The Plant Journal (2000) 24(2), 265273

TECHNICAL ADVANCE

An estrogen receptor-based transactivator XVE mediates highly inducible gene expression in transgenic plants
Jianru Zuo, Qi-Wen Niu and Nam-Hai Chua Laboratory of Plant Molecular Biology, The Rockefeller University, 1230 York Avenue, New York, NY 10021, USA
Received 2 May 2000; revised 1 August 2000; accepted 1 August 2000. *For correspondence (fax +1 212 327 8327; e-mail chua@rockvax.rockefeller.edu).

Summary We have developed an estrogen receptor-based chemical-inducible system for use in transgenic plants. A chimeric transcription activator, XVE, was assembled by fusion of the DNA-binding domain of the bacterial repressor LexA (X), the acidic transactivating domain of VP16 (V) and the regulatory region of the human estrogen receptor (E; ER). The transactivating activity of the chimeric XVE factor, whose expression was controlled by the strong constitutive promoter G10-90, was strictly regulated by estrogens. In transgenic Arabidopsis and tobacco plants, estradiol-activated XVE can stimulate expression of a GFP reporter gene controlled by the target promoter, which consists of eight copies of the LexA operator fused upstream of the 46 35S minimal promoter. Upon induction by estradiol, GFP expression levels can be eightfold higher than that transcribed from a 35S promoter, whereas the uninduced controls have no detectable GFP transcripts, as monitored by Northern blot analysis. Neither toxic nor adverse physiological effects of the XVE system have been observed in transgenic Arabidopsis plants under all the conditions tested. The XVE system thus appears to be a reliable and efcient chemical-inducible system for regulating transgene expression in plants. Keywords: gene expression, estradiol-inducible, transgenic plants, estrogen receptor, LexA.

Introduction Both in basic plant biology research and in biotechnological applications, it is highly desirable to express genes in a controllable fashion. Compared to constitutive promoters, inducible promoters offer numerous advantages and potentials in a variety of applications (reviewed by Gatz, 1997; Gatz and Lenk, 1998; Zuo and Chua, 2000). Under the control of an inducible system, a transgene can be expressed at a given developmental stage for a specic duration of interest. This exibility is particularly important for studying the function of a gene whose expression is developmentally or tissue-specically regulated. In addition, the use of inducible promoters will allow the generation of transgenic plants carrying a transgene whose constitutive expression is otherwise detrimental or even lethal to the host plants. To establish a reliable inducible expression system, several critical factors must be considered. First, the system must be highly inducible. Second, the system
2000 Blackwell Science Ltd

should be tightly controlled and respond only to specic inducers. Third, upon induction the system should specifically activate the target genes. Finally, activation of the system itself should not have any non-specic physiological or other undesirable effects on the host. Several chemical-inducible systems have been developed, in general based on transcriptional de-repression (Gatz et al., 1992), inactivation (Weinmann et al., 1994), and activation (Aoyama and Chua, 1997; Bruce et al., 2000; Caddick et al., 1998; Martinez et al., 1999) of the target gene. A dualcontrol system was recently developed in which a transactivator was subjected to both positive and negative regulation by two distinctive inducers (Bohner et al., 1999). Many of the above-mentioned inducible systems are based on transcriptional activation mediated by an animal steroid nuclear receptor. Thus far, the regulatory domains of the rat glucocorticoid receptor (GR; Aoyama and Chua, 1997; Bohner et al., 1999); the human estrogen receptor 265

266 Jianru Zuo et al. (ER; Bruce et al., 2000); and an insect ecdysone receptor (Martinez et al., 1999) have been used to construct chimeric transactivators whose transcriptional activities are regulated by specic hormones or structurally related compounds. These inducible systems have been successfully used in various studies, but they also show some drawbacks with respect to inducibility, a relatively high background expression level, or detrimental effects to the host plant. Here we report a detailed characterization of a reliable and efcient chemical-inducible system, termed XVE for LexA-VP16-ER. The XVE activator is strictly regulated by estradiol in transgenic plants, with undetectable transactivating activity in the absence of an inducer. Upon induction the activator is capable of stimulating expression of a GFP transgene over eight times higher than that obtained with a 35SGFP transgene. No apparent toxicity to Arabidopsis transgenic plants was detected with this system. In the second transcription unit, eight copies of the LexA operator sequence were fused to the 46 35S minimal promoter (Benfey et al., 1990) to control transcription of target genes. These two transcription units are separated by a selectable marker for plant transformation. High inducibility and tight control of the XVE system in transgenic plants To evaluate the system, we inserted a cDNA encoding the green uorescence protein (GFP) into the target expression cassette of an XVE vector (pER8; see Figure 1). The pER8GFP construct was then transformed into Arabidopsis and tobacco, and expression of the GFP gene was assessed. Similar results were obtained from both species. Here we present data obtained from a detailed analysis of 22 Arabidopsis transgenic lines. Transgenic plants were initially screened by visual inspection of seedlings germinated in the absence or presence of inducers (a mixture of 2 mM 17-b-estradiol and 0.5 mM 4hydroxyl tamoxifen) under a conventional uorescence microscope. In the absence of inducers, none of the 22 lines had detectable GFP signals, suggesting that the XVE system is tightly controlled. Whereas one line (9) did not have any detectable uorescence signals under all conditions tested, three other lines (11, 12, 20) showed a patchy expression pattern upon induction (Figure 2a), presumably caused by a positional effect on the transgene. These four lines were not investigated further. Upon induction, the remaining 18 lines showed uniform uorescence signals with varying intensities in young seedlings (Figure 2bd). The same observations were made by transferring transgenic plants from a non-inductive onto an inductive medium, and the GFP signal was usually detectable 5 6 h after the transfer. The above results were further conrmed in all positive lines by Northern blot analysis of the GFP transcript. Consistent with the visual inspection data, whereas all examined lines transcribed the reporter gene at varying levels upon induction, none had detectable GFP transcripts in the absence of inducers (see Figure 3a for examples). The GFP transcript level was well correlated with the uorescence intensity. The highest level of the GFP transcript was found in lines 22, 17, 1, 8 and 10. A prolonged incubation appeared to lead to a reduced transcript level (e.g. grown for 23 weeks in the presence of the inducer versus a 16 h induction; compare lanes b and c in Figure 3a). This time-dependent reduction was also observed in all other previously reported inducible systems (e.g. Aoyama and Chua, 1997; Martinez et al., 1999). To test whether this reduction of GFP expression was caused by inducer instability or by a desensitization of the XVE system, we transferred 3-week-old seedlings germinated and grown in the presence of the inducer
Blackwell Science Ltd, The Plant Journal, (2000), 24, 265273

Results Design and construction of the XVE inducible system The XVE vector consists of three transcription units (Figure 1). In the rst transcription unit G10-90, a strong constitutive promoter (Ishige et al., 1999), controls the XVE fusion gene. The latter encodes a chimeric transcriptional activator consisting of the DNA-binding domain (DBD) of the bacterial repressor LexA (X; residues 187); the acidic transactivating domain of VP16 (V; residues 403479; Dalrymple et al., 1985); and the carboxyl region of the human estrogen receptor (E or ER; residues 282595; Greene et al., 1986). As the DNA-binding domain and the dimerization domain of LexA are well separated (Horii et al., 1981; Miki et al., 1981), dimerization and subsequent binding to the target DNA of a LexA DBD-containing factor can be tightly regulated by heterologous regulatory elements fused to it. The ER sequence used in XVE contains the regulatory region, including the binding sites for a cellular regulatory complex and for estrogen hormones, overlapping with the transactivation function 2 (TAF2) domain (Tora et al., 1989; Webster et al., 1988). We wish to point out that the hormone-binding domain (HBD) alone is insufcient to provide any regulation to the fused heterologous DBD. Clearly, hormone inducibility requires the binding domain for the ligand as well as that for a regulatory complex consisting of several cellular proteins including HSP90, although some investigators refer to the receptor sequence in a chimeric factor as HBD. We suggest that it is more appropriate to use `the regulatory region/ domain' or specic domain names as recommended (Evans, 1988; Tsai and O'Malley, 1994) when referring to the receptor sequences.

The XVE inducible expression system 267

Figure 1. A schematic diagram of the XVE vector. Only the region between the right and left borders is shown (not to scale). PG1090, a synthetic promoter (Ishige et al., 1999) controlling XVE; XVE, DNA sequences encoding a chimeric transcription factor containing the DNA-binding domain of LexA (residues 187), the transcription activation domain of VP16 (residues 403479) and the regulatory region of the human estrogen receptor (residues 282595); TE9, rbcS E9 poly(A) addition sequence; Pnos, nopaline synthase promoter; HPT, hygromycin phosphotransferase II coding sequence; Tnos, nopaline synthase poly(A) addition sequence; OLexA, eight copies of the LexA operator sequence; 46, the 46 35S minimal promoter; MCS, multiple cloning sites for target genes; T3A, rbcsS3A poly(A) addition sequence. Arrows indicate the direction of transcription.

onto a fresh inductive medium, and the reporter gene activity was assessed by Northern blot analysis. Figure 3(b) shows that the second inducer treatment was able to fully reactivate the system, suggesting that the reduction in GFP expression during the rst induction was probably due to inducer instability. Some data presented above were obtained by using a mixture of two inducers, 17-b-estradiol and 4-hydroxyl tamoxifen; the latter is an estradiol agonist that acts as a strong inducer in certain animal cells. To distinguish which of the two is active or more active, transgenic plants (lines 1, 5 and 10) were separately treated with different concentrations of 17-b-estradiol or 4-hydroxyl tamoxifen (0.05, 0.2, 1, 2 and 5 mM, respectively), and the GFP transcript level was analyzed. Whereas both chemicals were capable of activating the reporter gene, estradiol appeared to produce a slightly higher transcript level (data not shown). This difference was probably due to the fact that the ER TAF2 function can be activated by estrogen but not 4-hydroxyl tamoxifen (Webster et al., 1988). Because 4hydroxyl tamoxifen is considerably more expensive than 17-b-estradiol, we used the latter in all subsequent experiments.

The XVE system responds to a broad range of estradiol concentrations To test the dose-response of the XVE inducible system, 2week-old transgenic plants germinated and cultured in the absence of inducers were transferred onto media containing varying concentrations of estradiol, and cultured for an additional 16 h. Two independent 35SGFP lines (Kost et al., 1998) were used as controls. Five pER8GFP lines (1, 8, 10, 17 and 22) were tested and similar results were obtained. Figure 4 shows results obtained from line 17. Under the assay conditions, the GFP transcript was not detected in either wild-type or untreated transgenic plants. The GFP transcript became detectable on treatment with 8 nM estradiol, and reached a level comparable to those of the 35SGFP lines around 0.2 mM estradiol. The system appeared to be saturated at 5 mM estradiol, as higher inducer concentrations did not signicantly increase the
Blackwell Science Ltd, The Plant Journal, (2000), 24, 265273

Figure 2. Estradiol-induced GFP expression in pER8GFP transgenic Arabidopsis plants. GFP expression was induced with 2 mM 17-b-estradiol and 0.5 mM 4hydroxyl tamoxifen (a), or with 2 mM 17-b-estradiol (bd). (a) Patchy expression pattern in roots (line 11). (b) A 4-day-old seedling germinated and grown in the inductive medium (line 22). (c) A closer examination of the seedling shown in (b). (d) GFP expression in a root (line 1) after a 16 h induction. All images were taken under a uorescence microscope.

268 Jianru Zuo et al. Rapid response of the XVE system to estradiol To investigate the induction kinetics, 2-week-old seedlings were transferred to a medium containing 2 mM 17-bestradiol, and cultured for varying periods. Three independent lines (lines 1, 17 and 22) were tested, and similar results were obtained. Figure 5 shows results obtained from line 17. Similar to wild type, the untreated transgenic plants (time 0 h) had undetectable GFP transcript levels under the assay conditions. The GFP transcript was detectable after a 30 min treatment, and reached a level similar to that of 35SGFP between 6 and 12 h. The highest transcript level was found at around 24 h after induction, when the activity of the inducible system was approximately 8.4-fold that of the 35S promoter (approximately 4.3- and 9.1-fold higher than the 35S promoter for lines 1 and 22, respectively). After this peak, the GFP transcript level gradually declined. The XVE system has no apparent toxic or physiological effects on plants Any reliable chemical-inducible system should have no toxic nor non-specic physiological effects on the host. We have carefully followed most growth and development stages (except embryogenesis) of ve pER8GFP Arabidopsis transgenic lines (1, 10, 16, 17 and 22) in the presence (10 mM estradiol) and absence of the inducer, and did not observe any apparent morphological or physiological alterations in these lines. Similarly, estradiolinduced expression of a number of other target genes was found to have no apparent non-specic or adverse physiological effects (J. Zuo, Q.-W. Niu, G. Frugis, H. Banno, S. Moller and N.-H. Chua, unpublished results). At the molecular level, we examined the expression of PDF1-2, a defense response-related gene which was recently shown to be induced by dexamethasone (dex) in some GVG-transformed Arabidopsis plants (Kang et al., 1999). Wild-type, XVE (pER8GFP, lines 1 and 17) and GVG (pTA7002-Luc, line 321; Aoyama and Chua, 1997) plants, germinated and grown in the absence of inducers for 2 weeks, were transferred onto media supplemented with or without estradiol or dex, and cultured for an additional 48 h. Whereas dex induced PDF1-2 expression in the GVG plants but not in wild-type or XVE plants, estradiol had no effects on PDF1-2 expression in all plants tested (Figure 6). Note that the low level of expression of PDF1-2 in some wild-type and XVE plants presumably resulted from a general stress response (e.g. during transferring plants), and this `basal' expression could reach considerably higher levels if the seedlings were not carefully handled (data not shown). The above results suggest that activation of the XVE system in transgenic Arabidopsis plants has no detectable non-specic effects on plant growth and development, or on PDF1-2 expression.
Blackwell Science Ltd, The Plant Journal, (2000), 24, 265273

Figure 3. Expression of the GFP reporter gene under the control of the XVE system. (a) Induction of GFP transcript levels in different transgenic lines. pER8 GFP transgenic plants (T2, heterozygous seeds) were germinated and cultured in the absence (lanes a, supplemented with 30 mg l1 hygromycin) or presence (lanes c) of inducers (2 mM 17-b-estradiol and 0.5 mM 4-hydroxyl tamoxifen, 30 mg l1 hygromycin) for 2 weeks. A portion of seedlings grown in the absence of the inducers were transferred onto an inductive medium (2 mM 17-b-estradiol and 0.5 mM 4hydroxyl tamoxifen, 30 mg l1 hygromycin), and incubated for 16 h (lanes b). Total RNA (10 mg) was analyzed by Northern blot hybridization using a GFP cDNA as a probe. Transgenic line numbers are given at the top. (b) Reactivation of the XVE system. pER8GFP transgenic plants (lines 1 and 17; T4, homozygous seeds) were germinated and grown in the absence or presence of 2 mM estradiol for 3 weeks. Lanes a, seedlings grown on the non-inductive medium were transferred onto the inductive medium and cultured for 16 h. Lanes b, seedlings germinated and grown on the inductive medium for 3 weeks. Lanes c, seedlings germinated and grown on the inductive medium for 3 weeks were cultured in the freshly prepared inductive medium for 16 h. Total RNA (5 mg) was analyzed by Northern blot hybridization using a GFP cDNA and an actin genomic fragment as probes. The blot shown for the GFP probe was exposed for 8 h (line 1) or 2 h (line 17).

expression levels. Under these assay conditions, the XVE target promoter activity was about four times higher than that of the 35S promoter. There were some variations among different transgenic lines in their responses to different inducer concentrations. Whereas the GFP transcripts in lines 1 and 22 could be detected in plants treated with 0.4 nM estradiol, a higher inducer concentration (up to 8 nM) was required to detect GFP transcription in lines 8, 10 and 17. Nevertheless, in all tested lines the system appeared to be saturated at approximately 5 mM estradiol.

The XVE inducible expression system 269

Figure 4. Doseresponse of the XVE inducible system to b-estradiol. Two-week-old pER8GFP transgenic plants (line 17; T4, homozygous seeds) grown in the absence of the inducer were transferred onto media containing various concentrations of 17-b-estradiol as indicated (top of each lane), and cultured for 16 h. Total RNA (5 mg each lane) prepared from untreated (lane 0) or estradiol-treated plants was analyzed by Northern blot hybridization using a GFP cDNA as a probe. RNA prepared from two independent lines of 35SGFP (35S-1 and 35S-2 on the left side; T5, homozygous; Kost et al., 1998) and wildtype plants (Col-0 on the right side) were included as controls. The blot was reprobed with an actin genomic DNA fragment. The GFP transcript level was normalized to that of actin. The average activity of the two 35SGFP lines was set as 100%, and the activity of each sample relative to this mean value was given in the histogram.

Discussion Here we report the development of a new chemicalinducible expression system for use in transgenic plants. A detailed characterization of the XVE system in transgenic Arabidopsis plants demonstrated that this system is tightly regulated and highly inducible without any detectable toxicity. In addition to GFP, the XVE system has been used for expression of a number of genes in transgenic Arabidopsis and tobacco plants. Among the transgenic lines examined by Northern blot analysis (over 500 lines), more than 75% showed a high transgene expression level upon induction, whereas none showed detectable background (J. Zuo, Q.-W. Niu, G. Frugis, H. Banno, S. Moller and N.-H. Chua, unpublished results). The XVE system has also been successfully used in stably transformed tobacco BY2 cells (Zuo et al., 2000). Recently, another ER-based inducible system was reported by Bruce et al. (2000). The chimeric transactivator ER-C was made by inserting the transactivation domain of the maize transcription factor C into the transactivation
Blackwell Science Ltd, The Plant Journal, (2000), 24, 265273

domain of the human ER. The target genes were controlled by a 35S minimal promoter fused to multiple copies of an estrogen-responsive element. In stably transformed maize Black Mexican Sweet (BMS) cells, ER-C could activate the expression of a luciferase reporter gene in an estradioldependent fashion (up to 14 000 relative light units) without detectable background expression. Using this inducible system, Bruce et al. (2000) overexpressed in BMS cells two transcription factors CRC (a fusion factor between C1 and R) and P, which are thought to be involved in the avonoid biosynthetic pathway, and identied a large number of downstream target genes. The ERC system, however, has not been tested in transgenic plants. In addition, the target promoter strength (for example, relative to the 35S promoter) is not known. The XVE system responds to a wide range of inducer concentrations (from 8 nM to 5 mM estradiol), and the target promoter strength can reach a level similar to that of a 35S promoter at a relatively low inducer concentration (0.2 mM). Moreover, transcription of target genes can be rapidly turned on by estradiol (in 30 min), and can reach

270 Jianru Zuo et al.

Figure 5. Induction time-course of the XVE system. Two-week-old pER8GFP transgenic plants (line 17; T4, homozygous seeds) grown in the absence of the inducer were transferred onto a medium containing 2 mM 17-bestradiol, and cultured for varying periods as indicated above each lane. Wild-type (Col-0) plants were harvested at the 96 h time point. 10 mg total RNA was loaded in each lane. See legend to Figure 4 for other technical details.

Figure 6. Expression of the PDF1-2 gene in wild-type and transgenic plants. Two-week-old wild-type (Col-0), pER8GFP (XVE-1, line 1; XVE-17, line 17; all T4, homozygous seeds) and GVG (GVG-Luc, line 321; T7, homozygous seeds; Aoyama and Chua, 1997) plants were grown in the absence of inducers. Seedlings were then transferred onto a medium without inducers (lanes a), with 10 mM dex (lanes b) or 10 mM 17-b-estradiol (lanes c), and incubated for 48 h. RNA (10 mg each lane) was prepared and analyzed by Northern blot hybridization using a PDF1-2 cDNA (EST clone 37F10T7 obtained from the Arabidopsis Biological Resources Center) and an actin genomic DNA fragment as probes. Induction of the transgene by estradiol in the XVE lines was conrmed by reprobing the blot with a GFP cDNA.

eightfold higher than that of a 35S promoter within 24 h of induction. These properties allow transgenes to be expressed at the desired level and period by properly

adjusting estradiol concentrations and incubation time. The advantages of the XVE system are presumably attributed to several factors. First, because ER binds to estradiol with a very high afnity (0.05 nM; Mueller-Fharnow and Egner, 1999), the XVE chimeric transactivator can be activated by a relatively low inducer concentration. Second, the DBD of LexA does not resemble structurally those of any known eukaryotic factors (Oertel-Buchheit et al., 1992), thus reducing the possibilities that XVE may bind to endogenous plant cis elements. Third, as mentioned before the DBD and the dimerization domain of LexA are physically separable (Miki et al., 1981), which allows the LexA DBD to be tightly regulated by the regulatory sequences fused to it without compromising its binding afnity. This property is probably the most critical factor for both low basal expression and high inducibility of the system. In addition to that of LexA, we have also tested two other DBDs in the course of developing chemical-inducible expression systems in plants. The DBD of the yeast transcription factor PUT3 (Marczak and Brandriss, 1991) was used for the construction a PVE (V: VP16; E: ER)
Blackwell Science Ltd, The Plant Journal, (2000), 24, 265273

The XVE inducible expression system 271 activator, which was found to be weakly active in transgenic Arabidopsis plants (P. Spielhofer and N.-H. Chua, unpublished results). In a second attempt, we used the bacterial lactose repressor (LacR) DBD (residues 164, including the linker or hinge sequence: Farabaugh, 1978; Lewis et al., 1996), which is also well separated from its regulatory and oligomerization domains, to replace the GAL4 DBD in GVG. The transactivating activity of the resulting chimeric activator LVG, however, appeared to be deregulated in transgenic Arabidopsis plants (J. Zuo and N.-H. Chua, unpublished results). It has been known that when a DBD is fused to a heterologous sequence, its DNAbinding activity may be signicantly altered, depending on the context and position of the fused sequences. Fusion of several non-bacterial proteins to the DBD of LexA, for example, severely inhibits the binding of these hybrid proteins to the LexA operator (Golemis and Brent, 1992). On the other hand, whereas several LexA-hHSF1 (human heat shock transcription factor 1) fusion proteins bound to the target DNA with very high afnity both in vitro and in vivo (Zuo et al., 1994; Zuo et al., 1995), fusing the LexA sequence to certain positions of the same factor completely abolished DNA-binding activity regardless of the presence or absence of the dimerization domain of LexA (J. Zuo and R. Voellmy, unpublished results). This sequence context and position-dependent effect may also be applicable to the PUT3 and LacR DBDs. An addition complication is that cofactors and/or post-translational modications may be required for the transactivating activity of a chimeric factor, as recently demonstrated for PUT3 itself (Huang and Brandriss, 2000). It is interesting to note that the tetracycline repressor (TetR), which employs a regulatory mechanism similar to that of LacR, binds to its operator in the absence of regulatory molecules (tetracycline), but dissociates from the target DNA on association with tetracycline. A TetR-derived chimeric factor, TGV (TetR-GR-VP16), which is structurally similar to LVG, appears to be tightly regulated by the adjoining GR moiety (Bohner et al., 1999). The different regulatory properties of TGV and LVG present additional examples of sequence context and position effects. It will also be interesting to determine at which level the TGV activity was regulated (e.g. dimerization, nuclear localization, DNA-binding and/ or transcription competence), which may provide invaluable clues for the future development of chemicalinducible systems in plants. In general, a heterologous DBD with a high DNA-binding afnity and tight regulation in vivo may not retain these characteristics in plant cells. In transgenic Arabidopsis plants, the XVE system did not appear to be toxic to the host nor to activate the PDF1-2 gene. However, expression of a transgene, controlled by either an inducible or a strong constitutive promoter (such as the 35S promoter), may have effects on the expression of endogenous genes in addition to the expected physio Blackwell Science Ltd, The Plant Journal, (2000), 24, 265273

logical effects. Moreover, introducing a foreign DNA causes alterations in the host genome which may have a substantial impact on plant growth and development. Therefore the XVE system may also have undesired effects on the host plants not revealed by our studies, and may affect expression of endogenous genes other than PDF1-2. In GVG-transformed plants, dex-dependent PDF1-2 expression in some plants was suggested to be caused by binding of the chimeric transcription factor to cis-elements in plant genome or a non-specic transactivating activity of the VP16 sequence (Kang et al., 1999). It appears that the second possibility can be excluded because the XVE fusion factor, which also contains the same VP16 sequence as in GVG, did not elicit any PDF1-2 expression. Whereas the XVE system is tightly regulated and highly inducible in Arabidopsis and tobacco transgenic plants, the system may not function as well in species with a high level of phytoestrogens, for example soybean. Conjugated phytoestrogens are known to be estrogenically inactive in plant cells, but the free form, which can be converted from the conjugated form in vacuoles, is active (reviewed by Kurzer and Xu, 1997). It will be interesting to determine if the XVE system can be regulated in species with a higher phytoestrogen concentration (e.g. soybean). Although we have not yet tested the XVE system in monocots, the activator appears to be potentially useful in monocots as ER-C can be tightly controlled and highly inducible in maize BMS suspension cells (Bruce et al., 2000). An additional concern is that the XVE system is unsuitable for eld applications because of the chemical nature of the inducers. However, it should be possible to replace the ER regulatory sequence with that of an insect steroid nuclear receptor, thus rendering the system under the control of an insecticide, as is the case with the GVGEc system recently developed by Martinez et al. (1999). Experimental procedures Plasmid constructions
The AvrII (mung bean nuclease treated) and EcoRI fragment of pTA7002 (Aoyama and Chua, 1997), which contains sequences encoding the VP16 transactivation domain and a portion of the regulatory region of the rat glucocorticoid receptor, was ligated to the XmnI (partially digested) and EcoRI digested p202/79 (Zuo et al., 1994). The resulting construct, pER0, contains the sequences encoding the DNA binding domain of LexA (residues 187) fused in frame to VP16 (403479)-GR. The GR sequence in pER0 was replaced with a PCR-generated DNA fragment encoding the carboxy-terminal region of human estrogen receptor (residues 282595; Greene et al., 1986) to yield pER1. A human EST clone (ID 1631712, accession number AI025006; Research Genetics, Inc., Huntsville, AL, USA) was used as the template for PCR reactions. An extra proline residue was introduced in the junctions of the fusion protein (X-V and V-E). The XVE chimeric genes, released from pER1 by XmnI or HindIII (blunted) and EcoRI (blunted), was placed under the control of the G10-90 promoter

272 Jianru Zuo et al.

(the SmaI/Ecl136II digested pLiG1090 vector; Ishige et al., 1999) to generate pER3 and pER4 (with an XmnI site between G10-90 and XVE), respectively. The G10-90XVE fragment (digested with Sse8337I and EcoRI) and the EcoRI (partially digested)/AII (blunted) fragment of pTA7002 (containing a portion of the GR sequence, the rbcS E9 poly(A) addition sequence and a portion of the hygromycin transcription unit or HPT) was ligated into the SmaI/PstI-digested pBlueScript vector to generate pER5 and pER5-X, respectively. The target promoter vector pER6, which contains eight copies of the LexA operator sequence fused to the 46 35S minimal promoter, was made by inserting the HindIII/ XbaI (blunted) fragment of pLexA-CAT (Zuo et al., 1995) into the HindIII/BglII (blunted) digested p6GF vector (Q. Wang and N.-H. Chua, unpublished results). The HindIII site of pER6 was converted to an SphI site to yield pER6-Sph. The pER7 vector was constructed by inserting the SphI fragment of pTA7002 (containing a portion of VP16, GR and the hygromycin transcription unit) into the SphI site of pER6-Sph in an orientation such that the nos poly(A) addition sequence of the hygromycin transcription unit faces the LexA operator (Figure 1). To construct pER8, the Sse8337I/NcoI (partially digested) fragment of pER5 (the G1090XVE transcription unit and the 5 half of the HPT transcription unit) and the NcoI/XhoI fragment of pER7 (the 3 half of the HPT transcription unit and the LexA-46 target promoter) were coligated into the Sse8337I/XhoI-digested pTA210 (P. Spielhofer and N.-H. Chua, unpublished results), a derivative of pPZP200 (Hajdukiewicz et al., 1994). The pER8GFP reporter construct was made by inserting an XhoI/SpeI fragment of pX-GFP (Kost et al., 1998) into the same sites of pER8. Northern blotting analysis was carried out using standard procedures (Sambrook et al., 1989). For quantitative analysis, the blots were scanned with a PhosphorImage scanner, and the data were collected using ImageQuant (Molecular Dynamics, Sunnyvale, CA, USA). See legend to Figure 4 for further details.

Acknowledgements
We would like to thank the Arabidopsis Biological Resources Center for providing DNA. We are grateful to Drs S. Moller, T. Kunkel, S. Hoth and S.A. Saxe for critically reading the manuscript and for their useful comments.

References
Aoyama, T. and Chua, N.-H. (1997) A glucocorticoid-mediated transcriptional induction system in transgenic plants. Plant J. 11, 605612. Bechtold, N., Ellis, J. and Pelletier, G. (1993) In planta Agrobacterium-mediated gene transfer by inltration of adult Arabidopsis thaliana plants. C. R. Acad. Sci. Paris, Life Sci. 316, 11941199. Benfey, P.N., Ren, L. and Chua, N.-H. (1990) Tissue-specic expression from CaMV 35S enhancer subdomains in early stages of plant development. EMBO J. 9, 16771684. Bohner, S., Lenk, I., Rieping, M., Herold, M. and Gatz, C. (1999) Transcriptional activator TGV mediates dexamethasoneinducible and tetracycline-inactivatable gene expression. Plant J. 19, 8795. Bruce, W., Folkerts, O., Garnaat, C., Crasta, O., Roth, B. and Bowen, B. (2000) Expression proling of the maize avonoid pathway genes controlled by estradiol-inducible transcription factors CRC and P. Plant Cell, 12, 6580. Caddick, M.X., Greenland, A.J., Jepson, I. et al. (1998) An ethanol inducible gene switch for plants used to manipulate carbon metabolism. Nature Biotech. 16, 177180. Dalrymple, M.A., McGeoch, D.J., Davison, A.J. and Preston, C.M. (1985) DNA sequence of the herpes simplex virus type 1 gene whose product is responsible for transcriptional activation of immediate early promoters. Nucl Acids Res. 13, 78657879. Evans, R.M. (1988) The steroid and thyroid hormone receptor superfamily. Science, 240, 889895. Farabaugh, P.J. (1978) Sequence of the lacI gene. Nature, 274, 765769. Gatz, C. (1997) Chemical control of gene expression. Ann. Rev. Plant Physiol. Plant Mol. Biol. 48, 89108. Gatz, C. and Lenk, I. (1998) Promoters that respond to chemical inducers. Trend. Plant Sci. 3, 352358. Gatz, C., Frohberg, C. and Wendenburg, R. (1992) Stringent repression and homogeneous de-repression by tetracycline of a modied CaMV35S promoter in intact transgenic tobacco plants. Plant J. 2, 397404. Golemis, E.A. and Brent, R. (1992) Fused protein domains inhibit DNA binding by LexA. Mol. Cell Biol. 12, 30063014. Greene, G.L., Gilna, P., Watereld, M., Baker, A., Hort, Y. and Shine, J. (1986) Sequence and expression of human estrogen receptor complementary DNA. Science, 231, 11501154. Hajdukiewicz, P., Svab, Z. and Maliga, P. (1994) The small, versatile pPZP family of Agrobacterium binary vectors for plant transformation. Plant Mol. Biol. 25, 989994. Horii, T., Ogawa, T. and Ogawa, H. (1981) Nucleotide sequence of the LexA gene of E. coli. Cell, 23, 689697. Blackwell Science Ltd, The Plant Journal, (2000), 24, 265273

Plant materials, growth conditions and plant transformation

The Columbia ecotype of Arabidopsis thaliana was used in all experiments except for the 35SGFP lines, which are in the Landsberg background (Kost et al., 1998). Plants were grown under continuous white light at 22C on solid A medium (1 3 MS salts, 3% sucrose, 0.8% agar) supplemented with appropriate antibiotics and inducers. Transformation of Arabidopsis plants was carried out by vacuum inltration (Bechtold et al., 1993). Transformation of tobacco plants was done as described previously (Kunkel et al., 1999). T2 heterozygous seeds were used in initial screens (Figures 2 and 3a). Segregation analysis indicated that most transgenic lines, including lines 1, 10, 17 and 22, contained single T-DNA insertions. These lines were selfed, and homozygous T3 or T4 seeds were used in experiments described in Figures 3(b) and 4 6. Line 8 probably contained two T-DNA insertions.

Chemicals and induction

4-hydroxyl tamoxifen and 17-b-estradiol were purchased from Research Biochemicals International (Natick, MA, USA), and dexamethasone from Sigma (St. Louis, MO, USA). The chemicals were prepared as 1, 4 and 10 mM stock solutions in dimethyl sulfoxide (DMSO), and stored at 20C in small aliquots. DMSO alone had no effects on transgene expression (data not shown). Chemical treatments were carried out by directly germinating seeds on a medium supplemented with inducers, or transferring seedlings from a non-inductive onto an inductive medium.

RNA preparation and Northern blotting analysis

Total RNA was prepared with the Plant RNA Prep Kit (Qiagen, Valencia, CA, USA), following the manufacture's instructions.

The XVE inducible expression system 273

Huang, H.L. and Brandriss, M.C. (2000) The regulator of the yeast proline utilization pathway is differentially phosphorylated in response to the quality of the nitrogen source. Mol. Cell Biol. 20, 892899. Ishige, F., Takaichi, M., Foster, R., Chua, N.H. and Oeda, K. (1999) A G-box motif (GCCACGTGCC) tetramer confers high-level constitutive expression in dicot and monocot plants. Plant J. 18, 443448. Kang, H.G., Fang, Y. and Singh, K.B. (1999) A glucocorticoidinducible transcription system causes severe growth defects in Arabidopsis and induces defense-related genes. Plant J. 20, 127133. Kost, B., Spielhofer, P. and Chua, N.-H. (1998) A GFP-mouse talin fusion protein labels plant actin laments in vivo and visualizes the actin cytoskeleton in growing pollen tubes. Plant J. 16, 393 401. Kunkel, T., Niu, Q.W., Chan, Y.S. and Chua, N.-H. (1999) Inducible isopentenyl transferase as a high-efciency marker for plant transformation. Nat. Biotechnol. 17, 916919. Kurzer, M.S. and Xu, X. (1997) Dietary phytoestrogens. Annu. Rev. Nutr. 17, 353381. Lewis, M., Chang, G., Horton, N.C., Kercher, M.A., Pace, H.C., Schumacher, M.A., Brennan, R.G. and Lu, P. (1996) Crystal structure of the lactose operon repressor and its complexes with DNA and inducer. Science, 271, 12471254. Marczak, J.E. and Brandriss, M.C. (1991) Analysis of constitutive and noninducible mutations of the PUT3 transcriptional activator. Mol. Cell Biol. 11, 26092619. Martinez, A., Sparks, C., Hart, C.A., Thompson, J. and Jepson, I. (1999) Ecdysone agonist inducible transcription in transgenic tobacco plants. Plant J. 19, 97106. Miki, T., Ebina, Y., Kishi, F. and Nakazawa, A. (1981) Organization of the lexA gene of Escherichia coli and nucleotide sequence of the regulatory region. Nucl. Acids Res. 9, 529543. Mueller-Fharnow, A. and Egner, U. (1999) Ligand-binding domain of estrogen receptors. Curr. Opin. Biotechnol. 10, 550556. Oertel-Buchheit, P., Porte, D., Schnarr, M. and Granger-Schnarr, M. (1992) Isolation and characterization of LexA mutant repressors with enhanced DNA binding afnity. J. Mol. Biol. 225, 609620. Sambrook, J., Fritsch, E.F. and Maniatis, T. (1989). Molecular Cloning: A Laboratory Manual. Cold Spring Harbor, NY: Cold Spring Harbor Laboratory Press. Tora, L., White, J., Brou, C., Tasset, D., Webster, N., Scheer, E. and Chambon, P. (1989) The human estrogen receptor has two independent nonacidic transcriptional activation functions. Cell, 59, 477487. Tsai, M.J. and O'Malley, B.W. (1994) Molecular mechanisms of action of steroid/thyroid receptor superfamily members. Annu. Rev. Biochem. 63, 451486. Webster, N.J., Green, S., Jin, J.R. and Chambon, P. (1988) The hormone-binding domains of the estrogen and glucocorticoid receptors contain an inducible transcription activation function. Cell, 54, 199207. Weinmann, P., Gossen, M., Hillen, W., Bujard, H. and Gatz, C. (1994) A chimeric transactivator allows tetracycline-responsive gene expression in whole plants. Plant J. 5, 559569. Zuo, J. and Chua, N.-H. (2000) Chemical-inducible systems for regulated expression of plant genes. Curr. Opin. Biotechnol. 11, 146151. Zuo, J., Baler, R., Dahl, G. and Voellmy, R. (1994) Activation of the DNA-binding ability of human heat shock transcription factor 1 may involve the transition from an intramolecular to an intermolecular triple-stranded coiled-coil structure. Mol. Cell Biol. 14, 75577568. Zuo, J., Rungger, D. and Voellmy, R. (1995) Multiple layers of regulation of human heat shock transcription factor 1. Mol. Cell Biol. 15, 43194330. Zuo, J., Niu, Q.-W., Nishizawa, N., Wu, Y., Kost, B. and Chua, N.H. (2000) KORRIGAN, an Arabidopsis endo-1,4-b-glucanase, localizes to the cell plate by polarized targeting and is essential for cytokinesis. Plant Cell, 12, 11371152.

Blackwell Science Ltd, The Plant Journal, (2000), 24, 265273