systems that detect alterations in O2 tension and transduce this signal into specific cellular functions (25). Regulation of TH gene expression by hypoxia in carotid body cells is part of this specific response (12). We propose that PC12 cells contain the same or a very similar mechanism for sensing oxygen tension and transducing this signal into augmented gene expression for TH. Additional evidence that PC12 cells respond to hypoxia similarly a s carotid body type I cells includes findings that hypoxia stimulates both TH activity (28) and dopamine release (29) in PC12 cells.
We found that the increase in TH gene transcription during hypoxia in PC12 cells is mediated by a fragment of the TH gene that extends from -272 to +27 relative to the transcription start site. This result was not unexpected since several r e platory elements (e.g. CRE, APl, AP2, OctPOU, MyoD, and SPI ) are located within this fragment (30.31). There is evidence that the CRE (CAMP responsive element) is involved in regulation of TH gene expression in response to pharmacological stimulation of the intracellular CAMP signal transduction pathway (32) and in response to membrane depolarization (33). Since CAMP levels increase in type I cells (34)(35)(36) and in PC12 cells2 during hypoxia, it seems reasonable to suspect that the CRE might be involved in mediating increased transcription of the TH gene during hypoxia. However, we found that elevation of intracellular CAMP by activation of adenylate cyclase with forskolin (10-200 p~) prior to hypoxia failed to attenuate the hypoxiainduced increase in expression of the -272/+27 TH-CAT fusion gene.2 In addition, stimulation of adenylate cyclase with forskolin caused rather weak increases in transcription of the -272/+27 TH-CAT fusion gene and of the wild type TH gene in nuclear runoff experiments as compared to the level of transcription evoked by hypoxia (not shown). These data suggest that CAMP is not the primary regulator of TH transcription during hypoxia. This is further supported by the observation that induction of Epo mRNA in the kidney during hypoxia does not involve CAMP (37).
In this study, we also demonstrated that hypoxia elicited a substantial increase in the stability of TH mRNA. Similar results were obtained with two different methods of measuring RNA stability. In one method, transcription of the TH gene was blocked with either actinomycin D or DRR prior to exposure to hypoxia. This method is sensitive and is used widely to study RNA half-life (38). However, drugs that block transcription have been reported to exert side effects on cellular metabolism in such a way that it might affect stability of the RNA of interest (38). We do not believe that this was the case in the present study, since the half-life of TH mRNA during normoxia was similar (-10 h ) to that measured with a variety of methods in PC12 cells (22,39 To our knowledge, this is the first report that describes regulation of TH gene expression at the level TH mFWA stability by a physiological stimulus. This appears to be a specific response, since increases in TH mRNA concentration that occur in response to dexamethasone or stimulation of adenylate cyclase are not accompanied by increase in TH mRNA half-life (22). Stimulation of the protein kinase C pathway in PC12 cells by phorbol esters regulates TH mRNA at the posttranscriptional level, in addition to an increase in transcription rate (39). However, prolongation of TH mRNA half-life was not measured in that study. It has also been reported that differentiation of mouse neuroblastoma cells with dimethyl sulfoxide (Me2SO) is accompanied by enhanced stability of TH mFWA (40). It was reported recently that stimulation of adrenal chromaflin cells with nicotinic receptor agonists was shown to affect TH mRNA at the posttranscriptional level (41). Although the mechanism that mediates the hypoxia-induced TH mRNA stability is unknown, preliminary results from our laboratory indicate that it may involve enhanced binding of a cytoplasmic protein to a specific sequence within the 3'-untranslated region of TH mRNA.
It is important to note that the time course for the increases in transcription and stability of TH mRNA during hypoxia are to some extent different. The increase in the rate of transcription is relatively fast with a peak that occurs at 6 h following the onset of hypoxia. The effect of hypoxia on TH mRNA stability, on the other hand, is much slower and was most evident during longer exposures (>12 h, a time at which the increase in transcription was less pronounced). We therefore speculate that different mechanisms are responsible for the early and late increases in TH gene expression during long term hypoxia. The increase in transcription of the TH gene is the major mechanism responsible for the enhancement of TH mRNA at the onset of hypoxia, whereas a combination of increased transcription and stabilization of TH mRNA is responsible for maintenance of the increased TH mRNA levels above control during longer exposures to hypoxia.
The signal transduction pathways that mediate the increases in transcription of the TH gene and stability of TH mRNA remain unknown. Preliminary results from our laboratory indicate that a heme protein may be involved in this pathway. However, it is important to realize that hypoxia is a very complex physiological stimulus that affects various intracellular metabolic pathways, which, in turn, might mediate different aspects of TH gene expression. In addition, the effects of hypoxia on processing and transport of the primary TH mRNA transcript are unknown. Such effects of hypoxia might decrease the efficiency of the transcription-mediated increase of mature mRNA and necessitate the need for increased stability to ensure an adequate level of TH enzyme during sustained hypoxia.
technical help. We are grateful to Dr. D. M. Chikaraishi for the TH-CAT