Protein Kinase Inhibitor H7 Blocks the Induction of Immediate-Early Genes zif268 and c-fosby a Mechanism Unrelated to Inhibition of Protein Kinase C but Possibly Related to Inhibition of Phosphorylation of RNA Polymerase II*

1-(5-Isoquinolinesulfonyl)-2-methylpiperazine (H7) has often been used in combination with protein kinase inhibitor (N-(2-guanidinoethyl)-5-isoquinolinesulfonamide) (HA1004) to assess the contribution of protein kinase C (PKC) to cellular processes, including the induction of gene expression. This use of H7 and HA1004 is based upon the fact that H7 inhibits PKC more potently than HA1004 in in vitro assays. Thus, although both compounds are broad spectrum protein kinase inhibitors, inhibition by H7, but not by HA1004, has often been interpreted as evidence for the involvement of PKC in the cellular process under study. Here we describe experiments that show that this interpretation is not correct with regard to the induction of two immediate-early genes,zif268 and c-fos, in PC12D cells. In these studies we confirmed that H7, but not HA1004, potently blocks the induction of zif268 and c-fos mRNA by nerve growth factor, carbachol, phorbol ester, Ca2+ionophore, or forskolin. Surprisingly, however, H7 has no effect on the ability of these agents to activate mitogen-activated protein kinase (MAPK), an upstream activator of zif268 and c-fos gene expression. H7 also does not inhibit preactivated MAPK in vitro. Taken together, these results suggest that H7 blocks gene expression by acting at a site downstream from MAPK. H7 has previously been shown to block transcription in vitro by blocking the phosphorylation of the carboxyl-terminal domain of RNA polymerase II (Yankulov, K., Yamashita, K., Roy, R., Egly, J.-M., and Bentley, D. L.(1995) J. Biol. Chem. 270, 23922–23925). In this study, we show that pretreating PC12D cells with H7, but not with HA1004, significantly reduces levels of phosphorylated RNA polymerase II in vivo. These results suggest that H7 blocks gene expression by inhibiting the phosphorylation of RNA polymerase II, a step required for progression from transcription initiation to mRNA chain elongation.

The immediate-early genes zif268 (also termed NGFI-A, egr-1, krox24, TIS8; reviewed in Ref. 1) and c-fos (2) encode transcription factors that have been proposed to function as "third messengers" in intracellular signal transduction cascades that convert information conveyed by extracellular stimuli into genomic responses that underlie growth, differentiation, and long term changes in the behavior of cells (3,4). We have previously shown that NGF 1 (1) and the carbachol (carbamylcholine) cause the rapid induction of zif268 mRNA in PC12D cells (5). Induction of zif268 mRNA by NGF is mediated by the high affinity NGF receptor, TrkA, which activates the Ras/MAPK cascade (6). Induction by carbachol is mediated by the m1 subtype of muscarinic acetylcholine receptor, which activates phospholipase C to produce the second messengers inositol 1,4,5-trisphosphate and diacylglycerol (5). Increased intracellular levels of inositol 1,4,5-trisphosphate trigger the release of Ca 2ϩ from internal stores, which in turn opens "capacitative influx" Ca 2ϩ channels in the cell membrane, resulting in a sustained influx of extracellular Ca 2ϩ . 2 Increased levels of diacylglycerol activate PKC. Both the sustained increase in intracellular Ca 2ϩ and the activation of PKC contribute to the induction of zif268 mRNA (5), at least in part by activating the MAPK cascade (81). Activation of the MAPK cascade is therefore a common element in the intracellular signaling events leading to gene expression that are initiated by NGF and carbachol in PC12D cells.
In the course of investigating the involvement of PKC in the induction of zif268 mRNA by NGF and carbachol, we compared the effects of pretreating PC12D cells with the protein kinase inhibitor H7 (7) with pretreatment of the cells with the related compound HA1004 (8). Both H7 and HA1004 are broad spectrum protein kinase inhibitors, but H7 inhibits PKC more potently than HA1004 (K i values ϭ 6 and 40 M for H7 and HA1004, respectively) in in vitro assays (7). Based upon this difference, many investigators have used these inhibitors in combination to evaluate the role of PKC in various cellular processes, including the induction of gene expression. In many 1 The abbreviations used are: NGF, nerve growth factor; PMA, phorbol 12-myristate, 13-acetate; PKC, protein kinase C; PKA, protein kinase A; MAPK, mitogen-activated protein kinase; SRE, serum response element; CRE, cAMP response element; CTD, carboxyl-terminal domain (of RNA polymerase II); TFIIH, transcription factor IIH; H7, [1-(5- of these studies, inhibition by H7 in the absence of inhibition by HA1004 was taken as evidence for a role for PKC in the process under investigation. The data presented in this paper, however, shows that inhibition of gene expression by H7 does not necessarily imply that PKC is involved. Rather, we found that although H7 potently inhibits the induction of zif268 and c-fos mRNAs following activation of PKC with phorbol ester, it fails to prevent activation of MAPK by phorbol ester. This shows that H7 can block the induction of gene expression without blocking PKC. Examination of the literature indicates that H7 blocks the induction of a broad spectrum of rapidly inducible genes by a variety of stimuli, including stimuli not previously associated with the activation of PKC. These observations suggest that H7 may block a site, different from PKC, that is universally required for the induction of rapidly inducible genes. A previous report that H7 blocks transcription in vitro by inhibiting the phosphorylation of RNA polymerase II by a TFIIH-associated kinase (9), led us to examine the effect of H7 on phosphorylation of RNA polymerase II in vivo. In the present study we show that pretreatment of PC12D cells with H7 significantly reduces levels of phosphorylated RNA polymerase II in vivo, suggesting that H7 blocks gene expression by directly inhibiting transcription.
Cell Culture-PC12D cells (12), a rapidly differentiating subline of rat pheochromocytoma-derived PC12 cells (13), were a gift from Mamoru Sano (Dept. of Biology, Faculty of Medicine, Kyoto Prefectural University of Medicine). PC12D cells were cultured in Dulbecco's modified Eagle's medium (DMEM, Nissui) supplemented with 5% fetal bovine serum, 5% horse serum, 0.16% sodium bicarbonate, 3.6 mM glutamine, 10 units/ml penicillin, 45 ng/ml streptomycin at 37°C under 5% CO 2 . Non-differentiated PC12D cells were used in all of the experiments. Unless noted, drugs were added directly to the culture medium and were present until the time at which the cells were harvested. The corresponding vehicle (water, Me 2 SO, or ethanol) was added to control cells.
MAPK Assays-MAPK assays were performed as described by Cook and McCormick (14) with some modifications. Briefly, PC12D cells grown to 80 -90% confluency in 3.5 cm in uncoated plastic culture dishes (Corning or Iwaki Glass Co.) were stimulated with various agents for 10 min and then lysed by addition of 200 l of lysis buffer containing 20 mM Tris-Cl (pH 8.0), 137 mM NaCl, 10% glycerol, 1% Triton X-100, 1.5 mM MgCl 2 , 1 mM EGTA, 50 mM NaF, 1 mM Na 3 VO 4 , 1 mM phenylmethanesulfonyl fluoride (PMSF), 20 M leupeptin, 10 g/ml aprotinin. After brief centrifugation to remove cellular debris, 0.1 g of anti-Erk-1 and 0.1 g of anti-Erk-2 antibodies were added to the supernatant fractions, and these were incubated for 1 h at 4°C with rotation to provide gentle mixing. Protein A-agarose (10 l of resin suspension, Santa Cruz Biotechnology catalog number sc-2001) was subsequently added to each sample and the incubation continued with rotation at 4°C for 1 h. The resin in each sample was then collected by centrifugation (2500 rpm) and washed twice with 200 l of lysis buffer and once with 200 l of 2ϫ reaction buffer. 1ϫ reaction buffer contained 25 mM MOPS (pH 7.2), 25 mM sodium ␤-glycerophosphate, 15 mM MgCl 2 , 1 mM EGTA, 0.1 mM NaF, 4 mM DTT, 1 mM Na 3 VO 4 . 22 l of reaction mix containing 25 M ATP, 1 Ci of [␥-32 P]ATP, 15 M myelin basic protein in 1ϫ reaction buffer was added to the resin, and the mixture was incubated at 30°C for 30 min. Reaction mixes were directly spotted on Whatman phosphocellulose filters, and filters were washed 6 times in 1% phosphoric acid for 5 min each wash. Radioactivity retained on the filters was quantified by liquid scintillation counting. During the course of this study, we determined that, unlike the anti-Erk-1 antibodies (Santa Cruz C-16), the anti-ErK-2 antibodies (Santa Cruz Biotechnology catalog number C-14) were not very effective in immunoprecipitating Erk-2 from cell lysates. 3 MAPK activities reported in this paper therefore reflect primarily Erk-1 activities. Qualitative assays of MAPK activation determined by measuring the activation-correlated shift-up in electrophoretic mobilities (15) showed, however, that Erk-1 and Erk-2 responded in the same manner to all treatments examined. 3 RNA Purification and Northern Blot Analysis-RNA was isolated from PC12D cells grown to 80 -90% confluency in 6-cm uncoated plastic dishes, and Northern analysis was carried out as described previously (5). The amount of RNA in each sample was determined by optical spectroscopy, and the integrity of the RNA in each sample was assessed by examining the ethidium bromide-stained RNA in denaturing gel used for Northern blot analysis. Unless noted, 10 g of total cellular RNA was electrophoresed in each lane. After blotting onto Pall Biodyne type B transfer membranes (0.45-m pore size), hybridization was carried out simultaneously using DNA probes prepared from zif268, c-fos (coding regions), and cyclophilin DNA fragments isolated from agarose gels and labeled using the Amersham Pharmacia Biotech Oligolabeling kit and [␣-32 P]dCTP. The intensities of bands in Northern blots were quantified using a Fuji Bioimaging analyzer BAS2000.
Plasmid Construction-An expression vector, pGLzif420, containing a firefly luciferase reporter gene linked to the rat zif268 promoter was constructed using pGL2 (Promega). Briefly, the zif268 promoter region (from Ϫ420 to 0 base pairs) containing 6 SRE sites and 2 CRE sites was amplified by polymerase chain reaction using synthetic oligonucleotide primers (forward primer corresponding to nucleic acid residues 121-149 of the rat zif268 promoter in the numbering system of Changelion et al. (Ref. 16; GenBank TM accession number J04154), 5Ј-AACAC-CATATAAGGAGCAGGAAGGATCCC-3Ј; backward primer containing a synthetic EcoRI site followed by nucleic acid residues 941-920 (14) of the rat zif268 gene, 5Ј-(GCGAATTC)TTGCTCAGCAGCATCATC-TCCT-3Ј). This polymerase chain reaction product was blunt-ended, digested with NruI, and the fragment containing the zif268 promoter sequences cloned into the blunt-ended HindIII site of pGL2. An expression vector containing the bacterial chloramphenicol acetyltransferase (CAT) gene under the control of the human elongation factor 1␣ promoter was constructed as follows: pBLCAT2 was digested with BamHI and BglII and then self-ligated. Digestion of the resulting plasmid with SalI and SmaI yielded a 1.5-kilobase pair fragment containing the CAT gene and a polyadenylation signal derived from SV40. This fragment was isolated, blunt-ended, and cloned into the XbaI site (after converting XbaI-cut ends to blunt ends) of pEF-BOS. The resulting vector, pEF-CAT, was used as an internal control in transfection experiments using zif268-luciferase expression vectors.
Transfection of PC12D Cells and Assay of Reporter Genes-Transfections were performed using LipofectAMINE TM reagent (Life Technologies, Inc.) essentially as recommended by the manufacturer. Cells were seeded in 6-cm plastic culture dishes (Corning or Iwaki Glass) at a density of 4 ϫ 10 6 cells/dish and cultured for 1 day prior to transfection. 0.92 g of pEF-CAT DNA, 2.3 g of luciferase expression vector DNA, 13.8 l of LipofectAMINE TM reagent were added to each dish of cells and incubated for 4 h, prior to adding the medium containing twice the normal concentration of serum. After incubation overnight, the cells in each 6-cm dish were resuspended and distributed into 12 ϫ 1.1-cm wells. The following day, the medium was replaced with normal DMEM, and the cells were cultured for 1 more day. Drugs were added directly to the culture medium, and cells were harvested after 4 h. Luciferase expression was carried out using the Promega Luciferase or Packard LucLite™, and luciferase activities were quantified using a Packard Tri-Carb or Top count scintillation counter as described in the manuals supplied by Promega and Packard. Background luciferase 3 E. Kumahara and D. Saffen, unpublished observations. expression was determined using cells transfected with pGL2, which lacks a promoter for luciferase gene expression. Transfection efficiency was determined by cotransfection with pEF-CAT. CAT activities were measured as described by Nordeen et al. (17), and these values were used to calculate normalized luciferase activities for each sample.
Assay of RNA Polymerase II Phosphorylation in Vivo-PC12D cells were labeled with 32 P in vivo as described (18). Briefly, PC12D cells grown to 80 -90% confluency in 10-cm culture plates were washed twice with 10 mM Tris (pH 7.0)-buffered saline (TBS) and overlaid with 5 ml of phosphate-free DMEM (Life Technologies, Inc.) supplemented with 25 mM HEPES, 5% dialyzed horse serum, and 5% dialyzed fetal bovine serum. 50 Ci of [ 32 P]orthophosphate was added to medium, and the cells were incubated at 37°C for 3 h in the absence of CO 2 . H7 (Seikagaku Kogyo) or other inhibitors were added 30 min prior stimulation with 5 ng/ml NGF for 15 min. Cells were washed twice with ice-cold phosphate-buffered saline supplemented with 0.2 mM PMSF, scraped into tubes, and centrifuged at 800 ϫ g for 5 min. Cells were resuspend in 1 ml of buffer containing 10 mM Tris-Cl (pH 7.9), 1 mM CaCl 2 , 1.5 mM MgCl 2 , 0.25 M sucrose, 0.2 mM PMSF, 0.5% Triton X-100, homogenized with Dounce homogenizer, and centrifuged at 800 ϫ g for 5 min. Nuclei pellets were washed once with 1 ml of the above buffer lacking Triton X-100. Pellets were resuspended in 100 l of buffer containing 25 mM Tris-Cl (pH 8.0), 2.5 mM magnesium acetate, 2 mM CaCl 2 , 0.05 mM EDTA, 0.1 mM DTT, 0.2 mM PMSF, and 12.5% glycerol and treated with 10 g of DNase and RNase on ice for 30 min. Nuclei were lysed by adding 100 l of 2% SDS, boiled for 3 min, and centrifuged at 15,600 ϫ g for 3 min to remove debri. Nuclei extracts (15 l/lane) were resolved by SDS-PAGE (5% polyacrylamide stacking gel containing 125 mM Tris-Cl (pH 6.8), 0.1% SDS; 5% polyacrylamide resolving gel (prepared from 29.5% acrylamide ϩ 0.5% bisacrylamide stock solution) containing 375 mM Tris-Cl (pH 8.8), 0.1% SDS; running buffer containing 25 mM Tris, 250 mM glycine, 0.1% SDS). Electrophoresis was carried out at 25 mA per gel until the tracking dye entered the resolving gel, after which the current was increased to 45 mA per gel. Proteins were then electrophoretically blotted (0.5 A, 45 min) onto polyvinylidene difluoride membranes (Immobilon TM transfer membrane, Millipore) using a Nihon Eido Western blotting apparatus (20 ϫ 20 cm) in buffer containing 100 mM Tris, 192 mM glycine, 10% methanol, and 0.02% SDS. Following transfer, the membranes were blocked by incubation in phosphatebuffered saline containing 5% skim milk and 0.5% Tween 20 overnight at room temperature. Phosphorylated proteins were detected by autoradiography using Fuji RX-U x-ray film. The membranes were then exposed to 0.02 g/ml antibodies that recognize the carboxyl-terminal domain (CTD) of RNA polymerase II (C-21, catalog number sc-900, Santa Cruz Biotechnology) in phosphate-buffered saline containing 0.5% skim milk and 0.05% Tween 20 for 2 h at room temperature, washed 3 times (10 min each) with buffer, and incubated in buffer containing anti-rabbit IgG antibodies cross-linked with horseradish peroxidase (Jackson ImmunoResearch, catalog number 111-035-003; 5000-fold final dilution) for 2 h at room temperature. After washing 3 times for 20 min/wash and 3 times for 10 min/wash with buffer, immune complexes were visualized by enhanced chemiluminescence (ECL kit, Amersham Pharmacia Biotech). Fig. 1, levels of both zif268 and c-fos mRNAs are low or undetectable in unstimulated cells but increase rapidly following exposure to NGF or carbachol, reaching high levels after 45 min. The fact that NGF-stimulated increases in zif268 and c-fos mRNA are blocked by the inhibitor K252a (19,20) indicates that these inductions require the activation of the tyrosine kinase of the high affinity NGF receptor, TrkA. Likewise, the ability of atropine to block the inductions of zif268 and c-fos mRNAs by carbachol indicates the involvement of muscarinic acetylcholine receptors. Pretreatment of the cells for 30 min with 100 M H7 completely blocks the induction of zif268 and c-fos mRNAs by NGF and carbachol. By contrast, pretreatment with 100 M HA1004 has essentially no effect on these inductions. Neither H7 nor HA1004 affect background levels of zif268 and c-fos mRNAs. 3 As shown in Fig. 2, H7 blocks the induction of zif268 and c-fos mRNAs in a dose-dependent manner, with complete inhibition observed at concentrations of 50 M and greater.

H7 Blocks the Induction of zif268 and c-fos mRNA by NGF and Carbachol-As shown in
The complete inhibition of zif268 and c-fos mRNA inductions by H7 is surprising for two reasons. First, although NGF has previously been reported to activate PKC in PC12 cells (21,22), the inductions of zif268 and c-fos mRNAs were not expected to require the activation of PKC, since down-regulation of PKC by prolonged exposure to phorbol ester has only a small effect on the induction of these mRNAs by NGF. 4 We have also previously observed (5) that induction of zif268 mRNA by NGF is not affected by the specific PKC inhibitor GF109203x (23,24). Second, although we have previously shown that PKC contributes to m 1 muscarinic acetylcholine receptor-mediated induction of zif268 mRNA in PC12D cells, this induction is only partially blocked by pretreatment with GF109203x (5). Thus, the total block of zif268 and c-fos gene induction by H7 seems to be too large an effect. To understand better how H7 blocks gene expression, we examined its effect on the induction of zif268 and c-fos mRNAs by additional agents.
H7 Also Blocks the Induction of zif268 and c-fos mRNAs by Phorbol Ester, Ca 2ϩ Ionophore, and Forskolin-Phorbol ester is expected to stimulate increases in immediate-early gene mRNA by activating PKC, and therefore this induction would be expected be inhibited by H7. By contrast, activation of immediate-early gene expression by elevated level of intracellular Ca 2ϩ in PC12 cells has not previously been suggested to require PKC. Rather, activation of Ca 2ϩ /calmodulin kinases (25,26) and/or activation of MAPK cascade (27,28) is thought to be sufficient. Similarly, forskolin, which increases intracellular levels of cAMP by stimulating adenylate cyclase, is thought to activate gene expression via activation of PKA (29) and/or the MAPK cascade (30 -33). The experiment depicted in Fig. 3, however, shows that H7 inhibits the induction of zif268 and c-fos mRNAs by each of these agents. These results strongly suggest that H7 blocks gene expression by acting at a site distinct from PKC.
MAPK Contributes to the Induction of zif268 and c-fos RNAs by NGF, Carbachol, Phorbol Ester, and Increases in Intracellular Ca 2ϩ or cAMP-MAPK is likely to play a central role in the induction of zif268 and c-fos by NGF and carbachol in PC12 cells. MAPK functions as part of the Ras/MAPK cascade (Ras-Raf-MEK-MAPK), which transmits signals from tyrosine kinase-and G protein-linked membrane receptors to the nucleus (6,34,35). Activated MAPK enters the nucleus where it phosphorylates and activates "ternary complex" transcription factors, e.g. Elk-1 or SAP-1, which stimulate gene expression by forming a complex with serum response factor bound to the serum response element (SRE) (36 -38). The c-fos promoter contains 1 SRE (2) and the zif268 promoter contains 6 SREs (16,39), which have been shown to play a role in the induction of zif268 mRNA by serum and NGF (16,40,41). In PC12 cells, MAPK has previously been shown to be strongly activated by NGF (42)(43)(44)(45), phorbol ester (44), and by elevated levels of intracellular Ca 2ϩ (27,28,46) or cAMP (30 -33). Together, these observations suggest that activation of MAPK plays a central role in the induction of c-fos and zif268 mRNAs by each of these agents in PC12D cells. Evidence supporting this inference is shown in the Northern blot depicted in Fig. 4, where induction of c-fos and zif268 RNAs by each of the agents tested was inhibited by pretreating PC12D cells with PD098059 (48, 49), a specific inhibitor of MEK, the immediate upstream activator of MAPK. Fig. 5 show that concentrations of H7 that block the induction of zif268 and c-fos mRNAs have no effect on the activation of MAPK. Pretreatment with HA1004 also does not block activation of MAPK. The inability of H7 to block the activation of MAPK by phorbol ester is particularly surprising, since H7 is expected to inhibit PKC. To determine if PKC can be effectively blocked in PC12D cells, we examined the effects of pretreatment with the specific PKC inhibitor GF109203x. The data in Fig. 6 show that pretreatment with GF109203x is effective in blocking both phorbol ester-mediated induction of zif268 and c-fos mRNAs and phorbol ester-mediated activation of MAPK. That phorbol ester activation of immediate-early gene expression and MAPK activation is mediated by PKC is supported by the observation that exposure to the same concentration of a less active phorbol ester, 4-␣-PMA, failed to increase zif268 and c-fos mRNA levels or activate MAPK. These results are also consistent with the conclusion that H7 blocks zif268 and c-fos gene expression by acting at a site distinct from PKC.

Concentrations of H7 Sufficient to Block the Induction of zif268 and c-fos mRNAs Do Not Block the Activation of MAPK-The data in
H7 Does Not Inhibit PKA in PC12D Cells-Although H7 has been reported to potently block PKA activity in vitro (K i ϭ 3 M) (7), the failure of H7 to block PKC in PC12D cells suggests that it may also not be effective in blocking PKA. To demonstrate that PKA in PC12D cells can be pharmacologically blocked, we pretreated the cells with the specific PKA inhibitor H89 (47). The results depicted in Fig. 7 show that pretreatment with 30 M H89 blocks both forskolin-stimulated induction of zif268 and c-fos mRNAs and forskolin-stimulated activation of MAPK. The specificity of the effects of forskolin is demonstrated by the lack of mRNA induction and MAPK activation by the related compound dideoxyforskolin, which does not activate adenylate cyclase. These results suggest that the ability of H7 to block the induction of c-fos and zif268 mRNA is unrelated to PKA.
H7 Does Not Inhibit Preactivated MAPK-The results presented so far show that although H7 potently inhibits the induction of zif268 and c-fos mRNAs, this is not caused by inhibition of PKC or PKA or by blocking the activation of MAPK. To determine if the block of mRNA induction is caused by inhibition of MAPK itself, we examined the ability of H7 to inhibit preactivated MAPK in vitro. As shown in Fig. 8, preactivated MAPK is essentially unaffected by H7, even at high 4 T. Ebihara and D. Saffen, unpublished observations. FIG. 4. MEK inhibitor PD098059 blocks the induction of zif268 and c-fos mRNAs by NGF, carbachol, PMA, thapsigargin,  concentrations of the inhibitor. Thus, it is unlikely that H7 blocks the induction of immediate-early gene RNAs by blocking MAPK. Rather, these results suggest that H7 acts at a site downstream from MAPK.
H7 Blocks NGF-mediated Induction of a Luciferase Reporter Gene Linked to zif268 Promoter-The data in Fig. 9 show that preincubation with H7, but not HA1004, blocks the induction of luciferase reporter gene linked to the zif268 promoter. By con-trast, GF109203x and H89 have only a small effect on the induction of the luciferase reporter by NGF. The ability of H7 to block the induction of gene expression mediated by the zif268 promoter in a heterologous system suggests that the mechanism of inhibition involves the functioning of the promoter rather than post-transcriptional controls such as transcriptional pausing, which is mediated by sequences in the first intron of the c-fos gene (50,51). The small effects of GF109203x and H89 on the expression of the reporter gene suggest that neither PKC nor PKA is essential for activation of the zif268 promoter by NGF.
H7 Inhibits the Phosphorylation of RNA Polymerase II in Vivo-The data presented up to this point show that H7 does not inhibit zif268 and c-fos gene expression by blocking the activation of MAPK or by directly inhibiting MAPK. Rather, the site at which H7 acts seems to be downstream from MAPK. One possibility we considered was that H7 blocks gene expression, not by inhibiting a step in the intracellular signaling cascade, but rather by inhibiting some step essential for transcription. A search of the literature for kinases known to be inhibited by H7 brought our attention to the general transcription factor TFIIH, which contains an associated kinase (cyclindependent kinase (cdk) MO15/Cdk-7) that phosphorylates the CTD of RNA polymerase II (9,(52)(53)(54). Phosphorylation of the CTD of RNA polymerase II is required for efficient RNA chain elongation (9, 54), and blocking CTD phosphorylation with H7 has been shown to inhibit transcription of human immunode-ficiency virus and c-myc mRNAs by RNA polymerase II in vitro (9). Based upon these results we decided to determine if H7 blocks the phosphorylation of RNA polymerase II in vivo. The experiment depicted in Fig. 10A shows that pretreatment of PC12D cells with H7 specifically reduces levels of phosphorylated RNA polymerase II, as evidenced by (i) a reduction in the intensity of 32 P-labeled RNA polymerase II detected by autoradiography (left panel, upper band within the bracket) and (ii) a reduction in levels of the "shifted-up" form of RNA polymerase II detected by immunochemical staining (right panel, upper band within the bracket). The slowly migrating, "shifted-up" form of RNA polymerase II has been previously shown to correspond to the phosphorylated form of the enzyme (18). By contrast, pretreatment of PC12D cells with HA1004 does not reduce levels of phosphorylated RNA polymerase II or change its electrophoretic mobility (Fig. 10, B and C). Inhibition of RNA polymerase II phosphorylation similar to that obtained with H7 is also observed following pretreatment of PC12D cells with two additional compounds known to block phosphorylation of the CTD of RNA polymerase II in vitro, the protein kinase inhibitor H8 (7,55) and the classic RNA polymerase II inhibitor DRB (9, 56) (Fig. 10B). By contrast, the MEK inhibitor PD098059 and broad spectrum protein kinase inhibitor staurosporine have no effect on phosphorylation of RNA polymerase II (Fig. 10B). Taken together, these data suggest that H7 blocks the induction of zif268 and c-fos mRNAs in vivo by inhibiting transcription elongation. DISCUSSION In this study we show that H7 blocks the induction of zif268 and c-fos mRNA in PC12D cells by NGF, carbachol, phorbol ester, and agents that increase intracellular Ca 2ϩ or cAMP (Figs 1-3 and 5) but fails to block the activation of MAPK, an enzyme that contributes to the induction of these mRNAs (Fig.  4). The inability of H7 to block the activation of MAPK by phorbol ester is surprising, since H7 is a potent inhibitor of PKC in vitro and is often used as a PKC inhibitor in in vivo experiments. Likewise, the inability of H7 to block the activation of MAPK by forskolin was not expected since H7 is also a potent inhibitor of PKA in vitro.
The inability of H7 to block PKC and PKA in vivo could be caused by an inability to efficiently penetrate the cells. Thus, H7 may not reach sufficient concentrations to block PKC or PKA, even though concentrations sufficient to block gene expression are attained. By contrast, the more potent PKC inhibitor GF109203x ( , fails to block the activation of MAPK or induction of zif268 and c-fos mRNA by forskolin in vivo suggests that HA1004 may also not efficiently permeate the cells. If it is true that H7 fails to block intracellular PKC and PKA because only a small amount enters the cells, then its ability to block mRNA induction may depend upon high affinity binding of H7. Inhibition by H7 apparently takes place very rapidly, since H7 was found to be equally effective at inhibiting the induction of c-fos and zif268 mRNAs when added 30 min before, at the same time, or up to 5 min after stimulation with NGF. 3 The fact that H7 does not effectively block PKC-and PKAmediated activation of MAPK in PC12D cells suggests that it inhibits the induction of zif268 and c-fos mRNAs by acting at a site unrelated to PKC or PKA. The observation that H7 also does not block preactivated MAPK (Fig. 8) suggests that H7 blocks the induction of zif268 and c-fos mRNAs by acting at a site downstream from MAPK or at a site unrelated to the FIG. 8. H7 does not inhibit preactivated MAPK in vitro. Activated forms of p42 and p44 MAPK were isolated by immunoprecipitation from PC12D cells that had been stimulated with water (W) or 5 ng/ml NGF for 5 min. The indicated concentrations of H7 or HA1004 were added to the standard in vitro MAPK assay reaction mix containing buffer, immunoprecipitated MAPK, [␥-32 P]ATP, and myelin basic protein (MBP). Reaction mixtures were incubated at 30°C for 30 min, and the incorporation of 32 P into myelin basic protein was determined by scintillation spectroscopy. H7, dark gray box; HA1004, dotted box. MAPK intracellular signaling cascade. Recent work has shown that MAPK activates two additional kinases that may be important for the induction of c-fos mRNA in PC12 cells. The serine/threonine kinases RSK-1 (p90 RSK ) and RSK-2 (CREB kinase) are activated upon phosphorylation by MAPK and then enter the nucleus and activate the transcription factors serum response factor (57) and CREB (58,59), respectively. Studies by Greenberg and co-workers (58) have shown that, although SRE and CRE function independently when linked to heterologous minimal promoters, both are required for activation of c-fos gene expression by the native c-fos promoter. Thus, it may be that activation of c-fos and zif268 gene expression requires the activation of RSK-1 and/or RSK-2 in addition to activation of MAPK. Significantly, Yin and Yang (60) have shown that H7 inhibits preactivated RSK-1 in vitro, although the effect of H7 on RSK-2 is still unknown. Taken together, the above observations suggest that H7 might inhibit gene expression by inhibiting RSK-1 and/or RSK-2.
Another possibility is that H7 interferes with gene expression by blocking the processing of mRNA. H7 has previously been reported to inhibit the splicing of c-fos mRNA in PC12 cells, as evidenced by an accumulation of prespliced c-fos mRNA in cells pretreated with high concentrations of H7 (61). (The "H7" used in that study was probably actually iso-H7 (10).) By contrast, we have never observed the accumulation of c-fos or zif268 precursor mRNAs in PC12D cells pretreated with authentic H7, and thus we have no evidence to suggest that H7 is blocking mRNA induction by inhibiting mRNA splicing. Nevertheless, our data do not rule out the possibility that H7 blocks gene expression by interfering with the processing of mRNA. The observation that H7 blocks the activation of the zif268 promoter linked to a luciferase reporter gene (Fig. 9), however, suggests that its effects are not restricted to the native zif268 gene.
A survey of published papers that use H7 to study gene expression reveals that H7 blocks the induction of a broad spectrum of rapidly inducible mRNAs by diverse stimuli. For example, H7 blocks the induction of mRNA by cytokines and lymphokines, including the induction of tumor necrosis factor mRNA by interleukin-1␤ (62) and the induction of immediateearly genes by interleukin-6 (60, 63-65), interleukin-1␤ (60, 66), ␣and ␥-interferons (67), leukemia inhibitory factor (60,68), and tumor necrosis factor (67). H7 also blocks mRNA induction by trophic/growth factors, including the induction of c-fos (69) and neuropeptide Y (70) mRNA by NGF, junB mRNA (71), and platelet-derived growth factor A and B chain mRNAs (72) by transforming growth factor (TGF)-␤, and c-fos mRNA by growth hormone (73). Additional examples, include the inhibition by H7 of the induction of c-fos and c-jun mRNAs by H 2 O 2 (74) and the induction of metallothionein mRNA induction by Cd 2ϩ or Zn 2ϩ (75).
In many studies using H7, inhibition of gene expression has been taken as evidence for the involvement of PKC. Whereas PKC certainly functions in the induction of some genes, there  (74) reported that down-regulation of PKC by long-term exposure to phorbol ester failed to block the gene expression under study, even though H7 was effective. Authors of other studies have also concluded that H7 blocks inducible gene expression by blocking a kinase distinct from PKC (60,66).
In a few studies, H7 had no effect on gene expression or, instead, stimulated gene expression. In most of these cases, the lack of inhibition can in retrospect be accounted for by the use of iso-H7, rather than authentic H7 (76,77), or very low concentrations of authentic H7 (78,79). In almost every study to date, authentic H7 used at concentrations between 50 and 100 M has been found to block gene induction. Together, these results suggest that H7 may be inhibiting some step that is generally required for the induction genes.
An attractive hypothesis to explain the breadth of the inhibitory activity of H7 is that it affects some aspect of the general transcriptional machinery, rather than the kinase cascades that activate specific transcription factors. Consistent with this idea is the observation that H7 blocks the kinase associated with the general transcription factor TFIIH, which phosphorylates the carboxyl-terminal domain of RNA polymerase II (9,52). TFIIH participates in the late stages of transcription initiation, and the associated kinase is required for efficient elongation mRNA by RNA polymerase II (9,52). Inhibition of TFIIH kinase with H7, in fact, blocks transcriptional elongation of human immunodeficiency virus and c-myc RNAs by RNA polymerase II in vitro (9).
Based upon these observations, we decided to determine whether pretreatment of PC12D cells with H7 affects the phosphorylation of RNA polymerase II. The results depicted in Fig.   10 show that a 30-min exposure to H7 is sufficient to reduce significantly levels of phosphorylated RNA polymerase II in both unstimulated cells and cells stimulated with NGF. By contrast, pretreatment of the cells with HA1004 has no effect on phosphorylation of RNA polymerase II. To our knowledge, this is the first demonstration that H7 inhibits the phosphorylation of RNA polymerase in vivo, although H7 has previously been shown to block c-myc transcriptional elongation in vivo (80). Together, these results suggest that H7 blocks the induction of gene expression by inhibiting elongation of RNA transcripts. A scheme indicating the site at which H7 inhibits gene expression is depicted in Fig. 11. Further work will be required to prove this model, but the data in this study provide a plausible explanation for the general inhibition of gene expression by H7 and suggest that caution must be taken when interpreting the effects of H7 on inducible gene expression. FIG. 11. Working model, H7 blocks transcription of c-fos and zif268 mRNAs by inhibiting phosphorylation of RNA polymerase II. PKC and PKA function upstream of MAPK, which activate gene expression by phosphorylating ternary complex factor. Ternary complex factor stimulates transcription by forming a ternary complex with the transcription factor serum response factor, which is prebound to the SRE. There is 1 SRE in the c-fos promoter and 6 SRE in the zif268 promoter (shown in diagram). H7 blocks elongation of nascent mRNAs by inhibiting phosphorylation of RNA polymerase II.