K+ channels in plants can currently be classified into six families with individual members being involved in nutrient uptake, loading of the xylem and the physiology of stomatal movement. In this study we have focused on akt2/3. This K+ channel, as shown by GUS-expression analysis, is expressed in the phloem and xylem of the aerial parts of Arabidopsis thaliana. Northern blot analyses revealed the highest akt2/3-concentrations in the flower stalk, followed by the leaf, flower and stem. During the light period (8 am to 4 pm), transcripts reached a peak around noon (11 am), decayed to almost 50% in the afternoon and reached a low background level in the following dark period. In continuous darkness, however, the K+ channel mRNA content had already decreased beyond the background level by noon. In leaves and flower stalk, the light-induced transcription of akt2/3 was suppressed by CO2-free air, indicating that gene activity is under the control of photosynthates. Additionally, when rosette leaves were illuminated and flower stalks shaded, akt2/3-mRNA transcription was still inhibited in the shaded region. This indicates that channel gene activation is sensitive to photosynthesis-derived factors from neighboring cells rather than factors mobile in the phloem. We propose that the coupling between sugar production and allocation involves the photosynthate- and light-dependent phloem K+ channel AKT2/3.
In order to determine the role of this K+ channel type ( Marten et al. 1999 ) in phloem physiology, we examined the cell-specific expression of akt2/3 in response to the developmental stage and day/night cycle. We found the expression of the phloem K+ channel under the control of CO2 assimilates and light, indicating that this K+ conductance is involved in the control of phloem transport.
Tissue- and cell-type-dependent expression
In order to analyse the organ-specific expression pattern of the K+ channel akt2/3 we performed Northern blot analyses. Although we cannot extrapolate RNA data to protein levels, Phillipar et al. 1999 ) have shown that an increase in K+ channel mRNA correlates with an increase in protein content. When compared to the root, stem, leaf and flower we identified the flower stalk (with leaves and flowers excised) of Arabidopsis thaliana as the organ with maximum K+ channel mRNA content ( Fig. 1b). A more detailed analysis of flower stalk segments revealed a gradual increase of akt2/3-mRNA levels from the apex to the rosette base ( Fig. 1b inset). When mRNA was isolated from dark grown and illuminated root cultures of Arabidopsis thaliana, from vascular-free callus or from the hypocotyl of etiolated or green seedlings, channel transcripts could not be detected in Northern experiments (data not shown). To follow akt2/3-gene activity during plant development we generated transgenic plants expressing GUS under the control of the akt2/3-promoter. β-glucuronidase activity was histochemically localized in the green parts (stems, flower stalks, leaves, sepals) of 38 out of 44 transgenic Arabidopsis plants, but not in the roots and hypocotyl of young seedlings ( Fig. 2). During the greening of the upper hypocotyl segment, activity of the K+ channel promoter emerged in the vascular tissue of some transgenic plants only ( Fig. 2a). Since the strongest signals always co-localized with the veins ( Fig. 2a,b), we analyzed 16 μm cross-sections from adult leaves to determine the cell type-specific activity of the akt2/3-promoter. In these experiments GUS-staining was associated with the phloem and xylem parenchyma cells, here marked in longitudinal ( Fig. 2c) and cross-sections ( Fig. 2d) through leaf veins.
Throughout the diurnal cycle with a light period from 8 am to 4 pm and darkness from 4 pm to 8 am levels of channel mRNA in leaves peaked before noon (11 am) about 3 h after the onset of illumination (8 am, Fig. 3a). Following a delay of 1 h transcripts increased gradually ( Fig. 3a) and saturated at around 10 to 11 am ( Fig. 3a inset). During the afternoon and dark period, akt2/3-transcripts dropped to background levels and when leaves were kept in continuous darkness (8 to 0 am) transcript levels con- tinuously decayed ( Fig. 3a). Since the akt2/3 gene is active in the vascular tissue it is possible that the channel activation signal generated by light is either restricted to the sites of phloem loading in source leaves, or travels along the phloem of the flower stalk. We anticipated a signal gradient underlying differential K+ channel expression because akt2/3-mRNA levels in flower stalks increased from the apex to the rosette leaf basis ( Fig. 1b inset). To test this hypothesis flower stalks were kept in the dark, while the rosette leaves were illuminated. Figure 3(b) shows that exposure of the entire plant to light for 3 h induced channel transcription in the leaves and flower stalk. However, when the flower stalk was kept in the dark, akt2/3-mRNA was almost undetectable, even though the leaves were illuminated. Furthermore, light-induction of K+ channel mRNA in the flower stalk was found even when the entire rosette was shaded (data not shown). These results indicate that channel transcription is triggered by a local light-derived signal ( Moore & Sheen, 1999; Umemura et al. 1998 for review) rather than by one originating from the leaves and translocated in the phloem with the mass flow of assimilates. Similar results were obtained when only one half (180°) of the rosette was illuminated; akt2/3-expression was observed in the illuminated half and not in the leaves kept in the dark (data not shown), thus again excluding the possibility of an exported signal. Possibly the phloem veins in the flower stalks, like those in the leaves, depend on the photosynthates received only from mesophyll cells in their close proximity.
In order to separate signals related to photosynthesis from direct light effects, experiments were performed in the presence and absence of CO2. Following a 3 h illumination with white light (8 to 11 am) in the presence of ambient CO2 the expression levels were maximal, while in the absence of CO2 during 3 h in the dark the amount of akt2/3-transcripts was largely reduced ( Fig. 3c). This behavior indicates that CO2-assimilation is required for phloem K+ channel expression. However, when comparing leaves derived from plants exposed to CO2-free air during 3 h of illumination, transcript levels were still higher than the shaded ones, indicating that light is also able to trigger channel expression through a non-photosynthetic pathway. Whether the latter induction of K+ channel transcripts is related to light boxes ( Ketchum & Slayman, 1996) in the akt2/3 promoter awaits further analyses.
We investigated the akt2/3-K+ channel mRNA content of Arabidopsis thaliana on the basis of Northern blot hybridization experiments as well as plants expressing Gus and found akt2/3 predominantly expressed in the vascular tissue of the shoot. Remarkably, hypocotyls which exhibit a vascular organization identical to roots lacked K+ channel transcripts. Although the mRNA for this K+ channel is light induced but did not increase even upon long-term white-light treatment of the hypocotyl and root cultures, akt2/3-mRNA is not generally expressed in phloem cells. Gene activity is restricted to the vascular tissue of photosynthetically active organs. In addition to recent localization studies based on whole mount in situ hybridisation ( Marten et al. 1999 ), GUS-staining revealed akt2/3-expression in xylem parenchyma cells as well ( Fig. 2). It is thus tempting to speculate that this K+ channel is involved in the transfer of potassium from the leaf xylem to the phloem of source leaves.
The akt2/3-gene is activated when the seedling is largely independent from carbohydrates imported by the seed. The decay of the akt2/3 transcript level in CO2-free air and the low K+ channel mRNA concentration in shaded leaves, indicate that photosynthesis elicit a signal required for K+ channel gene activation. On the level of the rosette leaves we did not observe pronounced variations in the amount of akt2/3-transcripts between leaves located at different positions within the rosette (data not shown). When the flower stalk emerges, and flowers and seeds develop, mRNA content of akt2/3 peaked in the flower stalk. The fact that K+ channel mRNA increased in the light, when assimilates are enriched at the source site and decay along the transport path towards the tip of the growing flower stalk (i.e. sink, Marschner et al. 1997 ), supports the hypothesis that products of CO2 assimilation in the rosette induce akt2/3-gene transcription. Experiments with shaded flower stalks, however, showed that phloem transport rates decay by only about 30% (G. Lohaus, personal communication) while akt2/3 transcription completely lacked activation. We therefore concluded that CO2 assimilates received from the flower stalk mesophyll rather than through the phloem trigger K+ channel transcription.
To gain insights into the light-dependent K+ channel regulation and phloem physiology, we will introduce phytochrome and sugar sensing mutants as well as akt2/3-knock out plants into our future experiments.
Arabidopsis thaliana (ecotype Wassiljewskija-2) was grown in soil (Type P, Gebr. Hagera Handelsges.m.b.H., Sinntal-Jossa, Germany) and exposed to continuous white light for 8 h at 160 μE, as measured with a Li-Cor Quantum Sensor (Model Li-185B, Lincoln, Nebraska, USA). Seeds from transgenic Arabidopsis plants were surface sterilized (5% hypochloride) before being grown in 0.8% agar medium (Sigma, Deisenhofen, Germany), containing 1 × MS salts (Murashige and Skoog, Sigma, Deisenhofen, Germany), 1% sucrose (Roth, Karlsruhe, Germany), pH 5.7, and 50 μg ml−1 kanamycin. All plants were kept in growth chambers at 22°C during the light and 16°C during the 16 h dark period. Air was filtered with dry CaCO3 for incubation of plants in CO2-free air. After 15 min of air circulation the infra-red gas analyzer (HCM-1000, Walz GmbH, Effeltrich, Germany) did not detect CO2 any more.
RNA extraction and Northern blot analysis
Total-RNA was isolated from Arabidopsis thaliana using the RNeasy-Extraction Kit (Qiagen, Hilden, Germany) according to the manufacturer's protocol. Ten μg of total-RNA were analyzed by Northern blot hybridization ( Sambrook et al. 1989 ). The hybridization probe was 32P labeled akt3-cDNA (Ready To Go, Pharmacia Biotech, Freiburg, Germany). Hybridization signals of three independent experiments were quantified using the Image Master VDS Software (Pharmacia Biotech, Freiburg, Germany) and normalized to the corresponding 18S rRNA. The highest value of the relative amount of akt2/3-mRNA of each experiment was set to 100%. Data represent the mean (± SD) of three assays.
Construction of the akt2/3-promoter-β-glucuronidase gene fusion
A genomic fragment containing 636 bp of the akt2/3-promoter was introduced into the promoterless Agrobacterium binary plasmid pGPTV-Kan ( Becker et al. 1992 ) via a XbaI restriction site. Using a triparental mating procedure with the helper plasmid pRK2013 ( Ditta et al. 1980 ), the akt2/3-pGPTV-Kan construct was transferred from Escherichia coli to Agrobacterium tumefaciens LBA 4404 ( Bevan, 1984). Transconjugates of A. tumefaciens were identified by resistance against rifampicin and kanamycin as well as PCR using akt2/3-promoter specific primers.
Transformation of A. thaliana
Arabidopsis plants (ecotype WS-2) were transformed with akt2/3-pGPTV-Kan by employing the vacuum infiltration method ( Bechtoldt et al. 1993 ) Transgenic plants were first grown in vitro on kanamycin-containing medium and later transferred to soil for optimal seed production.
GUS-analyses were carried out on primary transformants. Staining procedure with 5-bromo-4-chloro-3-indolyl β- d-glucuronide (X-Gluc) was performed as described by Gallagher (1992) with prefixed plant material. Samples were then incubated in 70% ethanol to remove the pigments. Semi-thin cross-sections (16 μm) were prepared from these samples with a microtome (Leica RM 2165) after complete dehydration, and by embedding in GMA (HEMA, Plano, Wetzlar).
We are grateful to Prof. R. Kaldenhoff for supplying the pGPTV-Kan vector, the Agrobacterium strain LBA4404 and pRK2013 helper plasmid. We thank J. Arnoldt for excellent technical assistance. This project was funded by DFG grants to Rainer Hedrich.