| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
Max Planck Institute of Psychiatry, Munich D-80804, Germany
Address all correspondence and requests for reprints to: Theo Rein, Max Planck Institute for Psychiatry, Kraepelinstrasse 10, Munich D-80804, Germany. E-mail: theorein{at}mpipsykl.mpg.de.
 |
ABSTRACT |
|---|
 TOP ABSTRACT INTRODUCTIONRESULTSDISCUSSIONMATERIALS AND METHODSREFERENCES |
|---|
|
INTRODUCTION |
|---|
|
TOPABSTRACTINTRODUCTIONRESULTSDISCUSSIONMATERIALS AND METHODSREFERENCES |
|---|
B (15, 16). It is a long-standing observation that glucocorticoids exert antiproliferative effects in most cellular contexts. Conversely, transactivation of GR-responsive promoters is presumed to be cell cycle-dependent since more than 30 yr. One of the first observations was a cell cycle-dependent induction of tyrosine aminotransferase (TAT) by glucocorticoids in hepatoma cells (17). While cells were glucocorticoid-responsive during the G1 and S cell cycle phases, cells in G2 as well as in M were reported to be completely resistant to glucocorticoids. Moreover, induction of endogenous alkaline phosphatase (18) or later on of epidermal growth factor receptors (19) in HeLa cells was shown to be most effective in late G1 and S-phase with an apparent lack of glucocorticoid responsiveness during G2/M and early G1. More recently, it was demonstrated that in fibroblasts synchronized in G2 the stably transfected mouse mammary tumor-virus (MMTV) promoter as well as the endogenous metallothionein-I (MT-I) promoter could not be activated by glucocorticoids (20). However, in accordance with the hypothesis of specific glucocorticoid resistance in G2 the MT-I promoter was still inducible by heavy metals in G2. Remarkably, inhibition of GR seemed to be confined to transactivation, because transrepression by GR was not affected in G2 cells (21).
Despite several observations relating to cell cycle-dependent activity of GR, including different hormone binding during the cell cycle (19, 22, 23) or decreased nuclear translocation of GR in G2 (19, 20), the molecular mechanisms leading to G2 silencing of GR function remained largely unclear. It has been speculated that differential phosphorylation of GR throughout the cell cycle (24, 25) might contribute to or account for cell cycle-dependent function of GR. Indeed, rat GR was shown to be a target for cyclin-dependent kinases and mitogen activated protein kinases in vitro (26). However, using site-specific mutations of phosphorylation sites of GR, it was not possible to identify a distinct phosphorylation pattern of GR that would lead to complete silencing of the receptor (27, 28). Interestingly, some functional consequences of GR phosphorylation were found, but these effects turned out to be promoter specific because they were apparent only at simple promoters containing just one to three GREs, but not at complex promoters like the MMTV promoter (29, 30, 31). Complex promoters are able to recruit additional cofactors, which themselves might be cell cycle-dependently regulated (32, 33, 34).
With the aim to identify molecular mechanisms explaining glucocorticoid resistance in G2, we sought to establish an experimental model to measure cell cycle-dependent glucocorticoid resistance. We tested cell cycle-dependent transactivation of endogenous as well as exogenous glucocorticoid-sensitive promoters, stably or transiently transfected in various cell lines. To our surprise, we found no silencing of GR function in G2 at all. Furthermore, mitotic repression of GR-induced transcription apparently is due to general chromatin condensation, and not to specific inactivation of GR.
|
RESULTS |
|---|
|
TOPABSTRACTINTRODUCTIONRESULTSDISCUSSIONMATERIALS AND METHODSREFERENCES |
|---|
). In contrast, cells collected after 1.5–4.5 h showed minimal incorporation of 3H-thymidine and, therefore, were regarded as cells in G2 between time zero and 1.5–4.5 h thereafter. These G2 cells showed only a slight, nonsignificant decrease of TAT induction by DEX compared with asynchronously proliferating cells stimulated under the same conditions in the presence of colcemid (Fig. 1).
|
, but not displayed).
The slight decrease in inducibility in G2 cells might either indicate an only partial repression of GR reactivity or, more likely, reflect simply the fact that these cells were synchronized in mitosis before preparation. To address this issue, H4-II-E-C3 cells were synchronized in G2 by incubation with nocodazole after 48 h of serum deprivation. For synchronization analysis of a population grown in parallel under the same conditions, mitotic cells were shaken off and discarded before cell harvesting to eliminate metaphase cells and to obtain cells synchronized in G2. Fluorescence- activated cell sorter (FACS) analysis confirmed synchronization of H4-II-E-C3 cells by nocodazole in G2 to about 77% (Fig. 2B) compared with about 14% cells in G2 within the asynchronously proliferating control population (Fig. 2A). Likewise, mitotic cells were collected by synchronization with colcemid and shaking off. FACS analysis showed more than 95% tetraploid mitotic cells in these preparations (Fig. 2C). After synchronization, cells were stimulated with DEX (5 nM) for 8 h, mitotic cells were discarded and the remaining cells harvested for determination of TAT activity. Asynchronously proliferating cells revealed 5.6-fold TAT activity compared with nonstimulated control cells (Fig. 2D). Cells synchronized in G2 showed almost the same TAT activity after DEX-stimulation as asynchronous cells but a slightly diminished inducibility (4.9-fold over basal). This is mainly due to a slightly increased TAT activity of nonstimulated cells in G2 (1.2-fold compared with asynchronous cells, Fig. 2D). In mitotic cells TAT induction by DEX was almost completely abolished (1.2-fold over basal); this is in accordance with the expected repression of DNA transcription during mitosis. These data suggest that there is only little reduction of GR inducibility in H4-II-E-C3 cells in G2, in stark contrast to data in the literature that indicate an almost complete inhibition (17).
|
shows the MMTV-driven GFP expression correlated with the DNA content of asynchronously proliferating HT-22-GFP cells before stimulation (Fig. 3A) and after stimulation with 1 µM DEX (Fig. 3, B and C). Stimulation with 1 µM DEX for 8 h resulted in an about 4-fold increase of mean GFP fluorescence (Fig. 3B). By gating the population of interest, we determined the DEX-induced increase of mean GFP fluorescence within the G1, S and G2 gate, respectively. With the inducibility of cells within the G1 gate set as 100%, we calculated the relative increase of GFP fluorescence of the S and G2 gate relative to that in G1. Figure 3D shows that there is no difference in glucocorticoid responsivity between the three populations after stimulation with DEX for 8 h. The fact that cells in G2 after the 8-h period of DEX-induction were in the S-phase at the beginning of the induction could obscure a potential G2 effect, if the reporter gene product was stable for at least 8 h. Therefore, we also performed DEX-inductions for only 3 h (Fig. 3C). The increase of mean GFP intensity was about 2-fold, but again stimulation was uniform for G1, S, and G2 (Fig. 3, C and D).
|
) and stimulated cell populations (Fig. 3, B and C). This increase can be explained either by a continuous accumulation of reporter between each cell division or by an increase in GR number in G2, which could give rise to an increase of both, ligand-dependent and -independent activity. We also noted an increase in the proportion of G1 cells after 8 h of incubation with DEX (Fig. 3B). This is consistent with an extended G1 phase, probably due to some unknown effect of DEX in this cell cycle phase. It is unlikely that this would affect our measurement of G2 cells. G1 cells also appear readily stimulated by DEX. Moreover, we did not observe significant changes in growth behavior of these cells upon incubation with DEX for 1–3 d (not shown), suggesting a transient effect of DEX on the G1 population. Taken together, our data of the MMTV promoter in nonsynchronized cells provide no support for transcriptional inactivity of GR in G2.
Transactivation of the MMTV Promoter in Cells Synchronized in G2
To address the possibility that a potential silencing of GR in G2 was obscured even at the short DEX-induction time of 3 h we set out to verify our data with the MMTV promoter by using G2 synchronizing drugs. HT-22-GFP cells were subjected to serum starvation (0.5% FCS) for 48 h followed by at least 18 h release in medium containing (10% FCS) in the absence (control) or presence of either nocodazole or taxol. Aliquots were taken to confirm synchronization using FACS analysis. Mitotic cells were discarded before as described above. Cells were treated with 1 µM DEX vs. ethanol for 8 h and then processed for measurement of MMTV-driven GFP expression by FACS. Treatment with nocodazole resulted in synchronization in G2 of on average 86% of the cell population (Fig. 4B) compared with 12% cells in G2 in the absence of synchronizing agent (Fig. 4A). With taxol, 80% of the cells were synchronized in G2 (Fig. 4C). By gating of either the entire cell population or cells in G2 (Fig. 4, A–C) we determined the cell cycle-dependent transactivation of the MMTV promoter. It is obvious from Fig. 4D that also in these G2 cell preparation there is no reduced response to DEX. The small, nonsignificant increase in stimulated G2 cells as compared with asynchronous cells is similarly observed in the GFP fluorescence of nonstimulated cells: nocodazole and taxol arrested cells each display a 1.5-fold higher reporter level. This is also reflected in a virtually unchanged stimulation after DEX (4.37 ± 0.34-fold for asynchronous cells, 3.86 ± 0.21-fold for nocodazole-arrested cells, and 3.54 ± 0.21-fold for taxol-arrested cells; data used for Fig. 4, but not displayed).
|
GR Activity in Mitosis on Transient Templates
Two explanations are provided, in general, for the nonspecific repression of DNA transcription in mitosis (36): condensation of chromatin or modification of distinct transcription factors (37, 38, 39). We were interested to find out what mechanism may apply to GR-dependent transcription.
First, we issued whether GR-mediated transcription is impaired during mitosis. We decided to use the MT-I promoter in HT-22 cells, because this endogenously active promoter is inducible by glucocorticoids and by heavy metals. This allowed us to check GR-independent inducibility of the promoter. Moreover, this promoter also was used to show GR silence in G2 (20). Therefore, we additionally checked for transactivation of MT-I by DEX in G2. HT-22 cells were synchronized in G2 using nocodazole or in mitosis using colcemid as described above. After synchronization, cells were stimulated either with 1 µM DEX or 5 µM CdCl2 for 8 h. Transcriptional activity of the MT-I promoter was determined using quantitative RT-PCR. The amount of amplified MT-I mRNA (280 bp) was normalized to the intensity of the corresponding ß-actin mRNA. Colcemid treatment resulted in over 95% accumulation of tetraploid cells (Fig. 5A). Transcription from the MT-I promoter in asynchronously proliferating cells increased on average 3.34-fold ± 0.78 (n = 5) over basal activity by DEX (representative gel in Fig. 5B, lane 1 and 2) and on average 5.47-fold ± 1.60 (n = 5) by CdCl2 (lane 3). Cells synchronized in G2 showed no significantly altered MT-I transcription neither by DEX nor by CdCl2 (DEX 2.43-fold ± 1.36, n = 5, and CdCl2 5.49-fold ± 1.79, n = 5, respectively, Fig. 5B, lanes 4–6). Cells synchronized in mitosis showed a negligible induction of MT-I by DEX, but still a noticeable induction by CdCl2 (DEX 1.15-fold ± 0.38 and CdCl2 3.28-fold ± 1.60, both n = 5 activity over basal value; Fig. 5B, lanes 7–9). Noninduced cells show no difference in the amount of MT-I mRNA before induction (G2 cells 1.07-fold ± 0.19 compared with asynchronous, M cells 0.99 ± 0.54; n = 5). These data confirm that cells in G2 are not glucocorticoid resistant, whereas induction of transcription by glucocorticoids is repressed in mitotic cells, probably through chromatin condensation.
|
Cells were synchronized in G2 or M. During the synchronization procedure, HT-22-GFP cells were transiently cotransfected with pMTV-Luc and pCMV (cytomegalovirus)-ß-galactosidase (gal). While in G2 both MMTV templates were inducible, in mitosis the transiently transfected MMTV promoter was inducible and the stably integrated one was silent (Fig. 6). The reporter activities of noninduced cells showed no significant difference (compare black bars). These data show that mitotic repression of GR-dependent transcription is likely due to chromatin condensation rather than to modification (e.g. by phosphorylation) of GR or any cofactor of GR.
|
). Noninduced G2 cells again showed a small, nonsignificant increase in reporter activity (1.3-fold for HT22 cells and 2-fold for H4-II-E-C3 cells, data used for Fig. 7, but not shown). This confirms that phosphorylation of GR has at least no direct effect on transactivation of glucocorticoid sensitive promoters. Moreover, GR is active in G2 on both complex and simple promoters.
|
Therefore, we began by testing the efficiency of HOE to synchronize our cells in G2. It first should be noted that HT-22-GFP cells we had used for this study were not efficiently synchronized in G2 by treatment with HOE (data not shown). The only cell line that in our hands readily synchronized in G2 after HOE treatment was Chinese hamster ovary (CHO)-TRex (Fig. 8A). These cells are a commercially available derivative of CHO cells, which have been reported to synchronize in G2 by treatment with HOE (49). After pretreatment with hydroxyurea and 18 h incubation with HOE, DEX-induced transactivation of the transiently transfected MMTV promoter was indeed not detectable in these cells (Fig. 8B). However, DEX response was not affected in G2 CHO-TRex cells synchronized by nocodazole (Fig. 8, A and B). To explore the differences between the synchronization procedures, we treated cells with different concentrations of HOE for 12 h without the preceding hydroxyurea incubation. Again, DEX response was not measurable at a concentration of 6 µg/ml HOE (Fig. 9A). However, FACS analysis revealed that these cells were not synchronized in G2 (Fig. 9B). This strongly suggested that the inhibition observed with HOE treatment is not due to synchronization in G2, but to some cell cycle-independent effect of HOE. Paradoxically, while at low concentrations of HOE there is still inducibility of the MMTV promoter, we observed a significant increase of the nonstimulated MMTV activity at 1 and 2 µg/ml of HOE (Fig. 9A).
|
|
), suggesting a somewhat specific effect of HOE on GR-dependent promoters.
To corroborate these findings, HT-22-GFP cells were coincubated with DEX and increasing concentrations of HOE ranging from 0.25 up to 4 µg/ml for 8 h. Although the cell cycle of these cells was not affected by HOE (Fig. 10A), induction of the MMTV promoter by DEX was dose dependently inhibited (Fig. 10B). In contrast to the transiently transfected MMTV promoter in CHO-Trex cells, the transcription from the stably integrated MMTV promoter in the absence of DEX was not significantly increased in HT-22-GFP cells (Fig. 10C). However, when we transiently transfected these cells with the MTV-Luc construct, again the promoter activity was significantly enhanced by HOE in the absence of DEX (Fig. 10D).
|
), transcription from transient templates is enhanced (Figs. 9A and 10D). With concentrations of HOE as high as in previous reports [7.5 µg/ml (20)] MMTV promoter activity was always down, with or without DEX and independent of synchronization (not shown). |
DISCUSSION |
|---|
|
TOPABSTRACTINTRODUCTIONRESULTSDISCUSSIONMATERIALS AND METHODSREFERENCES |
|---|
However, from our extensive studies of four different promoters in three different cell types, we conclude that GR is active throughout the entire cell cycle, thereby completely changing our view of GR.
Inactivity of GR in G2 cells was first reported in 1969 using the TAT promoter in rat hepatoma cells (17). In an effort to reproduce this previous report directly, we investigated the activity of the TAT promoter in H4-II-E-C3 rat hepatoma cells. While there was clear repression of TAT induction in mitosis (Fig. 2D), in G2 cells we found no difference in the TAT activity after induction with dexamethasone and only a marginal reduction of TAT inducibility compared with asynchronously proliferating cells (Figs. 1 and 2). This slight decrease of TAT inducibility in G2 cannot be explained by nonspecific effects of cytoskeleton disrupting agents like colcemid or nocodazol, because we found that they did not affect GR function neither in asynchronously proliferating H4-II-E-C3 cells nor in HT-22 cells (data not shown). This is in accordance with previous reports regarding nuclear translocation of GR (8, 50). We cannot rule out the possibility that this slight decrease reflects a partial impairment of GR in G2 in H4-II-E-C3 cells. However, we consider it more likely that this minor reduction is explained by the higher proportion of cells in mitosis in this population as compared with randomly proliferating cells.
We can only speculate why the previous report (17) is in contrast to our observations: for example, the TAT promoter is regulated not only by glucocorticoids via a GRE (51, 52), but also by other factors like cAMP (53, 54), or liver-specific hepatocyte nuclear factors (55, 56). Activation of TAT transcription by the cAMP pathway has been reported to be differentially sensitive in different liver-derived cell lines downstream of protein kinase A (57). Therefore, it is possible that cell type-specific and cell cycle-dependent variations in the cross-talk between these different signal transduction pathways may finally lead to impaired transactivation of the TAT promoter in G2. This would also mean that GR function itself is not directly linked to the cell cycle. An alternative, though less likely, explanation would be that it can be difficult to avoid contaminations of G2 cell preparations with mitotic cells in the way G2 cells have been prepared (17).
To either corroborate or disprove our conclusions, we used the MMTV promoter as another model system because it represents one of the best-studied GR-dependent promoters and because it has been used to show GR silence in G2 before (20, 21). We created HT-22 cells stably transfected with an MMTV-driven GFP gene to be able to directly correlate GR-dependent transactivation with the cell cycle for each cell in either synchronized cells or unsynchronized, randomly proliferating cells. A similar methodical approach was successfully used to show that heat shock protein 70 is cell cycle-dependently induced after heat stress in a limited number of cell lines (58). In our HT-22-GFP cells, FACS analysis revealed no cell cycle dependence of DEX-inducibility of GFP expression (Fig. 3). We noted, however, that cells with or without induction with DEX displayed an increasing amount of reporter product as they moved from G1 to S and to G2. We cannot completely rule out that this reflects a more efficient hormone binding and transactivational activity of GR during S-phase (17, 18, 19, 23), which may seemingly extend into G2, if GFP protein is stable. However, because the increase is independent of induction with DEX, the more likely explanation is that it is due to a steady accumulation of GFP between each cell division. Similarly, it is possible that the substantially greater mean variation of GFP expression of cells in G1 reflects differential DEX-responsiveness between early G1 cells and late G1 cells as suggested elsewhere (25). However, we consider it more likely that the higher variation in G1 is due to the higher proportion of G1 cells in the total population, i.e. cells spend more time in G1 which means that cells late in G1 had more time to accumulate GFP. The difference of our results to the data describing silencing of GR in G2 (20, 21) cannot be explained by the fact that GR silence previously was found in synchronized cells, because we also used nocodazole or taxol to synchronize HT-22-GFP cells in G2 and found no GR silencing at all.
Similarly, promoter specificity is unlikely to be the cue for explaining the difference between our data and those of others, because we also tested the endogenously expressed MT-I promoter and a simple promoter construct, again with no evidence for GR silencing in G2. The MT-I promoter was also used to show impaired GR function in G2 (20). The simple promoter containing only one GRE was of particular interest, because on the one hand GR is differentially phosphorylated during G1/S and during G2/M (24, 25) and on the other hand mutations of certain phosphorylation sites of GR affects transactivation only with simple glucocorticoid responsive promoters (25, 29, 30, 31). Thus, one would have predicted that a cell cycle-dependent activity of GR would be detectable only with a simple promoter construct, if differential phosphorylation was causal for it. Because we found no cell cycle dependence of GR function also on our simple promoter, we postulate that the cell cycle-dependent phosphorylation of GR has no effect on its ability to transactivate.
The discrepancy of our results to those obtained previously is, most likely, explained by the use of HOE in previous studies to synchronize cells in G2 (20, 21). We demonstrate that HOE itself interferes with GR-dependent transcription. The bisbenzimide HOE was reported to be suitable for reversible cell cycle synchronization of CHO cells (49, 59) and, more recently, of primary cultured porcine fibroblasts (60). CHO cells (49) and L-fibroblasts (20) apparently synchronized well in G2 at a concentration of HOE of 7.5 µg/ml. While our CHO-TRex cells readily synchronized in G2 at a concentration of HOE of 6 µg/ml, HT-22-GFP cells failed to synchronize efficiently by HOE treatment, even after presynchronization in S-phase by hydroxyurea (data not shown). Importantly, we found that HOE interfered with GR-mediated transactivation of MMTV in HT-22-GFP cells and CHO-TRex cells even in the absence of cell synchronization (Figs. 9 and 10). Strikingly, while HOE inhibited transcription from a stably integrated MMTV promoter, the same promoter transiently transfected was activated even in the absence of DEX (Fig. 10). Although we did not pursue this phenomenon further, we conclude that the use of HOE is not recommendable for assaying GR function.
It is well described that prolonged exposure to HOE is toxic to several cell types at nanomolar (48, 61) to micromolar concentrations (62). Using the MTT assay, we found no cytotoxicity after 8 or 12 h incubation of HT-22-GFP and CHO-TRex cells with HOE (not shown), but determined a half-maximal cytotoxic concentration of HOE for HT-22-GFP cells of about 3 µg/ml after 32 h of exposure (data not shown). While we observed inhibition of DEX-response already at nontoxic concentrations, one might speculate that cells are nevertheless already determined for an apoptotic process. This process could involve p53, which functionally inhibits GR function (63). Another potential mechanism explaining inhibition of DEX-induced transcription involves interference of HOE with DNA binding of transcription factors, which has been shown at several examples (45, 46, 47, 64). It may be for these reasons that HOE seems not to be used as the standard agent for cell synchronization in G2 (65, 66).
While the use of HOE can serve as a possible explanation for the difference between our conclusions and those using HOE, we can only speculate about other studies (17, 18, 19). We discussed Ref. 17 above. Citation (19) reports on the induction of EGF binding by DEX treatment throughout the cell cycle. Unfortunately, samples have been taken at the S/G2 transition and at the G2/M transition, but none clearly in G2 (Fig. 2 of Ref. 19). Because G2 and M cells were not differentiated, one might speculate that the cell preparation contained a significant fraction of M-phase cells. In addition, it cannot be excluded that EGF binding is down-regulated by ways other than reduced transcription. The work by Griffin and Ber (18) has been cited as support for inactivity of GR in G2 in numerous publications (e.g. Refs. 25 , 67, 68, 69, 70). However, although the authors analyzed induction of alkaline phosphatase by hydrocortisone for 48 h after mitotic cell selection, they did not analyze cells in the G2 phase. Inactivity in G2 may have been inferred from the 20 h lag period of induction when cells are exposed to DEX 12 h after mitosis (see Fig. 2 in Ref. 18). However, this is a rather indirect conclusion and the authors themselves make no claim about G2, they actually never mention G2 (18).
In most experiments, we observed a small, albeit nonsignificant increase in the amount of GR-dependent reporter in G2. When observed for each cell (Fig. 3), this may be explained by a steady increase in reporter between each cell division (see above). When normalized by total protein or ß-gal activity, this increase may reflect the reported increase of GR binding sites in S and G2 as compared with G1 (19), an explanation that would also apply to measurements per cell, of course. From Western blotting of protein extracts of asynchronous or G2/M-synchronized HT-22-GFP cells, we actually have preliminary evidence for a small increase in GR protein in G2 and M-phase (data not shown). However, given the about equally small increase in ligand-independent reporter activity in G2 and unchanged inducibility, we consider it unlikely that the increase in GR number in G2 compensates for a possible reduction in GR activity per se.
While we found no inhibition of GR-dependent transcription in G2, transcription of the endogenous MT-I promoter and the stably integrated MMTV promoter was not inducible by DEX in mitotic cells. With the endogenous MT-I promoter, we observed in mitosis only a reduced, but still evident inducibility by CdCl2. While we do not know the reason for this partial activity even during mitosis, we note that during mitosis DNase I hypersensitive sites can persist (38) and even TFIID-promoter complexes remain stable (71). Therefore, it may be possible that CdCl2, in contrast to DEX, activates factors that are able to activate the MT-I promoter preassembled with TFIID complexes even in mitosis.
In contrast to the stably integrated, chromosomal template, the MMTV promoter transiently transfected clearly was inducible by DEX even in mitosis. This model system of stable and transient cotransfection of the MMTV promoter at the same time has been successfully employed to elucidate the critical relevance of nucleosomal organization for GR-mediated transcription (42, 72, 73, 74, 75). We exploited this model system here for the first time to shed light on the mechanism of mitotic inhibition of GR-dependent transcription. The hypotheses put forward to explain transcriptional repression during mitosis fall into two general categories: hypotheses using chromosome condensation and hypotheses using inactivation of transcription factors as explanation (36, 37, 38, 39). Our findings clearly show that GR itself is not inhibited in mitotic cells, and therefore, repression of GR-dependent transcription in mitosis must be due to the nonspecific effect of chromatin condensation. Further, our results exclude the hypothesis that there is bulk inactivation of the basal transcriptional machinery during mitosis.
In summary, we propose a model of cell cycle-dependent activity of GR, in which GR-dependent transcription is silenced only during mitosis. This silencing, however, is due to chromatin condensation rather than to inactivation of GR itself. Therefore, chromatin-independent functions of GR are not suppressed in mitosis, e.g. transcription from viral templates or, possibly, DNA-independent activities of GR. Our model leaves open the possibility that certain GR-responsive promoters are regulated in a cell cycle-dependent manner due to cell cycle-dependent activity of promoter-specific cofactors.
|
MATERIALS AND METHODS |
|---|
|
TOPABSTRACTINTRODUCTIONRESULTSDISCUSSIONMATERIALS AND METHODSREFERENCES |
|---|
Plasmids and Transfection
The plasmid pRK5MLuc was obtained by cloning the PvuII-BamHI fragment from pMTV-Luc (76) encompassing the MMTV-LTR, the firefly luciferase structural gene and the SV40 polyA signal into the SpeI-AccI vector fragment from pRK5SV40PUR (77) containing the puromycin resistance gene. The Acc65I site was blunted and the SpeI site was ligated to the BamHI site via a linker. The plasmid pMTV-GFP was cloned by replacing the luciferase structural gene of pRK5Mluc (excised with XhoI and Acc65I) with the GFP gene linked to a sorting signal for localization to the endoplasmatic reticulum, which was amplified by PCR from the plasmid ER-GFP kindly provided by D. Pestov (78).
The plasmid pTK-GRE containing a single GRE from a truncated TK promoter linked to the luciferase gene was kindly provided by D. Spengler (79). Cells were transiently transfected using ExGene 500 (MBI Fermentas, St. Leon-Rot, Germany) according to the manufacturer’s instructions. For transfection of cells synchronized in G2 or M, transfection was at the end of serum starvation before incubation with either nocodazole, taxol or colcemid (all from Calbiochem, La Jolla, CA).
Electroporation was used to obtain HT-22 cells stably transfected with pMTV-GFP (80). Individual cell clones were grown in the presence of puromycin (10 µg/ml) and picked as described elsewhere (81). One of these subclones was further purified by the criterion of maximally DEX-induced GFP expression using a fluorescence-activated cell sorter (FACScalibur, Becton Dickinson and Co., Heidelberg, Germany) to obtain the clone HT-22-GFP. CHO-TRex cells stably expressing a tetracyclin sensitive repressor protein were transiently cotransfected with a construct of a tetracyclin inducible CMV promoter linked to the ß-gal gene and pMTV-Luc.
Cell Synchronization and Induction
Exponentially growing HT-22 and H4-II-E-C3 cells were maintained in DMEM supplemented with 0.5% charcoal-stripped FCS for 48 h. After serum starvation cells were released into DMEM supplemented with 10% FCS containing either 500 ng/ml nocodazole or 100 nM taxol for synchronization in G2 or containing 300 nM colcemid for synchro-nization in mitosis. Cells were maintained in the presence of each synchronizing agent for 18–24 h. After this time, aliquots of cells grown under identical conditions were either prepared for FACS analysis to verify synchrony or were stimulated with DEX or CdCl2. To obtain populations of cells in G2, mitotic cells were eliminated from adherent cell cultures treated with either nocodazole or taxol by shaking them off before FACS analysis or determination of luciferase or TAT activity. Conversely, for preparation of mitotic cells, metaphase cells were shaken off from adherent cell populations treated with colcemid before stimulation.
Routinely, all cell lines were cultured in DMEM containing charcoal-stripped steroid-free FCS for at least 24 h before hormone treatment. All treatments were done in steroid-free DMEM with either DEX dissolved at the concentrations indicated in ethanol or an identical volume of ethanol. Cadmium chloride was dissolved in water and used at a concentration of 5 µM. HT-22-GFP and H4-II-E-C3 cells were routinely exposed to DEX or CdCl2 for 8 h. CHO-TRex cells transiently transfected with either MTV-Luc or pTK-GRE-luc were routinely exposed to DEX for 12 h.
FACS Analysis
Cells were harvested by trypsinization and fixed in 70% ethanol at 4 C over night. Samples of fixed cells were resuspended and stained in PBS containing 20 µg/ml PI and 10 µg/ml RNAse A. FACS analysis of GFP emission at 525 nm (Fl1) and PI emission at 630 nm (Fl4) was performed using a Beckman Coulter (Krefeld, Germany) XL flow cytometer. After gating out doublets and clumps as described elsewhere (82), results of MMTV-driven GFP induction in HT-22-GFP cells were obtained from the mean GFP fluorescence of all events within a particular gate, e.g. a G2 gate or a gate spanning the entire cell cycle. Cell cycle analysis of DNA histograms was done using the Multicycle software (Phoenix Flow Systems, San Diego, CA).
Assays of Luciferase, ß-gal, and TAT Activity
Luciferase activity was determined according to manufacturer’s instructions (Roche Molecular Biochemicals, Mannheim, Germany) using a luminometer [Wallac, Inc. (Wildbad, Germany), victor2 multilabel counter]. Assays of ß-gal activity were performed using the Galacto-Light assay (Tropix, Inc., Bedford, MA) according to the manufacturer’s instructions. Luciferase activity obtained from pMTV-Luc transiently transfected in HT-22-GFP or H4-II-E-C3 cells was normalized to gal activity of cotransfected pCMV-ß-gal to correct for variations in transfection efficiency. Luciferase activity obtained from pMTV-Luc transiently transfected in CHO-TRex cells with or without stimulation by DEX was normalized to protein. In these cells, inducibility of general transcription was determined by induction of pCMV-ß-gal activity by tetracyclin (1 µg/ml for 12 h) after normalization to protein.
TAT activity was determined as described (83). In short, 10 µl samples of Triton-X 100-lyzed H4-II-E-C3 cells were incubated in 100 µl of a potassium phosphate buffer containing 125 mM K2HPO4/KH2PO4, pH 7.6; 4 mM L-tyrosine; 70 µM pyridoxal-5-phosphate; 13 mM 
-ketoglutaric acid; and 0.5% Triton-X-100 at 37 C for 40 min. After addition of 50 µl 1% hexatrimethyl-ammonium bromide dissolved in 2.8 M NaOH, samples were incubated at 37 C. After 30 min TAT activity of each sample was assayed by measuring the absorbance at 340 nm using a plate reader (Dynatech Corp. MR 7000). TAT activity was normalized to protein.
The relative activity of reporter enzymes was the ratio of stimulated activity divided by nonstimulated activity. The luciferase signal from pMTV-Luc transiently transfected in CHO-TRex cells was either normalized to protein or to tetracyclin-induced gal activity normalized to protein. All statistics were performed using the U test to determine significance.
Mitotic Shakeoff
Asynchronously proliferating H4-II-E-C3 cells were incubated in fresh DMEM without phenol red supplemented with 10% charcoal-stripped FCS containing 300 nM colcemid and 0.1 µCi 6-3H thymidine at time zero. Within intervals of 90 min after time zero, mitotic cells were shaken off from the adherent cell population by gently shaking the culture flask. Mitotic cells were collected by centrifugation and resuspended in PBS. For protein determinations and TAT assays, aliquots of cells were Triton X-100-lyzed in a buffer containing 125 mM K2HPO4 (pH 7.2). Incorporation of 6-3H-thymidine in mitotic cells was determined as described elsewhere (17, 66) using a scintillation-counter (Beckman Coulter LS 6500). TAT activity as well as 6-3H-thymidine incorporation were normalized to protein.
RT-PCR
Preparation of mRNA from samples of cells stored in RNAlater (QIAGEN, Hilden, Germany) were performed using the RNeasy kit (QIAGEN) according to the manufacturer’s instructions. Reverse transcription of 5 µg mRNA was performed as described elsewhere (84). Of each cDNA sample, 5 µl were used for PCR. PCR conditions were 94 C/45 sec, 60 C/45 sec and 72 C/2 min, 24 cycles. Primers for mouse-ß-actin were forward: GTG GGC CGC TCT AGG CAC CAA, reverse: CTC TTT GAT GTC ACG CAC GAT TTC, for mouse-MT-I forward: TTC ACC AGA TCT CGG AAT GGA C, reverse: TTC GTC ACA TCA GGC ACA GCA C. PCR products were separated by electrophoresis through a 2% agarose gel. After staining with ethidium bromide the bands corresponding to MT-I mRNA (280 bp) or ß-actin (540 bp) were quantified using a gel imaging system [Kodak (Rochester, NY) image station 440CF and Kodak 1D Image Analysis software] and the NIH image program. Using RNA from CdCl2-induced cells and the MT-I primers a linear increase in the amount of amplified DNA was observed between 1 and 6 µg mRNA used in the initial reverse transcription (data not shown). Amplification of ß-actin was linear up to 3 µg of input RNA in test assays (data not shown). Because we used 1 µg of RNA in our experimental assays and the variation in amplified ß-actin was at most 1.8-fold, our experimental conditions should have been in the linear response range.
MTT Assays and Protein Determination
Cell viability of cells grown in microtiter plates was assayed with the MTT assay as described (81). Protein was determined using the BCA assay (Pierce Chemical Co., Madison, WI).
|
ACKNOWLEDGMENTS |
|---|
|
FOOTNOTES |
|---|
Received for publication October 12, 2001. Accepted for publication February 8, 2002.
|
REFERENCES |
|---|
|
TOPABSTRACTINTRODUCTIONRESULTSDISCUSSIONMATERIALS AND METHODSREFERENCES |
|---|
, 
0.005.
