Inducible vectors for expression in mammalian cells Geoffrey T. Yarranton Celltech Ltd, Slough, U K The most recent developments in mammalian cell inducible expression systems have involved the use of bacterial gene control elements and viral transactivator proteins• The combination of hybrid viral transactivator and bacterial repressor proteins, and simple chemical inducers can provide induction ratios of over 1000-fold. These developments will have applications in both cell-based research and the generation of transgenic animals. Current Opinion in Biotechnology 1992, 3:506-511
Introduction Regulated expression of a gene product in mammalian cells is often desirable, especially w h e n the product is either cytotoxic or cytostatic. Early attempts to develop inducible gene expression vectors relied u p o n the use of heat shock control [1], metallothionine promoters and h e a v y metal ion control [2], or steroid regulatory promoters [3,4]. Although these systems give some level of regulation, they are often 'leaky' (i.e. uninduced levels of expression are clearly detectable) and induction ratios are low, usually about 10-20 fold (Table 1). An additional disadvantage of these systems is that the inducing treatment invariably leads to the co-ordinate induction of a variety of cellular genes, hence making conclusions concerning the physiological consequences of the expressed heterologous gene product difficult to interpret. The most promising developments in the area of inducible expression vectors for m a m m a l i a n cells have come from groups w h o have attempted to adapt bacterial transcriptional control systems to mammalian cells [2]. One advantage of these approaches is that specificity is achieved by the interaction of a bacterial protein with a bacterial DNA binding sequence (operator). The operator sequence is unlikely to occur in the promoter sequence of a mammalian gene, hence non-specific activation or repression of cellular gene expression is unlikely to occur. For this reason, as well as the fact that the induction ratios achieved with these chimeric systems tend to exceed those of the homologous systems, this review will focus mainly on the d e v e l o p m e n t of inducible expression systems in mammalian cells that e m p l o y bacterial elements.
Two recent developments in inducible gene expression systems based on mammalian cell or viral g e n e control elements which will be discussed in this review are the use of locus control regions (LCRs) for eliminating the problem of integration site position effects on expression [5,6"'], and exogenously a d d e d h u m a n immunodeficiency virus (HIV)-I Tat protein to transactivate the viral long terminal repeat (LTR) [7"q. Both approaches offer advantages over previous systems designed to achieve high level selective induction of target gene expression and the Tat protein system gives induction ratios similar to those reported for the lac and isopropyl-]l-D-thiogalactoside OPTG) systems (Table 1).
Regulated expression based on bacterial lac control sequences Regulated expression of chromosomal genes in b o t h bacteria and mammalian cells can be accomplished by the binding of either transcriptional repressor or activator proteins to particular DNA sequences. In this way, both types of cell can respond to environmental stimuli through changes in gene transcription. Inducible bacterial expression vectors based on the lac repressor-operator interaction have b e e n widely used. In this system, the lac repressor can bind to b o t h allolactose (or an analogue) and to a specific DNA sequence (operator). W h e n b o u n d to an operator sequence positioned within the promoter region of a gene, transcription of that gene is repressed. W h e n b o u n d to the allolactose analogue IPTG, however, the •
J
Abbreviations CAT---chloramphenicol acetyltransferase;DMSO--dimethylsulphoxide; GR---glucocorticoid receptor; HIV----human immunodeficiency virus; IFN--interferon; IPTG--isopropyl-ig-D-thiogalactoside; LCR-~locus control region; LTR--Iong terminal repeat; MEL--murine erythroleukemia; NLS--nuclear localization signals; PL--phospholipase; SV--simian virus; tTA--tetracycline-controlled transactivator; VP--virion protein.
506
© Current Biology Ltd ISSN 0958-1669
Inducible vectors for expression in mammalian cells Yarranton
Table 1. Inducible gene expression in permanent cell lines. Expression system
Inducer
Reporter gene
LAP348
- IPTG
CAT
24
50-100
LAP267
+IPTG (1 mM) +IPTG (5 mM) 32°C
CAT CAT CAT
24 24 48
100 1200 10
luciferase
24
100 000
CAT CAT CAT CAT
48 48 48 48
10 13 10 77
48 48
60-140 12 000
tTA hMT-11A lac
- TET +IPTG (20 mM) +DEX +Cd 2÷ +IPTG, DEX, Cd 2+
Induction
ratio
MTV/G R
+DEX +DEX
LCR/MEL
+DMSO (2%)
hGH
96
> 20
TAT (1-10 I.tg ml- l)
CAT
24
10 000
HIV LTR/TAT
[3-galactosidase IFN-~
Induction period (hrs)
CAT, chloramphenicol acetyltransferase; DEX, dexamethasone; DMSO, dimethylsulphoxide; GR, glucocorticoid receptor; hGH, human growth hormone; HIV, human immunodeficiency virus; hMT, human metallothionein; IFN, interferon; IPTG, isopropyl-~-Dthiogalactoside; LCR, locus control region; LTR, long terminal repeat; MEL, murine erythroleukemia; MTV, mammary tumor virus; TET, tetracycline; tTa, tetracycline-controlled transactivator.
repressor has a 300-fold lower affinity for the operator and transcription can proceed. In mammalian cells, this system has been shown to give induction ratios of 10-60-fold [2]. Modifications of the Escherichia coli lac repressoroperator system have been made, in order to improve its performance in mammalian cells. For example, the Lac repressor protein has been modified to improve its ability to localize within the nucleus [8]. Fusions of Lac repressor to either the simian virus (SV) 40 large T antigen or adenovirus Ela nuclear localization signals (NLS) produced functional proteins. Interestingly, gene fusions at the 3' end of the repressor gene produced a repressor that exhibited efficient nuclear accumulation (98% with NLS compared to 10% without NLS), strong repressor activity and greater sensitivity to IPTG induction than the native protein. These properties were dependent u p o n the length and composition of a threeamino-acid linker between the repressor and the NLS. This modification of the Lac repressor may provide better regulation and more rapid induction in mammalian ceils than observed with the native repressor. Hu and Davidson [2] have described additional modifications to the lac repressor--operator system in order to increase the induction ratio observed in mammalian cells. First, they selected a perfectly symmetrical 30base-pair operator sequence that has been s h o w n to bind to the repressor more strongly than the native
operator sequence. Second, they engineered the repressor protein to allow nuclear localization, using the minimal SV40 T-antigen NLS fused to the carboxyl terminus of the repressor, and third, they selected an inducible mammalian promoter, the metallothionein IIA gene promoter. This promoter is active in a range of cell types and is regulated by a wide variety of environmental stimuli, including steroids, growth factors and heavy metal ions. Transcriptional induction is mediated by specific regulatory proteins interacting with DNA elements within the promoter. Insertion of the symmetrical operator sequence close to the TATA b o x element gave efficient repression of transcription, greater than previously observed with the natural operator sequence in a chimeric Moloney sarcoma virusSV40 enhancer-promoter. This suggests that the symmetrical operator is preferred to the natural sequence for tight transcriptional control. Positioning of the lac operator is also crucial, with best regulation observed at positions - 35 and - 10 relative to the transcription start. No benefit was observed w h e n the lac repressor was employed with an NLS. However, the activity of this hybrid repressor relative to the natural repressor was not determined. Combining induction treatments in rat XC ceils expressing the lacI gene, an induction ratio of about 100-fold was obtained. The inducing agents used were IPTG, dexamethasone and Cd 2+. In the absence of IPTG but in the presence of dexamethasone and Cd 2+ the induction ratio was only 20-fold, so
507
508
Expression vectors
expressing LAP348 transactivator, IPTG p r o d u c e d a 50-100-fold reduction in reporter gene expression.
IPTG produces a further 5-fold induction w h e n combined with these agents (Table 1). Thus, this multiple induction approach provides a system that allows induction ratios of around 100-fold with relatively tight regulation of expression under non-inducing conditions.
Improvements in the induction ratio of the LAP system have been achieved by mutating the lacIgene so that it is both temperature sensitive for function and reverseregulated [10"]. Reverse regulated in this context means that IPTG stabilizes a temperature-sensitive repressor, thereby allowing it to bind to operator sequences. By inserting the VP16 coding region into the 3' end of the repressor gene (after c o d o n 267), the resulting chimeric protein (LAP267) possesses both temperature and IPTG conditional activity (Fig. la). In permanent m o u s e cell lines expressing the LAP267 chimeric protein, expression of a reporter gene from an SV40 early promoter with 14 lac operator sequences but no enhancer, was inducible either by a temperature switch or by the addition of IPTG (Fig. 2). Gene expression was activated at 32°C but not at 39.5°C giving an induction ratio of about 10-fold in the absence of IPTG. At 37°C or 39.5°C full induction of reporter gene expression was achieved with 5 mM IPTG, giving an induction ratio of about 1000fold. Maximum induction was achieved between 24-48 hours after treatment. The LAP267 system, therefore, is
Further modifications to the lac repressor-operator system have been reported recently by Labow et al. [9"] and Balm et al. [10-']. In this adaptation of the system, the specificity of the lac repressor for its operator sequence has been exploited in order to activate transcription rather than to repress it. A chimeric transactivator was made comprising the entire Lac repressor protein, the NLS of the SV40 large T antigen fused to the amino terminus of the Lac I protein, and the transactivator domain from the herpes simplex virus virion protein 16 (VP16) fused to the carboxy-terminus of the LacI protein (Fig. 1). The LAP348 transactivator protein, as it is called, activates expression from minimal promoter sequences containing spaced sets of lac operators. In contrast to the natural lac repressor system, IPTG inhibits LAP binding to the operators and hences inhibits activation of transcription. In routine cell lines (a)
DNA binding
Inducer binding
Lac, I
Tetramerization
I
I
NkS LAP267 I
I
267/369
488/268
I
VP16 NLS
3481
488/268
I
I VP16 SV40 minimal promoter i
Ip CAT
(b)
Operators tTA (c)
207
I
I VP16
(d)
-53
/\
-31 I AGGCCTATATAA
+1 ]
Luciferase
(TCTCTATCACTGATAGG GA) 7
Fig. 1. Schematic of the bacterial control systems employed in mammalian cells. (a) Structure of Lac repressor and LAP transactivator proteins. (b) Structure of SV40 minimal promoter with seven operator sequences (boxes) for transactivator binding. (c) Structure of the chimeric tetracycline-controlled transactivator (tTA) formed between the repressor of the tetracycline operon repressor and the VP16 transactivator. (d) Structure of the minimal CMV promoter with TATA box and seven operator sequences for transactivator binding. CAT, chloramphenicol acetyltransferase; NLS nuclear localization signals; VP, virion protein.
Inducible vectors for expression in mammalian cells Yarranton
one of the most powerful inducible systems currently available for use in permanent mammalian cell lines. • .
Relative 1000 CAT activity 100
... . . . . • ... . . . . . . . O- . . . .
toxicity are well documented. For cell-culture systems, however, the ability to induce expression by addition rather than removal of a chemical entity suggests that the LAP system will b e preferred.
•
i /'/
Regulated
10
control 1 l .... 0.1 0
expression
based on mammalian
gene
sequences
Metallothionein promoter systems "
I
24
'
I
I
I
48
96
120
Time (hours)
Fig. 2. Induction kinetics of the LAP348 expression system. CAT activity was measured using standard assay conditions at the times indicated following the addition of IPTG. - . - . 5mM IPTG at 37°C, - - - 32°C, -.-. 37°C no IPTG, o__
Introduction Regulated expression of a gene product in mammalian cells is often desirable, especially w h e n the product is either cytotoxic or cytostatic. Early attempts to develop inducible gene expression vectors relied u p o n the use of heat shock control [1], metallothionine promoters and h e a v y metal ion control [2], or steroid regulatory promoters [3,4]. Although these systems give some level of regulation, they are often 'leaky' (i.e. uninduced levels of expression are clearly detectable) and induction ratios are low, usually about 10-20 fold (Table 1). An additional disadvantage of these systems is that the inducing treatment invariably leads to the co-ordinate induction of a variety of cellular genes, hence making conclusions concerning the physiological consequences of the expressed heterologous gene product difficult to interpret. The most promising developments in the area of inducible expression vectors for m a m m a l i a n cells have come from groups w h o have attempted to adapt bacterial transcriptional control systems to mammalian cells [2]. One advantage of these approaches is that specificity is achieved by the interaction of a bacterial protein with a bacterial DNA binding sequence (operator). The operator sequence is unlikely to occur in the promoter sequence of a mammalian gene, hence non-specific activation or repression of cellular gene expression is unlikely to occur. For this reason, as well as the fact that the induction ratios achieved with these chimeric systems tend to exceed those of the homologous systems, this review will focus mainly on the d e v e l o p m e n t of inducible expression systems in mammalian cells that e m p l o y bacterial elements.
Two recent developments in inducible gene expression systems based on mammalian cell or viral g e n e control elements which will be discussed in this review are the use of locus control regions (LCRs) for eliminating the problem of integration site position effects on expression [5,6"'], and exogenously a d d e d h u m a n immunodeficiency virus (HIV)-I Tat protein to transactivate the viral long terminal repeat (LTR) [7"q. Both approaches offer advantages over previous systems designed to achieve high level selective induction of target gene expression and the Tat protein system gives induction ratios similar to those reported for the lac and isopropyl-]l-D-thiogalactoside OPTG) systems (Table 1).
Regulated expression based on bacterial lac control sequences Regulated expression of chromosomal genes in b o t h bacteria and mammalian cells can be accomplished by the binding of either transcriptional repressor or activator proteins to particular DNA sequences. In this way, both types of cell can respond to environmental stimuli through changes in gene transcription. Inducible bacterial expression vectors based on the lac repressor-operator interaction have b e e n widely used. In this system, the lac repressor can bind to b o t h allolactose (or an analogue) and to a specific DNA sequence (operator). W h e n b o u n d to an operator sequence positioned within the promoter region of a gene, transcription of that gene is repressed. W h e n b o u n d to the allolactose analogue IPTG, however, the •
J
Abbreviations CAT---chloramphenicol acetyltransferase;DMSO--dimethylsulphoxide; GR---glucocorticoid receptor; HIV----human immunodeficiency virus; IFN--interferon; IPTG--isopropyl-ig-D-thiogalactoside; LCR-~locus control region; LTR--Iong terminal repeat; MEL--murine erythroleukemia; NLS--nuclear localization signals; PL--phospholipase; SV--simian virus; tTA--tetracycline-controlled transactivator; VP--virion protein.
506
© Current Biology Ltd ISSN 0958-1669
Inducible vectors for expression in mammalian cells Yarranton
Table 1. Inducible gene expression in permanent cell lines. Expression system
Inducer
Reporter gene
LAP348
- IPTG
CAT
24
50-100
LAP267
+IPTG (1 mM) +IPTG (5 mM) 32°C
CAT CAT CAT
24 24 48
100 1200 10
luciferase
24
100 000
CAT CAT CAT CAT
48 48 48 48
10 13 10 77
48 48
60-140 12 000
tTA hMT-11A lac
- TET +IPTG (20 mM) +DEX +Cd 2÷ +IPTG, DEX, Cd 2+
Induction
ratio
MTV/G R
+DEX +DEX
LCR/MEL
+DMSO (2%)
hGH
96
> 20
TAT (1-10 I.tg ml- l)
CAT
24
10 000
HIV LTR/TAT
[3-galactosidase IFN-~
Induction period (hrs)
CAT, chloramphenicol acetyltransferase; DEX, dexamethasone; DMSO, dimethylsulphoxide; GR, glucocorticoid receptor; hGH, human growth hormone; HIV, human immunodeficiency virus; hMT, human metallothionein; IFN, interferon; IPTG, isopropyl-~-Dthiogalactoside; LCR, locus control region; LTR, long terminal repeat; MEL, murine erythroleukemia; MTV, mammary tumor virus; TET, tetracycline; tTa, tetracycline-controlled transactivator.
repressor has a 300-fold lower affinity for the operator and transcription can proceed. In mammalian cells, this system has been shown to give induction ratios of 10-60-fold [2]. Modifications of the Escherichia coli lac repressoroperator system have been made, in order to improve its performance in mammalian cells. For example, the Lac repressor protein has been modified to improve its ability to localize within the nucleus [8]. Fusions of Lac repressor to either the simian virus (SV) 40 large T antigen or adenovirus Ela nuclear localization signals (NLS) produced functional proteins. Interestingly, gene fusions at the 3' end of the repressor gene produced a repressor that exhibited efficient nuclear accumulation (98% with NLS compared to 10% without NLS), strong repressor activity and greater sensitivity to IPTG induction than the native protein. These properties were dependent u p o n the length and composition of a threeamino-acid linker between the repressor and the NLS. This modification of the Lac repressor may provide better regulation and more rapid induction in mammalian ceils than observed with the native repressor. Hu and Davidson [2] have described additional modifications to the lac repressor--operator system in order to increase the induction ratio observed in mammalian cells. First, they selected a perfectly symmetrical 30base-pair operator sequence that has been s h o w n to bind to the repressor more strongly than the native
operator sequence. Second, they engineered the repressor protein to allow nuclear localization, using the minimal SV40 T-antigen NLS fused to the carboxyl terminus of the repressor, and third, they selected an inducible mammalian promoter, the metallothionein IIA gene promoter. This promoter is active in a range of cell types and is regulated by a wide variety of environmental stimuli, including steroids, growth factors and heavy metal ions. Transcriptional induction is mediated by specific regulatory proteins interacting with DNA elements within the promoter. Insertion of the symmetrical operator sequence close to the TATA b o x element gave efficient repression of transcription, greater than previously observed with the natural operator sequence in a chimeric Moloney sarcoma virusSV40 enhancer-promoter. This suggests that the symmetrical operator is preferred to the natural sequence for tight transcriptional control. Positioning of the lac operator is also crucial, with best regulation observed at positions - 35 and - 10 relative to the transcription start. No benefit was observed w h e n the lac repressor was employed with an NLS. However, the activity of this hybrid repressor relative to the natural repressor was not determined. Combining induction treatments in rat XC ceils expressing the lacI gene, an induction ratio of about 100-fold was obtained. The inducing agents used were IPTG, dexamethasone and Cd 2+. In the absence of IPTG but in the presence of dexamethasone and Cd 2+ the induction ratio was only 20-fold, so
507
508
Expression vectors
expressing LAP348 transactivator, IPTG p r o d u c e d a 50-100-fold reduction in reporter gene expression.
IPTG produces a further 5-fold induction w h e n combined with these agents (Table 1). Thus, this multiple induction approach provides a system that allows induction ratios of around 100-fold with relatively tight regulation of expression under non-inducing conditions.
Improvements in the induction ratio of the LAP system have been achieved by mutating the lacIgene so that it is both temperature sensitive for function and reverseregulated [10"]. Reverse regulated in this context means that IPTG stabilizes a temperature-sensitive repressor, thereby allowing it to bind to operator sequences. By inserting the VP16 coding region into the 3' end of the repressor gene (after c o d o n 267), the resulting chimeric protein (LAP267) possesses both temperature and IPTG conditional activity (Fig. la). In permanent m o u s e cell lines expressing the LAP267 chimeric protein, expression of a reporter gene from an SV40 early promoter with 14 lac operator sequences but no enhancer, was inducible either by a temperature switch or by the addition of IPTG (Fig. 2). Gene expression was activated at 32°C but not at 39.5°C giving an induction ratio of about 10-fold in the absence of IPTG. At 37°C or 39.5°C full induction of reporter gene expression was achieved with 5 mM IPTG, giving an induction ratio of about 1000fold. Maximum induction was achieved between 24-48 hours after treatment. The LAP267 system, therefore, is
Further modifications to the lac repressor-operator system have been reported recently by Labow et al. [9"] and Balm et al. [10-']. In this adaptation of the system, the specificity of the lac repressor for its operator sequence has been exploited in order to activate transcription rather than to repress it. A chimeric transactivator was made comprising the entire Lac repressor protein, the NLS of the SV40 large T antigen fused to the amino terminus of the Lac I protein, and the transactivator domain from the herpes simplex virus virion protein 16 (VP16) fused to the carboxy-terminus of the LacI protein (Fig. 1). The LAP348 transactivator protein, as it is called, activates expression from minimal promoter sequences containing spaced sets of lac operators. In contrast to the natural lac repressor system, IPTG inhibits LAP binding to the operators and hences inhibits activation of transcription. In routine cell lines (a)
DNA binding
Inducer binding
Lac, I
Tetramerization
I
I
NkS LAP267 I
I
267/369
488/268
I
VP16 NLS
3481
488/268
I
I VP16 SV40 minimal promoter i
Ip CAT
(b)
Operators tTA (c)
207
I
I VP16
(d)
-53
/\
-31 I AGGCCTATATAA
+1 ]
Luciferase
(TCTCTATCACTGATAGG GA) 7
Fig. 1. Schematic of the bacterial control systems employed in mammalian cells. (a) Structure of Lac repressor and LAP transactivator proteins. (b) Structure of SV40 minimal promoter with seven operator sequences (boxes) for transactivator binding. (c) Structure of the chimeric tetracycline-controlled transactivator (tTA) formed between the repressor of the tetracycline operon repressor and the VP16 transactivator. (d) Structure of the minimal CMV promoter with TATA box and seven operator sequences for transactivator binding. CAT, chloramphenicol acetyltransferase; NLS nuclear localization signals; VP, virion protein.
Inducible vectors for expression in mammalian cells Yarranton
one of the most powerful inducible systems currently available for use in permanent mammalian cell lines. • .
Relative 1000 CAT activity 100
... . . . . • ... . . . . . . . O- . . . .
toxicity are well documented. For cell-culture systems, however, the ability to induce expression by addition rather than removal of a chemical entity suggests that the LAP system will b e preferred.
•
i /'/
Regulated
10
control 1 l .... 0.1 0
expression
based on mammalian
gene
sequences
Metallothionein promoter systems "
I
24
'
I
I
I
48
96
120
Time (hours)
Fig. 2. Induction kinetics of the LAP348 expression system. CAT activity was measured using standard assay conditions at the times indicated following the addition of IPTG. - . - . 5mM IPTG at 37°C, - - - 32°C, -.-. 37°C no IPTG, o__
