Biochimica et Biophysica Acre, 1!30 ( !992) 68- 74 © 1992 Elsevier Science Publishers B.V. All rights reserved 0167-4781/92/$05.00
68
BBAEXP 92340
Regulation of the Escherichia coli lac operon expressed in human cells Denis S.F. Biard, Michael R. James ~, Andr6 Cordier 2 and Alain Sarasin Laboratory of Molecular Genetics, UPR 42, CNRS, instimt de Recherches Sciemifiques sur le Cancer, V'dlejuif (France) (Received 5 August 1991)
Key words: Transcription regulation; EBV-vector; lac operon" (Human cell)
We have investigated the use of various Epstein-Barr virus (EBV)-based vectors bearing the two components of the Escherichia coil lac operator-represser (lacO, iacl) complex. Our aim was to develop a model system of gene expression by looking at the transcription of the bacterial #-galactosidase coding gene (lacZ) in 293 human embryonic kidney cells. Several vectors have been built carrying different promoters upstream of the lacl and lacZ genes and in which natural or synthetic operator sequences were inserted in the 5' part of the iacZ gene. In transient expression assays we achieved efficient !acZ gent repression which could be released by the specific inducer isopropyl //-D-thiogalactoside (IPTG). A stable transtormed cell line carrying two EBV-derived plasmids with the building blocks of the lac operator/represser system was established. This cell line allowed us to achieve a wide range of lacZ gene regulation. In this cell line IPTG alone could remove the repression to trigger a 5-fold increase of lacZ expression. Heavy metal ions, which induced the mouse metaliothionein I promoter located upstream of the iacZ gen¢, added together with IPTG gave rise to a 40-fold induction of lacZ expression.
Introduction in recent years Epstcin-Barr virus (EBV)-based shuttle vectors have emerged as powerful tools for genetic engineering. These vectors could replicate in a broad variety of cell lines such as human or monkey cells [1] without interfering with the expression of differentiated cell functions [2]. They appeared appropriate for efficient shuttling of large DNA fragments from mammalian cells to E. coli [3]. EBV vectors have successfully allowed foreign DNA expression in various recipient cells like the human tumor necrosis factor in human cells [3] and its receptor [4]; the rat Na-K ATPase a 2 isoform in human and monkey cells [5]; the human interferon 7 and extracellular domain of the human epidermal growth factor receptor in human
! Present address: Centre d'Etude du Polymorphisme humain, 27 rue Juliette Dodu, 75010 Paris, France, ~" Present address: lnstitut de Recherches sur la S6curit6 du M6dicament Rh6ne Poulenc-Rorer, 20 Quai de la R6volution, 94140 Aifortville, France. Correspondence: A. Sarasin, Laboratory of Molecular Genetics, UPR 42, CNRS, lnstitut de Recherches Scientifiques sur le Cancer, B.P. No. 8, 94801 Villejnif, France.
cells [6] or E. coli chloramphenicol acetyl transferase (CAT) gene in mammalian cells [1,7]. EBV vectors have been used for efficient expression of cDNA libraries [8] and effective antisense RNA-mediated inhibition of presclccted genes [9] in human cells. Moreover, these shuttle vectors proved to be suitable tools to study biological processes, such as mutagenesis [7,10,11] or recombination [12] of target genes after transfer into mammalian cells. In the present work, we aimed to devise a genetic assay composed of an easily quantifiable reporter gene which is regulated by various inducer and represser elements. In this goal, the regulation gene (represser) could be the target of external agents. To develop this approach, we have stably maintained and expressed the E. coli lac operator/represser complex in human cell lines by use of EBV vectors. This regulatory complex represents a paradigm for the study of control of gene expression in response to various environmental conditions [13]. The lacO/lacl system has been shown to be functional in mammalian cells both after transient or stable expression assays [14-18]. cis- and trans-acting elements of the E. coli lac operon have been used to regulate gene expression in vaccinia virus [19]. The repression could be reversibly removed using a specific inducer, isopropyl fl-D-thiogalactoside (IPTG). This in-
69 ducer appeared to bind to repressor therefore decreasing its affinity for the lac operator. In preliminary experiments, we have constructed different EBV vectors carrying the lacl gene on one hand and a lacO sequence on the other hand to obtain the best combination of plasmids which will allow appropriate regulation of the lacZ gene by the lacl repressor. Different promoters have be inserted downstream of the iacl and the lacO sequences. The functionality of the lacO/lacI complex was firstly assessed through transient expression assays, and ultimately in stable established cell lines. The specificity of repression by lacI repressor was apparent by the ability to be alleviated using IPTG. Our ultimate goal was to integrate one copy of the repressor gene (lad) into the genomic DNA in an effort to study disruptions of the transcriptional control of this regulatory gene. In view of these issues, we have established a cell line bearing one lacl gene constitutively expressed. These cells also carried a pEBVMTlacOlacZ vector giving rise to an efficient repression of the lacZ gene by the lacl repressor. The mouse metallothionein-I promoter (mMT-I) inserted upstream of lac operator has been shown to be active in almost all tissues and cell lines [20] and we have previously shown that it is efficiently induced by heavy metals when carried on an EBV episome [7,21]. In stable transformed lines, the lacZ expression was regulated by the lacO/lacl complex. The repression could be partially relieved by metal induction of the mMT-I promoter. In the presence of IPTG and metal cocktail (zinc and cadmium) we observed a synergistic effect on the lacZ transcription with a 40-fold increase.
AMPr pBr o
~ ' ~ ~
TK pr < f
~'~
.__
~hl~_
]J
G418r
p145 ~ pEBVMTlacOlacZ 12589bp
BNA'I
lacO lacZ
Fig. 1. Genetic maps of the pEBVAdyadCMVlacl (p189) and pEBVMTlacOlacZ (pl45). The elements of pig9 vector are in the inward circle with (i) the segment containing the EBV nuclear antigen ! (EBNA-I) and the repeat family (FR) (black-filled area): (it) the reporter gene transcriptional cassette comprising the CMV pronloler ((,'MV pr), the laci gene and the SV4(I late-region polyadenylation signal (SVpA + ) and oriented in the clockwise direction as EBV sequences (stippled area): (iii) the geneticin resistance gene (G418r) under control of the HSV-thymidine kinase promoter (TK pr) and the HSV-TK polyadenylation signal (TKpA'-) and oriented in the counterclockwise direction (open area): (iv) the oBR322 origin of replication (pBR ori) and the bacterial fl-lactamase gene (AMPr) in the same orientation as the eukaryotic resistance gene (hatched area). The pl45 plasmid (outward circle) carries the whole EBV replication origin with the repeat family (FR) and the dyad sequence (DYAD). Other building blocks are as follows: (i) the lacZ gene transcriptional cassette with the mMT-! promoter (mMT-! pr). the symmetrical lac operator (htcO) inserted into Xhal sites and the SVpA" in the clockwise direction: (it) the hygromycin B phosphotransferase gene (HYGROr) with TK pr and TKpA+ sequences in counterclockwise orientation.
Materialsand Methods Plasmid constructions All procedures for plasmid isolation, restriction endonuelease cleavage, ligation and transformation were carried out as described elsewhere [22]. In this work, vectors defective in the EBV dyad sequence were named pEBVAdyad. The vector pEBVAdyadCMVlacl (p 189 vector; Fig. l) was derived from the plasmid p205. It carried the pBR322 replication origin, the ampicillin resistance gene, the EBV EBNA-1 and FR sequences [23]. The EBV dyad palindromic sequence (between Spel and Hpal sites) was deleted to remove the putative replication origin [23,24]. Upstream of the laci gene lies the immediateearly human cytomegalovirus (CMV) promoter-enhancer [25]. The lacl gene was generously provided by M.C.T. Hu. It contained a modified initiation codon (ATG) with attendant Simian Virus 40 polyadenylation signal. The selection of cells carrying p189 plasmid, namely the 293/189 (lacl +) cell line, was by geneticin , (G~.18) resistance, pEBVMLPIacl plasmid was derived
from p205MLPCAT vectors [7] by the insertion of the lacl gene via the HindIII and BamH! restriction sites. To clone the iacZ gene downstream of the mouse metailothionein-i promoter (mMT-I), the adenovirus major late promoter (MLP) or the CMV promoter, we have isolated the HindlII/BamHI fragment from pRSVZ and inserted it into corresponding sites of pEBVCAT vectors to produce pEBVMTiacZ (p149), pEBVMLPiacZ and pEBVCMVlacZ plasmids. A synthetic operator with a 5' attendant Xbal restriction site (oligonucleotide 5'-CTAGATTGTGAGCGCTCACAAT-3') was inserted into Xbal site to obtain pEBVMTlacOCAT, pEBVMTlacOlacZ (p145; Fig. l), pEBVMLPlacOlacZ and pEBVCMVlacOlacZ plasraids. The lacl repressor possessed more affinity to the synthetic iac operator than the natural one [26]. For cloning, the selection of recombinant plasmids was done in WP3 bacteria in the presence of X-Gai chromogenic substrate [27] and verified by DNA sequence analysis. The natural and asymmetrical operator (oligonucleotide 5 '-CTAGAATTGTGAGCGGATAACAA-
70 TT-Y) was inserted into the same Xbal site in the two orientations giving rise to pEBVMTIacO1CAT, pEBVMTlacO2CAT, pEBVMTIacOIlacZ and pEBVMTlacO21acZ vectors. All these vectors conferred resistance to hygromycin B.
protein] x [CAT reference activity/CAT experimental activity]. For cell lines which only expressed CAT we followed exactly the Sleigh protocol [31]. CAT activity was given as CPM of [14C]acetyl-chloramphenicol/~g protein.
Celi cuhures and DNA transfection 293 human embryonic kidney cells were grown in Dulbecco's modified Eagle's medium supplemented with 10% foetal calf serum, 100 U / m l of penicillin and 100/zg/ml of streptomycin (DMEM/FCS medium) in a 5% COz incubator. For long term maintenance of cell lines, 5 #g of plasmid per 10 cm plate were introduced into the recipient cells by calcium phosphate coprecipitation [28]. The precipitates were left 6 h on cells. Selection was initiated on day two by cell tt),psination followed by replating at appropriate dilutions in a selective medium containing 125/zg/ml of hygromycin B (Sigma) and/or 500 ~ g / m l of (3418 (Sigma). To determine CAT and /~-galactosidase (/~gal) activities cells were seeded at appropriate dilutions. 3 days later, metal ions (100 p.M ZnCl., and 1 gM CdCI e) and/or IPTG (10 raM; Gibco-BRL) were added to the medium for 2 days unless otherwise mentioned and the cells were collected. For transient expression, 5 /zg (unless otherwise indicated) of pEBVlacl and pEBViacOlacZ plasmids and 1 /zg of a control plasmid of transfection (p205MLPCAT) were transformed into 10 cm plates containing exponentially growing cells. Medium was removed 6 h later and cells were refed with DMEM/FCS medium with or without 10 mM of IPTG. 48 h after, cells were collected and lysed to assess ~-galactosidase and CAT activities.
Results
~-Galactosidase and CAT assays The cells from stable transformed lines were scraped off dishes and pelleted in 300 tL! of buffer H (250 mM s u c r o s e / l ~ mM sodium phosphate, pH 7.5). Cells were lysed by four cycles of freeze-thawing then centrifuged. ~gal activities were determined according to Nielsen et al. [29] and expressed as units (nmol of orthonit%phenol formed per rain) per mg of protein. Purified E. coil ~l-galactosidase (Sigma) was used as a standard. Protein concentration was determined using the Bradford procedure [30] with bovine serum albumin as standard. For transient expression studies, cells were collected in 600 t~i of buffer H and enzymatic extracts were recovered as described above. An aliquot of each sample was heated at 65"C during 10 rain and assayed for CAT activity using a non chromatographic protocol [31]. To reduce the variability between each transfection experiment,/~lgal activities were expressed as '/3gai normalized values' corresponding to [/3gal units/rag
Transient expression assays Various transient expression assays have been carried out to assess the efficiency of repression of the iacZ gene by the lacl repressor into 293 human cells. To compare results from the same transfection or independent ones we have cotransfected the plasmid p205MLPCAT as a reference. This vector gave strong CAT expression 48 h after transfection. To characterize: the properties of the lacO/lacl complex in 293 cells, we have modified the expression rate of the two components (iacl and lacOlacZ sequences) in increasing the strength of the promoter or the amounts of DNA transfected. In absence of pEBVMLPIacl vector we observed a regular increase in ~gal normalized values when the amounts of pEBVMTlacOlacZ vectors increased up to 10/zg (Table I; lines 1-3). With 5 ~g of pEBVMLPlacl vector we reduced the lacZ expression. The repression rate decreased while the amount of transfected pEBVMTlacOlacZ vector increased (Table l; lines 4-6). Then, the magnitude of repression of the lac system could varied by transfecting more amounts of pEBVMTlacOlacZ vectors. A very high lacZ gene expression triggered by promoters stronger than mMT-! could also
TABLE !
Transient expression assay: rariation of amount of pEBVMTlacOlacZ ~'ector transfected with or without pEBVMLPlacl i'ector Different amounts of pEBVMTlacOlacZ and pEBVMLPlacl vector were cotransfected with I p,g of p205MLPCAT control plasmid. After removing the DNA-calcium precipitate the cells were fed or not with 10 mM of IPTG in DMEM/FCS medium for 48 h and assay was performed as in Materials and Methods./~gal normalized units correspond to (/3gal units/mg protein)×(CAT reference activity/ CAT experimental activity). These experiments were reproducible with a percentage of error of around 5%. Line
Transfected plasmids pEBVMTlacOlacZ
/]gal normalized pEBVMLPlacl
units IPTG (10 mM) -
1
2 3 4 5 6 7 8
2/~g
5/zg 10/.tg 2/zg 5/.tg 10 t~g 5 p,g 5 ~ttg
-
5/zg 5 ~g 5 p,g 2/.tg 10/.tg
130 232 631 56 68 3'78 94 39
+
132 255 617 130 246 857 238 91
71 flgal Pormalized units 800 600 I
400
! i
200 ,
Control
mMT-I
MLP
CMV AdyadCMV
Promoter upstream of the lacl gene Fig. 2./3-Galactosidase activities in transient expression assays: role of the strength of promoter upstream of the loci gene. l0 #g of pEBVMTlacOlacZ vector were transfected with 5/zg of pEBVMTlacl, pEBVMLPlacl, pEBVCMVlacl or pEBVJdyadCMVlacl vector and I/.tg of p205MLPCAT control plasmid. Assay ~,as perfi~rmed as in Materials and Melhods. Filled bars: - IPTG. Open bars: + IPTG. Control indicated the average level of the lacZ gene expression with the pEBVMTlacOlacZ vec!or alone without the pEBVlacl one.
bypass the lad repression. Therefore, when we cotransfected pEBVMLPlacl vector with another vector carrying the mMT-I, MLP or CMV promoters upstream of the lacZ gene, we obtained a repression rate more and more reduced (data not shown). The CMV promoter appeared to be the strongest and the mMT-I promoter the weakest. Since very strong promoters, such as MLP and CMV could overcome the l a d repression, we therefore focused our attention on the metal inducible mMT-I promoter. Enhanced flgal repression could be achieved in transfecting more DNA of pEBVMLPIacl vector. When amounts of pEBVMLPIacl vector reached 2, 5 and l0 /,tg we observed 60%, 71% and 83% of repression of /3gal activity (Table I, lines 7,5,8) as compared with transfection of pEBVMTiacOlacZ alone (Table l, line 2). The derepression rate by IPTG was lowered in increasing the amounts of pEBVMLPlacl vector. 2 and l0 ~g of pEBVMLPIacl vector gave respectively rise to a derepression by IPTG of 93% and 36% (Table !, lines 7, 8) of control values (Table I, line 2). To trigger an increased repression, we changed the promoter upstream of the lad gene for a stronger one. We observed a very important repression when the laci transcription was driven by the very strong promoter/ enhancer CMV (Fig. 2). For example the CMV promoter induced a repression 60% greater than with the mMT-I promoter (Fig. 2). In this experiment, the increased l a d gene expression triggered by strong promoters was not correlated with a decrease in the derepression level (with IPTG) as seen in experiments reported in the Table I. However, from different assays performed with various EBV vectors we observed that
the promoter/enhancer inserted upstream of the lacl gene could interfere with the iaeZ gene expression when the lad and lacZ genes were borne by two different vectors. We always obtained increased /3gal activities when the cotransfection where done with pEBVCMVlacl and pEBVMTIacZ vectors compared to transfection with pEBVMTIacl and pEBVMTlacZ vectors (data not shown). The CMV promoter seemed to be a good candidate to control the lad transcription in established cell lines. In the absence of lad gene or lacO sequence, IPTG did not modified the laeZ gene expression whatever the promoter used upstream of this gene (data not shown). In our ultimate aim, the integration of a lad repressor gene into the genomic DNA will be achieved in deleting the Dyad replication origin. As one step towards addressing this purpose, we have constructed the pEBVAdyadCMVlacl (p189)plasmid. In tran,;ient expression experiment, this deletion did not hinder the repression efficiency of the pEBVAdyadCMVlaci plasmid (Fig. 2). In the aforementioned experiments, we found an efficient regulation of the /3gal activity by the lad protein 48 h after transfection which could be partially relieved by IPTG.
Stably transformed cell lines with one EBV t'ector We have analyzed the behavior of our different plasmids in 293 cells after several weeks of mass culture. The success of establishment of cell lines which expressed one EBV vector depended on the plasmid construct. For vectors carrying the lacl gene when mMT-I or MLP promoters were present, all cell lines failed to express reporter gene activity after long term culture (more than 2-3 months) or they encountered difficulties in remaining stable, in contrast, the presence of the CMV promoter helped us to obtain cell lines with normal growth. In any construction where the CMV promoter was placed upstream of a reporter gene such as CAT, lad or lacZ gene, we always observed high transfection efficiencies of host cells related to a low vector copy number (data not shown). Using 50-day-old transformed polyclonal populations we checked the occurrence of hindrances in promoter activities arising from the insertion of a iac operator sequence upstream of CAT or lacZ gene (Table II). The presence of the symmetrical lacO downstream of the TATA-box entailed a 2.5-fold decrease in basal mMT-I promoter activities, whereas no inhibitory effect was observed with the natural lac operator (iacOl or lac02). However, the metal responsiveness of the mMT-I promoter was maintained with an induction rate of roughly 18-fold for pEBVlacOlacZ and pEBVlacOCAT constructs whatever the operator sequence inserted between the mMT-I promoter and the reporter gene. With the CAT or lacZ reporter
TABLE II
[3gai and CAT actirities in 50 days old 293 stably transformed cell lines and metal induction of mMT-! promoter EBV
ZnCd a
plasmid structure pEBVlacZ pEBVCAT
+ +
Reporter gene activities b promoter upstream of reporter gene mMT-I MTlacO MTlacOl
MTlacO2
MLP
CMV
186 2511 203 1557
340 2841 288 2616
0 0 639 624
2 366 2242 2189 2208
88 1507 71 1667
332 3600 118 2391
a ZnCd cocktail (100 ~M ZnCI 2 + 1/zM CdCI2)was added onto cells for 8 h. Enzymatic activities were assessed as in Materials and Methods. h//gal activities were expressed as ,Sgal units/rag protein while CAT activities were expressed in CPM of 14C-acetyl-chloramphenicol/tzg protein, lacO means the symmetrical lac operator, lacOI corresponds to the natural lac operator in the first orientation whereas lac02 was the natural lac operator in the second orientation,
genes, we found a higher expression driven by the CMV promoter than with MLP or mMT-I promoter without metal induction, The MLP promoter placed upstream of the lacZ gene led to no expression of this gene 50 days after the transfection, while the CMV promoter allowed an extensive lacZ transcription.
Stably transformed cell lines with two EBV cectors For the purpose to establish a cell line with the EBNA-1 +/lacl + phenotype, we have constructed the pEBVAdyadCMVlacI (p189)vector. This construction kept the entire EBNA-I transcriptional cassette. Moreover, the EBNA-1/FR sequences coupled with the CMV promoter upstream of the lacl gene led to a vector copy number per cell near one (data not shown). The 293 cells have been transfected with the p189 vector to give rise to the 293/189 (lacl +) cell line. 1 month later, p189 plasmid was found to be integrated into genomie DNA with one or two copies per cell in all 12 clones tested. All clones constitutively expressed the lacl gene (data not shown). The 293/189 (lacl +) cell population was retransfected with different vectors carrying the lacZ gene. When we assessed the transfection efficiencies of our EBV-derived vectors, we always TABLE ill
~-Galactosidase expressio.~ in stable transformed cell line days after transfection of 293/189 cells, we allowed cells to grow for 48 h in a DMEM/FCS medium in presence or not of 10 mM llrrG and 100/~M Z n C I : / l ~M CdCI: (ZnCd)./3gal activities were performed as mentioned in Materials and Methods. lacO sequence means symmetrical lac operator, whereas lacOI indicates natural lac o~rator in the first orientation. NT: not tested.
pEBVlacZvector carried by the 293/189 cell line
#gal units/rag protein -
IPTG
ZnCd
pEBVMTlacOlacZ pEBVMTlacOIlacZ pEBVMLPlacOiacZ pEBVCMVlacOlacZ
13 266 224 2125
66 239 189 2 451
131 2027 NT NT
IPTG ZnCd 536 1925 NT NT
reached higher transfection efficiencies after transfection into 293/189 (lacl +) cells than into 293 cells, with for example a 3.7 and a 2.8 fold-increase with pEBVMTlacOlacZ and pEBVMTlacZ vectors respectively (this experiment has been repeated six times and the increase was statistically significant). We have analyzed the regulatory expression of the lac complex in 30 days old cell lines carrying the two building blocks of the lac regulatory complex. In a mixed clone population carrying the pEBVAdyadCMViacl and pEBVMTlacOlacZ plasmids there was a 5-fold derepression triggered by 10 mM of IPTG for 48 h in culture medium. The metal induction of the mMT-I promoter partially overcame lacl repression (10-fold increase of iacZ expression). The presence of the two inducers (ZnCd and IPTG) triggered a tremendous lacZ expression with a 40-fold increase, as also found by Hu and Davidson with IPTG and glucocorticoid hormone [15]. In the cell line bearing the pEBVAdyadCMVlacl and pEBVMTlacOlacZ plasmids, the rescuing of the pEBVMTlacOlacZ into the bacteria shown a great stability of this EBV vector (data not shown). Since in transient and stable assays the CMV promoter led to a higher lacZ expression which could bypass the lacl repression, it turned out that this promoter was not suitable upstream of the lacOlacZ sequence (Table liD. The MLP promoter was also not suitable to drive the lacZ gene transcription since it entailed no repression in a 30-day-old cell line (Table lll). This cell population carrying the pEBVMLPiacOlacZ and expressing the lacl gene vector failed to express the lacZ gene after two-three months of culture. Contrary to the synthetic lacO, the wild-type lac operator (lacOl) proved to be a poor binding site with lacl repressor in mammalian cells (Table Ill). Similar results were obtained in transient transfection assays (data not shown). Discussion The main purpose of the experiments reported in this work, was to construct an experimental model with
73 the aim of considering the stringent regulation of an easily quantifiable reporter gene (lacZ) by the product of a unique iacl repressor gene. In this view, we have established a new cell line carrying a pEBV/tdyadCMVlacl plasmid integrated into the genomic DNA and several episomal pEBVMTlacOlacZ vectors. This cell population provides a powerful tool to study disruptions of the genetic control involved in the regulation of the lacZ gene. We could envisage the detection of mutagenic events occurring on the lacl gone in dosing the derepressed-lacZ gone expression. This could represent a breakthrough for assessing the mutagenie potential of xenobiotics. In this experimental model the unique lacI gene is the mutagenic target and the various pEBVMTlacOlacZ vectors serve to detect the aforementioned mutagenic events. We used 293 cells in which adenovirus E la gone products and especially the 289R protein appeared to be a widespread trans activator and repressor of transcription [32,33]. These factors have been found to trigger transcription driven by the CMV promoter [34] but they repressed the Ad2-Major Late Promoter [35]. In transient expression assays (Table I and Fig. 2) and in established cell lines (Table l i d we demonstrated that the E. coli lac operator-repressor complex could efficiently regulate the iacZ expression when the lacI repressor and the symmetrical lac operator were borne by EBV plasmids. Since a recent result has showed little difference between a lacl repressor carrying or not the SV40 T-antigen nuclear localization signal [15], we have used a repressor protein without this caryophylic signal. CMV promoter was not suitable upstream of the lacOlacZ sequence sirlce it led to high lacZ transcription overcoming the lacl repression. Conversely, this promoter was ver~ convenient to drive the lacI gene and to allow a good repression. It has been shown in an aforementioned study an efficient regulation of the CAT gene by the lacl repressor in using the pSViacOCAT and the pCMVlacl vectors [16]. The CAT expression was driven by the strong SV40 promoter upstream of the lacO sequence. The combination of these two vectors with a molar ratio 'placl/placZ' of about 3.8 (10/xg of pSVlacOCAT and 34/,tg of pCMVlacl) entailed a repression rate of 74%. In our system, where the weak mMT-I promoter was located upstream of the lacOlacZ sequence, we reached a repression of 71% after cotransfection of 5 /,tg of pEBVMTlacOlacZ vector with 5/zg of pEBVMLPlacl vector (molar ratio 'placl/placZ' : 1.2) and of 83% with 10 /zg of pEBVMLPlacI vector (molar ratio 'placl/placZ' :2.3) (Table I). Hence, our system appeared to be very efficient to transiently express the lac regulatory complex in human cells. In stable cell lines carrying one EBV vector, we observed a 2.5-fold decrease in the /3gal and CAT activities when the symmetrical iac operator was in-
serted upstream of the lacZ and CAT genes (Table 11). This result confirmed a previous study [14] where the insertion of the iac operator downstream the TATA box and upstream of the transcription start site reduced the transcription level. The lacl repressor did not modulate the iacZ transcription when the wild-type lac operator was placed upstream of the reporter gone (Table liD. In this natural operator the G.C pair at the center of the symmetry disturbs operator recognition and causes an 8-fold decrease of the iacl affinity [36]. To establish cell lines with a lacl + phenotype, we used an integrative form of plasmid to drive the lacl transcription. This pEBVAdyadCMVlacl plasmid carried the transactivation EBV E B N A - l / F R sequences without the Dyad replication origin. This protocol was designed to retain the potential for EBNA-I expression which may aid the stability of the subsequently introduced episomal EBV vectors, however we have not examined this effect in detail. Furthermore, we have carried out numerous transfection experiments with our EBV vectors. Several evidences suggested the occurrence of complex interactions between the transcriptional and replication elements carried hi cis by EBV vectors (data not shown). In particular, in the p189 plasmid, the EBNA-1/FR sequences (with or without the Dyad replication origin) coupled with the highly strong CMV promoter/enhancer usually led to a very high transfection efficiency coupled with a vector copy number per cell near one tdata not shown). The transfection of the 293 cells with the p189 plasmid gave rise to the 293/189 (lacl +) population. As expected, the lacl gene was integrated into the genomic DNA in one or two copies per cell in 12 clones analyzed (data not shown). Transfections of the 293/189 (lacl +) cell population with our EBV shuttle vectors were very efficient and the transfected cells contained and maintained very high amounts of pEBV vectors compared to control 293 cells (data not shown). These results might arise from a protective function of the laci protein. In a previous study, Hu and Davidson [14] considered that transfected DNA could acquire an apparent protection against nicking and cutting and remain in the cell nucleus for a longer period of time when iaci repressor was expressed. We showed that repression could be partially overcome in stable transformed lines when strong promoters such as MLP or CMV were placed upstream of the lac operator or after metal induction of the mMT-I promoter (Table 111). A synergistic effect on the lacZ expression was observed between metal and IPTG in a polyclonal cell line carrying pEBVAdyadCMVlacl and pEBVMTiacOlacZ vectors. In this cell line, repression was partially removed by IPTG to give a 5-told increase of iacZ expression, and a 40-fold induction was found after treatment of the cells with metal ions and IPTG. This cell line owns a gene regulation with several levels
74 of induction, in which the lacZ transcription could be stringently shut off by lacl protein and strongly switch on in the presence of different inducers. Furthermore, we have shown that this regulatory system works well when the target gene is carried on an episomal vector which represents a considerable advantage such as a ~ i b l e and easy recovery of the vectors in bacteria. Many applications of this genetic switch can be envisioned: As mentioned in the beginning of this discussien, we now intend to study disruption of gene transcription arising from environmental injuries, Alternatively, these cell constructs can provide a promising
approach for high production of foreign eukaryotic proteins in a natural mammalian cell context in presonce of heavy metal and IPTG. In this view, synthesis might be easily stopped by withdrawing inducers. This could be designed for production of proteins with high toxicity. Acknowledsments We thank J, Armier for synthesizing the lac operator oligonucleotides and Drs. Hu and Davidson for the generous gift of lacl repressor gone,. The authors are indebted to A, Stary for critical reading of this manuscript. This work has been supported by grants from Rh6ne-Poulenc-Rorer (Antony, France) and by the 'Association pour la Recherche sur le Cancer' (Villejuif, France). References I Jalanko, A., KaUio, A. and Ulmanen, !. (1988) Arch, Virol. 103, 157-166, 2 Lutfalla, G,, Armbruster, L,, Dequin, S. and Bertolotti, R. (1989) Gone 76, 27-39. 3 gioussi~ D., Wilson, F., Daniels, C,, Leveton, C., Taverne, J. and Playfair, J, (1987) EMBO J, 6, 355-361, 4 Holler, R,A,, Son6, K,, Villaret, D., Margolskee, R., Dunne, J,, Itayakawa, H, and Ringold, G,M. (1990) J. Biol. Chem. 265, :~708-5717. 5 .Canfield, V,, Emanu¢l, J.R., Spickofsky, N., Levenson, R. and Margolskee, R.F. (1990) Mol. Cell. Biol. 10, 1367-1372. 6 Young, J,M,, Cheadle, C., Foulke Jr, J.S., Drohan, W.N. and Sarver, N, (1988) Gene 62, 171-185. 7 James, M,R,, Stary, A,, Daya-Grosjean, L., Drougard, C. and Sarasin, A, (198o) Mut,, Res. 202, 169-185,
8 Margolskee, R.F., Kavathas, P. and Berg, P. (1988) Mol. Cell. Biol. 8, 2837-2847. 9 Groger, R.K., Morrow, D.M. and Tykocinski, M.L. (1989) Gene 81, 285-294. 10 DuBridge, R.B. and Calos, M.P. (1988) Mutagenesis 3, !-9. 11 Palombo, F. and Dogliotti, E. (1989) Biochim. Biophys. Acta 1009, 251-256. 12 Stary, A., James, M,R. and Sarasin, A. (1989) J. Virol. 63, 3837-3843. 13 Miller, J.H. and Reznikoff, W.A. (1980) The operon, Cold Spring Harbor Laboratory Press, Cold Spring Harbor. 14 Hu, M.C.T, and Davidson, N. (1987) Cell 48, 555-566. 15 Hu, M.C.T. and Davidson, N. (1990) Mol. Cell. Biol. 10, 61416151. !6 Brown, M., Figge, J., Hanson, U., Wright, C., Jeang, K,T., Khoury, G., Livingston, D.M. and Roberts, T.M. (1987) Cell 49, 603-612. 17 Liu, H-S., Feliciano, E.S. and Stambrook, P.J. (1989) Proc. Natl. Acad, Sci. USA 86, 9951-9955. 18 Figge, J., Wright, C., Collins, C J,, Roberts, T.M. and Livingston, D.M. (1988) Cell 52, 713-722. 19 Fucrst, T,R., Fermandez, M.P. and Moss, B. (1989) Proc. Natl. Acad. Sci. USA 86, 2549-2553. 20 Durnam, D.M. and Palmiter, R.D. (1981) J. Biol. Chem. 256, 5712-5716. 21 James, M.R., Sarasin, A., Perricaudet, M. and Joab, !. (1990) Gene 86, 233-239. 22 Maniatis, T., Fritsch, E.F. and Sambrook, J. (1982) Molecular cloning: a laboratory manual, Cold Spring Harbor Laboratory Press, Cold Spring Harbor. 23 Reisman, D., Yates, J. and Sugden, B. (1985) Mol. Cell. Biol. 5, t822-1832. 24 Chittenden, T., Lupton, S. and Levine, A.J. (1989) J. Virol. 63, 3016-3025. 25 Boshart, M., Weber, F., Jahn, G., Dorsch-Hiisler, K., Fleckenstein, B. and Schaffner, W (1985) Cell 41,521-530. 26 Sadler, J,R., Sasmor, H. and Betz, J.L. (1983) Proc. Natl. Acad. Sci. USA 80, 6785-6789. 27 Madzak, C., Menck, C.F., Armier, J. and Sarasin, A. (1989) J. Mol. Biol. 205, 501-509. 28 Wigler, M., Pellicer, A., Silverstein, S. and Axel, R. (1978) Cell 14, 725-731. 29 Nielsen, D.A., Chou, J., MacKrel, A.J, Casadaban, M.J. and Steiner, D.F. (1986) Proc. Natl. Acad. Sci. USA 80, 5198-5202. 71) [3~,tdtor0, M. (i976) Anal. Biochem. 72, 248-254. I Sleigh, M.J. (1986) Anal. Biochem. 156, 251-256. ;)2 Nevins, J.R. (1987) Microbiol. Rev. 51,419-430. 2,3 Velcich, A. and Ziff, E.B. (1989) Oncogene 4, 707-713. 3J, Gorman, C.M., Gie, D., McCray, G. and Huang, M. (1989) Virology 17I, 377-385. 35 BorreUi, E., Hen, R. and Chambon, P. (1984) Nature 317, 608612. 36 Simons, A., Tils, D., Von Wilcken-Bergmann, B. and Miiller-Hill, B. (1984) Proc. Natl. Acad. Sci. USA 81, 1624-1628.
68
BBAEXP 92340
Regulation of the Escherichia coli lac operon expressed in human cells Denis S.F. Biard, Michael R. James ~, Andr6 Cordier 2 and Alain Sarasin Laboratory of Molecular Genetics, UPR 42, CNRS, instimt de Recherches Sciemifiques sur le Cancer, V'dlejuif (France) (Received 5 August 1991)
Key words: Transcription regulation; EBV-vector; lac operon" (Human cell)
We have investigated the use of various Epstein-Barr virus (EBV)-based vectors bearing the two components of the Escherichia coil lac operator-represser (lacO, iacl) complex. Our aim was to develop a model system of gene expression by looking at the transcription of the bacterial #-galactosidase coding gene (lacZ) in 293 human embryonic kidney cells. Several vectors have been built carrying different promoters upstream of the lacl and lacZ genes and in which natural or synthetic operator sequences were inserted in the 5' part of the iacZ gene. In transient expression assays we achieved efficient !acZ gent repression which could be released by the specific inducer isopropyl //-D-thiogalactoside (IPTG). A stable transtormed cell line carrying two EBV-derived plasmids with the building blocks of the lac operator/represser system was established. This cell line allowed us to achieve a wide range of lacZ gene regulation. In this cell line IPTG alone could remove the repression to trigger a 5-fold increase of lacZ expression. Heavy metal ions, which induced the mouse metaliothionein I promoter located upstream of the iacZ gen¢, added together with IPTG gave rise to a 40-fold induction of lacZ expression.
Introduction in recent years Epstcin-Barr virus (EBV)-based shuttle vectors have emerged as powerful tools for genetic engineering. These vectors could replicate in a broad variety of cell lines such as human or monkey cells [1] without interfering with the expression of differentiated cell functions [2]. They appeared appropriate for efficient shuttling of large DNA fragments from mammalian cells to E. coli [3]. EBV vectors have successfully allowed foreign DNA expression in various recipient cells like the human tumor necrosis factor in human cells [3] and its receptor [4]; the rat Na-K ATPase a 2 isoform in human and monkey cells [5]; the human interferon 7 and extracellular domain of the human epidermal growth factor receptor in human
! Present address: Centre d'Etude du Polymorphisme humain, 27 rue Juliette Dodu, 75010 Paris, France, ~" Present address: lnstitut de Recherches sur la S6curit6 du M6dicament Rh6ne Poulenc-Rorer, 20 Quai de la R6volution, 94140 Aifortville, France. Correspondence: A. Sarasin, Laboratory of Molecular Genetics, UPR 42, CNRS, lnstitut de Recherches Scientifiques sur le Cancer, B.P. No. 8, 94801 Villejnif, France.
cells [6] or E. coli chloramphenicol acetyl transferase (CAT) gene in mammalian cells [1,7]. EBV vectors have been used for efficient expression of cDNA libraries [8] and effective antisense RNA-mediated inhibition of presclccted genes [9] in human cells. Moreover, these shuttle vectors proved to be suitable tools to study biological processes, such as mutagenesis [7,10,11] or recombination [12] of target genes after transfer into mammalian cells. In the present work, we aimed to devise a genetic assay composed of an easily quantifiable reporter gene which is regulated by various inducer and represser elements. In this goal, the regulation gene (represser) could be the target of external agents. To develop this approach, we have stably maintained and expressed the E. coli lac operator/represser complex in human cell lines by use of EBV vectors. This regulatory complex represents a paradigm for the study of control of gene expression in response to various environmental conditions [13]. The lacO/lacl system has been shown to be functional in mammalian cells both after transient or stable expression assays [14-18]. cis- and trans-acting elements of the E. coli lac operon have been used to regulate gene expression in vaccinia virus [19]. The repression could be reversibly removed using a specific inducer, isopropyl fl-D-thiogalactoside (IPTG). This in-
69 ducer appeared to bind to repressor therefore decreasing its affinity for the lac operator. In preliminary experiments, we have constructed different EBV vectors carrying the lacl gene on one hand and a lacO sequence on the other hand to obtain the best combination of plasmids which will allow appropriate regulation of the lacZ gene by the lacl repressor. Different promoters have be inserted downstream of the iacl and the lacO sequences. The functionality of the lacO/lacI complex was firstly assessed through transient expression assays, and ultimately in stable established cell lines. The specificity of repression by lacI repressor was apparent by the ability to be alleviated using IPTG. Our ultimate goal was to integrate one copy of the repressor gene (lad) into the genomic DNA in an effort to study disruptions of the transcriptional control of this regulatory gene. In view of these issues, we have established a cell line bearing one lacl gene constitutively expressed. These cells also carried a pEBVMTlacOlacZ vector giving rise to an efficient repression of the lacZ gene by the lacl repressor. The mouse metallothionein-I promoter (mMT-I) inserted upstream of lac operator has been shown to be active in almost all tissues and cell lines [20] and we have previously shown that it is efficiently induced by heavy metals when carried on an EBV episome [7,21]. In stable transformed lines, the lacZ expression was regulated by the lacO/lacl complex. The repression could be partially relieved by metal induction of the mMT-I promoter. In the presence of IPTG and metal cocktail (zinc and cadmium) we observed a synergistic effect on the lacZ transcription with a 40-fold increase.
AMPr pBr o
~ ' ~ ~
TK pr < f
~'~
.__
~hl~_
]J
G418r
p145 ~ pEBVMTlacOlacZ 12589bp
BNA'I
lacO lacZ
Fig. 1. Genetic maps of the pEBVAdyadCMVlacl (p189) and pEBVMTlacOlacZ (pl45). The elements of pig9 vector are in the inward circle with (i) the segment containing the EBV nuclear antigen ! (EBNA-I) and the repeat family (FR) (black-filled area): (it) the reporter gene transcriptional cassette comprising the CMV pronloler ((,'MV pr), the laci gene and the SV4(I late-region polyadenylation signal (SVpA + ) and oriented in the clockwise direction as EBV sequences (stippled area): (iii) the geneticin resistance gene (G418r) under control of the HSV-thymidine kinase promoter (TK pr) and the HSV-TK polyadenylation signal (TKpA'-) and oriented in the counterclockwise direction (open area): (iv) the oBR322 origin of replication (pBR ori) and the bacterial fl-lactamase gene (AMPr) in the same orientation as the eukaryotic resistance gene (hatched area). The pl45 plasmid (outward circle) carries the whole EBV replication origin with the repeat family (FR) and the dyad sequence (DYAD). Other building blocks are as follows: (i) the lacZ gene transcriptional cassette with the mMT-! promoter (mMT-! pr). the symmetrical lac operator (htcO) inserted into Xhal sites and the SVpA" in the clockwise direction: (it) the hygromycin B phosphotransferase gene (HYGROr) with TK pr and TKpA+ sequences in counterclockwise orientation.
Materialsand Methods Plasmid constructions All procedures for plasmid isolation, restriction endonuelease cleavage, ligation and transformation were carried out as described elsewhere [22]. In this work, vectors defective in the EBV dyad sequence were named pEBVAdyad. The vector pEBVAdyadCMVlacl (p 189 vector; Fig. l) was derived from the plasmid p205. It carried the pBR322 replication origin, the ampicillin resistance gene, the EBV EBNA-1 and FR sequences [23]. The EBV dyad palindromic sequence (between Spel and Hpal sites) was deleted to remove the putative replication origin [23,24]. Upstream of the laci gene lies the immediateearly human cytomegalovirus (CMV) promoter-enhancer [25]. The lacl gene was generously provided by M.C.T. Hu. It contained a modified initiation codon (ATG) with attendant Simian Virus 40 polyadenylation signal. The selection of cells carrying p189 plasmid, namely the 293/189 (lacl +) cell line, was by geneticin , (G~.18) resistance, pEBVMLPIacl plasmid was derived
from p205MLPCAT vectors [7] by the insertion of the lacl gene via the HindIII and BamH! restriction sites. To clone the iacZ gene downstream of the mouse metailothionein-i promoter (mMT-I), the adenovirus major late promoter (MLP) or the CMV promoter, we have isolated the HindlII/BamHI fragment from pRSVZ and inserted it into corresponding sites of pEBVCAT vectors to produce pEBVMTiacZ (p149), pEBVMLPiacZ and pEBVCMVlacZ plasmids. A synthetic operator with a 5' attendant Xbal restriction site (oligonucleotide 5'-CTAGATTGTGAGCGCTCACAAT-3') was inserted into Xbal site to obtain pEBVMTlacOCAT, pEBVMTlacOlacZ (p145; Fig. l), pEBVMLPlacOlacZ and pEBVCMVlacOlacZ plasraids. The lacl repressor possessed more affinity to the synthetic iac operator than the natural one [26]. For cloning, the selection of recombinant plasmids was done in WP3 bacteria in the presence of X-Gai chromogenic substrate [27] and verified by DNA sequence analysis. The natural and asymmetrical operator (oligonucleotide 5 '-CTAGAATTGTGAGCGGATAACAA-
70 TT-Y) was inserted into the same Xbal site in the two orientations giving rise to pEBVMTIacO1CAT, pEBVMTlacO2CAT, pEBVMTIacOIlacZ and pEBVMTlacO21acZ vectors. All these vectors conferred resistance to hygromycin B.
protein] x [CAT reference activity/CAT experimental activity]. For cell lines which only expressed CAT we followed exactly the Sleigh protocol [31]. CAT activity was given as CPM of [14C]acetyl-chloramphenicol/~g protein.
Celi cuhures and DNA transfection 293 human embryonic kidney cells were grown in Dulbecco's modified Eagle's medium supplemented with 10% foetal calf serum, 100 U / m l of penicillin and 100/zg/ml of streptomycin (DMEM/FCS medium) in a 5% COz incubator. For long term maintenance of cell lines, 5 #g of plasmid per 10 cm plate were introduced into the recipient cells by calcium phosphate coprecipitation [28]. The precipitates were left 6 h on cells. Selection was initiated on day two by cell tt),psination followed by replating at appropriate dilutions in a selective medium containing 125/zg/ml of hygromycin B (Sigma) and/or 500 ~ g / m l of (3418 (Sigma). To determine CAT and /~-galactosidase (/~gal) activities cells were seeded at appropriate dilutions. 3 days later, metal ions (100 p.M ZnCl., and 1 gM CdCI e) and/or IPTG (10 raM; Gibco-BRL) were added to the medium for 2 days unless otherwise mentioned and the cells were collected. For transient expression, 5 /zg (unless otherwise indicated) of pEBVlacl and pEBViacOlacZ plasmids and 1 /zg of a control plasmid of transfection (p205MLPCAT) were transformed into 10 cm plates containing exponentially growing cells. Medium was removed 6 h later and cells were refed with DMEM/FCS medium with or without 10 mM of IPTG. 48 h after, cells were collected and lysed to assess ~-galactosidase and CAT activities.
Results
~-Galactosidase and CAT assays The cells from stable transformed lines were scraped off dishes and pelleted in 300 tL! of buffer H (250 mM s u c r o s e / l ~ mM sodium phosphate, pH 7.5). Cells were lysed by four cycles of freeze-thawing then centrifuged. ~gal activities were determined according to Nielsen et al. [29] and expressed as units (nmol of orthonit%phenol formed per rain) per mg of protein. Purified E. coil ~l-galactosidase (Sigma) was used as a standard. Protein concentration was determined using the Bradford procedure [30] with bovine serum albumin as standard. For transient expression studies, cells were collected in 600 t~i of buffer H and enzymatic extracts were recovered as described above. An aliquot of each sample was heated at 65"C during 10 rain and assayed for CAT activity using a non chromatographic protocol [31]. To reduce the variability between each transfection experiment,/~lgal activities were expressed as '/3gai normalized values' corresponding to [/3gal units/rag
Transient expression assays Various transient expression assays have been carried out to assess the efficiency of repression of the iacZ gene by the lacl repressor into 293 human cells. To compare results from the same transfection or independent ones we have cotransfected the plasmid p205MLPCAT as a reference. This vector gave strong CAT expression 48 h after transfection. To characterize: the properties of the lacO/lacl complex in 293 cells, we have modified the expression rate of the two components (iacl and lacOlacZ sequences) in increasing the strength of the promoter or the amounts of DNA transfected. In absence of pEBVMLPIacl vector we observed a regular increase in ~gal normalized values when the amounts of pEBVMTlacOlacZ vectors increased up to 10/zg (Table I; lines 1-3). With 5 ~g of pEBVMLPlacl vector we reduced the lacZ expression. The repression rate decreased while the amount of transfected pEBVMTlacOlacZ vector increased (Table l; lines 4-6). Then, the magnitude of repression of the lac system could varied by transfecting more amounts of pEBVMTlacOlacZ vectors. A very high lacZ gene expression triggered by promoters stronger than mMT-! could also
TABLE !
Transient expression assay: rariation of amount of pEBVMTlacOlacZ ~'ector transfected with or without pEBVMLPlacl i'ector Different amounts of pEBVMTlacOlacZ and pEBVMLPlacl vector were cotransfected with I p,g of p205MLPCAT control plasmid. After removing the DNA-calcium precipitate the cells were fed or not with 10 mM of IPTG in DMEM/FCS medium for 48 h and assay was performed as in Materials and Methods./~gal normalized units correspond to (/3gal units/mg protein)×(CAT reference activity/ CAT experimental activity). These experiments were reproducible with a percentage of error of around 5%. Line
Transfected plasmids pEBVMTlacOlacZ
/]gal normalized pEBVMLPlacl
units IPTG (10 mM) -
1
2 3 4 5 6 7 8
2/~g
5/zg 10/.tg 2/zg 5/.tg 10 t~g 5 p,g 5 ~ttg
-
5/zg 5 ~g 5 p,g 2/.tg 10/.tg
130 232 631 56 68 3'78 94 39
+
132 255 617 130 246 857 238 91
71 flgal Pormalized units 800 600 I
400
! i
200 ,
Control
mMT-I
MLP
CMV AdyadCMV
Promoter upstream of the lacl gene Fig. 2./3-Galactosidase activities in transient expression assays: role of the strength of promoter upstream of the loci gene. l0 #g of pEBVMTlacOlacZ vector were transfected with 5/zg of pEBVMTlacl, pEBVMLPlacl, pEBVCMVlacl or pEBVJdyadCMVlacl vector and I/.tg of p205MLPCAT control plasmid. Assay ~,as perfi~rmed as in Materials and Melhods. Filled bars: - IPTG. Open bars: + IPTG. Control indicated the average level of the lacZ gene expression with the pEBVMTlacOlacZ vec!or alone without the pEBVlacl one.
bypass the lad repression. Therefore, when we cotransfected pEBVMLPlacl vector with another vector carrying the mMT-I, MLP or CMV promoters upstream of the lacZ gene, we obtained a repression rate more and more reduced (data not shown). The CMV promoter appeared to be the strongest and the mMT-I promoter the weakest. Since very strong promoters, such as MLP and CMV could overcome the l a d repression, we therefore focused our attention on the metal inducible mMT-I promoter. Enhanced flgal repression could be achieved in transfecting more DNA of pEBVMLPIacl vector. When amounts of pEBVMLPIacl vector reached 2, 5 and l0 /,tg we observed 60%, 71% and 83% of repression of /3gal activity (Table I, lines 7,5,8) as compared with transfection of pEBVMTiacOlacZ alone (Table l, line 2). The derepression rate by IPTG was lowered in increasing the amounts of pEBVMLPlacl vector. 2 and l0 ~g of pEBVMLPIacl vector gave respectively rise to a derepression by IPTG of 93% and 36% (Table !, lines 7, 8) of control values (Table I, line 2). To trigger an increased repression, we changed the promoter upstream of the lad gene for a stronger one. We observed a very important repression when the laci transcription was driven by the very strong promoter/ enhancer CMV (Fig. 2). For example the CMV promoter induced a repression 60% greater than with the mMT-I promoter (Fig. 2). In this experiment, the increased l a d gene expression triggered by strong promoters was not correlated with a decrease in the derepression level (with IPTG) as seen in experiments reported in the Table I. However, from different assays performed with various EBV vectors we observed that
the promoter/enhancer inserted upstream of the lacl gene could interfere with the iaeZ gene expression when the lad and lacZ genes were borne by two different vectors. We always obtained increased /3gal activities when the cotransfection where done with pEBVCMVlacl and pEBVMTIacZ vectors compared to transfection with pEBVMTIacl and pEBVMTlacZ vectors (data not shown). The CMV promoter seemed to be a good candidate to control the lad transcription in established cell lines. In the absence of lad gene or lacO sequence, IPTG did not modified the laeZ gene expression whatever the promoter used upstream of this gene (data not shown). In our ultimate aim, the integration of a lad repressor gene into the genomic DNA will be achieved in deleting the Dyad replication origin. As one step towards addressing this purpose, we have constructed the pEBVAdyadCMVlacl (p189)plasmid. In tran,;ient expression experiment, this deletion did not hinder the repression efficiency of the pEBVAdyadCMVlaci plasmid (Fig. 2). In the aforementioned experiments, we found an efficient regulation of the /3gal activity by the lad protein 48 h after transfection which could be partially relieved by IPTG.
Stably transformed cell lines with one EBV t'ector We have analyzed the behavior of our different plasmids in 293 cells after several weeks of mass culture. The success of establishment of cell lines which expressed one EBV vector depended on the plasmid construct. For vectors carrying the lacl gene when mMT-I or MLP promoters were present, all cell lines failed to express reporter gene activity after long term culture (more than 2-3 months) or they encountered difficulties in remaining stable, in contrast, the presence of the CMV promoter helped us to obtain cell lines with normal growth. In any construction where the CMV promoter was placed upstream of a reporter gene such as CAT, lad or lacZ gene, we always observed high transfection efficiencies of host cells related to a low vector copy number (data not shown). Using 50-day-old transformed polyclonal populations we checked the occurrence of hindrances in promoter activities arising from the insertion of a iac operator sequence upstream of CAT or lacZ gene (Table II). The presence of the symmetrical lacO downstream of the TATA-box entailed a 2.5-fold decrease in basal mMT-I promoter activities, whereas no inhibitory effect was observed with the natural lac operator (iacOl or lac02). However, the metal responsiveness of the mMT-I promoter was maintained with an induction rate of roughly 18-fold for pEBVlacOlacZ and pEBVlacOCAT constructs whatever the operator sequence inserted between the mMT-I promoter and the reporter gene. With the CAT or lacZ reporter
TABLE II
[3gai and CAT actirities in 50 days old 293 stably transformed cell lines and metal induction of mMT-! promoter EBV
ZnCd a
plasmid structure pEBVlacZ pEBVCAT
+ +
Reporter gene activities b promoter upstream of reporter gene mMT-I MTlacO MTlacOl
MTlacO2
MLP
CMV
186 2511 203 1557
340 2841 288 2616
0 0 639 624
2 366 2242 2189 2208
88 1507 71 1667
332 3600 118 2391
a ZnCd cocktail (100 ~M ZnCI 2 + 1/zM CdCI2)was added onto cells for 8 h. Enzymatic activities were assessed as in Materials and Methods. h//gal activities were expressed as ,Sgal units/rag protein while CAT activities were expressed in CPM of 14C-acetyl-chloramphenicol/tzg protein, lacO means the symmetrical lac operator, lacOI corresponds to the natural lac operator in the first orientation whereas lac02 was the natural lac operator in the second orientation,
genes, we found a higher expression driven by the CMV promoter than with MLP or mMT-I promoter without metal induction, The MLP promoter placed upstream of the lacZ gene led to no expression of this gene 50 days after the transfection, while the CMV promoter allowed an extensive lacZ transcription.
Stably transformed cell lines with two EBV cectors For the purpose to establish a cell line with the EBNA-1 +/lacl + phenotype, we have constructed the pEBVAdyadCMVlacI (p189)vector. This construction kept the entire EBNA-I transcriptional cassette. Moreover, the EBNA-1/FR sequences coupled with the CMV promoter upstream of the lacl gene led to a vector copy number per cell near one (data not shown). The 293 cells have been transfected with the p189 vector to give rise to the 293/189 (lacl +) cell line. 1 month later, p189 plasmid was found to be integrated into genomie DNA with one or two copies per cell in all 12 clones tested. All clones constitutively expressed the lacl gene (data not shown). The 293/189 (lacl +) cell population was retransfected with different vectors carrying the lacZ gene. When we assessed the transfection efficiencies of our EBV-derived vectors, we always TABLE ill
~-Galactosidase expressio.~ in stable transformed cell line days after transfection of 293/189 cells, we allowed cells to grow for 48 h in a DMEM/FCS medium in presence or not of 10 mM llrrG and 100/~M Z n C I : / l ~M CdCI: (ZnCd)./3gal activities were performed as mentioned in Materials and Methods. lacO sequence means symmetrical lac operator, whereas lacOI indicates natural lac o~rator in the first orientation. NT: not tested.
pEBVlacZvector carried by the 293/189 cell line
#gal units/rag protein -
IPTG
ZnCd
pEBVMTlacOlacZ pEBVMTlacOIlacZ pEBVMLPlacOiacZ pEBVCMVlacOlacZ
13 266 224 2125
66 239 189 2 451
131 2027 NT NT
IPTG ZnCd 536 1925 NT NT
reached higher transfection efficiencies after transfection into 293/189 (lacl +) cells than into 293 cells, with for example a 3.7 and a 2.8 fold-increase with pEBVMTlacOlacZ and pEBVMTlacZ vectors respectively (this experiment has been repeated six times and the increase was statistically significant). We have analyzed the regulatory expression of the lac complex in 30 days old cell lines carrying the two building blocks of the lac regulatory complex. In a mixed clone population carrying the pEBVAdyadCMViacl and pEBVMTlacOlacZ plasmids there was a 5-fold derepression triggered by 10 mM of IPTG for 48 h in culture medium. The metal induction of the mMT-I promoter partially overcame lacl repression (10-fold increase of iacZ expression). The presence of the two inducers (ZnCd and IPTG) triggered a tremendous lacZ expression with a 40-fold increase, as also found by Hu and Davidson with IPTG and glucocorticoid hormone [15]. In the cell line bearing the pEBVAdyadCMVlacl and pEBVMTlacOlacZ plasmids, the rescuing of the pEBVMTlacOlacZ into the bacteria shown a great stability of this EBV vector (data not shown). Since in transient and stable assays the CMV promoter led to a higher lacZ expression which could bypass the lacl repression, it turned out that this promoter was not suitable upstream of the lacOlacZ sequence (Table liD. The MLP promoter was also not suitable to drive the lacZ gene transcription since it entailed no repression in a 30-day-old cell line (Table lll). This cell population carrying the pEBVMLPiacOlacZ and expressing the lacl gene vector failed to express the lacZ gene after two-three months of culture. Contrary to the synthetic lacO, the wild-type lac operator (lacOl) proved to be a poor binding site with lacl repressor in mammalian cells (Table Ill). Similar results were obtained in transient transfection assays (data not shown). Discussion The main purpose of the experiments reported in this work, was to construct an experimental model with
73 the aim of considering the stringent regulation of an easily quantifiable reporter gene (lacZ) by the product of a unique iacl repressor gene. In this view, we have established a new cell line carrying a pEBV/tdyadCMVlacl plasmid integrated into the genomic DNA and several episomal pEBVMTlacOlacZ vectors. This cell population provides a powerful tool to study disruptions of the genetic control involved in the regulation of the lacZ gene. We could envisage the detection of mutagenic events occurring on the lacl gone in dosing the derepressed-lacZ gone expression. This could represent a breakthrough for assessing the mutagenie potential of xenobiotics. In this experimental model the unique lacI gene is the mutagenic target and the various pEBVMTlacOlacZ vectors serve to detect the aforementioned mutagenic events. We used 293 cells in which adenovirus E la gone products and especially the 289R protein appeared to be a widespread trans activator and repressor of transcription [32,33]. These factors have been found to trigger transcription driven by the CMV promoter [34] but they repressed the Ad2-Major Late Promoter [35]. In transient expression assays (Table I and Fig. 2) and in established cell lines (Table l i d we demonstrated that the E. coli lac operator-repressor complex could efficiently regulate the iacZ expression when the lacI repressor and the symmetrical lac operator were borne by EBV plasmids. Since a recent result has showed little difference between a lacl repressor carrying or not the SV40 T-antigen nuclear localization signal [15], we have used a repressor protein without this caryophylic signal. CMV promoter was not suitable upstream of the lacOlacZ sequence sirlce it led to high lacZ transcription overcoming the lacl repression. Conversely, this promoter was ver~ convenient to drive the lacI gene and to allow a good repression. It has been shown in an aforementioned study an efficient regulation of the CAT gene by the lacl repressor in using the pSViacOCAT and the pCMVlacl vectors [16]. The CAT expression was driven by the strong SV40 promoter upstream of the lacO sequence. The combination of these two vectors with a molar ratio 'placl/placZ' of about 3.8 (10/xg of pSVlacOCAT and 34/,tg of pCMVlacl) entailed a repression rate of 74%. In our system, where the weak mMT-I promoter was located upstream of the lacOlacZ sequence, we reached a repression of 71% after cotransfection of 5 /,tg of pEBVMTlacOlacZ vector with 5/zg of pEBVMLPlacl vector (molar ratio 'placl/placZ' : 1.2) and of 83% with 10 /zg of pEBVMLPlacI vector (molar ratio 'placl/placZ' :2.3) (Table I). Hence, our system appeared to be very efficient to transiently express the lac regulatory complex in human cells. In stable cell lines carrying one EBV vector, we observed a 2.5-fold decrease in the /3gal and CAT activities when the symmetrical iac operator was in-
serted upstream of the lacZ and CAT genes (Table 11). This result confirmed a previous study [14] where the insertion of the iac operator downstream the TATA box and upstream of the transcription start site reduced the transcription level. The lacl repressor did not modulate the iacZ transcription when the wild-type lac operator was placed upstream of the reporter gone (Table liD. In this natural operator the G.C pair at the center of the symmetry disturbs operator recognition and causes an 8-fold decrease of the iacl affinity [36]. To establish cell lines with a lacl + phenotype, we used an integrative form of plasmid to drive the lacl transcription. This pEBVAdyadCMVlacl plasmid carried the transactivation EBV E B N A - l / F R sequences without the Dyad replication origin. This protocol was designed to retain the potential for EBNA-I expression which may aid the stability of the subsequently introduced episomal EBV vectors, however we have not examined this effect in detail. Furthermore, we have carried out numerous transfection experiments with our EBV vectors. Several evidences suggested the occurrence of complex interactions between the transcriptional and replication elements carried hi cis by EBV vectors (data not shown). In particular, in the p189 plasmid, the EBNA-1/FR sequences (with or without the Dyad replication origin) coupled with the highly strong CMV promoter/enhancer usually led to a very high transfection efficiency coupled with a vector copy number per cell near one tdata not shown). The transfection of the 293 cells with the p189 plasmid gave rise to the 293/189 (lacl +) population. As expected, the lacl gene was integrated into the genomic DNA in one or two copies per cell in 12 clones analyzed (data not shown). Transfections of the 293/189 (lacl +) cell population with our EBV shuttle vectors were very efficient and the transfected cells contained and maintained very high amounts of pEBV vectors compared to control 293 cells (data not shown). These results might arise from a protective function of the laci protein. In a previous study, Hu and Davidson [14] considered that transfected DNA could acquire an apparent protection against nicking and cutting and remain in the cell nucleus for a longer period of time when iaci repressor was expressed. We showed that repression could be partially overcome in stable transformed lines when strong promoters such as MLP or CMV were placed upstream of the lac operator or after metal induction of the mMT-I promoter (Table 111). A synergistic effect on the lacZ expression was observed between metal and IPTG in a polyclonal cell line carrying pEBVAdyadCMVlacl and pEBVMTiacOlacZ vectors. In this cell line, repression was partially removed by IPTG to give a 5-told increase of iacZ expression, and a 40-fold induction was found after treatment of the cells with metal ions and IPTG. This cell line owns a gene regulation with several levels
74 of induction, in which the lacZ transcription could be stringently shut off by lacl protein and strongly switch on in the presence of different inducers. Furthermore, we have shown that this regulatory system works well when the target gene is carried on an episomal vector which represents a considerable advantage such as a ~ i b l e and easy recovery of the vectors in bacteria. Many applications of this genetic switch can be envisioned: As mentioned in the beginning of this discussien, we now intend to study disruption of gene transcription arising from environmental injuries, Alternatively, these cell constructs can provide a promising
approach for high production of foreign eukaryotic proteins in a natural mammalian cell context in presonce of heavy metal and IPTG. In this view, synthesis might be easily stopped by withdrawing inducers. This could be designed for production of proteins with high toxicity. Acknowledsments We thank J, Armier for synthesizing the lac operator oligonucleotides and Drs. Hu and Davidson for the generous gift of lacl repressor gone,. The authors are indebted to A, Stary for critical reading of this manuscript. This work has been supported by grants from Rh6ne-Poulenc-Rorer (Antony, France) and by the 'Association pour la Recherche sur le Cancer' (Villejuif, France). References I Jalanko, A., KaUio, A. and Ulmanen, !. (1988) Arch, Virol. 103, 157-166, 2 Lutfalla, G,, Armbruster, L,, Dequin, S. and Bertolotti, R. (1989) Gone 76, 27-39. 3 gioussi~ D., Wilson, F., Daniels, C,, Leveton, C., Taverne, J. and Playfair, J, (1987) EMBO J, 6, 355-361, 4 Holler, R,A,, Son6, K,, Villaret, D., Margolskee, R., Dunne, J,, Itayakawa, H, and Ringold, G,M. (1990) J. Biol. Chem. 265, :~708-5717. 5 .Canfield, V,, Emanu¢l, J.R., Spickofsky, N., Levenson, R. and Margolskee, R.F. (1990) Mol. Cell. Biol. 10, 1367-1372. 6 Young, J,M,, Cheadle, C., Foulke Jr, J.S., Drohan, W.N. and Sarver, N, (1988) Gene 62, 171-185. 7 James, M,R,, Stary, A,, Daya-Grosjean, L., Drougard, C. and Sarasin, A, (198o) Mut,, Res. 202, 169-185,
8 Margolskee, R.F., Kavathas, P. and Berg, P. (1988) Mol. Cell. Biol. 8, 2837-2847. 9 Groger, R.K., Morrow, D.M. and Tykocinski, M.L. (1989) Gene 81, 285-294. 10 DuBridge, R.B. and Calos, M.P. (1988) Mutagenesis 3, !-9. 11 Palombo, F. and Dogliotti, E. (1989) Biochim. Biophys. Acta 1009, 251-256. 12 Stary, A., James, M,R. and Sarasin, A. (1989) J. Virol. 63, 3837-3843. 13 Miller, J.H. and Reznikoff, W.A. (1980) The operon, Cold Spring Harbor Laboratory Press, Cold Spring Harbor. 14 Hu, M.C.T, and Davidson, N. (1987) Cell 48, 555-566. 15 Hu, M.C.T. and Davidson, N. (1990) Mol. Cell. Biol. 10, 61416151. !6 Brown, M., Figge, J., Hanson, U., Wright, C., Jeang, K,T., Khoury, G., Livingston, D.M. and Roberts, T.M. (1987) Cell 49, 603-612. 17 Liu, H-S., Feliciano, E.S. and Stambrook, P.J. (1989) Proc. Natl. Acad, Sci. USA 86, 9951-9955. 18 Figge, J., Wright, C., Collins, C J,, Roberts, T.M. and Livingston, D.M. (1988) Cell 52, 713-722. 19 Fucrst, T,R., Fermandez, M.P. and Moss, B. (1989) Proc. Natl. Acad. Sci. USA 86, 2549-2553. 20 Durnam, D.M. and Palmiter, R.D. (1981) J. Biol. Chem. 256, 5712-5716. 21 James, M.R., Sarasin, A., Perricaudet, M. and Joab, !. (1990) Gene 86, 233-239. 22 Maniatis, T., Fritsch, E.F. and Sambrook, J. (1982) Molecular cloning: a laboratory manual, Cold Spring Harbor Laboratory Press, Cold Spring Harbor. 23 Reisman, D., Yates, J. and Sugden, B. (1985) Mol. Cell. Biol. 5, t822-1832. 24 Chittenden, T., Lupton, S. and Levine, A.J. (1989) J. Virol. 63, 3016-3025. 25 Boshart, M., Weber, F., Jahn, G., Dorsch-Hiisler, K., Fleckenstein, B. and Schaffner, W (1985) Cell 41,521-530. 26 Sadler, J,R., Sasmor, H. and Betz, J.L. (1983) Proc. Natl. Acad. Sci. USA 80, 6785-6789. 27 Madzak, C., Menck, C.F., Armier, J. and Sarasin, A. (1989) J. Mol. Biol. 205, 501-509. 28 Wigler, M., Pellicer, A., Silverstein, S. and Axel, R. (1978) Cell 14, 725-731. 29 Nielsen, D.A., Chou, J., MacKrel, A.J, Casadaban, M.J. and Steiner, D.F. (1986) Proc. Natl. Acad. Sci. USA 80, 5198-5202. 71) [3~,tdtor0, M. (i976) Anal. Biochem. 72, 248-254. I Sleigh, M.J. (1986) Anal. Biochem. 156, 251-256. ;)2 Nevins, J.R. (1987) Microbiol. Rev. 51,419-430. 2,3 Velcich, A. and Ziff, E.B. (1989) Oncogene 4, 707-713. 3J, Gorman, C.M., Gie, D., McCray, G. and Huang, M. (1989) Virology 17I, 377-385. 35 BorreUi, E., Hen, R. and Chambon, P. (1984) Nature 317, 608612. 36 Simons, A., Tils, D., Von Wilcken-Bergmann, B. and Miiller-Hill, B. (1984) Proc. Natl. Acad. Sci. USA 81, 1624-1628.
