Isolation of AtSUC2 promoter-GFP-marked companion cells for patch-clamp studies and expression profiling


  • Natalya Ivashikina,

    1. Lehrstuhl für Molekulare Pflanzenphysiologie und Biophysik, Julius-von-Sachs-Institut für Biowissenschaften, Universität Würzburg, Julius-von-Sachs-Platz 2, 97082 Würzburg, Germany,
    Search for more papers by this author
      These authors contributed equally to this work.
  • Rosalia Deeken,

    1. Lehrstuhl für Molekulare Pflanzenphysiologie und Biophysik, Julius-von-Sachs-Institut für Biowissenschaften, Universität Würzburg, Julius-von-Sachs-Platz 2, 97082 Würzburg, Germany,
    Search for more papers by this author
      These authors contributed equally to this work.
  • Peter Ache,

    1. Lehrstuhl für Molekulare Pflanzenphysiologie und Biophysik, Julius-von-Sachs-Institut für Biowissenschaften, Universität Würzburg, Julius-von-Sachs-Platz 2, 97082 Würzburg, Germany,
    Search for more papers by this author
  • Erhard Kranz,

    1. Institut für Allgemeine Botanik, Zentrum für Angewandte Molekularbiologie der Pflanzen, Universität Hamburg, Ohnhorststr. 18, 22609 Hamburg, Germany, and
    Search for more papers by this author
  • Benjamin Pommerrenig,

    1. Lehrstuhl Botanik II, Molekulare Pflanzenphysiologie, Universität Erlangen-Nürnberg, Staudtstrasse 5, 91058 Erlangen, Germany
    Search for more papers by this author
  • Norbert Sauer,

    1. Lehrstuhl Botanik II, Molekulare Pflanzenphysiologie, Universität Erlangen-Nürnberg, Staudtstrasse 5, 91058 Erlangen, Germany
    Search for more papers by this author
  • Rainer Hedrich

    Corresponding author
    1. Lehrstuhl für Molekulare Pflanzenphysiologie und Biophysik, Julius-von-Sachs-Institut für Biowissenschaften, Universität Würzburg, Julius-von-Sachs-Platz 2, 97082 Würzburg, Germany,
      For correspondence (fax +49 931 8886157; e-mail
    Search for more papers by this author

For correspondence (fax +49 931 8886157; e-mail

These authors contributed equally to this work.


K+ channels control K+ homeostasis and the membrane potential in the sieve element/companion cell complexes. K+ channels from Arabidopsis phloem cells expressing green fluorescent protein (GFP) under the control of the AtSUC2 promoter were analysed using the patch-clamp technique and quantitative RT-PCR. Single green fluorescent protoplasts were selected after being isolated enzymatically from vascular strands of rosette leaves. Companion cell protoplasts, which could be recognized by their nucleus, vacuole and chloroplasts, and by their expression of the phloem-specific marker genes SUC2 and AHA3, formed the basis for a cell-specific cDNA library and expressed sequence tag (EST) collection. Although we used primers for all members of the Shaker K+ channel family, we identified only AKT2, KAT1 and KCO6 transcripts. In addition, we also detected transcripts for AtPP2CA, a protein phosphatase, that interacts with AKT2/3. In line with the presence of the K+ channel transcripts, patch-clamp experiments identified distinct K+ channel types. Time-dependent inward rectifying K+ currents were activated upon hyperpolarization and were characterized by a pronounced Ca2+-sensitivity and inhibition by protons. Whole-cell inward currents were carried by single K+-selective channels with a unitary conductance of approximately 4 pS. Outward rectifying K+ channels (approximately 19 pS), with sigmoidal activation kinetics, were elicited upon depolarization. These two dominant phloem K+ channel types provide a versatile mechanism to mediate K+ fluxes required for phloem action and potassium cycling.


In higher plants, photosynthates formed in source leaves are translocated to sink tissues via the sieve element/companion cell (SE/CC) complexes. Functional SEs are interconnected by plasmodesmata, which provide a low-resistance pathway for solute transport. Functioning of the enucleate SE is sustained by metabolically active CCs connected to SE by branched plasmodesmata (van Bel and Kempers, 1997; Oparka and Turgeon, 1999). Symplasts of SE/CC complexes in Arabidopsis are isolated from adjacent cells (van Bel and van Rijen, 1994) so that loading and unloading of the phloem require an apoplastic step controlled by the plasma membrane transporters of the phloem cells.

Because of high K+ concentration in the phloem sap and the stimulation of sugar translocation by K+ (reviewed by Marschner et al., 1997; Pate and Jeschke, 1995), it has been proposed that cycling of K+ and other mineral nutrients between source and sink tissues is required to maintain both the membrane potential of SE/CC and the osmotic potential of the phloem (van Bel and Kempers, 1997; Marschner et al., 1996, 1997, and references therein). From changes in the phloem electrical potential in response to different apoplastic K+ concentrations, potassium channels were supposed to provide a major pathway for K+ entry and release in the phloem (Ache et al., 2001). Among the five plant Shaker K+ channel subfamilies, members of the AKT2 subfamily were found to be expressed in phloem tissues of Arabidopsis thaliana (Lacombe et al., 2000; Marten et al., 1999), Zea mays (Bauer et al., 2000; Philippar et al., 1999), Vicia faba (Ache et al., 2001) and Populus (Langer et al., 2002). The AKT2 gene is under the control of light and sugars in line with a role in regulating the entry of photosynthates into the phloem (Ache et al., 2001; Deeken et al., 2000, 2002). Deeken et al. (2002) recently demonstrated that the control of the SE/CC membrane potential and sugar loading was impaired in AKT2-deficient plants and that the sucrose content of their sieve tubes was only 50% of that in wild-type plants.

Heterologously expressed AKT2/3 channels are blocked by protons and Ca2+ (Lacombe et al., 2000; Marten et al., 1999) and display two distinct gating modes characterized by time-dependent and instantaneous current components (Dreyer et al., 2001). In a yeast two-hybrid screen, the protein phosphatase AtPP2CA was identified as an interacting partner to AKT2/3 (Vranováet al., 2001). Co-expression of AtPP2CA with AKT2 in animal cells increased inward rectification of the channel (Cherel et al., 2002). In addition to AKT2/3, KAT2 channels have also been found in the phloem of A. thaliana Columbia ecotype (Pilot et al., 2001). When expressed in Xenopus oocytes, KAT2 mediates inward rectifying currents with a single channel conductance of approximately 7 pS. In contrast to AKT2/3, KAT2 homomers are proton-activated and Ca2+-insensitive (Pilot et al., 2001; Lacombe, personal communication).

Previous studies showed that Arabidopsis lines expressing GFP can be used to study the electrical properties of certain cell types of the root (Kiegle et al., 2000; Maathuis et al., 1998). In our investigation into the nature and properties of phloem-expressed K+ channels, we applied RT-PCR and patch-clamp techniques to phloem protoplasts that had been isolated from transgenic Arabidopsis plants expressing GFP under control of the companion cell-specific AtSUC2 promoter (Imlau et al., 1999). Individual GFP-expressing protoplasts were selected to analyse their K+ channel transcript composition and K+-dependent electrical properties after being isolated enzymatically from vascular strands. We identified members of the A. thaliana K+ channel superfamily capable of controlling phloem potassium uptake, release and cycling. In addition to ion channel genes, we also discuss the unique companion cell expression profile (EST collection) with respect to SE/CC autonomy, stress tolerance and hormone action.


Isolation of phloem- and mesophyll-specific cDNAs

The successful application of the laser capture microdissection technique to plant cells has been reported recently (Asano et al., 2002; Kerk et al., 2003; Nakazono et al., 2003). Differentially expressed genes were identified in vascular tissues from maize and rice phloem tissues composed of functionally different cell types. These microarray studies and EST collections, however, lacked expression profiles for the companion cell-specific proton ATPase and sucrose/proton symporter, as well as for the phloem-localized potassium channels. To bridge this gap, we applied the laser microdissection and pressure catapulting (LMPC) technique to the vascular-rich flower stalk of Arabidopsis. Following the excision of about 150 individual phloem sectors (Figure 1a), mRNA was isolated and probed for the presence of transcripts for a phloem K+ channel, sucrose carrier and H+ pump. Using RT-PCR with primers specific for the well-known phloem transporters, we identified three phloem-specific expressed genes SUC2 (sucrose carrier), AKT2 (K+ channel) and AHA3 (H+ pump; Figure 1b; Imlau et al., 1999; Marten et al., 1999; Truernit and Sauer, 1995). These results were in line with the ones of Doering-Saad et al. (2002), who reported on mRNA in barley phloem sap. In addition to these transcripts, we also detected expression of SUC3, a gene encoding another sucrose transporter (Figure 1b), which has been previously found in the phloem periphery (Meyer et al., 2000). The presence of SUC3 suggested that the mRNA originated from different phloem cell types such as sieve elements, companion cells and parenchyma cells. Because companion cell transcripts could not be isolated by LMPC without contamination by mRNA from other cell types, we then enzymatically isolated companion cells expressing GFP under the control of the AtSUC2 promoter. Vascular strands were excised from major veins of fully developed rosette leaves from transgenic A. thaliana AtSUC2 promoter-GFP plants (Figure 2a). Release of protoplasts from phloem tissue was observed 1 h after enzyme application (Figure 2b). With blue light excitation, 20–30% of the protoplast population showed green fluorescence (Figure 2c,d). One hundred and forty-five individual fluorescent protoplasts containing chloroplasts and/or vacuole(s) were collected under an epifluorescence microscope and, after washing, were transferred to PCR lysis buffer using glass micropipettes. As a control, we collected 150 non-fluorescing mesophyll protoplasts. Real-time RT-PCR was used to quantify marker gene transcripts in the individual cDNAs in order to both characterize the protoplast type and to test for contamination of phloem protoplasts by mesophyll cells (and vice versa). Transcripts of the sucrose transporter SUC2 (Truernit and Sauer, 1995) and the proton ATPase AHA3 (De Witt and Sussman, 1995) were detected in the phloem fraction only, whereas SUC3 transcripts were found in both phloem and mesophyll protoplasts (cf. Figures 1b and 3a). Quantification, however, revealed that SUC3 expression was most prominent in the mesophyll cells, while only background levels (7%) were found in the phloem (Figure 3b, cf. Meyer et al., 2000). From the distribution of these marker gene transcripts, we concluded that the preparations of phloem and mesophyll cells were not cross-contaminated.

Figure 1.

RT-PCR analysis of phloem samples obtained by LMPC technique.

(a) Arabidopsis phloem (left) was excised from inflorescence stalk cross-sections by a laser beam (middle) and catapulted into a reaction cap (right) for further analysis.

(b) RT-PCR products amplified from RNA of LMPC-excised phloem segments. Identification of SUC2, AKT2 and AHA3 transcripts in phloem samples only. M = λPstI marker, C = reference transcript.

Figure 2.

Isolation of phloem protoplasts.

(a) GFP fluorescence in the vascular tissue of a rosette leaf from AtSUC2 promoter-GFP plants.

(b) Release of fluorescent protoplasts from phloem tissue during enzymatic digestion.

(c, d) Protoplasts derived from vascular strands. Bright field image (c) and epifluorescence image (d).

Figure 3.

Marker transcripts of phloem and mesophyll cells.

(a) RT-PCR products of mesophyll cell (MC) and companion cell (CC) protoplasts separated in an 1% tris-borate-ethylenediaminetetraacetic acid (TBE)–agarose gel.

(b) Relative transcript rates calculated from quantitative real-time PCR.

Note the RT-PCR product of SUC3 in the phloem shown in (a) represents 7% of the expression in the mesophyll after quantification by real-time PCR.

Companion cell cDNA library and EST collection

We used the mesophyll-free companion cell mRNA to generate a cDNA library and partially sequenced 2000 individual clones. About 56% of the Arabidopsis gene sequences were identified and they formed the foundation for a steadily increasing EST collection. Within this group, 33% encoded unknown proteins while others encoded previously described phloem-expressed genes (Nakazono et al., 2003), as well as a unique selection of genes most likely required for sieve tube function and survival, hormone action and pathogen defence. Singlets as well as contigs of up to 40 identical sequences were present within the latter fraction (Table 1). Putative functions were assigned to the cDNAs when predictions and scores were identical in all three data bases used for analysis: blastx against Swissprot plant proteins, blastn against Arabidopsis coding sequences (−introns, −UTRs) and blastn against Arabidopsis genes (+introns, +UTRs). Finally, genes were subgrouped into 10 functional clusters, which were then related to the number of identified cDNAs (Figure 4): redox regulation (R = 19.4%), stress (S = 11.0%), defence (D = 2.3%), metabolism (M = 10.1%), transcription and translation (TT = 10.1%), hormones and signalling (HS = 9.1%), transport and membranes (TM = 1.9%), cell wall (CW = 0.9%), photosynthesis (PS = 0.9%) and cytoskeleton (CS = 0.5%). Among the 563 genes analysed in detail, 454 were singlets, 59 were doublets, 17 were triplets, 22 genes appeared 4–9 times and 11 genes appeared 10–40 times (Table 1). Six types of sequences were the most abundant among the cDNAs analysed. They showed homology with metallothionin 2b (40×), a water stress-induced protein (26×), thioredoxin h (18×), a translation initiation factor (18×), dihydrofolate reductase (17×), 12-oxophytodienoate reductase (OPR1; 12×), a low temperature and salt-responsive protein (11×) and the heat shock protein 17 (10×).

Table 1.  Representative genes selected from an EST collection of Arabidopsis companion cells grouped into 10 functional clusters
Singlets and contigsESTsPutative gene identificationReferences
  1. EST numbers indicate identical cDNA clones within a contig. References confirming the phloem specificity of the genes listed in the table: (a), Asano et al. (2002); (b), Nakazono et al. (2003); (c), Walz et al. (2002); (d), Hoffmann-Benning et al. (2002); (e), Husebye et al. (2002); (f), Chen et al. (2001); (g), Sasaki et al. (2001); (h), Ruiz-Medrano et al. (1999).

 ARAB.90.C12Actin de-polymerizing factor 6(a)
 A011-a12.TEx5_085.ab11Tubulin beta-2/beta-3 chain 
 A012-e09.TEx5_067.ab11Microtubule-associated protein 
 A016-c10.TEx5_071.ab11Dynein light subunit lc6 
Cell wall
 A010-b01.T3_009.ab12Proline-rich protein M14 
 ARAB.57.C12Pollen coat protein 
 ARAB.61.C11Harpin-induced protein 
 A014-d09.TEx5_074.ab11Bax inhibitor-1 
 ARAB.107.C13Myrosinase(e), (f)
 A015-g08.TEx5_056.ab12Lectin PP2(c)
 ARAB.102.C11Component of aniline 
 A012-h01.TEx5_012.ab12Disease-resistance protein(a)
 A001-f01.T3_011.ab12Cysteine proteinase inhibitor B(c), (b)
Hormones and signalling
 A007-d11.TEx5_091.ab11ACC synthase (AtACS-6) 
 ARAB.22.C13ACC oxidase 
 ARAB.30.C17Gibberellin-responsive protein 
 A019-f07.TEx5_059.ab11Ethylene-type zinc finger protein 
 ARAB.40.C15Ethylene-responsive transcriptional co-activator 
 ARAB.31.C12Ethylene-responsive protein 
 ARAB.47.C11212-oxophytodienoate reductase (OPR1) 
 ARAB.39.C14Ripening-related protein(c)
 ARAB.77.C12Calcium-binding protein 
 A006-c03.T3_018.ab11S-adenosylmethionine decarboxylase 
 A003-a04.T3_021.ab11S-adenosylmethionine synthase 2 
 A014-d01.TEx5–010.ab11Molybdopterin synthase (CNX2) 
 ARAB.93.C12Subtilisin-like serine protease 
 A017-c07.TEx5_051.ab11Two-component phosphorelay mediator 
 A016-h07.TEx5_061.ab11Mitogen-activated protein kinase 
 A004-b11.T3_089.ab11Receptor Ser/Thr protein kinase(b)
 ARAB.50.C12Protein phosphatase 2A 
 A017-h05.TEx5_044.ab11Protein phosphatase 2C 
 A013-a07.TEx5_049.ab11COP1-interacting protein CIP8 
 A013-b05.TEx5_041.ab11Symbiosis-related protein 
 ARAB.63.C111Low temperature and salt-responsive protein(a)
 ARAB.65.C14Cold and ABA inducible protein kin1(b)
 ARAB.75.C13Pheromone receptor 
 A007-f06.TEx5_048.ab11NAC-domain protein(h)
 ARAB.9.C117Dihydrofolate reductase 
 ARAB.83.C19Steroid 5-alpha reductase 
 A006-b01.T3_009.ab11Tropinone reductase 
 ARAB.18.C12Cytosolic triosephosphate isomerase 
 ARAB.105.C14Peptidylprolyl isomerase ROC1 
 A012-a12.TEx5_085.ab11Anthranilate synthase, alpha subunit 
 ARAB.74.C12Sucrose-UDP glucosyltransferase 
 A005-c07.T3_050.ab11Aspartate aminotransferase (Asp3) 
 A019-c05.TEx5_034.ab11Glycosyl transferase 
 A006-c06.T3_038.ab11Steroid-binding protein 
 A018-a07.TEx5_049.ab11Steroid sulfotransferase 
 A008-g08.TEx5_065.ab12Glyceraldehyde-3-phosphate dehydrogenase(b)
 A002-f06.T3_047.ab12Mitochondrial proline oxidase 
 A015-h11.TEx5_092.ab12Ubiquinol cytochrome-c(b)
 A002-h01.T3_012.ab11Anthocyanidin synthase 
 A013-g11.TEx5_084.ab11Mitochondrial ATP synthase delta chain 
 ARAB.52.C12Epoxide hydrolase (ATsEH) 
 A002-d12.T3_094.ab11Flavanone 3-hydroxylase 
 A007-e04.TEx5_023.ab11Gamma glutamyl hydrolase 
 A007-f03.TEx5_027.ab11Pyruvate kinase(b)
 A017-h09.TEx5_077.ab11Choline kinase GmCK2p 
 A012-d03.TEx5_026.ab11Inorganic pyrophosphatase 
 A019-g04.TEx5_024.ab11Nicotianamine synthase 
 A003-f02.T3_015.ab11ADP-ribosylation factor 
 A015-d12.TEx5_094.ab12Acyl CoA-binding protein(b)
 A012-d07.TEx5_058.ab11Glutamate-/aspartate-binding peptide 
 ARAB.37.C12Photosystem II reaction center (6.1 kDa) 
 A017-c05.TEx5_034.ab11PSI type III chlorophyll a/b-binding protein 
 A018-h06.TEx5_048.ab11Protochlorophyllide reductase 
 ARAB.97.C12Ribulose bisphosphate carboxylase, SSU 
Redox regulation
 ARAB.1.C140Metallothionein 2b(b)
 ARAB.60.C17Glutathione-S-transferase (GST6)(b)
 A018-h05.TEx5_044.ab15Quinone oxidoreductase 
 ARAB.5.C14Blue copper-binding protein 
 A019-h09.TEx5_076.ab11Superoxidase dismutase(c), (d)
 A011-h07.TEx5_060.ab11Copper/zinc superoxide dismutase(c)
 A003-g02.T3_008.ab11Cytochrome P450 monooxygenase 
 A012-h10.TEx5_080.ab11NADH dehydrogenase 
 A019-g12.TEx5_088.ab11Delta 9 desaturase 
 ARAB.28.C126Water stress-induced protein(b)
 ARAB.43.C15Dehydrin Xero2 
 ARAB.27.C110Heat shock protein 17 
 ARAB.27.C23Heat shock protein 18(b)
 A014-h05.TEx5_044.ab12Heat shock protein DnaJ 
 A015-a08.TEx5_053.ab12Heat shock protein 70 
 A017-b12.TEx5_094.ab11Heat shock protein 81-2 
 A003-g07.T3_052.ab17Small heat shock protein 
 ARAB.99.C13Cytosolic cyclophilin (ROC3) 
 A014-g07.TEx5_052.ab11GTP-binding protein GB2 
 ARAB.23.C12Stress-induced protein 
Transport and membranes
 A004-a10.T3_069.ab12Sugar transporter 
 A011-f10.TEx5_079.ab11Amino acid transport protein AAP2(a)
 A011-e09.TEx5_067.ab11Putative AAA-type ATPase 
 A011-f01.TEx5_011.ab11Mitochondrial phosphate translocator(d)
 A018-a12.TEx5_085.ab11Metal ion transporter 
 A008-g11.TEx5_085.ab12Outer membrane lipoprotein 
 A006-f06.T3_047.ab11Coated vesicle membrane protein 
 A011-c07.TEx5_050.ab11Coatomer delta subunit 
 A012-f02.TEx5_015.ab11Geranylgeranylated protein 
Transcription and translation
 A002-c04.T3_022.ab11Transcription initiation factor TFIID-1(a)
 A003-c07.T3_050.ab12Transcription factor GT-3a 
 A008-b12.TEx5_094.ab11bZIP transcription factor 
 A018-e01.TEx5_003.ab11G-box binding bZIP transcription factor 
 A011-f07.TEx5_059.ab11Nucleic acid-binding protein 
 ARAB.110.C12RNA polymerase II(b)
 ARAB.103.C13Zinc finger protein (PMZ)(b)
 A011-a10.TEx5_069.ab11DNA helicase TPS1 
 A001-d01.T3_010.ab11Ribonuclease, RNS1 
 A019-e01.TEx5_003.ab11Histone H2B 
 A008-e03.TEx5_019.ab11Histone H2A 
 A012-f10.TEx5_079.ab11Histone H3.3 
 A013-a01.TEx5_001.ab11Histon H3 
 A011-e03.TEx5_019.ab11DNA damage-inducible protein 
 A008-a07.TEx5_050.ab11DNA-3-methyladenine glycosidase 
 A012-g12.TEx5_088.ab11Glutamyl-tRNA reductase 
 ARAB.10.C118Translation initiation factor(b)
 ARAB.45.C14Eukaryotic initiation factor 5A 
 A003-b05.T3_041.ab11Translation elongation factor eEF-1 
 A007-f02.TEx5_015.ab11Ribosomal protein S3a 
 A003-e06.T3_039.ab1130S ribosomal protein S20 
 A017-c02.TEx5_006.ab1160S ribosomal protein L36 
 A012-d12.TEx5_094.ab1140S ribosomal protein S20-like 
 A019-c06.TEx5_038.ab1150S ribosomal protein L33(b)
 ARAB.96.C1360S ribosomal protein L10 
 A010-g01.T3_004.ab1160S acidic ribosomal protein P2 
 A014-a01.TEx5_001.ab1160S acidic ribosomal protein P0 
 ARAB.80.C14Polyubiquitin (UBQ14)(c), (d)
 ARAB.67.C14Ubiquitin-conjugating enzyme E2(b)
 A002-b05.T3_041.ab11ATP-dependent protease, proteolytic Clp 
 A015-a01.TEx5_001.ab11PCI domain protein proteasome 
 A012-f06.TEx5_047.ab11Multicatalytic endopeptidase complex 
 A013-f03.TEx5_027.ab11Signal peptidase subunit 
Figure 4.

Relative distribution of functional gene clusters.

Functional gene clusters of companion cell ESTs grouped by their putative gene function: N, unknown function; R, redox regulation; S, stress; D, defence; M, metabolism; TT, transcription and translation; HS, hormones and signal transduction; TM, transport and membranes; CW, cell wall; PS, photosynthesis; CS, cytoskeleton.

In addition to the phloem markers SUC2, AKT2 and AHA3, which were identified by RT-PCR and did not yet appear in the EST collection, we also searched for rare transcripts of differentially expressed genes involved in signal transduction and allocation. These included auxin transporters, ethylene and brassinosteroid receptors and a putative lipid transfer protein defective in induced resistance (DIR1). In the phloem-free mesophyll mRNA fraction, we found ethylene insensitive4 (EIN4; Chang and Stadler, 2001, and references therein), brassinosteroid insensitive1 (BRI1), involved in the signal transfer of brassinosteroids (Wang et al., 2001) and PIN3, a component of the lateral auxin transport system regulating tropic growth (Friml et al., 2002). In contrast, DIR1, involved in systemic acquired resistance (Maldonado et al., 2002), and the polar auxin transport-related PIN6 were detected only in companion cells. The expression patterns of these two PIN genes were in line with the differential expression of PIN3 and PIN6 deduced from the analysis of Arabidopsis mutants defective in interfascicular fibre differentiation (Zhong and Ye, 2001).

K+ channel transcripts in companion cells

The companion cell EST collection did not contain ion channel sequences (Table 1), so we used quantitative RT-PCR to analyse the K+ channel transcript profile in two cDNA populations derived from non-cross-contaminated companion and mesophyll cell protoplast samples. Among the Shaker-like K+ channel transcripts, KAT1 and AKT2 dominated the companion cell fraction, ATKC1 appeared in mesophyll protoplasts and AKT1 transcripts were rare in both cell types of the Arabidopsis ecotype C24 (Figure 5a). It should be noted that in ecotype Col-0, KAT2 has also been identified as a phloem K+ channel by KAT2 promoter-GUS studies (Pilot et al., 2001). Further analyses are required to clarify whether it is KAT1 or KAT2, which shares 72% identical amino acids and exhibits identical electrical properties, that is expressed in the phloem of other Arabidopsis ecotypes. Transcripts of the protein phosphatase AtPP2CA, which has been shown to interact with the AKT2/3 channel (Cherel et al., 2002; Vranováet al., 2001), were detected in companion cells and also in non-AKT2 expressing cells such as mesophyll, hypocotyl cortex, root hairs and A. thaliana tumours (our unpublished data). The expression of KCO1, with small amounts of KCO5 and KCO6, in mesophyll protoplasts was in line with the results of Schönknecht et al. (2002) for the dominant members of the KCO family. KCO6, the only member of this channel family in the phloem, showed the highest KCO expression level so far measured in any tissue type (Figure 5a and our unpublished data). If the actin-based transcript abundance of AKT2, KCO6 and KAT1 are compared for rosette leaves, phloem-rich flower stalks and companion cell protoplasts, then AKT2 and KAT1 transcripts increased with the number of companion cells in a given fraction (Figure 5b). KCO6 mRNA was also most abundant in the companion protoplast fraction, but was lower in the stalks than in the leaves.

Figure 5.

Phloem- and mesophyll-specific ion channel gene profile.

(a) Relative transcript number of Shaker and KCO channels, and phosphatase PP2CA quantified by real-time-PCR on protoplast cDNA. GORK and SKOR transcripts were not detectable. One representative of three separate experiments is shown.

(b) Expression level of phloem channels quantified by real-time PCR obtained from RNA of whole leaves, phloem-rich inflorescence stalks and isolated companion cell protoplasts (CC).

Electrical properties of K+ channels in the phloem

We characterized the electrical properties of K+ channels by performing patch-clamp measurements on KAT1-, AKT2- and KCO6-expressing companion cell protoplasts in the whole-cell and outside-out mode. Among the companion cell protoplasts, there were two major populations: some cells dominated by inward currents and others dominated by outward currents. When, in the whole cell configuration, protoplasts were clamped at −48 mV, with 150 mm K+ in the pipette and 30 mm K+ in the bath, hyperpolarizing voltages, negative to −108 mV, elicited slowly activating inward currents (n = 11; Figure 6a). Tail K+ currents reversed direction around the Nernst equilibrium potential for potassium (EK = 41 mV; Figure 6b). Time-dependent single channel fluctuations were observed at hyperpolarizing voltages in outside-out patches excised from protoplasts dominated by inward rectifying currents (Figure 6c). With 150 mm K+ in the pipette and 30 mm K+ in the bath, inward channels were characterized by a unitary conductance of about 4 pS. Under these conditions, the single channel current reversed direction at −40 mV, close to the EK = −41 mV (Figure 6d). When protoplasts were exposed to increasing Ca2+/K+ ratios in the bath, a pronounced voltage-dependent block of the inward rectifier was observed (Figure 6e). Ca2+-dependent decrease in K+ current amplitudes at hyperpolarizing voltages was accompanied by a shift in voltage dependence of the inward rectifier towards more positive potentials. Voltage-dependent Ca2+ block increased when K+ concentration was lowered from 30 to 10 mm (Figure 6e). A similar block of inward K+ channels has previously been described for Z. mays, V. faba, Solanum tuberosum, Nicotiana tabacum and A. thaliana guard cells (Dietrich et al., 1998; Fairley-Grenot and Assmann, 1992). However, when compared to inward rectifiers from root hairs and guard cells, phloem K+ channels exhibited an opposite pH-sensitivity. A change in the pH of the external solution from 7.0 to 5.6 shifted the voltage dependence of the phloem inward rectifier towards more negative potentials (Figure 6f). A decrease in external pH caused a −20 ± 8 mV shift in half activation potential (V1/2) of the Boltzmann curve. Under these conditions, the voltage dependence of the guard cell inward rectifier shifts towards more positive potentials (Brüggemann et al., 1999). Among the plant Shaker K+ channels so far identified, only AKT2/3-type channels from Arabidopsis, maize and poplar were blocked by protons (Bauer et al., 2000; Lacombe et al., 2000; Langer et al., 2002; Marten et al., 1999). It should be noted that proton-blocked inward rectifiers have also been recorded in Samanea saman pulvinus protoplasts, which also express AKT2/3-like channels (Moshelion et al., 2002; Yu et al., 2001). The companion cell inward rectifier thus seems to share properties with AKT2/3 (H+ and Ca2+ inhibition) and KAT1 (strong inward rectification).

Figure 6.

Electrical properties of the phloem K+ inward rectifier.

(a) Time- and voltage-dependent whole-cell inward K+ currents in fluorescent protoplasts. Voltage pulses were applied from a holding potential of −48 mV in 20-mV steps in the range from +52 to −188 mV, with subsequent pulse to −88 mV.

(b) De-activation (tail currents) in response to a double-pulse protocol starting from a holding potential of −48 mV to a pre-pulse voltage of −168 mV and followed by voltage steps from +12 to −148 mV. Pipette and bath solutions contained 150 and 30 mm K-gluconate, respectively.

(c) Single channel fluctuations induced by hyperpolarizing voltages in an outside-out membrane patch excised from a protoplast with a macroscopic inward current. O, open state of the channel; C, closed state of the channel.

(d) Single channel amplitudes plotted versus the membrane voltage. The data represent mean channel amplitudes ± SD from three patches.

(e) Voltage-dependent Ca2+-block of the phloem inward rectifier. Voltage dependence of inward K+ currents at different Ca2+ to K+ ratio. Current amplitudes were sampled at the end of 1-sec pulses to voltages in the range from −8 to −188 mV. External solutions contained: 20 mm CaCl2 + 10 mm K-gluconate (▪), 20 mm CaCl2 + 30 mm K-gluconate (○), 1 mm CaCl2 + 30 mm K-gluconate (●) and 0.1 mm CaCl2 + 30 mm K-gluconate (▿).

(f) pH-dependent shift in voltage activation of inward K+ channels. Relative conductance at different external pH values was fitted by Boltzmann function. External solutions contained 1 mm CaCl2, 30 mm K-gluconate and 10 mm Mes/Tris (pH 5.6) or 10 mm Hepes/Tris (pH 7.0). Data points represent means ± SE for three protoplasts.

When depolarizing voltages were applied to companion cell protoplasts in the whole cell configuration, an activation of time-dependent outward currents positive to −28 mV was observed (n = 10; Figure 7a). Tail K+ currents reversed direction around EK (−41 mV; Figure 7b). Single channel fluctuations with a unitary conductance of approximately 19 pS were recorded in cell-free outside-out patches excised from protoplasts with prominent time-dependent outward currents (Figure 7c). The single channel currents reversed direction around −40 mV, close to EK (Figure 7d). Time- and voltage-dependent parameters, as well as the unitary conductance of the phloem K+ outward rectifier (Figure 7a–e), were reminiscent of guard cell outward rectifying K+ channel (GORK) expressed in Xenopus oocytes, Arabidopsis guard cells and root hairs (Ache et al., 2000; Ivashikina et al., 2001). In contrast to the latter Arabidopsis cell type, phloem outward K+ channels did not inactivate in response to prolonged (10 sec) de-polarization (Figure 7f). Inactivation of K+ outward rectifier has previously been described in Arabidopsis guard cells (Pei et al., 1998) and root hairs, both being GORK-expressing cell types (Ache et al., 2000; Ivashikina et al., 2001). RT-PCR analysis of channel transcripts (Figure 5a), however, showed that neither GORK nor SKOR was expressed in GFP-tagged protoplasts. We may thus assume that the phloem outward rectifier represents either the product of the KCO6 gene, or another yet non-identified K+ channel involved in the repolarization of the phloem potential.

Figure 7.

Electrical properties of the phloem K+ outward rectifier.

(a) Time- and voltage-dependent whole-cell outward K+ currents in fluorescent protoplasts. Voltage pulses were applied from a holding potential of −48 mV in 20-mV steps in the range from −108 to +52 mV, with subsequent step to −88 mV.

(b) De-activation currents in response to a double-pulse protocol starting from a holding potential of −48 mV to a prepulse voltage of +52 mV and followed by voltage steps from +32 to −88 mV.

(c) Single channel fluctuations induced by de-polarization in an outside-out patch excised from a protoplast with a macroscopic outward current. O, open state of the channel; C, closed state of the channel.

(d) Single channel amplitudes plotted versus the membrane voltage. The data represent mean channel amplitudes ± SD from five patches.

(e) Voltage dependence of outward current sampled at the end of 1-sec pulses in the range from −108 to + 52mV and normalized in respect to +32 mV. Data points represent means ± SE for three protoplasts. Iss, the (quasi) steady-state current.

(f) Comparison of time-dependent outward K+ currents measured in protoplasts isolated from companion cell and root hair protoplasts in response to a 10-sec voltage pulse from −48 to +52 mV.


Molecular mechanism of phloem K+ loading and release

In our search for companion cell channels involved in K+ retrieval and release in phloem cells, we focused on the nine Shaker-like and six KCO channels encoded by the Arabidopsis genome (reviewed by Mäser et al., 2001). A remarkable feature of Shaker channels is their ability to promiscuously form heterotetramers with different α-subunits (Baizabal-Aguirre et al., 1999; Dreyer et al., 1997; Paganetto et al., 2001; Pilot et al., 2001; Reintanz et al., 2002). Inward rectifying channels result from heteromers between KAT1 and AKT1, KAT1 and AtKC1 (Dreyer et al., 1997), KAT1 and AKT2/3 (Baizabal-Aguirre et al., 1999), as well as KAT1 and KAT2 (Pilot et al., 2001), but not between KAT1 and the outward rectifier GORK (Ache et al., 2000). Formation of heteromultimeric structures has been shown to modify the sensitivity of K+ channels to voltage, Ca2+ and pH (Dreyer et al., 1997; Paganetto et al., 2001; Reintanz et al., 2002). Numerous K+ channel subunits, KAT1, KAT2, AKT1, AtKC1, GORK and AKT2/3 are expressed in Arabidopsis guard cells (Szyroki et al., 2001), while root hairs express just three Shaker K+ channel α-subunits, AKT1, AtKC1 and GORK (Reintanz et al., 2002). We also found different K+ channel transcripts in companion cell protoplasts (Figure 5a). Among these channels, only AKT2/3 was characterized by a pronounced Ca2+ sensitivity and block by protons (Hoth et al., 2001; Lacombe et al., 2000; Marten et al., 1999), pointing to a contribution by the AKT2/3 subunit to the Ca2+-sensitive K+ conductance of the phloem. These properties, however, require further testing using direct genetic analyses such as loss-of-function mutants or RNAi lines.

The Arabidopsis Shaker superfamily contains two outward rectifying K+ channels: SKOR, localized in the root pericycle and stelar parenchyma cells (Gaymard et al., 1998), and GORK, expressed in guard cells (Ache et al., 2000) and the root epidermis (Ivashikina et al., 2001). In this paper, we suggest that the phloem outward rectifier could be the product of the KCO6 gene (Figure 5a). The latter is also expressed in mesophyll and hypocotyl cortex cells, which lack GORK and SKOR transcripts, but contain a non-inactivating delayed outward rectifier (our unpublished data). The phloem outward rectifier did not undergo time-dependent inactivation (Figure 7f). The identification of the companion cell outward rectifier awaits the availability of GORK and KCO6 loss-of-function plants expressing GFP under the control of the SUC2 promoter.

Taken together, the inward and outward K+ rectifiers characterized in this study may provide for a mechanism to control K+ cycling, the membrane potential and, consequently, H+-driven assimilate translocation in the phloem. Assuming that K+ concentration in the phloem varies in the range of 50–150 mm (Marschner et al., 1996) and apoplastic K+ from 1 to 10 mm, inward K+ channels can activate negative to −40 mV and outward K+ channels positive to −120 mV. Both channels can therefore operate in the voltage range recorded for SE/CC (between −100 and −185 mV; van Bel and van Rijen, 1994; Deeken et al., 2002). Future electrophysiological and molecular studies will focus on the nature and regulation of phloem-localized Ca2+ and Cl channels, as well as on electrogenic carriers and pumps, in order to gain insights into the formation of complex electrical signals travelling along the phloem.

Gene expression of companion cell protoplasts

Membrane transport

In addition to the well-known phloem transporters SUC2, AKT2 and AHA3, two channels, KAT1 and KCO6, were found differentially expressed in companion cells, while ATKC1 and SUC3 appeared predominately in the mesophyll cell fraction. We studied the expression of genes encoding plant signalling components in order to increase the number of potential companion cell and mesophyll markers. In addition to quantitative RT-PCR analyses with transporter-specific primers, we, so far, found in the companion cell EST collection, a sugar transporter, a metal ion transporter and components of the membrane sorting/trafficking system (Table 1, ‘transport and membranes’), as well as the previously identified phloem-localized amino acid transporter AAP2 (Okumoto et al., 2002).

Companion cell identity

In good agreement with the organelle composition of companion cells, the EST collection contained chloroplastic (Table 1, ‘photosynthesis’) and mitochondrial genes, as well as a relative large number of nuclear and ribosomal genes involved in transcription and translation, together with genes required for protein folding (HSPs) and protein degradation (ubiquitin). There was even a translation initiation factor (contig of 18) among the few high-copy genes (Table 1). This profile further underlines the role of the companion cell as the ‘work horse’ of the SE/CC complex. Proteins residing in the nucleus- and ribosome-free sieve tubes are produced in the companion cells. The characteristic tubular structures found in the phloem sap of sieve tubes and based on actin-, tubulin-, dynein- and microtubule-associated proteins (Schobert et al., 1998, 2000) might be gene products of cytoskeleton genes expressed in companion cells (Table 1, ‘cytoskeleton’).

Signal transduction

Among the auxin transporter transcripts tested by RT-PCR analyses (data not shown), we found PIN6 in companion cells and PIN3 in mesophyll cells. Transcripts of the ‘so-called’ auxin-binding protein ABP1 were found in both mRNA pools. This differential expression points to the phloem as a bi-directional, long-distance pathway for auxin transport. Future studies on the nitrilase found in the EST collection and other potential auxin synthesis genes are still needed to confirm that companion cells are also sites of auxin production. In this respect, it was somewhat unexpected that genes of the M cluster were dominated by those with a known function in secondary metabolism. Among them, the EST collection harbours genes involved in ethylene, jasmonate, ABA, gibberellin and steroid (possibly brassinosteroid) synthesis, perception, transduction and response (for respective receptor kinases, calcium-binding proteins, protein kinases and phosphatases and MAP kinase; see Table 1). Future studies on candidate genes will help to link hormone and phloem action.


Taken together, the genes encoding R, S and D comprised up to 33% of identified genes (Figure 4). This raises the question as to whether this expression pattern reflects stress imposed by protoplast isolation (including loss of turgor), exposure to fungal cell wall-degrading enzymes (and thus release of cell wall oligosaccharide with potential elicitor-like function) or the companion cell biology? Among the defence genes, we identified a phloem-localized myrosinase, which, together with phloem-mobile glycosinolates, has been described before and linked to pathogen defence (Chen et al., 2001; Husebye et al., 2002), possibly directed against phloem feeding insects. Furthermore, the lectin PP2 and cystatin were identified as major components of phloem exudates for all species analysed so far (Schobert et al., 1998; Walz et al., 2002). The stress gene fraction was dominated by water stress-induced proteins (cf. kin1, ‘hormones and signalling’ in Table 1) and heat shock proteins. Future loss-of-function studies will clarify whether these genes are required for protein folding and shuttling into the sieve elements and therefore survival of the nucleus- and ribosome-free sieve tubes. It should be noted that the large number of HSP transcripts correlated with the large abundance of members of the TT gene cluster (Table 1). Similarly, some of the defence gene members (e.g. lectin PP2, see ‘defence’ in Table 1) and of the R cluster have also been identified as a major protein fraction of the phloem sap (Walz et al., 2002, and references cited). Among them, metallothionein, thioredoxin, glutaredoxin and glutathione-S-transferase appeared in contigs with up to 40 copies. Future studies based on transgenic plants expressing promoter–reporter gene constructs will have to clarify whether a ‘stress’ gene is constitutively expressed in companion cells or induced upon interaction with pathogens/symbionts, meristem development or fruit ripening (see, e.g. nodulin ENOD 40, symbiosis-related protein, ripening-related protein and NAC-domain protein under ‘hormones and signalling’ in Table 1).


We are currently generating a saturating EST collection with the present founder ESTs as a base. Together with genome array data, future studies will take advantage of a substantial phloem marker pool. Intact phloem samples gained using LMPC (Figure 1) will allow companion cell-specific genes to be distinguished within the phloem-specific ones. Ongoing bioinformatic analyses of the respective genes will provide the backbone for genome-wide predictions about proteins involved in phloem action, and improve our understanding of sink–source regulation and control of flowering and ripening.

Experimental procedures

Plant material

Transgenic A. thaliana AtSUC2 promoter-GFP plants were grown in soil in a growth chamber with a 8-h day/16-h night regime, 21°C day/16°C night temperature and a photon flux density of 120 µmol m−2 sec−1.

Laser microdissection and pressure catapulting

Arabidopsis inflorescence stalks were cut into 5–15-mm pieces and fixed for 4 h in 3 : 1 ethanol:acetic acid, and subsequently dehydrated and embedded in Paraplast plus (Sigma, Steinheim, Germany) according to Kerk et al. (2003). Cross-sections of 10–15 µm were cut on a rotary microtome (RM2165, Leica, Bensheim, Germany), floated in water on membrane-coated glass slides (PALM, Bernried, Germany) at 42°C and air-dried. Slides were de-paraffinized two times in xylene for 5 min each and air-dried. LMPC of phloem was carried out using the PALM Laser-MicroBeam System (PALM, Bernried, Germany). One hundred and fifty phloem regions were collected in 10-µl DEPC-treated water containing 40 units RNase inhibitor (MBI, St Leon-Rot, Germany).

Protoplast isolation

Vascular strands were excised from fully developed rosette leaves and incubated for 1.5 h at 30°C in enzyme solution containing 0.8% (w/v) cellulase (Onozuka R-10, Yakult Itorisha, Tokyo, Japan), 0.1% pectolyase (Sigma), 0.5% bovine serum albumin, 0.5% polyvinylpyrrolidone, 1 mm CaCl2 and 10 mm Mes/Tris (pH 5.6). The osmolarity of the enzyme solution was adjusted to 630 mosmol kg−1 with d-sorbitol. Protoplasts released from vascular-enriched tissues were filtered through a 20-µm nylon mesh and washed two times in 1 mm CaCl2 buffer (osmolarity 580 mosmol kg−1 (pH 5.6)). For isolation of mesophyll protoplasts, the osmolarity of all solutions was adjusted to 400 mosmol kg−1 and protoplasts were filtered through a 100-µm nylon mesh. The protoplast suspension was stored on ice, and aliquots were used for patch-clamp measurements or separation of single protoplasts to generate cDNA libraries.

Collection of contamination-free protoplasts

Individual phloem and mesophyll protoplasts were collected under an epifluorescence inverted microscope (Axiovert 35 M Carl Zeiss, Oberkochen, Germany) from 3-cm plastic dishes containing leaf or stem protoplast suspension. Fluorescing protoplasts were visualized by short-wave blue light. Protoplasts were transferred by microcapillaries with a tip opening of approximately 50 µm (CC protoplasts) and 200 µm (mesophyll) using a computer-controlled micropump (dispenser/diluter, Microlab-M; Hamilton, Darmstadt, Germany) as described by Koop and Schweiger (1985) and Kranz (1999). For selection and washing, protoplasts were transferred into 2000-nl microdroplets of washing solution (1 mm CaCl2 and 10 mm Mes/Tris (pH 5.6), osmolarity 580 mosmol kg−1), covered by mineral oil. After washing, 145 phloem protoplasts and 150 mesophyll protoplasts were transferred with a microcapillary into a 0.5-ml reaction tube.

As an alternative less time-consuming approach for isolating mesophyll protoplasts, we examined the protocol of Cherel et al. (2002), which was used by the authors to ‘semiquantify’ the level of expression of AKT2, KAT1, KAT2 and AtPP2CA in mesophyll cells. In contrast to the approach by Cherel et al. (2002), we used AtSUC2 promoter-GFP plants to verify the purity of mesophyll protoplasts. As a result, we found the mesophyll preparation contaminated by green fluorescent protoplasts, and detected transcripts of the phloem-specific marker genes SUC2 and AHA3 by RT-PCR (not shown).

RNA extraction, cDNA synthesis and sequencing

RNA from LMPC samples was extracted using the Gentra-Purescript-RNA-Isolation-Kit (Biozym, Hess. Oldendorf, Germany). Poly(A) RNAs from mesophyll and CC protoplasts were isolated and purified twice with the Dynabeads mRNA Direct kit (Dynal, Oslo) to prevent contamination with genomic DNA. The SMART cDNA Library Construction Kit (BD Biosciences Clontech, Heidelberg, Germany), which is designed for limited amounts of mRNA and includes a PCR-based protocol, was used for cDNA synthesis and amplification. The resulting λTriplEx2 library was converted to a plasmid library for preparation of the ESTs. Plasmid DNA was subjected to a standard sequencing procedure using the λTriplEx 5′ LD-Insert Screening Amplimer (5′-CTCGGGAAGCGCGCCATTGTGTTGGT-3′), the Epicentre-SequiTherm-EXCEL II Kit (Biozym, Oldendorf, Germany) and the Li-Cor-dna-Analyzer-Gene-Readir 4200 Sequencer (Li-Cor, Bad Homburg, Germany).

Sequencing of 2000 individual ESTs was performed in cooperation with Syngenta Biotechnologies Inc. (Research Triangle Park, NC, USA) using an ABI PRISM® 3700 DNA Analyzer from Applied Biosystems, Foster City, CA, USA. Alignment and blast procedures for the sequenced ESTs were also performed by Syngenta following standard algorithms. All kits were used according to the manufacturers' protocols.

RT-PCR experiments

First-strand cDNA was prepared with RNA of LMPC and protoplasts by using Superscript RT (Gibco/BRL, Karlsruhe, Germany). Qualitative PCR was carried out using 1 µl of 1 : 10 water-diluted cDNA in a standard 50 µl reaction. For quantitative real-time PCR, the cDNA was diluted 20-fold in water and amplified in a LightCycler (Roche Molecular Biochemicals, Mannheim, Germany) with the LightCycler-FastStart DNA Master SYBR Green I kit (Roche Molecular Biochemicals), according to the manufacturers' protocol. All primers were chosen to amplify fragments not exceeding 500 base pairs. Primers and gene accession numbers used are as follows: KAT1fwd (5′-ACT TCC GAC ACT GC-3′), KAT1rev (5′-CCC AAA TGA CAT CTA A-3′); KAT2fwd (5′-ATA TTG ATA TGG GGT CA-3′), KAT2rev (5′-ATC TAT TTC TGC GTT TT-3′); AKT1fwd (5′-CCA ACT GTT GCG TAT-3′), AKT1rev (5′-CTG CGT GGT ACT CC-3′); AKT2fwd (5′-AAA ATG GCG AAA ACA C-3′), AKT2rev (5′-CGC TGC TTC ACA TAG AA-3′); AKT5fwd (5′-AGG CCA CAG TTG TTC-3′), AKT5rev (5′-CGC CAT TTT CTG ATA A-3′); AKT6fwd (5′-GCC AGT GCG GTT AC-3′), AKT6rev (5′-GAC TCA ATC GCT TGG TA-3′); ATKCfwd (5′-ATA TTG CGA TAC ACA AG-3′), ATKCrev (5′-GAC CTA ACT TCG CTA AT-3′); GORKfwd (5′-CCT CCT TTA ATT TAG AAG-3′), GORKrev (5′-GCT CCA TCC GAT AG-3′); SKORfwd (5′- TGA AAC GGC TTC TTA-3′), SKORrev (5′- GAG CCA CTC GGA AAC-3′); KCO1fwd (5′-GTT GGC ACG ATT TTC-3′), KCO1rev (5′-GCT TCG CAA GAT GAT-3′); KCO2fwd (5′-GAT CGG GAC AAAGTG-3′), KCO2rev (5′-ACG CAG CCA TTA CAG-3′); KCO3fwd (5′-CTT TAC CAG AAC ACA ACG-3′), KCO3rev (5′-GCA CAA TTA AAA AGC CAC-3′); KCO4fwd (5′-GCA AGA TAA GGT TAA AGT G-3′), KCO4rev (5′-CAT GAC AGT AGT ACG AT-3′); KCO5fwd (5′-AGA CGA CAA AGA AGA-3′), KCO5rev (5′-CCG GTG AGA ATC ATA-3′); KCO6fwd (5′-ACC CAA TTC GTC AAA A-3′), KCO6rev (5′-CCG CTT AGC AGA GTC T-3′); PP2CAfwd (5′-AAT TGT TGC TGA CTC C-3′), PP2CArev (5′-AAC TCT TAA CCA TCG T-3′); SUC2fwd (5′-CTT ATG CTT AAC GCT ATT-3′), SUC2rev (5′-GAC AAT GGC TAG ATT-3′); SUC3fwd (5′-CAC TAT ATG TAC TCT TGT C-3′), SUC3rev (5′-CAT CAA CGT AGG TCT C-3′); AHA3fwd (5′-GAG TCC ACT CTA CAA TC-3′), AHA3rev (5′-GTC TTT GTG TTT ACC GA-3′); DIR1fwd (5′-TAT GTT GGT CGA TAC ATC A-3′), DIR1rev (5′-CAT GGA GAG TTC TTG TAA-3′); EINfwd (5′-GAA GTA ACT GCT GTC TCG-3′), EINrev (5′-CTT TCC CTA ACA TGA TCT-3′); BRI1fwd (5′-CTT ACT ATG CTT ACG GA-3′), BRI1rev (5′-GTT AGC AGT TCT ATC GC-3); PIN3fwd (5′-GAA TGA TGA TGC CAA C-3′), PIN3rev (5′-GTT ACC CGA ACC TAA T-3′); PIN6fwd (5′-ATC AAT GGA TCA GTG C-3′), PIN6rev (5′-CCC ACG ACT GTT AGT A-3′). Fragment length: KAT1 = 379 bp, KAT2 = 392 bp, AKT1 = 347 bp, AKT2 = 353 bp, AKT5 = 481 bp, AKT6 = 428 bp, AtKC1 = 373 bp, GORK = 496 bp, SKOR = 253 bp, KCO1 = 500 bp, KCO2 = 401 bp, KCO3 = 234 bp, KCO4 = 281 bp, KCO5 = 456 bp, KCO6 = 344 bp, PP2CA = 431 bp, SUC2 = 351 bp, SUC3 = 239 bp, AHA3 = 305 bp, DIR1 = 174 bp, EIN4 = 203 bp, BRI1 = 360 bp, PIN3 = 252 bp, PIN6 = 197 bp. GenBank, EMBL or MIPS-code numbers: KAT1 (X93022), KAT2 (CAA16801), AKT1 (X62907), AKT2/3 (U40154), AKT5 (AJ249479), AKT6 = SPIK (CAC85283), AtKC1 (U81239), GORK (AJ279009), SKOR (AJ223358), KCO1 (X97323), KCO2 (AJ131641), KCO3 (CAB40380), KCO4 (AT1G02510.1), KCO5 (AJ243456), KCO6 (AT4G18160.1), SUC2 (AY050986), SUC3 (AJ289165), PP2CA (P49598), AHA3 (AY072153), DIR1 (AAL76110), EIN4 (NP_187108) and BRI1 (NP_195650), PIN3 (NP_177250), PIN6 (AAD52696). cDNA quantities were calculated using LIGHTCYCLER 3.1 (Roche, Mannheim, Germany) and were all normalized to 10 000 molecules of actin cDNA fragments (An et al., 1996) amplified by AtACT2/8fwd (5′-GGTGAT GGT GTG TCT-3′) and AtACT2/8rev (5′-ACT GAG CAC AATGTT AC-3′). Each transcript was quantified using an individual standard. To enable detection of contaminating genomic DNA, PCR was performed with RNA as template. These DNA-free RNA samples were subsequently used for cDNA synthesis.

Patch-clamp recordings

Patch-clamp recordings were performed in the whole-cell mode using an EPC-7 amplifier (List-Medical-Electronic, Darmstadt, Germany). Data were low-pass-filtered with an eight-pole Bessel filter (Compu Mess Electronic GmbH, Garching, Germany) with a cut-off frequency of 2 kHz and sampled at 2.5 times the filter frequency. Data were digitized (ITC-16, Instrutech Corp., Elmont, New York, USA) and analysed using pulse and pulsefit software (HEKA Elektronik, Lambrecht, Germany), as well as igorpro (Wave Metrics Inc., Lake Oswego, OR, USA). Patch pipettes were prepared from Kimax-51 glass capillaries (Kimble products, Vineland, NY, USA) and coated with silicone (Sylgard 184 silicone elastomer kit, Dow Corning GmbH, Wiesbaden, Germany). The command voltages were corrected off-line for liquid junction potential (Neher, 1992). The pipette solution (cytoplasmic side) contained 150 mm K-gluconate, 2 mm MgCl2, 10 mm EGTA, 2 mm Mg-ATP and 10 mm Hepes/Tris (pH 7.4). The standard external solution contained 30 mm K-gluconate, 1 mm CaCl2 and 10 mm Mes/Tris (pH 5.6). Osmolarity of all solutions was adjusted to 580 mosmol kg−1 with d-sorbitol. Modifications to solute compositions are included in the figure legends. Chemicals were obtained from Sigma-Aldrich (Taufkirchen, Germany).


We thank Bernd Müller-Röber for providing the KCO1 cDNA clone and P. Dietrich, D. Becker and T.G.A. Green for critical reading of the manuscript. This work was funded by DFG Grants to R.H.