Logo ZMBH - Dr. Hermann Bujard

Controlling gene activities via the Tet regulatory systems:

- A Trouble Shooting Guide (Jan.1996) -

Table of Contents

  1. General Remarks
  2. Essential Components of the Tc-controlled Regulatory System
  3. To Get Started
  4. Establishing Cell Lines that Stably Express the tTA or rtTA Gene
  5. Establishing Double Stable Cell Lines
  6. Activation of Expression
  7. Is the VP16 Domain Poisonous? - Concuding Remarks
  8. References

1. General Remarks

Leakiness, tightness and regulation factors.

Although we have discussed these issues in our various publications (ref 1 - 5), letters we receive as well as some recently published work (ref 6 - 9) demonstrate that they may have to be clarified once more. We define "leakiness" of a minimal promoter-tet operator construct such as PhCMV*-1 (1), as the intrinsic activity of such a sequence upon transfer into cells. When, for example pUHC13-3 encoding the luciferase gene under the control of PhCMV*-1 is transiently transfected into HeLa cells, the luciferase activity observed depends on (a) the intrinsic residual activity of PhCMV*-1, (b) the number of copies in the cell (which depends on the amount of DNA used for transfection). The intrinsic activity of the minimal promoter may vary in different cell lines, it may also change when additional sequence elements, which could function as enhancers, are introduced into the vector. Thus, to examine the suitability of a minimal promoter in a given vector and in a given cell line, transient experiments should be performed in which residual activities obtained under defined conditions are compared with those monitored in HeLa cells. Should the residual activities of the minimal promoter drastically exceed the values observed with pUHC 13-3 in HeLa cells, one might have to modify the vector or switch to a different minimal promoter sequence.

When a transcription unit controlled by a proper minimal promoter-tet operator sequence is integrated into the chromosome, the situation changes profoundly. After packaging into chromatin suppression, the residual activity of such promoters is drastically reduced. On the other hand, since such minimal promoters function also as enhancer traps, they may be activated by nearby enhancers. There may also be transcriptional read-through from outside promoters. Thus, in stable cell lines the so-called leakiness is primarily a function of a particular integration site.

This is demonstrated by the finding that cell lines like X1 (1) constitutively producing tTA and containing the PhCMV*-1-luciferase unit stably integrated show no measurable luciferase activity in presence of tetracycline (Tc). In absence of Tc, luciferase can be highly stimulated by tTA. Thus, it can be concluded that, by these criteria, PhCMV*-1 is not leaky in HeLa cells. The regulation factors observed in such cell lines can exceed values of 105. It has to be kept in mind, however, that such cell lines were identified after screening for low or no luciferase background in the non-induced state.

Quantitation of luciferase activity in our X1 HeLa cell line (1) indicates that in the uninduced state there are less than 7 molecules of the enzyme (detection limit) present in the cell. This shows that the system, if properly set up, is not only very tight but can also be highly active (induction factor > 105). We like to emphasize these facts again since several misleading reports were published recently (ref. 6,7,8) in which the regulation factors observed in transient experiments were compared to those described above and in ref. 1. Comparing the activity of the fully activated PhCMV*-1 in transient expression experiments with other promoters shows that it is a strong promoter exceeding e. g. the activity of he hCMV promoter in HeLa cells and B cell lines (unpublished).Thus, the PhCMV*-1 is a tTA/rtTA-responsive promoter with a particularly broad range of regulation which is also suitable when high amounts of RNA are required as e. g. in anti-message approaches.

The tTA versus the rtTA system.

Although we are aware of experimental approaches where the use of the reverse tetracycline-controlled transactivator (rtTA) is advantageous, we do not share the views expressed in many letters we received recently. We do not feel that the rtTA system is "much better" than the tTA system, we also have no reason to assume that it is "much tighter" than the authentic (tTA) system. Even the "problem of having continuously tetracycline in the cell culture" is not really an issue. For example, using doxycycline at 1 ng/ml (which is about 1000 fold below cytotoxic concentrations) is sufficient to inactivate tTA. Without knowing, you may actually have used since years sera in your cell culture media which contain much higher tetracycline concentrations (see below)!

We would like to emphasize again that both systems are truly complementary and that the decision which one to use depends on the particular experimental strategy. To keep a gene switched off and to induce it rapidly at a given time, the rtTA system may be preferable. On the other hand, to keep a gene active and to turn it off occasionally, the tTA system may be better. Thus, don't throw out your tTA-producing cell lines or transgenics - only because molecular biologists like to induce gene activities by adding an effector substance following the paradigm of the E.coli lac operon!


2. Essential Components of the Tc-controlled Regulatory System

Regulator Plasmids.

The tTA-encoding vectors.
The basic plasmid (pUHD15-1) [1] encodes a CMV promoter/enhancer, the coding sequence for tTA and a SV40 polyadenylation site. An alternative plasmid, pUHG15-1 [10] contains the %szlig;-globin intron/poly (A) site instead of the SV40 poly (A) site. This may be particularly useful when tTA expression is examined by RNA analysis. In our experience, the presence or absence of an intron does not result in different levels of tTA synthesis.

We recommend the hCMV promoter/enhancer for tTA expression since it has been used successfully in a number of cell lines. Depending on the specific application, particularly when tissue-specific expression in transgenic organisms is the goal, one has to consider more appropriate expression signals. The optimal promoter/enhancer can only be defined in the context of an experiment with respect to a particular cell line or organism.

We strongly recommend not to modify the tTA ORF! Several attempts (including our own) e. g. insertion of nuclear localization sequences have resulted in a hampered regulatory system.

The rtTA-encoding vectors.
Two versions of rtTA have been described (2). rtTA as encoded in pUHD17-1 or pUHG17-1 differs from tTA only by 4 amino acid exchanges, resulting in the reverse phenotype. Beside the respective nucleotide exchanges pUHD17-1 is identical to pUHD15-1 (see above). For analytical purpose, the two plasmids can be distinguished by an additional HindIII site in the rtTA coding region. rtTA-nls as encoded by pUHD172-1neo contains an additional N-terminal nuclear localization sequence (2). In addition, this plasmid contains a neomycin resistance cassette.

Response plasmids.

The response plasmids such as pUHD10-3 (5) and its derivatives can be used for the tTA as well as for the rtTA system. They have a tTA-dependent promoter in front of a multiple cloning site or of a gene of interest. We strongly recommend to examine tTA or rtTA function with a tTA-controlled reporter unit. Our prefered reporter system is luciferase (pUHC13-3). If luciferase is not a suitable reporter, pUHG16-3 (10) encoding tTA-controlled %szlig;-galactosidase can be used.

The Effector Substance.

We now use preferentially doxycycline hydrochloride (Dox-HCl, which, like tetracycline hydrochloride, is water soluble!) which functions with tTA and rtTA. We use a freshly made stock solution of Dox-HCl for 2 weeks (1 mg/ml; filter-sterilized and stored at 4oC in the dark). Alternatively, stock solutions can be frozen in small aliquots at -20oC, for long term storage. Dox-HCl in PBS forms a precipitate after a few days; medium supplemented with Dox-HCl should therefore not be stored for prolonged times. Tetracycline-HCl is only marginally functional in the rtTA system. The most potent rtTA effector so far identified by us is doxycycline-HCl. To feed doxycycline to transgenic animals, we routineously dissolve this substance at a final concentration of 200 µg/ml in deionized water (water bottle should be light protected). The drinking water which should be renewed every 3 days contains 5 % succrose.

Culture media.

In our as well as in several other laboratories, it has been observed that highly regulatable cell lines suddenly do not express the gene of interest anymore when incubated in "Tc-free" medium. A thorough analysis of several of these incidents revealed that in each case a change in the batch of calf serum or fetal calf serum had occurred and we know today that commercial serum preparations can shut off tTA responsive promoters. We suspect that they contain Tc or one of its derivatives. We are in the process of identifying such tetracyclines in some batches of serum. A good way to examine the quality of sera in this respect is to grow and regulate the X1 cell line described in our 1992 PNAS-paper. This cell line will soon be available through the ECACC (European Collection of Cell Cultures, CAMR, Salisbury, Wiltshire SP4 0JG, UK, Tel. +44-1980-612512, Fax +44-1980-611315). 

3. To Get Started

The tTA System.

We recommend to test the system first in a transient assay, making cotransfections with e.g. The ratio of the tTA producing plasmid vs. the response plasmids is critical for these transient experiments. An excess of pUHD15-1 over pUHC13-3 (up to 100-fold) is advisable. These conditions assure high intracellular concentrations of tTA and consequently high occupancies of the tet operator sites by tTA. At the same time, these plasmid ratios give a low background while maintaining high expression potential. In the presence of Tc, the background synthesis of reporter enzymes decreases proportionally when the amount of the response plasmid is lowered. This is in contrast to activated levels (i.e. in the absence of Tc) which are affected to a much lesser extent by the relative amount of the plasmid.

For Ca-phosphate transfections or lipofections, one transfection reaction should be split between two tissue culture dishes. Similarly, after electroporation the cells should be divided between two dishes. In either case, Dox-HCl should be added to one of the plates (2 ng/ml for initial experiments; this concentration may be lowered in subsequent experiments) (1, 6).

From our experience it is not necessary to preincubate cells with Dox-HCl prior to transfection since the uptake of the antibiotic is very fast. However, it cannot be ruled out that some cell lines may behave differently.

Since the transactivator has to be synthesized in order to initiate expression of the gene of interest, the time period required for the detection of the respective protein may be longer as compared to constitutive expression.Thus, the time period for maintaining expression may have to be adjusted accordingly.

The residual activity of the CMV minimal promoter sequence (-53 to +75) located between the tet operators and the gene of interest may vary to some extent in different cell lines. Transient expression experiments should be performed for each particular cell line in parallel with HeLa cells to examine the intrinsic activity in each cellular context. In our experience, cotransfection of HeLa cells with pUHC13-3 and pUHD15-1 yields regulation factors between 100 and 1000-fold in luciferase activities after 24 h +/- Dox-HCl. For the respective double stable cell lines see (1). In case tTA dependent regulation in the cell line of interest is not as good as in HeLa cells (due to elevated luciferase activity in the presence of Dox-HCl), switching to a different minimal promoter should be considered. In NIH 3T3 cells we observed a rather high basal activity of PhCMV*-1. This problem could be circumvented by using a Tk-based minimal promoter. Concommitant, however, with a lower basal activity (in presence of Dox-HCl), is in this case a lower activation of the promoter upon Dox-HCl withdrawal.

When examining the results of the transient expression experiments, it has to be kept in mind that higher regulation factors are usually found in double stable cell lines due to the reduction of background synthesis after chromosomal integration of the response unit. If background expression (i. e. expression in the presence of Dox-HCl) is too high, one might consider exchanging the minimal promoter sequence. Our plasmids are constructed such that this exchange can be readily achieved (see sequence print-outs).

The rtTA System.

rtTA-dependent gene expression can be demonstrated in double transient experiments using in absence or presence of 1 µg/ml doxycycline.

It should be emphasized that transient experiments with rtTA are less straight forward than with tTA.

We have observed major differences between rtTA and rtTAnls , whereby rtTAnls yielded a higher basal activity (absence of doxycycline). So far, we have not analyzed whether this is due to the elevated accumulation of the protein in the nucleus upon transient overexpression, or to a somehow altered affinity of the protein to its cognate binding sequence. It is clear, however, that this effect is not that pronounced after integration of the tet operator containing response units into the chromosome. Therefore, for transient experiments a broad titration of the regulator vs the responsive plasmids (we routineously use 1:1 down to .001:1) is initially recommended. These findings are based on only a very limited number of cell lines tested, since our own focus with the rtTA system is on transgenic animals. Here rtTA is superior to rtTAnls.

As with tTA the full potential of the rtTA regulation system can only be exploited in stably transfected cells. 


4. Establishing Cell Lines that Stably Express the tTA or rtTA Gene.

We recommend to establish the regulatory system in two steps: a stable cell line expressing tTA or rtTA should be constructed and characterized first; in a second step, this line should be used for the transfer of the gene of interest. This approach will yield tTA- and rtTA-positive cell lines that can serve as a defined genetic background. These lines will then allow the direct comparison of different clones containing a subsequently introduced gene of interest. Moreover, a well defined tTA or rtTA positive cell line allows the insertion of a variety of genes under control of a tTA-responsive promoter. By contrast, after cotransfection of the regulator and the response plasmid, quantitative differences will also be due to different expression levels or integration loci of the tTA or rtTA construct. Thus, a strict comparison of the different clones is not possible. Moreover, during cotransfection, the expression unit of the transactivator may integrate at the same chromosomal locus (as it is frequently observed in cotransfection experiments). In this way, the enhancer of PhCMV driving tTA or rtTA is brought into proximity of the minimal promoter which can result in increased basal activity of the minimal promoter.

We have successfully used two strategies: cotransfection of pUHD15-1 or pUHD17-1 with a selection marker (generally SV2neo) or integration of the respective resistance cassette into the tTA or rtTA encoding vectors. In the latter case, the percentage of G418 resistant clones showing the expected tTA or rtTA phenotype is definitively higher. Still we prefer the first approach since it seems to give higher-regulatable cell lines. However this is an observation from a limited number of experiments and we have no good explanation for the difference.

Once resistant clones are isolated they should be examined for tTA or rtTA expression via transient supertransfection with pUHC13-3 or pUHG16-3 +/- Tc (see above). So far, we observed only limited expression of the tTA or rtTA gene after stable transfection of mammalian cells with pUHD15-1 or pUHD17-1. The low amount of transactivator in those cells makes its direct detection by band shift assays or immunoblotting in most cases difficult and not suitable for screening many clones. In addition, the indirect assay by supertransfection will provide a result on the functionality of the clones rather than just demonstrating the presence of the protein in the respective cells.

The identification of rtTA-positive clones has, of course, to be carried out in presence of 1 µg/ml of Dox.

Should the transactivator-positive cell lines be used for transient experiments, whereby the gene of interest is controlled by a tTA- or rtTA-responsive promoter, it is important to use low amounts of the respective expression plasmid for transfection. In this way, it is prevented that the low intracellular concentration of tTA or rtTA becomes limiting (to meet transfection requirements, unspecific DNA may be added). 


5. Establishing Double Stable Cell Lines.

Upon integration of a gene of interest into pUHD10-3, the resulting vector should be transfered into the tTA or rtTA-positive cell lines by cotransfection together with a second selection marker. Selection markers or other constructs having proven or suspected enhancer activity, should be avoided. Since the cotransfected DNAs will frequently cointegrate and the minimal promoter may serve as an enhancer trap. Therefore, enhancerless selection markers are recommended to obtain cells with a low basal expression level. This should, however, be a less critical concern when the goal is only a conditional overexpression. 

6. Activation of Expression

Induction by withdrawal or by addition of Dox-HCl may take some time to be completed. This is not necessarily an intrinsic limitation of the regulatory system itself, but rather primarily a function of the half-lives of the mRNA and the protein under study: it will take some time before stably expressed proteins accumulate to equilibrium levels.

In any case, suitably regulatable clones should be subcloned. This has on occasion resulted in the identification of clones with even further improved regulatory properties. 


7. Is the VP16 Domain Poisonous? - Concluding Remarks

Numerous people have expressed concerns about the "poisonousness" of the VP16 domain. Just recall Paracelsus (1493 - 1541) who already pointed out that whether a compound may be a precious drug or a poison is often a question of concentration! It is common knowledge that many gene products, and in particular regulatory proteins such as transcription factors, function "physiologically" only when their intracellular concentration is limited to a defined window. Due to their high activation potential, VP16 fusion proteins like tTA/rtTA may require a lower concentration window than other transcription factors. However, random integration of the tTA/rtTA expression unit into cellular genomes and careful screening for the proper stable clones yield cell lines which fulfill exactly this requirement. For example, our HtTA-1 and X1 HeLa cell lines contain less than 10.000 tTA molecules per cell (probably around 6000) which, however, are sufficient to activate PhCMV*-1 more than 105 fold. Both cell lines grow in our lab without selection pressure since more than 5 years.Another about 15 cell lines established in our and other laboratories appear to behave analogously.

Similarly, we keep a number of tTA and rtTA mouse lines whose members live happily since generations, thanks to the proper concentration window of tTA/rtTA. Modifying the activating domain for lower "poisonousness" would most likely result in higher intracellular concentrations as compared e. g. to tTA. But for physiological function there would again be an upper limit. Thus, unless there are special reasons for changing the activating domain (e. g. to control specific interactions or to achieve tissue specificity, etc.), we do not see a need so far for abandoning the VP16 domain.

Is autoregulation of tTA (A. Bonin, M. Gossen and H. Bujard, unpublished; ref. 9) advantageous? As long as an autoregulatory system as described in ref. 9 is used, a rather complex situation is generated. Here, PhCMV*-1 - a very strong promoter when fully activated - controls tTA/rtTA synthesis. Upon induction, the transactivator may be quickly overexpressed to "poisonous" levels, which we feel lie > 10.000 molecules/cell. Thus, unless one likes to generate a "run-a-way" system which does not need to remain physiological, one has to limit the tTA/rtTA production via Dox-HCl. However, this may limit the expression of the gene of interest - also controlled by PhCMV*-1 - to concentrations not appropriate for the intended study. Therefore by autoregulating tTA by the same promoter which drives the gene of interest, the regulatory range of the latter is limited. This may be overcome by adjusting the individual strength of the two tTA-responsive promoters involved which, however, may turn out not to be as simple as it may appear at first glance.

In conclusion, although there may be cell lines where stable expression of tTA/rtTA to a proper level is not tolerated (our experience does not support this view), in general there appears to be no need to control tTA/rtTA expression in cultured cell lines. Moreover, in transgenics the overwhelming number of approaches will require tissue specificity of tet control which we feel cannot be achieved readily by an autoregulatory system of the type described in ref. 9.

And here a notion of encouragement: In a time where clever combinations of commercially available kits and receipies appear to make experimental skills and an understanding of a system unnecessary, just keep in mind that the tet-system is not (yet?) a kit and that it is worthwhile to spend some thoughts on the basics - its fun and may be rewarding ! 


8. References

  1. Gossen, M. and Bujard, H. (1992) Tight Control of Gene Expression in Mammalian Cells by Tetracycline Responsive Promoters. Proc. Natl. Acad. Sci. USA 89, 5547-5551.
  2. Gossen, M., Freundlieb, S., Bender, G., Müller, G., Hillen, W. and Bujard, H. (1995) Transcriptional activation by tetracycline in mammalian cells. Science 268, 1766-1769.
  3. Gossen, M., Bonin, A. and Bujard, H. (1993) TIBS 18, 471-475.
  4. Gossen, M., Bonin, A.L., Freundlieb, S. and Bujard, H. (1994) Inducible gene expression systems for higher eukaryotic cells. Curr. Opin. Biotech. 5, 516-520.
  5. Gossen, M. and Bujard, H. (1995) Efficiency of Tetracycline-controlled Gene Expression is influenced by cell type: A commentary.. Biotechniques 19, No. 2.
  6. Ackland-Berglund, C.E. and Leib, D.A. (1995) Efficiency of a tetracycline-controlled gene expression is influenced by cell-type. Biotechniques 18, 196-200
  7. Howe, J.R. et al (1995) The Responsiveness of a Tetracycline-sensitive expression system differs in Different Cell Lines. J. Biol. Chem. 270, 14168-14174
  8. Miller, K. and Rizzino, A. (1995) The Function of inducible Promotor Systems in F9 Embryonal Carcinoma Cells. Exp. Cell Res. 218, 144-150
  9. Shockett, P. et al (1995) A modified tetracycline-regulated system provides autoregulatory inducible gene expression in cultured cells and transgenic mice. Proc. Natl. Acad. Sci. 92, 6522-6526
  10. Resnitzky, D., Gossen, M., Bujard, H. and Reed, S.I. (1994) Acceleration of the G1/S phase transition by expression of cyclins D1 and E using an inducible system. Mol.Cell.Biol. 14, 1669-1679.
  11. Gossen, M. and Bujard, H. (1993) Anhydrotetracycline, a novel effector for tetracycline controlled gene expression systems in higher eukaryotic cells. Nucleic Acids Res. 21, 4411-4412.