Stimulus-transcription coupling in pheochromocytoma cells. Promoter region-specific activation of chromogranin a biosynthesis.

To explore stimulus-transcription coupling in pheochromocytoma cells, we studied the biosynthetic response of chromogranin A, the major soluble protein co-stored and co-released with catecholamines, to chromaffin cells' physiologic nicotinic cholinergic secretory stimulation. Chromogranin A mRNA showed a time-dependent 3.87-fold response to nicotinic stimulation, and a nuclear run-off experiment indicated that the response occurred at a transcriptional level. Transfected chromogranin A promoter/luciferase reporter constructs were activated by nicotinic stimulation, in time- and dose-dependent fashions, in both rat PC12 pheochromocytoma cells and bovine chromaffin cells. Cholinergic subtype agents indicated that nicotinic stimulation was required. Promoter deletions established both positive and negative nicotinic response domains. Transfer of candidate promoter domains to a heterologous (thymidine kinase) promoter conferred region-specific nicotinic responses onto that promoter. A proximal promoter domain (from −93 to −62 base pairs) was activated in copy number- and distance-dependent fashion, and thus displayed features of a promoter element. Its activation was sufficient to account for the overall positive response to nicotine. Within this proximal region, a cAMP response element (CRE) was implicated as a major nicotinic response element, since a CRE point-gap mutation decreased nicotinic induction, transfer of CRE to a thymidine kinase promoter augmented the promoter's response to nicotine, and nicotine activated the CRE-binding protein CREB through phosphorylation at serine 133. We conclude that secretory stimulation of pheochromocytoma cells also activates the biosynthesis of the major secreted protein (chromogranin A), that the activation is transcriptional, and that a small proximal domain, including the CRE box, is, at least in part, both necessary and sufficient to account for the positive response to nicotine.

To explore stimulus-transcription coupling in pheochromocytoma cells, we studied the biosynthetic response of chromogranin A, the major soluble protein co-stored and co-released with catecholamines, to chromaffin cells' physiologic nicotinic cholinergic secretory stimulation. Chromogranin A mRNA showed a time-dependent 3.87-fold response to nicotinic stimulation, and a nuclear run-off experiment indicated that the response occurred at a transcriptional level. Transfected chromogranin A promoter/luciferase reporter constructs were activated by nicotinic stimulation, in timeand dose-dependent fashions, in both rat PC12 pheochromocytoma cells and bovine chromaffin cells. Cholinergic subtype agents indicated that nicotinic stimulation was required. Promoter deletions established both positive and negative nicotinic response domains. Transfer of candidate promoter domains to a heterologous (thymidine kinase) promoter conferred region-specific nicotinic responses onto that promoter. A proximal promoter domain (from ؊93 to ؊62 base pairs) was activated in copy number-and distance-dependent fashion, and thus displayed features of a promoter element. Its activation was sufficient to account for the overall positive response to nicotine. Within this proximal region, a cAMP response element (CRE) was implicated as a major nicotinic response element, since a CRE pointgap mutation decreased nicotinic induction, transfer of CRE to a thymidine kinase promoter augmented the promoter's response to nicotine, and nicotine activated the CRE-binding protein CREB through phosphorylation at serine 133. We conclude that secretory stimulation of pheochromocytoma cells also activates the biosynthesis of the major secreted protein (chromogranin A), that the activation is transcriptional, and that a small proximal domain, including the CRE box, is, at least in part, both necessary and sufficient to account for the positive response to nicotine.
When chromaffin cells and sympathetic axons discharge their vesicular stores of peptides and catecholamines after secretory stimuli, how do they resynthesize the released components? Does the same signal that causes secretion also trigger transcriptional activation of secretory peptide genes?
Since the quantitatively major soluble protein co-stored and co-released by exocytosis with catecholamines is chromogranin A (6, 7), we studied the response of the chromogranin A gene to activation of PC12 pheochromocytoma cell secretion by nicotinic cholinergic stimulation (4,5,8,9). Our results suggest that chromogranin A biosynthesis is specifically activated by nicotinic receptor-subtype cholinergic stimulation, that the response occurs at a transcriptional level, and that multiple promoter regions contribute to the overall pattern of transcriptional activation.

MATERIALS AND METHODS
Cell Culture-Primary cultures of bovine chromaffin cells were harvested and incubated following the method described previously (7).
The majority of these studies were performed in PC12 rat pheochromocytoma cells (10), obtained from David Schubert, Salk Institute, La Jolla, CA. They were cultured in high-glucose Dulbecco's modification of Eagle's medium with 10% heat-inactivated horse serum, 5% heat-inactivated fetal bovine serum, and penicillin/streptomycin. Cell passage number (since initiation of the line (10)) was between 10 and 25 in these experiments.
Chromogranin A Promoter/Luciferase Reporter Plasmids-Plasmid pHJLD5, in which a functional 1133-bp 1 mouse chromogranin A promoter drives expression of a luciferase reporter, has been described (11). The mouse chromogranin A promoter region in this plasmid extends from Ϫ1133 bp upstream of the transcription initiation (cap) site, to ϩ42 bp downstream of the cap site. A series of 5Ј deletions (12) of this 1133-bp promoter, also fused to luciferase, was prepared by restriction digestion or by polymerase chain reaction. All promoter fragments had the same 3Ј terminus. Correct orientation of the deletion mutants was confirmed by asymmetric double restriction digest, or by DNA sequencing through crucial fragment boundaries. Each fragment prepared by polymerase chain reaction was verified by dideoxy sequencing and studied as two independent clones.
Several domains of the mouse chromogranin A promoter were prepared by restriction digestion or polymerase chain reaction, and ligated either just upstream of the heterologous herpes simplex virus thymidine kinase (TK) promoter, or in the AatI site, 1.2 kbp away from the promoter, in the TK promoter/luciferase reporter plasmid pTK-LUC, as described previously (12). Orientation and multiplicity of promoter domain inserts were determined as described previously (12). The mouse chromogranin A promoter's CRE domain (TGACGTAA) (12) as well as a consensus CRE (TGACGTCA) (13) were also synthesized (both strands) and ligated just upstream of the TK promoter in pTK-LUC. A point-gap mutation (mutant M41) in the chromogranin A promoter's CRE region, abolishing the CRE consensus (to TGA-GTAA; gap (indicated by hyphen) induced by mutation), was created (by a mutant polymerase chain reaction primer) in a Ϫ77 to ϩ42 bp chromogranin A promoter/luciferase reporter (12), and verified by dideoxy sequencing. A 1.2-kbp rat chromogranin A promoter, fused to the luciferase reporter, was also prepared by screening and selecting rat genomic cosmid clones in the vector sCos-1 (11).
A 726-bp human chromogranin A promoter/chloroamphenicol acetyltransferase (CAT) reporter plasmid (pCAT-726), whose promoter fragment (spanning residues -726 bp to ϩ14 bp) had been inserted between the HindIII and PstI sites of the pCAT-BASIC reporter vector (Promega, Madison, WI), was obtained from Lee Helman, National Cancer Institute, Bethesda, MD (14). A CRE box containing somatostatin promoter (Ϫ71 bp to ϩ55 bp)/luciferase reporter plasmid was obtained from Marc Montminy, Salk Institute, La Jolla, CA. Its CRE box (TGACGTCA) was at Ϫ41 to Ϫ48 bp.
Transfections and Chromogranin A Promoter Expression-One day before transfection, PC12 cells were split onto poly-D-lysine-(Sigma) coated 6-cm plastic plates at 40 -50% confluence. Cells were transfected with 2.5-10 g of supercoiled luciferase reporter plasmid DNA per plate, using the efficient lipofection (16) method (Lipofectin ® , Life Technologies, Inc., 5 g/ml). Supercoiled plasmids were twice banded through gradients of cesium chloride in ethidium bromide. When the reporter or transactivator plasmids did not total the requisite amount of DNA, the balance was composed of supercoiled pBluescript (Stratagene, La Jolla, CA).
To control for transfection efficiency differences between plasmids, some transfections were accompanied by co-transfection of another reporter plasmid, pRSV-CAT (17), expressing the CAT reporter driven by the strong Rous sarcoma virus (RSV) promoter. In a series of transfections of pRSV-CAT into PC12 cells, in the absence (n ϭ 48) or presence (n ϭ 48) of nicotine (10 Ϫ3 M), neither total cell protein (71.3 Ϯ 1.04 versus 71.0 Ϯ 1.47 g/plate) nor CAT reporter activity (6.12 Ϯ 0.132 versus 6.45 Ϯ 0.137 cpm of 14 C-acetylated chloramphenicol/g of protein) was changed by nicotine.
Bovine chromaffin cells (1.5 ϫ 10 6 cells/ml) were also transfected by lipofection (plasmid DNA and Lipofectin ® concentrations both 5 g/ml) while in suspension just after harvest by collagenase digestion (7), then plated at 1 ml/well (in 6-well plates) and changed to serum-containing medium the next morning. On the third morning, nicotine (10 Ϫ3 M) or mock (vehicle) stimulation was begun and continued for the next 48 h.
In a time course experiment, nicotine (10 Ϫ3 M) was added to transfected PC12 cells (all transfected at t ϭ minus 48 h before cell harvest) at a series of times after transfection but before harvest (time 0). The control condition was established without nicotine addition at any time.
RNA Extraction and Northern Blots-Total RNA was prepared from PC12 cells by RNA Zol II (Ref. 20, TEL-TEST Inc., Friendswood, TX). 20 g of total cellular RNA were fractionated on formaldehyde-agarose gels and transferred to a nylon filter. The integrity of the RNA was judged by the appearance of 28 S and 18 S rRNA bands on the ethidium bromide-stained gel. Northern blots were done as described previously (21).
Hybridizations and washes were done at 65°C. Between probes, filters were stripped by boiling for 30 min in 0.1 ϫ SSC, 0.5% SDS, prior to rehybridization.
Transcriptional Activation (Nuclear Run-on Assay)-This method was modified from an established protocol (27). After stimulation by nicotine or vehicle for different time periods, PC12 cells were suspended in phosphate-buffered saline and lysed by adding Nonidet P-40 buffer (10 mM Tris, pH 7.4, 10 mM NaCl, 3 mM MgCl 2 , 1% Nonidet P-40). Lysed cells were kept on ice for 5 min and centrifuged for 5 min at 500 ϫ g. The nuclear pellet was washed once with Nonidet P-40 lysis buffer, then resuspended in 200 l of glycerol storage buffer (50 mM Tris, pH 8.3, 40% glycerol, 5 mM MgCl 2 , 0.1 mM EDTA), and frozen in liquid N 2 . To begin the run-on reaction, frozen nuclei (200 l) were thawed, and 200 l of 2 ϫ reaction buffer (10 mM Tris, pH 8, 5 mM MgCl 2 , 0.3 M KCl), 10 l of 100 mM rATP, 10 l of 100 mM rCTP, 10 l of 100 mM rGTP, 5 l of 1 M dithiothreitol, and 10 l of [␣-32 P]rUTP (800 Ci/mmol, 10 mCi/ml; DuPont NEN) were added immediately to each milliliter of thawed nuclear extract. After 30 min of incubation at 30°C with shaking, newly radiolabeled hnRNA was isolated by adding 1 ml of RNA Zol II method. Samples were votexed, incubated on ice for at least 15 min, and microcentrifuged for 15 min at 4°C. After isopropyl alcohol precipitation of the aqueous phase, the pellet was dispersed in 200 l of TES buffer (10 mM TES, pH 7.4, 10 mM EDTA, 0.2% SDS). Unincorporated label was removed by NucTrap probe purification column (Stratagene La Jolla, CA), followed by a 100-l TES wash. Total radioactivity of each extraction was ϳ10 7 cpm. The hnRNA solution was mixed with an equal volume of TES/NaCl buffer (10 mM TES, 10 mM EDTA, 0.2% SDS, 0.6 M NaCl), and then 1 l of the mixture from each extraction was used for Cerenkov counting of 32 P. A vial from each nuclear extract was counted, and RNA buffer (see above) was added as needed to achieve similar (within 5%) concentrations of radiolabeled hnRNA (counts/min/l) in each reaction. The 3.3-kbp rat chromogranin A genomic DNA probe (28) spanned a region from 260 bp upstream of the cap site to 3.04-kbp downstream from the cap site. This genomic DNA fragment included exon 1, exon 2, intron A, and part of intron B. Mouse cyclophilin cDNA (23) was used as a control (housekeeping transcript) probe. Target cloned DNA fragments (5 g of either the rat chromogranin A genomic DNA fragment, or the cyclophilin cDNA) were immobilized on strips of nitrocellulose, using vacuum filtration on a slot-blot apparatus. The nitrocellulose strips were then placed in the vials with newly labeled hnRNA, and then incubated in a Hybaid oven (National Labnet Inc., Woodbridge, NJ) for 24 h at 65°C. After hybridization, filters were washed twice with 2 ϫ SSC for 1 h at 65°C, then incubated with 2 ϫ SSC containing 10 g/ml RNase A (Boehringer Mannheim) for 30 min at 37°C, followed by washing once more with 2 ϫ SSC for 1 h at 37°C. Filters were exposed to x-ray film (Kodak X-Omat AR5, with intensifier screen, at Ϫ70°C), and radioactive bands were excised for liquid scintillation counting of probe-immobilized 32 P.
Statistics-Results are reported as the mean value Ϯ 1 S.E. Data were analyzed by either t test (two groups) or analysis of variance (3 or more groups), using the software packages Statworks on Macintosh. Differences were considered significant if p Ͻ 0.05.

Response of Chromogranin A mRNA and Chromogranin A Transcription to Nicotinic Cholinergic Stimulation-Exposure
of PC12 pheochromocytoma chromaffin cells to nicotine caused a time-dependent augmentation of the chromogranin A mRNA, with a maximal (3.87-fold) increase at 16 h (Fig. 1).
The chromogranin A mRNA response to nicotine occurred at the level of transcriptional activation: a nuclear run-on (Fig. 2) demonstrated a newly transcribed chromogranin A heterogeneous nuclear RNA (hnRNA) response to nicotinic stimulation within 6 h, with maximal (2.23-fold) response by 16 h. The time course for transcriptional activation of chromogranin A (Fig. 3) paralleled the time course for accumulation of chromogranin A mRNA by chromaffin cells (Fig. 1).
Time Course of Responses of Chromogranin A, Chromogranin B, and Tyrosine Hydroxylase mRNAs to Nicotinic Stimulation-Nicotinic cholinergic stimulation of PC12 pheochromocytoma cells caused time-dependent augmentation of mRNAs for not only chromogranin A, but also chromogranin B (up to 2.13-fold), and tyrosine hydroxylase (up to 1.80-fold). For both chromogranin A and chromogranin B, mRNA signals were stronger at 24 than at 6 h of nicotinic stimulation; by contrast, the tyrosine hydroxylase mRNA response was greatest at 6 h, returning to basal levels after 24 h of stimulation (Fig. 4).
Transactivation of a Transfected Chromogranin A Promoter by Nicotinic Cholinergic Stimulation-In PC12 pheochromocytoma cells, an 1133-bp mouse chromogranin A promoter/luciferase reporter construct was transfected, along with pRSV-CAT as a transfection efficiency control. Nicotinic stimulation transactivated this promoter, and the degree of stimulation was dependent on dose of reporter plasmid transfected, being maximal for plasmid doses of 2.5 g (or less) of supercoiled DNA per 6-cm cell culture plate (Fig. 5). This result suggested that very high transfected doses of the DNA element responding to nicotine in cis eventually saturate and deplete the factors in trans which mediate the nicotinic response. In primary cul- Chromogranin A promoter/luciferase reporter constructs containing promoters from mouse, rat, or human chromogranin A genomic clones were similarly transactivated by nicotine in PC12 cells: 3.02-fold for mouse, 2.69-fold for rat, and 1.98-fold for human (Table I). Subsequent transfection experiments were done with mouse chromogranin A promoter/luciferase reporter constructs in PC12 cells. A time course of transfected mouse chromogranin A promoter exposure to nicotine (Fig. 3) indicated that the transcriptional response was detectable by 8 h of exposure, and was maximal after 40 h. In subsequent transfection experiments, cells were typically harvested 48 h after transfection.
Dose-response relationships for nicotine effects in PC12 cells on both the transfected chromogranin A promoter and catecholamine secretion are shown in Fig. 6. The transfected promoter's response to nicotine was maximal at 10 Ϫ3 M, while the catecholamine secretory response was maximal at a lower dose of 10 Ϫ4 M.
When both nicotinic and muscarinic cholinergic agonists were tested in PC12 cells on chromogranin A transcription, nicotine more effectively activated transcription than muscarine (3.06-versus 1.39-fold; Table II); furthermore, the nicotinic effect was more completely blocked by a nicotinic antagonist (hexamethonium; 90% inhibition), than by a muscarinic antagonist (atropine; 31% inhibition). The L-type voltage-gated calcium channel antagonist verapamil virtually abolished nicotinic activation of the transfected chromogranin A promoter (Table III). Chromogranin A Promoter Regions Responding to Nicotinic Stimulation-The nicotinic response was tested on a series of transfected chromogranin A promoter deletion/luciferase reporter constructs in PC12 pheochromocytoma cells (Fig. 7). The 1133-bp promoter showed 2.9-fold stimulation by nicotine, while a 425-bp promoter showed an even greater (4.1-fold) nicotinic response. Deletion of two promoter regions caused substantial decrements in the nicotine response: Ϫ425 to Ϫ328 bp, and Ϫ147 to Ϫ43 bp. Deletion beyond (3Ј of) Ϫ43 bp (which is just 5Ј of the TATA box at Ϫ22 bp) the nicotinic response was abolished. Deletion of two other promoter regions caused increments in the nicotine response: Ϫ818 to Ϫ425 bp and Ϫ181 to Ϫ147 bp.
When several large promoter domains were transferred to a heterologous TK promoter (Table IV), they generally transferred responses to nicotine qualitatively similar to their con-

TABLE I Activation of mouse 1133 bp or rat 1.2-kbp chromogranin A (CgA) promoter/luciferase (LUC) reporter constructs, or a 726-bp human CgA promoter/chloramphenicol acetyltransferase (CAT) reporter construct, by nicotinic cholinergic stimulation (nicotine, 1 mM, 48 h) in PC12 cells
Results are the mean value Ϯ 1 S.E. (n ϭ 4 transfections/condition). To correct for differences in transfection efficiency, a control co-transfection was performed, with a plasmid expressing a different reporter (CAT or LUC) under the control of a strong, constitutively active promoter (the Rous sarcoma virus (RSV) long terminal repeat). Units are ratios of luciferase activity (integrated light units) and CAT activity (cpm incorporated into 14 C-acetylated chloramphenicol). Species tributions (positive or negative effect of nicotine) within the entire chromogranin A promoter (Fig. 7). Regions Ϫ93 to Ϫ62 bp, Ϫ128 to Ϫ99 bp, and Ϫ159 to Ϫ128 bp all responded positively to nicotine. The positive response of region Ϫ93 to Ϫ62 bp was copy number-dependent, and was lost at a distance of 1.2 kbp. The 2.72-fold nicotine response of this Ϫ93 to Ϫ62 bp region was similar in magnitude to the 3-fold response of the entire promoter, either in the endogenous gene ( Figs. 1 and 2) or in the transfected promoter/reporter construct (Table I, Fig. 7). Since the proximal positive response region (Fig. 7) contains a CRE homology ((Ϫ71 bp)-TGACGTAA-(Ϫ64 bp)), we tested the effect of a CRE box point-gap mutation (TGA-GTAA) in a transfected 77-bp chromogranin A promoter/luciferase reporter construct (Table V). The nicotinic induction ratio (nicotine stimulation/mock stimulation) fell from 2.58-fold in the wildtype promoter down to 1.64-fold (i.e. a 60% fall, p Ͻ 0.05) in the promoter with the CRE box point-gap mutation.
We then tested whether a CRE box alone could transfer nicotinic responsiveness to a heterologous (TK) promoter (in pTK-LUC). The chromogranin A CRE (TGACGTAA) insert conferred 3.97-fold nicotinic induction (Table VI), while a consensus CRE (TGACGTCA) was induced 4.94-fold by nicotine; each of these nicotinic responses was significantly (p Ͻ 0.05) greater than that of the original TK promoter vector (Table VI). The nicotinic induction was blunted (down to 1.31-fold) by pointgap mutation of the CRE box (TGA-GTAA). Even a CRE box (TGACGTCA) in the transfected promoter of a gene (somatostatin) not ordinarily expressed in chromaffin cells responded 2.03-fold (p Ͻ 0.05) to nicotine (from 134 Ϯ 10.1 to 272 Ϯ 19.6 luciferase light units/mg protein; n ϭ 4 transfected wells/condition).
In electrophoretic gel mobility shift assays, we tested the effect of PC12 nicotine exposure (at a dose (10 Ϫ3 M) and time (16 h) which stimulated the transfected promoter (Figs. 3 and 6)) on nuclear protein binding to four promoter regions: three that responded positively to nicotine during transfection (Ϫ432 to Ϫ312, Ϫ147 to Ϫ77, and Ϫ71 to Ϫ64 bp (the CRE box)), and one that seemed to respond negatively (Ϫ181 to Ϫ147 bp). None of these regions showed any significant change in band retardation pattern after nicotine (Fig. 8).
Because CREB-bound CRE boxes are only transcriptionally transactivated when CREB itself has been activated by phosphorylation at serine 133 (33), we undertook antibody supershift assays (Fig. 9), using not only an antibody directed against CREB (epitope: the CREB kinase-inducible domain), but also an antibody which specifically recognized CREB only in its activated, phosphorylated form (pS133-CREB). Cell exposure to nicotine did not alter the anti-CREB supershift by PC12 nuclear proteins, but did result in a qualitatively new anti-pS133-CREB supershift band. Competition experiments with a 100-fold molar excess of unlabeled CRE (Fig. 9) indicated that both the anti-CREB and the anti-pS133-CREB supershifts represented CRE-binding complexes. DISCUSSION When chromaffin cells secrete the complex mixture of components in their secretory granule cores by exocytosis, is resynthesis of those components, both amines and peptides, somehow coupled to the exocytotic event? In particular, does the stimulus which causes exocytosis (a nicotinic cholinergic agonist (6)) also trigger resynthesis of just-secreted granule core components, and does this stimulation occur at a transcriptional level?
To address these questions, we focused on regulation of the biosynthesis of chromogranin A, the major soluble protein in chromaffin granule cores (7). Nicotinic stimulation augmented the chromogranin A message in PC12 pheochromocytoma cells (Fig. 1), and a nuclear run-off (Fig. 2) confirmed a nicotinic effect on the rate of initiation of new chromogranin A transcripts.
After nicotinic stimulation, chromogranin A, chromogranin B, and tyrosine hydroxylase transcripts accumulated in PC12 cells, indicating that biosynthesis of multiple peptides, as well as catecholamines, responded to the nicotinic signal. However, accumulation of these 3 transcripts by PC12 cells after nicotine followed different time courses (Fig. 3), suggesting different controls on the biosynthesis or degradation of the messages.
Transactivation of a transfected chromogranin A promoter (Figs. 3 and 5) by nicotine is further evidence for transcriptional activation; transactivation of mouse, rat, and human transfected chromogranin A promoters (Table I) suggests that the response is of general importance, rather than simply species-specific. Nicotinic and muscarinic cholinergic agonist and antagonist studies (Table II) confirmed specific or preferential participation of the nicotinic cholinergic receptor in transcriptional stimulation.
Both the secretory and transcriptional responses to nicotine were biphasic, with a decline in response at the highest dose employed (Fig. 6); such a pattern is typical of nicotinic responses, in which a decline in response at highest agonist doses may reflect densitization of the receptor (34). A maximal re- a Significantly (p Ͻ 0.05) different from the mock (no agonist) value for the same antagonist (same row). b Significantly (p Ͻ 0.05) different from the control (no antagonist) value for the same agonist (same column). Results are mean value Ϯ 1 S.E. (n ϭ four determinations/condition).

TABLE III
Effect of blockade of L-type voltage-gated cell surface calcium channels on nicotinic activation of chromogranin A transcription PC12 cells were transfected with the mouse 1133-bp chromogranin A promoter/luciferase reporter (see Table II  sponse to 100 M nicotine is typical for the secretory effect of nicotine in chromaffin cells (35); by contrast, the transcriptional effects of nicotine upon the transfected CgA promoter were maximal at a higher dose of 1 mM. It is worthwhile to note that the time course of the catecholamine secretory response to nicotinic exposure is well under 30 min (6), while nicotinic transactivation of the transfected promoter/reporter plasmids is progressive over 48 h; further time-dependent receptor desensitization is perhaps predictable during this very long period of nicotine exposure, thereby increasing the dose necessary for a maximal effect on transcription in this setting.
A series of promoter deletions partly localized transcriptional elements in cis responding to nicotinic activation (Fig. 7). Promoter domains responding positively to nicotine (that is, domains whose deletion diminished the nicotinic response) included Ϫ425 to Ϫ328 bp, Ϫ147 to Ϫ77 bp, and Ϫ77 to Ϫ43 bp; the domain from Ϫ181 to Ϫ147 bp responded in an apparent negative way to nicotine (i.e. its deletion augmented the nicotinic response). Several large promoter domains could also transfer their nicotinic induction onto a heterologous promoter (Table IV); thus, such domains were apparently sufficient to mediate nicotinic induction.
Since the CRE element in the chromogranin A promoter is conserved across all species characterized (mouse (12), rat (28), bovine (36), and human (37)), and since it is also crucial in transcriptional responses of other genes, such as tyrosine hydroxylase (38,39), to secretory stimuli in chromaffin cells, we evaluated the chromogranin A CRE's ((Ϫ71 bp) 5Ј-TGACG-TAA-3Ј (Ϫ64 bp)) role in the nicotinic response, both by pointgap mutation (to TGA-GTAA) and by transfer to a heterologous TK promoter. Each experiment gave evidence of a substantial role for the CRE element in the nicotinic response, the CRE point mutation reduced the response by 60% (p Ͻ 0.05), while the mouse chromogranin A CRE domain transfer augmented the response by 1.85-fold (ϭ3.97/2.15) over basal (Table VI), and even a CRE box in the transfected promoter of a gene (somatostatin) not ordinarily expressed in chromaffin cells was stimulated 2.03-fold (p Ͻ 0.05) by nicotine. However, it should be noted that the CRE element is apparently not the sole site mediating the nicotinic response, since its mutation did not completely abolish nicotinic activation, nor did its transfer in isolation completely reproduce the 3-4-fold nicotinic activation of the full chromogranin A promoter (Figs. 1, 2, 4 -7 and Tables  I and II).
Even the TK promoter alone, without full CRE oligonucleotide inserts (Tables IV and VI), displayed a modest (1.46 -2.15fold) response to nicotinic stimulation; it is worth noting that

TABLE VI CRE (cAMP-response element) domain transfers
Effect of several CRE-related oligonucleotide domains on the response of a heterologous (thymidine kinase (TK)) promoter to nicotinic stimulation. Each domain was inserted into the luciferase reporter vector pTK-LUC, just 5Ј of (adjacent to) the TK promoter. Transfected PC12 pheochromocytoma cells were exposed to nicotine (10 Ϫ3 M) or vehicle (mock), and harvested after 48 h for luciferase reporter assay. Values shown are luciferase light units/mg protein, mean Ϯ 1 S.E. for n ϭ 4 transfections (plates). Ratio, each nicotine stimulated value in a group/mean unstimulated value in this group.  this TK promoter element (which spans Ϫ100 bp to ϩ52 bp in the herpes simplex virus TK promoter) contains two CRE "halfsites" (GTCA) at Ϫ89 bp (plus strand) and Ϫ15 bp (minus strand) (40). When the nicotine-responsive promoter domains (Fig. 7) were tested for interactions in vitro with PC12 cell nuclear proteins (Fig. 8), each DNA domain was bound by one or more factors on electrophoretic gel mobility shift assay, but none showed qualitatively new shifted bands after nicotine. A qualitatively new change after nicotine, phosphorylation of the CRE-binding protein CREB on serine 133, was revealed by specific antibody supershifts (Fig. 9); thus, nicotine induced post-translational activation of a particular trans-acting factor complementary to the crucial CRE domain in cis.
Simon et al. (41) have shown that exposure to nicotine augments the biosynthesis of new chromogranin A protein in chromaffin cells. Our results indicate that this nicotinic response has a number of characteristic features: it occurs at the level of transcription (Fig. 2); is general across species (Table I); is specifically mediated by the nicotinic-subtype cholinergic receptor (Table II); is not temporally coupled to the transcriptional response of the catecholamine biosynthetic enzyme tyrosine hydroxylase (Fig. 5); maps to local domains in cis on the chromogranin A promoter (Fig. 7, Table VI), including the CRE box; and activates the CRE binding factor CREB in trans.
Past studies of chromogranin A biosynthesis after secretory stimulation in chromaffin cells have yielded variable results: while Simon et al. (41) found nicotinic stimulation of [ 35 S]methionine incorporation into newly biosynthesized chromogranin A, several reports suggest that secretory stimuli have only minimal effects on the chromogranin A steady-state mRNA or immunoreactive protein in PC12 pheochromocytoma (42) or bovine chromaffin cells (43)(44)(45). Indeed, prior findings of "constitutive" regulation of chromogranin A biosynthesis contributed to the hypothesis that chromogranin A may have a role in secretory granule biogenesis (46). How do our studies differ? First, our studies focused directly on transcription, by evaluating not only steady-state chromogranin A mRNA levels ( Fig. 1) but also the rate of initiation of new chromogranin A transcripts (Fig. 3) and trans-activation of the transfected chromogranin A promoter (Figs. 5-7). Second, we examined a broad spectrum of nicotine dose effects on both transcription and secretion (Fig. 6). Finally, we used early passage number (pas- FIG. 8. Electrophoretic gel mobility shift assays of the interactions of different 32 P-labeled chromogranin A promoter regions with PC12 pheochromocytoma cell nuclear proteins. Promoter domains were chosen based on positive or negative responses to nicotinic stimulation (see Fig. 7). In each group, lane 1 is the free probe (whose mobility is indicated by an asterisk (*)) incubated with bovine serum albumin alone; lane 2 is the band incubated with and shifted by a PC12 control (cells untreated) nuclear extract; and lane 3 is the band incubated with and shifted by a PC12 nuclear extract after treatment of the cells with nicotine (10 Ϫ3 M, 16 h). No significant differences (new shifted bands) were found between nicotine and control stimulation for any promoter region.
FIG. 9. Interaction of the chromogranin A promoter's CRE (cAMP response element, (؊71 bp)-TGACG-TAA-(؊64 bp)) box with PC12 cell nuclear proteins: electrophoretic gel mobility shift and antibody supershift assays, before and after exposure of PC12 cells to nicotine. Synthetic CRE box oligonucleotides (annealed plus and minus strands) were 32 P-labeled by fill-in, and then incubated with PC12 nuclear proteins harvested from cells treated with nicotine (10 Ϫ3 M, 16 h) or vehicle. Antibodies were directed against either the kinase-inducible domain of the rat CRE-binding protein CREB, or more specifically against phosphoserine residue 133 in CREB (pS-133-CREB). Specificity of nuclear protein binding to the CRE in shifted and supershifted bands is shown from displacement of the labeled probe by a 100-fold molar excess of unlabeled probe. sage 8 through 25) PC12 cells, and perhaps avoided the problem of PC12 phenotype de-differentiation at high passage number (47).
The chromogranin A steady-state mRNA level (Fig. 1) and rate of initiation of new transcripts (Fig. 3) responded over rather prolonged ϳ16-h time courses to nicotinic stimulation; this is perhaps not surprising, since chromogranin A is a quantitatively unusual protein-it may account for ϳ7% of total chromaffin cell protein (48), its transcripts may represent ϳ5% of total chromaffin cell mRNA (49), and its t1 ⁄2 in chromaffin cells is ϳ79 h (50).
Transcriptional activation of the tyrosine hydroxylase gene by nicotinic stimulation has been studied in cultured chromaffin cells, as well as during trans-synaptic induction in vivo (1)(2)(3); in different studies, the nicotinic response maps to either an AP1 site (2, 3) or a CRE site (38,39) in the tyrosine hydroxylase promoter. The mouse chromogranin A promoter has partial matches for both sites. The mouse chromogranin A CRE ((Ϫ71 bp) 5Ј-TGACGTAA-3Ј (Ϫ64 bp)) homology, which has a 7/8-bp match with the CRE consensus TGACGTCA (13), lies in the proximal nicotinic response positive domain (Fig. 7), while an AP1 ((Ϫ418 bp] 5Ј-TGAGTCAG-3Ј (Ϫ425 bp)) partial (6/8-bp match) homology with an AP1 consensus TGAGTCAG (51) is in the distal positive domain (Fig. 7). Whether the distal AP1 homology specifically participates in the chromogranin A transcriptional response to nicotinic stimulation can only be resolved by additional mutagenesis.
Other chromaffin cell-expressed genes besides chromogranin A display up-regulation of mRNA after stimulation of secretion. Chromaffin granule soluble components whose mRNAs respond to secretory stimuli include chromogranin B (Fig. 4), proenkephalin A, and carboxypeptidase H (5). Catecholamine biosynthetic enzymes that respond to such stimuli include tyrosine hydroxylase (1-3, 38, 39), phenylethanolamine-N-methyltransferase (9), and dopamine-␤-hydroxylase (52). In some of these cases (38,39), the promoters' CRE boxes are implicated, but whether transcriptional responses of these several nicotinic-inducible genes are mediated by the same or different signal transduction pathways remains to be fully explored.