Fluorescent tagging of proteins and confocal imaging techniques have become methods of choice in analysing the distributions and dynamic characteristics of proteins at the subcellular level. In common use are a number of strategies for transient expression that greatly reduce the preparation time in advance of imaging, but their applications are limited in success outside a few tractable species and tissues. We previously developed a simple method to transiently express fluorescently-tagged proteins in Arabidopsis root epidermis and root hairs. We describe here a set of Gateway-compatable vectors with fluorescent tags incorporating the ubiqutin-10 gene promoter (PUBQ10) of Arabidopsis that gives prolonged expression of the fluorescently-tagged proteins, both in tobacco and Arabidopsis tissues, after transient transformation, and is equally useful in generating stably transformed lines. As a proof of principle, we carried out transformations with fluorescent markers for the integral plasma membrane protein SYP121, a member of the SNARE family of vesicle-trafficking proteins, and for DHAR1, a cytosolic protein that facilitates the scavenging of reactive oxygen species. We also carried out transformations with SYP121 and its interacting partner, the KC1 K+ channel, to demonstrate the utility of the methods in bimolecular fluorescence complementation (BiFC). Transient transformations of Arabidopsis using Agrobacterium co-cultivation methods yielded expression in all epidermal cells, including root hairs and guard cells. Comparative studies showed that the PUBQ10 promoter gives similar levels of expression to that driven by the native SYP121 promoter, faithfully reproducing the characteristics of protein distributions at the subcellular level. Unlike the 35S-driven construct, expression under the PUBQ10 promoter remained elevated for periods in excess of 2 weeks after transient transformation. This toolbox of vectors and fluorescent tags promises significant advantages for the study of membrane dynamics and cellular development, as well as events associated with environmental stimuli in guard cells and nutrient acquisition in roots.
If you can't find a tool you're looking for, please click the link at the top of the page to "Go to old article view". Alternatively, view our Knowledge Base articles for additional help. Your feedback is important to us, so please let us know if you have comments or ideas for improvement.
The strong 35S (P35S) promoter of Cauliflower mosaic virus has long been a favoured choice to drive constitutive expression (Benfey and Chua, 1990), both in studies relying on stable and transient transformation. The use of a constitutive promoter has advantages when the gene of interest is normally expressed at low levels, and when tissue-specific characteristics of expression are not of primary concern. However, overexpression can present difficulties of its own, including co-suppression or gene silencing (Elmayan and Vaucheret, 1996; Elmayan et al., 1998; Mishiba et al., 2005), and long-term developmental effects of expressing the gene of interest. Transient expression methods avoid problems associated with development and, at least initially, they circumvent problems of gene silencing, as well as the substantial lead-in time needed to generate transgenic plants. Nonetheless, both stable and transient expression methods can give rise to mislocalisation of target proteins when expressed ectopically at high levels and, even in transient studies, expression driven by the 35S promoter shows suppression over a period of days (cf. Campanoni et al., 2007).
In search for alternatives we settled on the Arabidopsis thaliana ubiquitin-10 gene promotor (PUBQ10) that facilitates moderate expression in nearly all tissues of Arabidopsis (Norris et al., 1993). We developed a vector set that combines the advantages of previous systems using the PUBQ10 promoter (Geldner et al., 2009), adding the advantages of Gateway compatibility for fast and efficient cloning, the ease of selection using Basta and hygromycin resistance in stable lines, and incorporates the option to generate both N- and C-terminal fusion constructs for expression. The latter option is especially important, as N- or C-terminal fusions can mask the targeting sequences of proteins and lead to mis-localisation of the fusion protein product. Thus, it is a standard requirement to prepare and test fusion constructs in both configurations. Finally, we extended recent methods for transient expression based on co-cultivation of Agrobacterium with Arabidopsis seedlings (Campanoni et al., 2007; Li et al., 2009). Here, we describe applications of the PUBQ10-based vector set in generating fluorescent fusion proteins, and their application to expression in Arabidopsis in stable lines, and after co-cultivation in Arabidopsis and tobacco. We show that the vector set gives moderate levels of expression of intrinsic membrane and soluble proteins, both in tobacco and Arabidopsis, but with a temporal stability after transient transformation that is apparently limited only by the maintenance of the plants. We report, too, that these transient methods enable reliable transformation of guard cells as well as epidermal cells and root hairs in Arabidopsis.
Results and discussion
The Gateway-compatible, pUB-Dest vector set (Figure 1; see also Figure S1) was designed to enable both N- and C-terminal fusions of proteins of interest with several fluorophores, including eGFP, mRFP, CFP and YFP (Fricker et al., 2006), the bimolecular fluorescence complementation (BiFC) components nYFP and cYFP (Walter et al., 2004), as well as the photoconvertible EOS fluorescent protein (Wiedenmann et al., 2004; Gurskaya et al., 2006). These binary vectors incorporated the nucleotide sequence corresponding to the first 634 base pairs of the promoter (PUBQ10) immediately upstream of the ubiquitin-10 gene from Arabidopsis (At4g05320). They also included sequences encoding phosphinothricin-N-acetyltransferase and aminoglycoside-3-adenyltransferase for resistance to Basta or hygromycin in the plant, and to streptomycin and spectinomycin in Escherichia coli, respectively.
To determine the efficiency of transformation and to characterize expression, we made use both of immunochemical methods and confocal laser scanning microscopy. The analysis was based in part on comparative measurements with the same vectors incorporating the standard 35S (P35S) promoter, both in tobacco and in Arabidopsis after transient transformations. For this purpose we used two markers: (i) the Arabidopsis SNARE protein SYP121 (At3g11820; 38 kDa), an integral plasma membrane protein that is known to express throughout the plant (Lipka et al., 2007; Bassham and Blatt, 2008; Enami et al., 2009), and overlaps functionally with its tobacco homologue (Sutter et al., 2006); and (ii) the Arabidopsis dehydroascorbate reductase DHAR1 (At1g19570; 24 kDa), a redoxin that is found primarily in the cytosol, but has also been associated with membranes (Dixon et al., 2002; Elter et al., 2007). We also carried out analyses using BiFC constructs of SYP121 and its interacting partner, the K+ channel KC1 (At4g32650; 78 kDa) (Honsbein et al., 2009). As described below, the analysis also summarizes standardized parameters for co-cultivation of Arabidopsis with Agrobacterium tumefaciens GV3101 that give consistent and prolonged expression of fusion proteins in root, hypocotyl, shoot, and cotyledon epidermis, as well as in guard cells of Arabidopsis seedlings.
Agrobacterium growth and transgene expression
Bacterial preparation proved an important determinant for the efficient transformation and expression, both of tobacco and Arabidopsis. We compared the relative efficiency of transformation with Agrobacterium tumefaciens harvested from liquid culture following a single 16-h growth cycle (Campanoni et al., 2007), and following a subculture cycle for an additional 6–8-h period after the bacteria were diluted 1:10 in fresh medium, fllowing the method of Grefen et al. (2008), before harvesting (final OD600, 1–2). After harvest and pre-treatment with acetosyringone, Agrobacterium carrying DHAR1 and SYP121 fusion constructs were infiltrated into tobacco leaves and co-cultivated with 4-day-old Arabidopsis seedlings in half-strength MS salts plus 0.003% Sylwet, as described previously (Campanoni et al., 2007; Li et al., 2009). The concentration of Agrobacterium used for injections and in co-cultivation proved to have no measurable effect on transformation efficiency, whereas increasing the pH of the co-cultivation medium, especially, promoted transformation and expression, much as we reported previously for Agrobacterium rhizogenes-mediated transformation (Campanoni et al., 2007); by contrast, the fluorescence signals were enhanced between three- and five-fold when the bacteria were passed through the additional subculture cycle (Grefen et al., 2008), and we therefore standardized conditions, based on this method, with a final OD600 of 0.2 and pH of 7 that yielded over 90% transformation effciency of Arabidopsis with Agrobacterium tumefaciens.
Yield and temporal characteristics of PUBQ10-driven expression
Differences associated with the P35S and PUBQ10 promoters were most notable in construct expression in tobacco. Quantified on the basis of fluorescence yield, leaf infiltrations of tobacco with Agrobacterium carrying the SYP121-GFP construct gave roughly 20-fold greater signal driven by the P35S promoter compared with the same construct under PUBQ10 control (Figures 2a and 3); however, this difference in fluorescence signals was greatly attenuated when expressed in Arabidopsis (Figure 2a). The difference between tobacco and Arabidopsis expression was confirmed by western blot analysis showing the presence of the doublet of bands (Geelen et al., 2002; Tyrrell et al., 2007), recognized by polyclonal antibodies to SYP121 (Figure 2c). Expressed under both promoters, the vesicle trafficking protein showed a strong fluorescence associated with the cell periphery (Figure 3), consistent with its known localization to the plasma membrane where it facilitates the final stages of membrane fusion between vesicle and target membranes (Lipka et al., 2007; Bassham and Blatt, 2008), as well as controlling channel-mediated K+ transport (Honsbein et al., 2009). We noted in every case the absence of any appreciable fluorescence from internal structures with PUBQ10-driven expresssion. Expressed under the P35S promoter, SYP121-GFP fluorescence was frequently associated with mobile punctate structures in the cytosol (Figure 3a–d; see also Video Clips S1, S2 and S3) and, occasionally, with the nucleus in tobacco. These latter observations are consistent with retention in the secretory pathway (Campanoni et al., 2007), and may indicate the formation of inclusion bodies associated with the high levels of P35S-driven expression. This interpretation was supported by fluorescence bleaching experiments that showed a pronounced mobility of SYP121 fluorescence within cells when expressed under the control of the P35S promoter, but not under PUBQ10 control (not shown; see Figure 5 below).
The SYP121 construct exhibited prolonged transgene expression that remained substantially elevated even 2 weeks after transformation, both in tobacco and Arabidopsis, when driven by the PUBQ10 promoter. Under our standardized conditions, the fluorescence yield for SYP121-GFP reached a maxiumum, independent of the promoter, approximately 72 h after adding the bacteria. Expression remained essentially stable thereafter for the PUBQ10-driven construct, but decayed significantly after 6–8 days when driven by the P35S promoter, whether expressed in Arabidopsis (Figure 2c) or in tobacco (not shown). Analysis of DHAR1-GFP expression led to a similar conclusion. Finally, we noted that SYP121-GFP fluorescence under PUBQ10 control compared favourably with that of the SYP121 transgene when stably expressed and driven by its own promoter (PSYP121Figure 2A; see also Figures 4–7), although transcriptional analyses (see Figure S2; Winter et al., 2007) suggest a 5–10-fold enhancement would be expected with the PUBQ10-driven construct. Indeed, translational and post-translational factors are likely to contribute to the temporal characteristics of transgene expression (Grefen et al., 2008). Thus, the similarity in fluorescence yields in this case is fortuitous. Nonetheless, we suspect that the moderate activity of the PUBQ10-driven system means that its control of expression is less prone to gene silencing over time.
Cellular characteristics of PUBQ10-driven expression
Co-cultivation of Arabidopsis seedlings and Agrobacterium with T-DNA carrying fluorescent transgene fusions invariably led to the transformation of the root tissues, hypocotyl, stem and cotyledons. Within these tissues, transformation was limited to the outer (epidermal) cell layer, much as was reported previously (Campanoni et al., 2007), indicating that all cells exposed to the Agrobacterium culture during co-cultivation are subject to transformation. Figure 4 illustrates results typical of transformations with PUBQ10-driven SYP121-GFP and its distribution in the root epidermis, as well as the advantages for confocal image analysis of expression in a tissue that shows little or no background fluorescence. Driven by the PUBQ10 promoter, SYP121-GFP fluorescence was observed along the length of the root, both in epidermal cell files forming root hairs (trichoblasts) and those that did not (atrichoblasts). SYP121-GFP fluorescence was strongest, however, at the base of trichoblasts and at the tips of root hairs, especially younger – presumably still growing – root hairs (see Figure 4a–d; Video Clip S4). Analysis of individual root hairs (Figure 4e–n) showed the fluorescence to be associated with the cell periphery, not with the cytosol or tonoplast, and we noted the absence of any appreciable fluorescence from internal structures. Furthermore, SYP121 showed little evidence of mobility within individual root hairs, consistent with its situation as an integral plasma membrane protein. In the latter case, we used fluorescence recovery after photobleaching (FRAP) and fluorescence lifetime in photobleaching (FLIP) experiments to assess lateral movement of SYP121-GFP when expressed in root hairs, and we repeated these measurements with experiments using SYP121-EOS to follow the dynamics of the SNARE after photoactivation of the photochromic EOS protein (Figure 5).
Using fluorescence yield, we compared the PUBQ10-driven expression of SYP121 with that of the SNARE when stably expressed and driven by its own promoter (PSYP121). Under PSYP121 control, SYP121-GFP expression has been reported throughout the vegetative plant, especially in epidermal tissues of leaves, stems, the root and root hairs (Enami et al., 2009). Within the root epidermis and root hairs, we found the PUBQ10 promoter to faithfully reproduce the protein distribution, with expression driven by the SYP121 promoter (Figure 6A). In each case, the fluorophore tag was strongest at the basal end of trichoblasts and at the tips of younger root hairs (Figure 6B; see also Figure 4a–d,i–n). We noted a pronounced elevation in fluorescence around the dome of the root hair that dropped off steeply within the first 10–20 μm of the tip apex (Figure 6b,c). Quantified on the basis of peripheral fluorescence at the apex relative to the base of the root hair, SYP121 showed the greatest difference in its distribution in root hairs shorter than 40–60 μm, with the fluorescence ratio declining to approximately 2:1 in root hairs longer than 140–160 μm, irrespective of the promoter driving expression (Figure 6a). The same distribution was reflected in YFP fluorescence following PUBQ10-driven expression of the BiFC partners fused with SYP121 with the K+ channel subunit KC1 (Figure S3; Video Clip S8). This enhanced localization to the root hair tip suggests a corresponding spatial distribution to the channel and SNARE functionalities, and may bear on their roles in K+ nutrition (Honsbein et al., 2009).
Finally, it is of particular interest that our use of the PUBQ10 promoter and the strategies based on co-cultivation also lead to transient transformation of stomatal guard cells in Arabidopsis. In the past, studies of protein dynamics in guard cells have been hampered by the recalcitrance of this cell type to transient transformation by Agrobacterium, leaving biolistic methods as the only alternative (Meckel et al., 2005; Mikosch et al., 2006; Sutter et al., 2007). Figure 7 shows DHAR1-EOS expressed in epidermal and guard cells of an Arabidopsis cotyledon, and compares favourably with DHAR1-EOS expressed stably in Arabidopsis (not shown) when driven by the PUBQ10 promoter. Images were taken before and after photoactivation of the EOS protein, and they clearly show the local rise in EOS emission at wavelengths of 560–615 nm, with exciation at 543 nm following photoactivation, and its spread within those cells targeted in photoactivation. Also visible is a concurrent decline in emission between 505 and 530 nm, with excitation at 488 nm, that is a common feature of EOS photoactivation (Wiedenmann et al., 2004). DHAR1 is predominantly a cytosolic protein like other members of the redoxin/GST transferase protein family (Noctor and Foyer, 1998; Frova, 2006), so its mobility within the cell is expected. The EOS photoconversion in this case offers a check against interference from background fluorescence associated with these photosynthetic tissues, as chloroplast fluorescence is also enhanced by the high-intensity light used in photoconversion. The observations thus underscore the utility of the transformation strategy in studies of protein dynamics in guard cells, for example in relation to environmental stimuli.
In conclusion, we present an improved vector set for transformation, as well as extended methods for transient expression analysis in Arabidopsis. These new tools will help accelerate studies of endosomal organization, protein localization and dynamics, and will facilitate work on root hairs and in guard cells that, in the past, have proven recalcitrant to transient Agrobacterium-mediated transformation as a strategy for transient gene expression. Our vector set offers straightforward access to both N- and C-terminal fusion constructs with a promoter that gives moderate levels of expression of both intrinsic membrane and soluble proteins. Expression based on the PUBQ10 promoter gives exceptional temporal stability that, in conjunction with a co-cultivation strategy, enables reliable transformation of guard cells as well as epidermal cells and root hairs in Arabidopsis. These tools should find wide application in studies, for example, of polar development, root–rhizosphere interactions, ion transport and interactions with symbionts, and may also find applications in high-throughoutput screens directed to characterizing candidate gene products and their partners.
A 634 bp genomic DNA fragment immediately upstream of the ATG start codon was identified, based on the UBQ10 promoter sequence (Norris et al., 1993) and was amplified from Arabidopsis Col-6 genomic DNA using forward (5′-gcgaagcttGTCGACGAGTCAGTAATAAACG-3′) and reverse (5′-gcgctcgagCTGTTAATCAGAAAAACTCAG-3′) primers. The amplified fragment was digested by HindIII and XhoI and cloned into pGPTVII.GFP (Walter et al., 2004) to verify its ability to drive GFP expression after Agrobacterium-mediated transformation in Nicotiana benthamiana. Sequencing of several clones showed a C to A nucleotide exchange at position -29 in the Col-6 promoter sequence when compared to that of the published Col-0 sequence (Norris et al., 1993). The promoter was introduced into the pUGT1 vector (K. Schumacher, unpublished) from which it was excised as a starting point in developing the vectors described here.
The vector pB7WGR2 (Karimi et al., 2002) was digested using SpeI and AflII, and the opened segment amplified by PCR using a forward primer generating 5′SpeI and PsiI sites upstream of the T35S, and a 3′AflII site to incorporate the Basta resistance gene (see Figure S1 for details). Ligation of this PCR product into the open vector created an intermediate construct. The 35S promoter was removed by digestion with PmeI and SpeI, and was replaced with a corresponding fragment from the vector pUGT1, containing the UBQ10 promotor followed by the Gateway cassette, thereby creating pUB-Dest. Fluorescence tags mGFP6, mYFP, mRFP and EOS, as well as the split-YFP variants nYFP and cYFP (Walter et al., 2004; Wiedenmann et al., 2004; Fricker et al., 2006), were synthesized via PCR to introduce 5′SpeI and 3′PsiI sites, and were subsequently ligated into the vector creating the C-terminal vector set in pUBC-Dest.
To generate the corresponding vector set with N-terminal fluorophores, two PCR fragments were created, using pUB-Dest as a template. The fragments contained the vector backbone, the first stretching from an AvrII site between the pVS replication and pBR322 origin to the 3′ end of the UBQ10 promotor, adding SpeI and MfeI sites, and a second stretching from the AvrII site to the 5′ end of the 35S terminator, and adding MfeI and PsiI sites. Ligation of these fragments resulted in a circular vector lacking a Gateway cassette. This intermediate vector was opened using the blunt end enzyme PsiI and the Gateway cassette A from the Gateway Conversion kit (Invitrogen, http://www.invitrogen.com) was introduced, creating the pUBN-Dest vector. The fluorescence tags were amplified by PCR to introduce SpeI and MfeI sites at the 5′ and 3′ ends, respectively, and were ligated into pUBN-Dest to generate the N-terminal vector set. A schematic of both pUBN-Dest and pUBC-Dest is shown in Figure 1.
Gateway cloning and bacterial transformation
All Gateway vectors were cloned and amplifed using E. coli ccdB-survival™ cells (Invitrogen), and selected using spectinomycin (100 μg ml−1) and chlorampenicol (30 μg ml−1). Vector backbones were verified by restriction digest analysis and complete sequencing. Entry clones encoding SYP121, KC1 and DHAR1 were constructed by PCR amplification using primers that contained attB1 and attB2 sites as 5′ modifications. Gel-purified PCR products were introduced into pDONR207 (Invitrogen) using BP-clonase II according to the manufacturer’s instructions. Recombinant entry clones were amplified in Top10™ cells (Invitrogen), and were verified by restriction digest analysis and sequencing. LR recombination reactions were performed by mixing 1 μl (=150 ng) of pUBC-Dest or pUBN-Dest vector, including the fluorophores, 1 μl (=150 ng) of the entry clone and 0.5 μl of LR-clonase II (Invitrogen) to generate SYP121 and DHAR1 N-terminally and KC1 C-terminally tagged with the fluorophores. The reaction was incubated at room temperature (20°C) for 1 h. Aliquots of 1 μl were then used to transform Top10 bacterial cells, which were grown overnight at 37°C on 2% agar plates with Luria Bertani (LB)media containing 100 μg μl−1 spectinomycin. Selected colonies were isolated and the clones were verified through restriction digest analysis. Destination clones were transformed subsequently in Agrobacterium tumefaciens strain GV3101 pMP90 (Koncz and Schell, 1986). Successful transformations were verified by plasmid rescue in E. coli and restriction digest analysis.
Agrobacterium carrying clones of interest were grown in a first cycle overnight (16 h, 220 rpm, 28°C) in 5 ml LB medium with 50 μg ml−1 rifampicin, 25 μg μl−1 gentamycin and 100 μg ml−1 spectinomycin. A second cycle of growth was started by innoculating an aliquot from the overnight growth at a 1:10 dilution in fresh medium, and then cultured for 6–8 h (220 rpm, 28°C) to give a final OD600 of 1–2. The bacteria were harvested and resuspended in 10 mm MgCl2 with 150 μm acetosyringone (3,5-dimethoxy-acetophenone; Fluka, now part of Sigma-Aldrich, http://www.sigmaaldrich.com) and 10 mm 2-(N-morpholine)-ethanesulphonic acid (MES), pH 5.5, before dilution in the same medium to give a final OD600 0.1 or 0.2 for tobacco leaf infiltration. For co-cultivation with Arabidopsis, the bacteria were resuspended in half-strength MS basal salts medium (Sigma-Aldrich), pH 7, to a final OD600 0.2–0.3 with 150 μm acetosyringone and 0.003% Sylwet-77 (Lehle Seeds, http://www.arabidopsis.com) (Campanoni et al., 2007; Li et al., 2009).
Plant material and co-cultivation
Nicotiana benthamiana and Nicotiana tabacum were grown as described previously (Sutter et al., 2006). A. thaliana L. (ecotype Colombia) seeds were sterilized for 10 min in 10% NaHClO3 with 1% Triton-X100 and stratified at least 2 days at 4°C in the dark before incubation in six-well plates containing 3 ml half-strength MS medium. Arabidopsis seeds were germinated under constant light at 80 μmol s−1 m−2 light and 22°C. After germination, the medium was exchanged with fresh half-strength MS plus 150 μm acetosyringone, 0.003% Sylwet-77, including a suspension of Agrobacterium carrying the transgene of interest.
Confocal microscopy and quantification
Expression was determined by confocal laser imaging, using a Carl Zeiss CLSM510-META-UV confocal laser scanning microscope (Carl Zeiss, Inc., http://www.zeiss.co.uk), with argon ion and helium/neon lasers. For visualizing GFP and non-photoactivated EOS, excitation at 488 nm was used and fluorescence collected after reflection off an NFT545 dichroic mirror and passage through a 505–530-nm bandpass filter. EOS was photoactivated with 351- and 364-nm light from an Enterprise II UV laser, and the photoactivated EOS was visualized with excitation at 543 nm after passage through an NFT545 dichroic mirror and 560–615-nm bandpass filter. Chloroplast fluorescence was separated using a CFT635 dichroic mirror and was collected using the META detector set for a bandwidth of 632–700 nm. Laser power, dichroics, filters, detector gains and offsets were kept fixed for all the different specimens, to allow comparison between the different samples. As necessary, only the laser attenuation was varied within the pre-calibrated linear range of the tuneable optical filters of the instrument. Transformation, distributions and fluorescence tracking were quantified using both the Carl Zeiss LSM 510 AIM software (v3.2) and ImageJ (http://www.rsbweb.nih.gov/ij). Fluorescence values are reported as the means ± SEs of at least three independent experiments or 18 seedlings after subtracting background fluorescence measured in controls co-cultivated with non-transformed Agrobacterium.
Total protein extraction and western blot analysis
For expression analysis by western blot, seedlings were harvested after examination under the confocal microscope, flash-frozen and ground under liquid nitrogen. Total proteins were extracted in denaturing ‘lyse and load’ buffer (50 mm Tris–HCl, pH 6.8, 4% sodium dodecylsulfate, 8 m urea, 30% glycerol, 0.1 m dithiothreitol and 0.005% bromophenol blue), and SDS-PAGE and western blot analyses were performed as described previously (Grefen et al., 2009). Polyclonal rabbit anti-GFP primary antibody (Abcam, http://www.abcam.com) was diluted 1:5000. Bound antibodies were detected using goat, anti-rabbit IgG alkaline phosphatase (Sigma-Aldrich) and staining solution of 66 μl MBT (50 mg ml−1 nitro-blue-tetrazolium chloride in 70% dimethylformamide) and 33 μl BCIP (50 mg ml−1 5-bromo-4-chloro-3-indoylphosphate-p-toluidine in 100% dimethylformamide) in 10 ml of staining buffer containing 100 mm Tris-HCl, pH 9.5, 100 mm NaCl and 5 mm MgCl2.
All non-commercial vector constructs described in this article are available for academic and non-profit use on request.
We are grateful to Amparo Ruiz-Pardo for help in the glasshouse, and to Annegret Honsbein for comments on the manscript. This work was supported by grants BB/F001630/1 and BB/F001673/1 from the UK Biotechnology and Biological Sciences Research Council and a VIP Award grant from the Wellcome Trust to MRB, and by grants from the DFG (FOR 964) to KS and JK.