In vivo analysis of plant promoters and transcription factors by agroinfiltration of tobacco leaves


For correspondence (fax +1 501 575 7601; e-mail


A convenient, Agrobacterium-mediated transient expression assay has been evaluated for rapid analysis of plant promoters and transcription factors in vivo. By simple infiltration of Agrobacterium cells carrying appropriate plasmid constructs into tobacco leaves in planta, reproducible expression assays could be conducted in as little as 2–3 days without using expensive equipment (e.g. biolistic gun or electroporation apparatus) or complicated procedures (e.g. preparation of protoplasts). Biotic and abiotic treatments could be applied to the intact plant to examine their influence on promoter activity and gene expression. Using this method, we have tested the stress-responsive as-1 and heat shock elements, yeast GAL4 transactivation system, two promoters of pathogenesis-related (PR) genes as well as a heat shock promoter. Through deletion analyses of tobacco PR1a promoter and a novel myb1 promoter, we have also successfully identified the cis-regulatory regions in these promoters that are responsive to salicylic acid treatment or tobacco mosaic virus infection. Together, our results demonstrate that Agrobacterium-mediated transient expression is a simple and efficient method for in vivo assays of plant promoters and transcription factors.


The interaction between transcription factors and cis-acting regulatory sequences presented in plant promoters is a key step involved in the regulation of plant gene expression. As a result, much effort has been made to dissect various plant promoters, to identify different transcription factors, and to study their specific interactions (reviewed by Katagiri & Chua, 1992; Menkens et al. 1995 ; Rushton & Somssich, 1998; Zhu et al. 1996 ). Commonly, in vivo characterization of plant promoters and transcription factors is carried out by generation and analysis of transgenic plants expressing promoter/reporter gene fusion and/or corresponding trans-factors. This process is labor-intensive, time-consuming, and usually takes several months. In addition, transgene expression in transgenic plants often varies significantly due to insert position and other effects, thus complicating data analysis. To overcome some of those disadvantages, methods for transient expression analysis have been developed as convenient alternatives to stable transformation. For example, biolistic transient transformation of suspension cells or plant organs has often been used to dissect plant promoters ( Baum et al. 1997 ; Tuerck & Fromm, 1994) and to characterize transcription factors ( Goff et al. 1990 ; Gubler et al. 1999 ; Sainz et al. 1997 ). PEG-mediated gene transfer and electroporation of protoplasts has also frequently been used to identify cis-acting elements in promoters activated by external stimuli ( Abel & Theologis, 1994; Hattori et al. 1992 ; Sablowski et al. 1994 ; Solano et al. 1995 ). However, biolistic transient transformation requires very expensive instruments and often has large sample variations. In most cases, a control construct (e.g. LUC gene) must be cotransformed with test constructs to standardize the variation. On the other hand, protoplast transient transformation involves a care-intensive, complicated procedure of isolating protoplasts from cell cultures or leaf mesophyll. These limitations thus impede the widespread use of transient expression assays for dissection of plant promoters and in vivo characterization of cis-elements and trans-factors.

To facilitate in vivo analysis of plant promoters and transcription factors, we have adopted a convenient and efficient transient expression assay based on Agrobacterium-mediated transformation of tobacco leaves in planta. By simple infiltration of agrobacteria carrying plasmid constructs into tobacco leaves (agroinfiltration), in vivo expression analysis of promoters and transcription factors can be conducted in as little as 2–3 days. In addition, the effects of external stimuli such as pathogen infections and environmental stresses on promoter activity can be evaluated in intact plants when combining agroinfiltration with biotic or abiotic treatments. We have tested the suitability of this simple system using the SA-responsive as-1 element, the heat shock element (HSE), and the promoters of tobacco PR1a and PR2d genes. We have also applied this method to the rapid dissection of a novel tobacco myb1 promoter, and identified a cis-acting regulatory region conferring the majority of promoter activity and inducibility in response to SA treatment and TMV infection. The suitability of this simple method for studying in vivo cis-element/trans-factor interaction and transcriptional regulation was verified using the yeast GAL4/VP16 transactivation system.


Reproducibility of the transient expression assay

A reliable transient expression assay requires small variation and good reproducibility. To test if this is the case for the Agrobacterium-mediated transient expression assay, the pathogen- and SA-inducible PR2d promoter (−607, Hennig et al. 1993 ; Shah & Klessig, 1996) was fused with an intron-containing β-glucuronidase (GUS) reporter gene on a high-copy binary vector (pCAMBIA 1391Z). After the resulting construct was introduced into hypervirulent A. tumefaciens strain EHA105, the transient transformation was carried out by agroinfiltration in combination with salicylic acid (SA) treatment or tobacco mosaic virus (TMV) infection. Little variation on transient expression efficiency was observed within a series of transformations performed in the same tobacco leaf. The standard deviation of GUS activities from eight independent transformations was approximately 12% of the mean value. Good reproducibility was also obtained using near fully expanded leaves from different 6-week-old plants grown under the same conditions. Among the transient assays conducted on four tobacco plants, the standard deviation of GUS activities is less than 20%. We found that leaf or plant age could significantly affect the transient transformation. Generally, fully expanded, old leaves gave poor transformation efficiency. Use of half-expanded, young leaves often resulted in variable efficiency, even for transformations performed in the same leaf; this may be largely due to the difficult and uneven infiltration of agrobacteria into the intercellular space of young leaves. By using near fully expanded leaves located at the middle position of 6-week-old tobacco plants, we were able to obtain reproducible transient expression data with small variations.

Analysis of stress-responsive cis-elements

The as-1 and HSEs were used to test the suitability of the transient expression assay for analysis of stress-responsive cis-acting sequences. The as-1 element was initially identified in cauliflower mosaic virus (CaMV) 35S promoter as well as Agrobacterium ocs and nos gene promoters, and is responsive to SA, jasmonates and other signals ( Qin et al. 1994 ). Subsequently, as-1 was also found in promoters of some stress-responsive genes such as those encoding tobacco PR1a and soybean glutathione S-transferase ( Ellis et al. 1993 ; Strompen et al. 1998 ; Ulmasov et al. 1994 ). In this study, the −87 35S promoter with as-1 located from −83 to −63 was compared with −49 35S minimal promoter as well as −87 35S promoter with mutated as-1 element. The −49 35S promoter exhibited a low basal level of transcription activity and was not responsive to SA treatment or TMV infection ( Figure 1a). Addition of the as-1-containing sequence (−87 to −50) to the −49 35S resulted in a significant increase of GUS activity in response to SA treatment (eightfold induction) or TMV infection (13-fold induction). But the mutation of the as-1 element (TGACGTAAGGGATGACGCAC was changed to TGctGTAAGGGATctCGCAC) completely eliminated SA and TMV inducibility of the −87 35S promoter.

Figure 1.

Stress responsiveness of the as-1 and heat shock elements.

Mean values and standard errors of four independent transformations are shown.

(a) Transient transformations were conducted by agroinfiltration of −87 35S:GUS (with as-1 element), −87m 35S:GUS (with mutated as-1 element) or −49 35S:GUS (without as-1) constructs. Their GUS activities were determined after H2O, SA, mock or TMV treatments.

(b) Transient transformations were conducted by agroinfiltration of HSE/ −49 35S:GUS or −49 35S:GUS constructs. Their GUS activities were determined after 37°C heat treatment for 20 h, followed by a 6 h recovery period. In control experiments (CK), tobacco plants were grown under the normal condition without heat treatment.

The thermoinducibility of heat shock genes is attributed to the conserved cis-acting HSEs. In plants, the optimal HSE core consensus was shown to be 5′-aGAAg-3′ ( Barros et al. 1992 ). The efficient binding of HSE by heat stress transcription factors requires at least three units of HSE. In our experiments, an oligonucleotide (5′-tcgaggaTTCtcGA-AgtTTCcaGAAgtTTCtaGAAgcaaatc-3′), which contains a trimer of the consensus binding site, was fused with −49 35S and a GUS reporter gene. In comparison with the −49 35S promoter, addition of the HSEs produced a 66-fold increase in GUS activity in response to heat shock treatment ( Figure 1b). These results indicated that Agrobacterium-mediated transient transformation is an appropriate method for in vivo analysis of cis-acting elements including stress-responsive regulatory sequences.

Interaction between cis-element and trans-factor in vivo

To evaluate whether the transient expression assay is suitable for studying cis-element/trans-factor interactions, we have tested the yeast GAL4 transactivation system that is known to be functional in plant cells ( Ma et al. 1988 ). A plasmid construct was made to contain both activator and reporter genes ( Figure 2a). Driven by a 35S constitutive promoter, the activator gene encodes the GAL4/VP16 fusion protein, which combines the DNA-binding domain of yeast GAL4 protein ( Carey et al. 1989 ; Giniger et al. 1985 ) with the activation domain of herpes simplex virus protein VP16 ( Triezenberg et al. 1988 ). The reporter gene consists of a 10× GAL4 binding site linked with −49 35S and GUS. In comparison with control constructs (A and B in Figure 2a) that produced only basal levels of GUS expression, transient transformation of the test construct (C in Figure 2a) carrying both reporter and activator gene led to a more than 18-fold increase in GUS activity ( Figure 2b). Therefore the agroinfiltration-based transient assay is capable of detecting the interaction between GAL4/VP16 transcription factor and the GAL4 binding site based on activated GUS reporter gene expression in plant cells.

Figure 2.

Activation of GUS expression by yeast GAL4/VP16 trans-factor.

(a) Fusion constructs used in the transient assay: A, B, control constructs; C, test construct. The DNA fragments in the constructs are not proportionally drawn. 35S, full-length 35S promoter; GAL4/VP16, GAL4/VP16 coding sequence; Nos, Agrobacterium Nos terminator; GBS, GAL4 binding site; −49, −49 35S minimal promoter; GUS, β-glucuronidase coding sequence.

(b) Levels of GUS expression resulting from transient transformation with constructs A–C: mean values and standard errors of four independent transformations.

Induction of PR and heat shock gene promoters

The pathogen-responsive promoters of tobacco PR1a and PR2d and the heat shock promoter of soybean Gm17.5 were previously characterized using stable transformation and/or transient transformation based on protoplast electroporation ( Hennig et al. 1993 ; Shah & Klessig, 1996; Strompen et al. 1998 ; Treuter et al. 1993 ; Uknes et al. 1993 ; Van de Rhee & Bol, 1993; Van de Rhee et al. 1990 ). To assess whether the transient expression assay is useful for analysis of stress-responsive promoters, these promoters and some of their deletion derivatives were fused to the intron-containing GUS reporter gene and tested for their promoter activity and inducibility in response to biotic and abiotic stresses.

The PR1a promoter (−893 to +29) yielded strong activity and inducibility of GUS expression in response to SA treatment (approximately fourfold induction) and TMV infection (approximately sixfold induction; Figure 3a). Although the −679 deletion derivative exhibited significant reduction of promoter activity, it still retained similar levels of inducibility in response to SA treatment and TMV infection (approximately fourfold and sevenfold induction, respectively). However, further deletion to −499 and −237 abolished almost all activity and inducibility. These data suggest that SA- and TMV-responsive sequence(s) of PR1a promoter are mainly localized in the region between −893 and −499, which is in agreement with previous analysis using transgenic plants ( Strompen et al. 1998 ; Uknes et al. 1993 ; Van de Rhee et al. 1990 ).

In addition to PR-1a promoter, the transient assay of PR2d promoter (−607) also showed significant induction (four- to fivefold) of GUS activity in response to SA treatment and TMV infection ( Figure 3b). Both PR1a and PR2d promoters exhibited high levels of GUS activity (150–200 pmol mg−1 protein min−1) in response to water or mock treatments in our transient assays ( Figure 3a,b). In contrast, the heat-responsive Gm17.5 promoter was not affected by agroinfiltration and exhibited very low levels of GUS activity under normal conditions ( Figure 3c). The heat shock treatment (37°C for 20 h), however, resulted in 28-fold induction of GUS activity. We also found that the edge of tobacco leaf exhibited consistently higher levels of induction than the center of the leaf in the Gm17.5 promoter assay, suggesting that the periphery of tobacco plants was more strongly affected by, and sensitive to, heat shock treatment.

Figure 3.

Induction of PR and heat shock gene promoters.

Mean values and standard errors of three independent transformations are shown.

(a) Transient transformations were conducted by agroinfiltration of tobacco PR1a promoter deletion constructs (−893, −679, −499, −237). GUS activities were determined after H2O, SA, mock or TMV treatments.

(b) Transient transformations were conducted by agroinfiltration of tobacco −607 PR2d:GUS construct. GUS activities were determined after H2O, SA, mock or TMV treatments.

(c) Transient transformations were conducted by agroinfiltration of soybean Gm17.5:GUS construct. GUS activities were determined after 37°C heat treatment for 20 h, followed by a 6 h recovery period.

Dissection of a novel pathogen-inducible myb1 promoter

Tobacco myb1 gene was previously shown to be induced by TMV infection and SA treatment, and probably encodes a transcription factor involved in plant defense responses ( Yang & Klessig, 1996). Recently, we have isolated and sequenced a 2 kb EcoRI fragment containing the myb1 promoter (GenBank accession number AF248962). The transcription start site of myb1 was determined using primer extension and found to be located at 163 bp upstream of the translational start codon (Y. Yang, unpublished results). In this study, the Agrobacterium-mediated transient assay was employed to rapidly dissect the myb1 promoter and to identify the SA- and TMV-responsive element. Our initial tests showed that the myb1 promoter yielded very low levels of GUS expression. This is not unexpected, as myb1 is a very weakly expressed regulatory gene. To facilitate the analysis of the weak promoter, the myb1 promoter and its deletion derivatives were cloned into pCAMBIA1391Z, so that they are positioned downstream of a reverse-orientated 35S promoter in the T-DNA region of the vector. Previous reports and our experience showed that the 35S promoter, even in reverse orientation, could affect the downstream promoter sequences and significantly increase the basal levels of GUS expression. In most cases, the high background of GUS activity interferes with promoter analysis. In our transient assays, however, it allowed the detection of promoter activity and facilitated the dissection of the weak myb1 promoter. As shown in Figure 4, most constructs yielded induced GUS expression in response to SA or TMV. Overall, induction of the myb1 promoter by TMV infection (four- to sevenfold) is stronger than that by SA treatment (two- to threefold). Deletion analyses have shown that −1531, −1122, −832 and −519 had similar levels of activity and inducibility. Although slightly lower, −351 retained the majority of the myb1 promoter activity and inducibility. Therefore the SA- and TMV-responsive element(s) probably reside in the 351 bp region upstream of the transcription start site. Interestingly, −2073 exhibited significantly lower levels of activity and inducibility than many deletion derivatives (e.g. fragments −1531, −1122, −832, −519, −351), suggesting that negative regulatory sequence(s) may be present in the region between −2073 and −1531.

Figure 4.

Deletion analysis of tobacco myb1 promoter.

Transient transformation was conducted by agroinfiltration of the vector control (pCAMBIA1391Z) as well as tobacco myb1 promoter (−2073) and its deletion derivatives (−1531, −1122, −832, −519, −351, −116). Mean values and standard errors of three independent transformations are shown. GUS activities were determined 24 h after H2O and SA treatments (a) or 48 h after mock and TMV infections (b).


Agroinfiltration has been widely used in stable transformation of A. thaliana and some other plant species. Recently, agroinfiltration-based transient transformation has also been employed to analyse foreign gene expression ( Kapila et al. 1997 ); gene silencing ( Baulcombe, 1999; Schöb et al. 1997 ); and the gene-for-gene interaction between host resistance and pathogen avirulence genes ( Frederick et al. 1998 ; Scofield et al. 1996 ; Tang et al. 1996 ; Van den Ackerveken et al. 1996 ). In most of these studies, only qualitative analysis was conducted based on GUS histochemical staining or hypersensitive reaction resulting from the gene-for-gene interaction. To our knowledge, however, the Agrobacterium-mediated transient expression has not been evaluated for its suitability and potential applications in quantitative analysis of plant promoters and cis-element/trans-factor interaction in vivo.

In our experiments, tobacco was selected as the plant material for agroinfiltration because it gave excellent transformation efficiency and allowed multiple transient expression assays on a single large leaf. Typically, eight to 16 individual transient expression assays could be conducted within a single tobacco leaf with little variation on transient transformation efficiency. To obtain consistent results, near fully expanded leaves from uniformly grown plants should be used to minimize assay variations among leaves from different plants. Reproducible results were obtained with the agroinfiltration conducted in different leaves as well as in the same leaf. In contrast, transient assays based on particle bombardment often have large variations and generally require cotransformation of an internal control construct to standardize the variation.

Using well characterized cis-elements and transactivation system, we have demonstrated that the Agrobacterium-mediated transient expression is a rapid and reliable method for studying cis-elements, trans-factors, and their interactions. The as-1 and HSEs have been previously defined and are induced by pathogen/SA and heat shock, respectively. In our transient assays, the −87 35S promoter that carries as-1 element exhibited significant induction by SA treatment (eightfold induction) and TMV infection (13-fold induction), whereas the −49 35S promoter that lacks as-1 element exhibited no induction ( Figure 1a). Furthermore, site-directed mutations of as-1 element in −87 35S promoter completely eliminated its SA- and TMV-inducibility. These data are in agreement with a previous study by Qin et al. (1994) , who observed five- to 10-fold induction of GUS activity by SA treatment using the −90 35S:GUS transgenic tobacco plants. The HSE also yielded a significant induction (66-fold) of GUS expression in response to heat treatment in our assays ( Figure 1b). Since the heat shock treatment was carried out on an intact tobacco plant in a growth chamber (a slow heat shock process), the induced GUS expression is relatively low when compared with rapid heat shock treatments using protoplasts ( Treuter et al. 1993 ). In addition, we have shown that the Agrobacterium-mediated transient transformation of the plasmid construct carrying both GAL4 binding site/−49 35S:GUS reporter gene and GAL4/VP16 activator gene resulted in 18-fold induction of GUS gene expression ( Figure 2). In contrast, transformation of the control constructs without the GAL4 binding site or activator gene did not induce GUS expression. This result is consistent with a previous report by Schwechheimer et al. (1998) , who observed a 10-fold activation when testing the GAL4/VP16 transactivation system in tobacco mesophyll protoplasts.

The Agrobacterium-mediated transient assay may also facilitate the dissection and rapid analysis of native plant promoters, including stress-responsive promoters. In this study, we tested two PR gene promoters, one heat shock promoter and a novel myb gene promoter. The −893 promoter of tobacco PR1a yielded high levels of induced GUS gene expression in response to SA treatment and TMV infection. The levels of induced GUS activity are comparable to the data obtained in previous studies using transgenic plants ( Uknes et al. 1993 ; Van de Rhee et al. 1990 ). The transient assay indicated that deletion of the PR1a promoter to −679 resulted in decreased but still inducible expression of the reporter gene. Further deletion to −499 or −237 eliminated the promoter activity and inducibility by SA and TMV. These results are in agreement with previous analyses of tobacco PR1a promoter ( Grüner & Pfitzner, 1994; Strompen et al. 1998 ; Uknes et al. 1993 ; Van de Rhee & Bol, 1993; Van de Rhee et al. 1990 ). Previous studies using transgenic tobacco demonstrate that −900 PR-1a promoter has very high levels of activity and inducibility in response to SA treatment or TMV infection. The −689 or −661 promoters were shown to retain significant levels of inducibility, whereas a further deletion to −643 or −600 abolished induced reporter gene expression. Recently, Strompen et al. (1998) have shown that SA inducibility was mediated by a 139 bp region (from −691 to −553) in the PR1a promoter. Their data suggest that a regulatory element located from −691 to −625 may contribute to the induction of PR1a gene expression, whereas an as-1-like sequence located from −593 to −574 was identified to control the level of PR1a expression.

We have recently isolated a 2 kb promoter of tobacco myb1 gene that is induced by SA and TMV during the hypersensitive response and development of systemic acquired resistance ( Yang & Klessig, 1996). Using the agroinfiltration-based transient assay, we were able within 2 weeks to localize the region (−351) that was responsible for the majority of myb1 promoter activity and inducibility. Sequence analysis of this region indicated the presence of an elicitor-responsive element that can interact with pathogen-inducible WRKY transcription factors ( Rushton & Somssich, 1998).

We have noticed that the Agrobacterium-mediated transient assay of PR promoters had high levels of GUS activity in water and mock treatments. For example, Uknes et al. (1993) showed that transgenic lines expressing −903 PR1a:GUS had an average of 48 GUS units (1 unit = pmol MU/mg protein min−1) in the water treatment. In our assays, the GUS activities for the water and mock treatments were about 200 units. Because PR genes can be induced by wounding and/or compatible interactions with virulent bacteria, infiltration-associated wounding as well as agrobacterial infection probably contributed to the activation of PR promoters to some extent. Although high basal levels of GUS expression in water and mock treatments resulted in reduced induction by SA and TMV, the transient assay still allowed us to identify the SA- and TMV-responsive region in the PR-1a and myb1 promoter. Other inducible promoters, such as those responsive to salt, drought, cold or heat, are likely to be less affected by agroinfiltration. Our study showed that agroinfiltration did not affect the expression of heat-responsive promoters such as Gm17.5. Only heat shock treatment significantly activated GUS expression.

A number of limitations should be considered with regard to our transient assay system. Like protoplast-based transient assays, the agroinfiltration-based transient assay is probably not suitable for detecting tissue-specific or developmentally regulated promoter activity. Although biolistic bombardment may allow limited analysis of tissue specific promoters ( Baum et al. 1997 ), stable transformation is probably required for studying tissue-specific and developmentally regulated promoters. In addition, potential effects of agroinfiltration itself must be taken into account, particularly if pathogen-induced promoters are being studied. Despite its limitations, the agroinfiltration-based assay does not require expensive apparatus and is a much simpler process than existing transient assays such as protoplast electroporation and particle bombardment. Furthermore, it allows the analysis of external effects (such as pathogen infection, salt stress and heat shock) on promoter expression in intact plants, rather than in detached tissues or protoplasts. Therefore, it more closely reflects actual patterns of gene expression in plants under various biotic and abiotic stresses. During the past several years there has been increasing interest in developing artificial transactivation systems for regulated gene expression and creating synthetic promoters with high inducibility or tissue specificity ( Aoyama & Chua, 1997; Böhner et al. 1999 ; Ishige et al. 1999 ; Martinez et al. 1999a; 1999b ). This transient assay may provide a simple alternative to biolistic bombardment or protoplast electroporation for quick screening of DNA constructs and evaluation of synthetic promoters or transactivation systems before conducting time-consuming stable transformation.

Experimental procedures

Plasmid construction

Binary vector pCAMBIA1391Z (Cambia, Australia) and its modified form were used to make all plasmid constructs. The T-DNA region of pCAMBIA1391Z contains a promoterless GUS gene interrupted by a catalase intron and a CaMV 35S:hgp gene fusion encoding the hygromycin B phosphotransferase. A nearby 35S promoter (positioned in a reverse orientation) was removed to eliminate its influence on promoters being analysed. This was done by excision of the BstXI–XhoI fragment containing 35S:hgp gene from pCAMBIA1391Z, resulting in the modified vector p1391X.

To construct plasmids carrying cis-elements, DNA fragments with and without the as-1 element (−87 35S and −49 35S, respectively) were amplified by high-fidelity polymerase chain reaction (PCR) using the 35S promoter as the template. In addition, PCR-mediated site-directed mutagenesis was performed in −87 35S promoter to knock out the as-1 element (mutation of TGACGTAAGGGATGACGCAC to TGctGTAAGGGATctCGCAC). Subsequently, these fragments were cloned into the EcoRI–NcoI site of p1391X, producing plasmids p35S-87, p35S-87m and p35S-49, respectively. A heat shock cis-element containing a trimer of HSE oligonucleotide (5′-cgaggaTTCtcGAAgt TTCcaGAAgtTTCtaG-AAgcaaatc-3&prime, Treuter et al. 1993 ) was also cloned into the EcoRI–NcoI site of p1391X.

For in vivo assays of DNA–protein interaction, three plasmids containing different combinations of the GUS reporter gene, GAL4 binding site and/or GAL4/VP16 activator gene were constructed ( Figure 2a). An EcoRI fragment containing 10 × 17mer GAL4 binding site was synthesized and ligated into the EcoRI site of p35S-49. The resulting reporter gene construct was used as a control (A in Figure 2a). A PstI–EcoRI fragment containing a 35S promoter linked with the GAL4/VP16 activator gene was cloned into p35S-49 (B in Figure 2a), which contains the activator gene and −49 35S:GUS reporter gene, but lacks the GAL4 binding site. The test construct (C in Figure 2a) was made by further addition of a 10 × 17mer GAL4 binding site upstream of −49 35S:GUS reporter gene.

To test stress-responsive promoters, tobacco PR1a promoter and its deletion derivatives (from −893, −679, −499, −237 to +29) were generated by high-fidelity PCR and cloned into EcoRI–NcoI sites of p1391X. The PR2d promoter (−607) in pGA482 ( Hennig et al. 1993 ; Shah & Klessig, 1996 ) was excised with HindIII and NcoI and cloned into p1391X. The soybean Gm17.5 promoter ( Czarnecka et al. 1989 ) in pBI221 was excised with HindIII and BamHI and fused to the GUS reporter gene in p1391X.

The 5′ region of tobacco myb1 cDNA was used as a probe to screen a Nicotiana tabacum genomic library (Clontech, CA, USA) for genomic clones containing the myb1 promoter. A 2 kb myb1 promoter was subcloned into pUC19 plasmid and sequenced in both directions (Genbank accession number AF248962). A series of 5′ deletions of the myb1 promoter was generated using exonuclease III and S1 nuclease digestion. After confirming their deletion positions by DNA sequencing, these 5′ deletion derivatives were linked to the GUS reporter gene by ligation into the EcoRI–HindIII sites of pCAMBIA1391Z.

Plant materials

Tobacco (Nicotiana tabacum var. Xanthi nc) plants were grown in a greenhouse at 22°C for 6 weeks. Two days before agroinfiltration they were transferred into a Conviron growth chamber and maintained at 22°C under 16 h light.

Preparation of Agrobacterium suspension

Agrobacterium tumefaciens strain EHA 105 containing individual constructs was streaked on YEP solid medium (10 g l−1 Bacto-peptone, 10 g l−1 yeast extract, 5 g l−1 NaCl, 15 g l−1 agar) supplemented with rifampicin (60 μg ml−1) and kanamycin (50 μg ml−1), and grown at 28°C for 2 days. Agrobacteria were then inoculated in 20 ml induction medium ( Winans et al. 1988 ) containing AB salts (1 g l−1 NH4Cl, 0.3 g l−1 MgSO4−7H2O, 0.15 g l−1 KCl, 0.01 g l−1 CaCl2, 0.0025 g l−1 FeSO4−7H2O), 2 m m phosphate, 1% glucose, 20 m m 2-(N-morpholino)ethanesulfonic acid (MES, pH 5.5), 100 μM acetosyringone as well as rifampicin and kanamycin. After overnight culture at 28°C, agrobacterial cells were collected by centrifugation for 15 min at 3000 g and resuspended in 10 m m MES (pH 5.5) plus 10 m m MgSO4 solution or 10 m m MES (pH 5.5) plus MS basal medium ( Murashige & Skoog, 1962). After supplemented with 100 μM acetosyringone, bacterial suspension was adjusted to a final OD600 of 0.8 for agroinfiltration. Although addition of 10 m m MES (pH 5.5) in suspension solution appears to slightly increase transient expression efficiency (about 10%), no significant difference in transformation was observed between agrobacterial cells suspended in 10 m m MgSO4 solution versus MS basal medium.

Agroinfiltration of tobacco leaves in planta

Agrobacterium-mediated transient transformation was conducted on near fully expanded leaves that still attached to the intact plant. Bacterial suspension was infiltrated into intercellular spaces of intact leaves using a 1 ml plastic syringe. By infiltrating 100 μl of bacterial suspension into each spot (typically 3–4 cm2 in infiltrated area), eight to 16 spots separated by veins could be arranged in a single tobacco leaf. After agroinfiltration, tobacco plants were covered with transparent plastic bags and maintained in a growth chamber at 22°C under 16 h light for 24–48 h.

Biotic and abiotic treatments

TMV infection was carried out 24 h after agroinfiltration. Agroinfiltrated leaves were inoculated with TMV strain U1 at a concentration of 1 μg ml−1 in 50 m m phosphate buffer (pH 7.0) by rubbing with carborundum. Mock-inoculated leaves were rubbed with carborundum and buffer only. For chemical treatment, tobacco leaves were infiltrated with 1 m m SA (pH 6.5) 48 h after agroinfiltration. Leaf discs (approximately 1.5 cm2) were sampled for GUS assays 48 h after TMV inoculation and 24 h after SA treatment, respectively. Heat shock treatment was conducted 48 h after agroinfiltration by placing tobacco plants at 37°C in a growth chamber for 20 h. Following heat shock, tobacco plants were allowed to recover in a 22°C growth chamber for 6 h before leaf discs were sampled for GUS assays. All the experiments described in this study were repeated independently at least two times.

Determination of GUS activity

Leaf discs from each infiltration site were collected in 1.5 ml eppendorf tube and ground in 500 μl GUS extraction buffer (50 m m NaHPO4 pH 7.0, 10 m m 2-mercaptoethanol, 10 m m Na2EDTA, 0.1% sodium lauryl sarcosine, 0.1% Triton X-100). After centrifugation for 10 min (12 000 g) at 4°C, 50 μl of supernatant was mixed with 250 μl of GUS assay solution (2 m m 4-methylumbelliferyl- d-glucuronide in extraction buffer) and 200 μl of GUS extraction buffer. A 50 μl aliquot was removed immediately and added into 2 ml stop buffer (0.2 m sodium carbonate) to be used as control. The rest of the mixture was incubated at 37°C for 60 min. GUS activity was determined using the DyNAQuant 200 fluorometer (Hoefer, CA, USA) and protein concentration of tissue homogenates was determined with the Bradford reagent (BioRad, CA, USA).


We thank Drs Daniel Klessig, Joyti Shah, Lutz Nover and William Gurley for providing PR1, PR-2, heat shock elements and promoters, respectively. This study was supported by grants from United State Department of Agriculture NRICGP (98-35303-6987) and Arkansas Science and Technology Authority (99-B-14).