Investigating Glutamate Receptor-Like Gene Co-Expression in Arabidopsis Thaliana

Plant, Cell and Environment (2008) 31, 861–871 doi: 10.1111/j.1365-3040.2008.01801.x

Investigating glutamate receptor-like gene co-expression in Arabidopsis thaliana

S. J. ROY1*, M. GILLIHAM1†, B. BERGER1‡, P.A. ESSAH1, C. CHEFFINGS1§, A. J. MILLER2, R. J. DAVENPORT1¶, L.-H. LIU1**, M. J. SKYNNER3, J. M. DAVIES1, P. RICHARDSON3, R. A. LEIGH1† & M. TESTER1

1Department of Plant Sciences, University of Cambridge, Downing Street, Cambridge, CB2 3EA, UK, 2Rothamsted Research, Harpenden, Hertfordshire, AL5 2JQ, UK and 3Cambridge Biotechnology Limited, Babraham Research Campus, Cambridge, CB2 4AT, UK

ABSTRACT Abbreviations: Em, membrane potential; GLR, glutamate receptor-like gene; MEX, micro-EXpression amplification. There is increasing evidence of the important roles of glutamate receptors (GLRs) in plant development and in INTRODUCTION adaptation to stresses. However, the studies of these putative ion channels, both in planta and in Xenopus oocytes, may Plant GLRs are so named because of the similarity of their have been limited by our lack of knowledge of possible GLR amino acid coding sequences to those of the ionotropic heteromer formation in plants. We have developed a modi- glutamate (iGluR) receptors in animals (Lam et al. 1998). In fication of the single-cell sampling technique to investigate animals, iGluRs function as glutamate- and glycine-gated GLR co-expression, and thus potential heteromer forma- non-selective cation channels (NSCCs), which are involved, tion, in single cells of Arabidopsis thaliana leaves. Micro- most notably, in signalling across neurosynaptic gaps catal- EXpression amplification (MEX) has allowed us to amplify ysing K+,Ca2+ and Na+ entry into cells (Dingledine et al. gene transcripts from a single cell, enabling expression of up 1999). The N-methyl-d-aspartic acid (NMDA) subclass of to 100 gene transcripts to be assayed. We measured, on iGLuR is composed of four subunits, two of which bind average, the transcripts of five to six different AtGLRs in a glycine and two glutamate.These subunits are thought to be single cell. However, no consistent patterns of co-expression arranged in pairs to form the complete heteromeric tet- or cell-type-specific expression were detected, except that ramer, which acts as the functional ion channel (Colquhoun cells sampled from the same plant showed similar expression & Sivilotti, 2004). profiles. The only discernible feature was the detection of In plants, the function of GLRs is still unclear. Nonethe- AtGLR3.7 in every cell examined,an observation supported less, some evidence suggests the in vivo mechanism of GLR by GUS staining patterns in plants stably expressing action in plants is similar to that of animal iGluRs, leading promoter::uidA fusions. In addition, we found AtGLR3.7 to speculation that GLRs are a candidate family that may expression in oocytes induces a Ba2+-, Ca2+- and Na+- encode plant Ca2+-permeable NSCCs (Lacombe, Becker permeable plasma membrane conductance. & Hedrich 2001; Davenport 2002; Demidchik, Davenport & Tester 2002; Dubos et al. 2003; Meyerhoff et al. 2005).

Key-words: AtGLR3.7; glutamate receptor-like genes; Notably, a strong and transient depolarization of the Em, micro-EXpression amplification; single-cell sampling; two- concurrent with a rise in cytosolic free Ca2+, has been electrode voltage clamp; Xenopus oocytes. observed in Arabidopsis hypocotyl cells and root cells upon exogenous application of low concentrations of a limited number of amino acids, including glutamate and glycine Correspondence: M. Tester. Fax: +61 8 8303 7102; e-mail: (Dennison & Spalding 2000; Dubos et al. 2003; Meyerhoff [email protected] et al. 2005; Gilliham et al., unpublished data). The response *Present address: The Australian Centre for Plant Functional to low micromolar concentrations of glutamate was signifi- Genomics and the University of Adelaide, PMB1, Glen Osmond, cantly reduced by the genetic knockout of AtGLR3.3 (Qi, SA 5064, Australia. †Present address: School of Agriculture, Food and Wine, The Uni- Stephens & Spalding 2006). In addition, the constitutive versity of Adelaide,Waite Campus, PMB1, Glen Osmond, SA 5064, overexpression of AtGLR3.2 in Arabidopsis led to symp- Australia. tomsofCa2+ deficiency, even though Ca2+ accumulation in ‡Present address: Institute of Botany II, University of Cologne, these plants was not affected (Kim et al. 2001). Gyrhofstrasse 15, D-50931 Cologne, Germany. Through the application of pharmacological agents, the § Present address: Joint Nature Conservation Committee, putative function of GLRs have been linked to a variety of Monkstone House, City Road, Peterborough, PE1 1JY, UK. processes: root elongation (Filleur et al. 2005), hypocotyl ¶Present address: Oxford Institute of Ageing, Manor Rd Building, Manor Rd, Oxford, OX1 3UQ, UK. elongation (Brenner et al. 2000), floral stem and vascular **Present address: College of Resources and Environmental bundle development (Kim et al. 2001; Turano et al. 2002), Sciences, China Agriculture University, 100094, Beijing, China. senescence (Zimmermann et al. 2004; Meyerhoff et al. © 2008 The Authors Journal compilation © 2008 Blackwell Publishing Ltd 861 862 S. J. Roy et al.

2005), carbon and nitrogen metabolism (Kang & Turano observed expressed in shoots at the whole tissue level. 2003), and salt and cold stress (Maathuis, Filatov & Herzyk Promoter::uidA fusion (under the control of the AtGLR3.7 2003; Meyerhoff et al. 2004; see Gilliham et al. 2006, for a promoter) revealed AtGLR3.7 expression in a variety of recent review). It has also been hypothesized that kanamy- cell types present in embryos through to mature plants. cin can rescue the de-etiolated 3 (det3) mutant phenotype by AtGLR3.7 expressed heterologously in Xenopus oocytes working as an agonist for AtGLRs (Dubos et al. 2005). resulted in a glutamate-insensitive cation-permeable plasma Most characterization of AtGLR function to date has membrane conductance. Together the results suggest a been indirect and inferred from their possible role as fundamental role for AtGLR3.7 and point to groupings of targets in planta. Although mentioned in conference AtGLRs that could function together in planta. abstracts and in passing in the literature (e.g. Chefﬁngs 2001; Davenport 2002; Meyerhoff et al. 2005), no data have yet been published from experiments that demonstrate the MATERIALS AND METHODS functional activity of any AtGLR in heterologous systems. Plant growth conditions Even optimizing expression in heterologous systems can be unsuccessful. For example, expression of AtGLR2.1 cRNA All chemicals were supplied by Sigma-Aldrich (Gillingham, Xenopus oocytes did not result in any detectable current, Dorset, UK) unless otherwise stated. A. thaliana even following fusion with the signal peptide from the rat Columbia-0 and GAL4-VP16 UAS-GFP enhancer trap lines GluR6, which facilitated the functional characterization of (see http://www.plantsci.cam.ac.uk/Haseloff/) were grown the previously non-functional Synechocystis GluR0 (Chen on full-strength Murashige and Skoog (MS) medium with et al. 1999; Davenport et al., unpublished data). 1% (w/v) sucrose (pH 5.8) and 0.9 % bactoagar (Melford, One reason for the apparent recalcitrance of GLRs to a Ipswich,Suffolk,UK ).After 14 d,seedlings were transferred deﬁnitive unmasking of their function in planta, or in het- to a hydroponic growth solution comprising (in mm): erologous systems, is the lack of information regarding 5 KNO3,1KH2PO4, 0.5 MgSO4, 1 CaSO4, 0.05 NaFeEDTA,

GLR heteromer formation. Indeed, it is not known whether 0.05 H3BO3, 0.012 MnCl2 and 1 mm CuCl2,1mm ZnCl2,30nm it occurs and, if it does, what GLRs are likely to form the (Na)6 Mo7O24, adjusted to pH 5.7. Seedlings were grown for heteromers. Sequencing of the A. thaliana genome has a further 14 d with the nutrient solution refreshed after 7 d. identified 20 genes that encode GLRs, which can be divided The enhancer trap lines expressing green fluorescent protein into three distinct clades by parsimony analysis (Lacombe (mGFP5-ER) in leaf epidermal cells (KS019) and mesophyll et al. 2001; Davenport 2002). However, expression profiling cells (JR112) were kindly supplied by Dr Jim Haseloff and has been able to show only that all 20 GLRs are expressed Dr Kate Wilson (University of Cambridge), respectively. in root tissue, 14 in shoot and 12 in flowers (Chiu et al. Plants used for GUS assays were grown under the same 2002). No attempt has been made to characterize the com- conditions as described above, but mature plant organs, position of GLR co-expression in single cells. flowers and embryos were obtained for staining from plants It is increasingly clear that many proteins that catalyse grown on soil.All plants were grown under an 8 h light/16 h ion transport are only functional, or alter their transport dark photoperiod, at an irradiance of 250 mm m-2 s-1, and properties, when in heteromeric associations or in the pres- at 20 °C. ence of endogenous regulators. For example, these include the homomeric associations of members of the Shaker super family of voltage-dependent K+ channels (Dreyer Whole tissue RNA extraction et al. 2004); the heteromeric association of AtAKT1 and the Seedlings were dissected into root, stems and petioles, and ‘silent subunit’ AtKC1 (Reintanz et al. 2002); and AKT1 shoots. Total RNA was extracted using TriPure Isolation with CBL1/9, CPIK23 (Li et al. 2006). Therefore, the possi- Reagent (Roche, Welwyn Garden City, UK), following the bility that AtGLRs form multimers is worthy of investiga- protocol described by Chomczynski (1993). Genomic DNA tion (especially in light of their structural similarity to contamination was removed using Promega’s RQ1-RNase iGluRs; Dubos et al. 2003). Free DNase (Promega, Madison, WI, USA). Qiagen’s In this study, we characterized the expression patterns of Omniscript kit was used to synthesize cDNA from the the AtGLR gene family in three different cell types of extracted mRNA (Qiagen, Valencia, CA, USA). mature Arabidopsis leaves.To provide information at a high enough spatial resolution to indicate possible heteromeriza- tion, a modified single-cell sampling technique was devel- Whole tissue gene-specific PCR amplification oped, allowing uncontaminated samples to be obtained. The MEX amplification system (Richardson et al. 2000; PCR was performed using Qiagen’s HotStarTaq kit. Briefly, Rodrigues et al. 2005;Dixon & Skynner 2006) allowed detec- 50 mL reactions were set up according to the manufacturer’s tion of transcripts for up to 100 genes from a single-cell guidelines using 1 mL of RT product. PCR primers were sample.Expression of five to six AtGLRs could be measured designed to lie within 300 bases of the 3′ poly A tail of the in a single leaf cell, with AtGLR3.7 being expressed in every original mRNA to be compatible with MEX amplified cell type sampled. However, there were no discernible pat- products (see below). The sequences of the primers used terns of expression of the other 15 AtGLRs that had been to amplify specific GLRs are given in Supplementary © 2008 The Authors Journal compilation © 2008 Blackwell Publishing Ltd, Plant, Cell and Environment, 31, 861–871 Glutamate receptor-like gene co-expression in Arabidopsis thaliana 863

Table S1. PCR products were visualized on 2% (w/v) dNTPs, 1 mL each of primers (FAP,TAP) specific for gene agarose gels stained with ethidium bromide. sequences within the FAP-T3 and TAP-RT primers (100 ng mL-1) and 1 mL Advantage 2 polymerase mix (BD Single-cell sampling of RNA Clontech). This reaction was cycled using the following parameters: one cycle of 95 °C for 1 min, then 29 cycles of Single-cell sampling (Tomos & Leigh 1999) and single-cell 95 °C for 30 s, 68 °C for 3 min, concluded by one cycle at RNA extractions (Karrer et al. 1995; Brandt et al. 1999; 68 °C for 3 min. Aliquots (1 mL) of this MEX amplification Laval et al. 2002) have been described elsewhere. A steril- product were used directly for downstream gene-specific ized and silanized microcapillary was inserted into the PCR. desired leaf cell; to increase the amount of mRNA extracted, gentle suction was applied through a tube connected to the capillary, which removed residual cytoplasmic Single-cell gene-specific PCR contents in addition to the vacuolar sap that was forced into One microlitre of MEX product was added to a 25 mL the tip through turgor pressure. Sap was immediately Qiagen HotStarTaq PCR reaction using the same primers expelled into 5 mL of nuclease-free water containing 10 U and conditions as those for the whole tissue amplification, of RNasin (Promega) on ice. To avoid contamination from except the number of amplification cycles was increased to other cells, when sampling from subsurface mesophyll cells, 40.PCR products were run on 2% agarose gels,and presence salmon testes DNA was sucked up into the tip of the cap- or absence of transcripts was determined. Because of the illary for 10–20 s before being expelled immediately prior number of cycles used in the PCRs, all bands were clearly to sampling. A blank control was always run, where air was resolved so any score regarding the presence of a gene’s expelled from an empty salmon sperm-dipped microcapil- transcript was safely within detection limits.All MEX prod- lary. No plant-specific PCR product could ever be amplified ucts were first screened for the presence of actin2 before from this sample. transcript profiling for AtGLRs. In addition, during MEX optimization, single-cell RT-PCR (as described by Laval MEX amplification et al. 2002) was performed on the same cell sap to test for the MEX amplification (Dixon & Skynner 2006; patent number presence of genes of particular interest (AtGLR3.1, 3.2, 3.7 US2006240420) was carried out by Cambridge Biotechnol- and actin2),and consistent amplification from MEX samples ogy Ltd (Cambridge, UK). cDNA was produced from was found (n = 3; data not shown). mRNA using Sensiscript (Qiagen) reverse transcriptase and an oligo dT primer containing a unique heel sequence, GUS transformations which could be recognized later by a sequence specific primer.The single stranded cDNA was converted to double A 979 bp region of genomic DNA immediately 5′ of the stranded cDNA by priming second-strand synthesis with AtGLR3.7 start codon was cloned into a pCR2.1 TOPO a semidegenerate primer containing an additional unique vector and transformed into competent TOP10 F’ Escheri- sequence. This degeneracy allowed the second-strand chia coli cells (Invitrogen, Carlsbad, CA, USA). After incu- primer to bind every 300–700 bases along the Arabidopsis bation overnight, colonies containing the promoter region genome. Use of these degenerate primers and primers to were selected and the plasmid extracted using Qiagen’s the unique 3′ heel in a 35 cycle PCR would produce enough Miniprep kit. The AtGLR3.7 promoter was excised from products of similar sizes on which 100 gene-specific PCRs the TOPO vector, purified using Qiagen’s QIAquick gel could be carried out. extraction kit and inserted into the PBI101 binary vector (BD Clontech) upstream of the GUS uidA gene.The binary Reverse transcription vector was transformed into competent E. coli cells for bulking up overnight, before the PBI101 construct contain- Cell harvests, in approximately 2 mL of buffer, were com- ing both the AtGLR3.7 promoter and GUS into Agrobac- bined with 0.5 mL first-strand buffer (Qiagen), 0.5 mLcDNA terium (AGL1) cells by electroporation. After selection for synthesis primer (TAP-RT, 100 ng mL-1), 0.5 mL second- colonies containing the vector, cultured transformed cells strand synthesis primer (FAP-T3, 100 ng mL-1) and sterile were used for Agrobacterium-mediated plant transforma- H O added to a final volume of 4.3 mL. Samples were heated 2 tions (Clough & Bent 1998; http://www.cropsci.uiuc.edu/~a- to 70 °C for 5 min and snap-cooled on ice. To each tube, bent/protocol.html). 0.5 mL dNTPs mix (10 mm each of dATP, dCTP, dGTP, dTTP) and 0.2 mL Sensiscript reverse transcriptase (Qiagen) were added. Tubes were mixed briefly using a bench-top GUS assays vortex and centrifuged,followed by a 2 h incubation at 37 °C. Whole seedlings, embryos or plant organs were submerged

in staining solution containing 0.1 m NaH2PO4 (pH 7.0), Pre-amplification 0.5 mm K3[Fe(CN)6], 0.5 mm K4[Fe(CN)6], 10 mm NaEDTA, Reverse-transcribed product (5 mL) was added to a 50 mL 0.5% (v/v) Triton X-100 and 0.96 mm X-Gluc (5-bromo- PCR reaction containing 5 mL Advantage II PCR buffer 4-chloro-3-indolyl-b-glucuronide). The samples were infil- (BD Clontech, Mountain View, CA, USA), 1 mL10mm trated three times using a desiccator connected to a vacuum © 2008 The Authors Journal compilation © 2008 Blackwell Publishing Ltd, Plant, Cell and Environment, 31, 861–871 864 S. J. Roy et al. pump and incubated at 37 °C overnight. The staining solu- for 2 weeks on MS–bactoagar plates, followed by 2 weeks in tion was removed and the samples washed with 70% (v/v) hydroponics. Under our growth conditions, all 20 AtGLRs ethanol until all the chlorophyll was removed. GUS-stained were expressed in roots, while only 16 were seen to be tissues were photographed using a digital camera attached expressed in leaves and stems plus petiole. AtGLRs 2.2, 2.3, to a Leica MZ FLIII stereo-microscope (Leica, Wetzlar, 2.4 and 2.6 were not expressed in stems plus petioles or leaf Germany). samples (Fig. 1). These results are similar to those observed in 8-week-old Arabidopsis plants (Chiu et al. 2002). Xenopus electrophysiology Single-cell sampling from leaf epidermis AtGLR3.7 was cloned from A. thaliana ecotype and mesophyll Columbia-0 and inserted between the 5′ and 3′ UTR Xenopus laevis b-globin sequences and behind a strong A concern with single-cell sampling is whether sap from Kozak sequence (Kozak 1996) of a plasmid modified to overlying cells contaminates samples taken from subsurface enhance protein production in Xenopus oocytes (supplied cells (Kehr 2003). To examine whether contamination by Prof. John Wood, Department of Biology, University occurred during sampling, sap from mesophyll cells was College London, UK). AtGLR3.4 and 2.1 were inserted collected from Arabidopsis lines that were expressing GFP into the oocyte expression vector ‘pOO2’ (Ludewig, von in leaf adaxial epidermal cells (Fig. 2a). Microcapillaries Wiren & Frommer 2002). cRNA was made using the mMes- were inserted through the adaxial epidermal layer to the sage Machine T7 kit (Ambion Inc, Austin, TX, USA). mesophyll cells below by piercing the overlying cell. To X. laevis oocytes were enzymically dissociated from dis- ensure the tip of the capillary was clear of any sap before it sected ovarian lobe using 1.5 mg mL-1 type IA collagenase diluted in sterile-filtered (0.22 mm filter; Millipore, Bed- ford, MA, USA) 0-Ca2+, Mg-OR2 solution [containing (in mm) 82 NaCl, 2 KCl, 2 MgCl2,5mm 4-(2-hydroxyethyl)-1- piperazineethanesulfonic acid (HEPES) and 2.8 NaOH to pH 7.50]. After dissociating, stage V–VI oocytes were selected and maintained in sterile-filtered ND-96 media

[containing (in mm) 96 NaCl, 2 KCl, 1 MgCl2, 1.8 CaCl2,5 HEPES and 2.8 NaOH to pH 7.50)] plus PGH solution (2.5 mm Na pyruvate, 100 mg mL-1 gentamycin and 1% w/v horse serum at 18 °C). Oocytes were injected with 40 nL of -1 cRNA (1 ng nL unless stated) or nuclease free H2O using a Nanoject II (Drummond Scientific Company, Broomall, PA, USA) and incubated in ND96 + PGH for a further 48–72 h. Two-electrode voltage clamping, as specified by the manufacturers’ instructions (OC-725C amplifier; Warner Instruments Corp., Hamden, CT, USA), was performed in the solutions stated in the figure legends. Before recording, oocytes were injected with 4 mm 1,2-bis(o-aminophenoxy)ethane-N,N,N′,N′-tetraacetic acid (BAPTA) unless stated. Oocytes were continuously irrigated at 2 mL min-1 when held in the recording chamber. Liquid junction potentials were measured and corrected appropriately. Data were acquired and analysed using pClamp 8.2 (Axon Instruments, Union City, CA, USA).

RESULTS Expression of AtGLRs in 4-week-old hydroponically grown Arabidopsis plants Because of different growth conditions used in our experiments to those already reported in the literature, such as Chiu et al. (2002) and Zimmermann et al. (2004), it was ﬁrst Figure 1. Transcription patterns of all 20 AtGLRs,andactin 2, necessary to characterize the expression patterns of the 20 as revealed by RT-PCR. Samples were taken from roots (R), AtGLRs at the whole tissue level in our system. Total RNA stem and petioles (SP), and leaves (L) of 4-week-old plants was extracted from shoot, stem and petiole, and root tissues grown in hydroponics. The experiment was repeated twice with of 4-week-old Arabidopsis seedlings that had been grown consistent results observed. © 2008 The Authors Journal compilation © 2008 Blackwell Publishing Ltd, Plant, Cell and Environment, 31, 861–871 Glutamate receptor-like gene co-expression in Arabidopsis thaliana 865

(a) (e) always negative. Under these conditions, when mesophyll sap was collected, no GFP transcripts were detected, while they were detected from adaxial epidermal cells (Fig. 2d). The reverse experiment was also performed using enhancer trap lines expressing GFP specifically in mesophyll cells (Fig. 2e) to test whether epidermal sap could also be sampled uncontaminated by mesophyll sap. As was (b) expected, only the sap extracted from mesophyll cells showed GFP expression, while the epidermal cell sap showed no contamination by GFP mRNA (Fig. 2f). With (c) truly cell-specific sampling in place, it was possible (using MEX amplification) to fragment and amplify cDNA from (d) the sampled mRNA, resulting in enough cDNA to complete up to 100 gene-specific PCRs (data not shown), currently much greater than standard single-cell assays.

Cell-specific expression patterns of AtGLRsin Figure 2. Avoiding contamination of overlying epidermal leaf tissue cells when sampling from mesophyll cells. Line KS019. (a) Autofluorescence image showing expression of GFP Expression profiling of the 20 AtGLRs was carried out on specifically in adaxial epidermal cells. (b) GFP transcripts 16 adaxial epidermal, 16 mesophyll and 14 abaxial epider- amplified from samples taken from single adaxial epidermal cells mal cells. With the exception of AtGLR3.7, there were no expressing GFP. (c) GFP amplified from samples taken from discernible patterns of expression of any AtGLR gene in single mesophyll cells that have been sampled after the pipette any specific cell type (Fig. 3a). In addition, although on has passed through adaxial epidermal cells from the same plant average there was co-expression of five to six GLRs per as in (b). (d) Use of salmon testes DNA prevents contamination of mesophyll samples with RNA from the upper epidermis. GFP cell, no overall pattern of co-expression was discernible. can be amplified from samples taken from epidermal cells, and Interestingly, there appears to be more similarity in the not from mesophyll cells of the same plant. Line JR112. expression patterns of AtGLRs between cell types from the (e) Autofluorescence image showing expression of GFP same plant than between similar cell types of different specifically in mesophyll cells. (f) Demonstration that samples plants. In addition, the number of co-expressed AtGLRs in from mesophyll cells can be taken and mRNA amplified the adaxial epidermal cells varies considerably, from two to successfully. GFP transcripts amplified from mesophyll cells and nine, when compared with mesophyll and abaxial epidermal not epidermal cells of line JR112. Samples were taken from the last fully expanded leaf in 4-week-old plants. U, upper epidermal cells, which are more likely to have four to six AtGLRs cell, M, mesophyll cell. expressed within a cell (Fig. 3b). However, in all cases, AtGLR3.7 was found to be expressed in every cell sampled, suggesting a fundamental role for this gene. Expression of was moved down into the mesophyll layer, the capillary was actin2 was also confirmed in every MEX amplified sample removed from the epidermal cell; any sap that had entered (Fig. 3a), confirming integrity of the sampling and MEX the tip was expelled and the capillary reinserted through amplification procedures. the same hole. Under standard sampling conditions, it was Two AtGLRs, AtGLR2.2 and 2.3, were expressed in found that GFP mRNA could be detected in the sample single leaf epidermal and mesophyll cells that were not collected from both adaxial epidermal and mesophyll initially seen at the whole tissue level. However, additional cells (Fig. 2b,c), with further investigation showing that the PCR amplification on whole tissues, using 45 cycles, showed primary source of contamination was mRNA attaching to low level expression of these genes (data not shown). the glass capillary (results not shown). As a control, additional amplification using primers for Although microcapillaries were both sterilized and AtGLR2.6 was also carried out in whole leaf tissue extracts, silanized, it appeared that there were still sites to which as this gene was initially found to be not expressed in both mRNA from the epidermal cells could attach when the leaf tissue and single cells. No gene product could be found capillary was being moved towards the mesophyll layer. A with additional amplification. technique to remove the remaining mRNA binding sites As the ubiquitous expression of AtGLR3.7 in every was therefore developed. It was found that the use of other mesophyll and leaf epidermal cell sampled suggested an nucleotide containing compounds, such as salmon testes important central role for this gene, it was important to DNA, resolved contamination problems by binding to the examine whether it was constitutively expressed in other capillary and removing any remaining mRNA binding sites cell types throughout the plant to which cellular transcripts could attach. Thus, microcapillaries were dipped into salmon testes DNA for 10–20 s AtGLR3.7 promoter::uidA fusion immediately before sampling. Control samples using only salmon testes DNA were taken to test whether signals To confirm the expression patterns of AtGLR3.7, Arabidop- could come from this source. Results for salmon DNA were sis promoter::uidA fusions were created by fusing the © 2008 The Authors Journal compilation © 2008 Blackwell Publishing Ltd, Plant, Cell and Environment, 31, 861–871 866 S. J. Roy et al.

(a) (b)

8 Upper epidermal cells

Frequency 2

0 0246810 Number of expressed GLRs per cell

8 Mesophyll cells

Frequency 2

0 0246810 Number of expressed GLRs per cell

8 Lower epidermal cells

Frequency 2

0 0246810 Number of expressed GLRs per cell

Figure 3. Transcription profiles of all AtGLR gene family members in different cell types in the leaf. (a) Expression profile of AtGLRs in different cell types. mRNA was sampled from individual cells of different plants, amplified by MEX and gene-specific PCR for each GLR and actin2 (act2) was carried out on every sample. Grey blocks represent the presence of transcript in that cell sample for that particular gene. (b) Frequency histograms showing the number of detected AtGLR transcripts per cell. Ad Ep, adaxial epidermis; Mes, mesophyll, Ab Ep, abaxial epidermis. promoter of AtGLR3.7 to uidA, transforming this stably differences in the extent of vacuolation of different into Arabidopsis and staining seedlings of several indepen- cells. The possibility that AtGLR3.7 functions as a channel dent transformants for the location of GUS (Fig. 4). Analy- subunit was examined by expression in Xenopus oocytes. sis of a variety of different tissues, at different stages of development, demonstrated that not only was AtGLR3.7 AtGLR3.7 expression induces a expressed in all leaf cells but it was expressed in cell types of glutamate-insensitive cation-permeable developing embryos, seedlings and fully mature plants and conductance in Xenopus oocytes flowers. In Fig. 4, it should be noted that staining varies greatly between different cell types, but longer staining did Varying amounts (10–100 ng) of AtGLR3.7 cRNA were show expression throughout tissues (data not shown). The injected into oocytes. After 48–86 h incubation, current- differences in staining intensities may reflect differences voltage analyses were performed using a two-electrode in levels of transcriptional activity, or they may reflect voltage clamp. No ligand-gated currents that were not also © 2008 The Authors Journal compilation © 2008 Blackwell Publishing Ltd, Plant, Cell and Environment, 31, 861–871 Glutamate receptor-like gene co-expression in Arabidopsis thaliana 867

(a) (b) (c)

(d) (e) (f)

(g) (h) (i)

(j) (k) (l) Figure 4. Histochemical GUS staining of embryos and organs of pAtGLR3.7::uidA lines. Blue staining indicates activity of the reporter gene product. (a–c) Embryos at heart (a), torpedo (b) and late torpedo (c) stages. Primary leaves (d), root–shoot junction (e) and root apex (f) of 7-day-old seedlings. Secondary leaf with trichomes (m) (n) (o) (g), shoot (h) and emerging root buds (i) of 14-day-old seedlings. Full expanded leaf (j), root–shoot junction and petioles (k), and young lateral root (l) of 4-week-old plants. Mature leaf with emerging ﬂoral bud (m), silique (n) and ﬂower (o) of 5- to 6-week-old plants.

present in water-injected controls were activated across the H2O-injected control oocytes) were observed in AtGLR3.7 oocyte plasma membrane by l-glutamate (10 mm–10 mm; cRNA-injected oocytes, regardless of ligand application Fig. 5a; n = 23), glycine (10 mm–10 mm; n = 21), or a combi- (Fig. 5b,c) even following injection of oocytes with BAPTA nation of both (n = 16), at any voltage, at various external (BAPTA injected n = 86; non-BAPTA injected n = 38). pH, with a range of external cations. In both AtGLR3.7 Currents resembling the native oocyte Ca2+-activated Cl- cRNA-injected and water-injected controls, an inward conductance induced by AtGLR3.7 cRNA injection were current in the nA range was infrequently seen at hyperpo- abolished by BAPTA injection. larized potentials when the oocytes were irrigated with Attempts were made to delineate the residual conduc- glutamate, which may have been attributable to XenU1 or tance from that of a native hyperpolarization-activated Ca2+ similar endogenous subunits (Soloviev & Barnard 1997). conductance, which can be activated by the overexpression Additional potential ligands or blockers of AtGLR identi- of foreign proteins (Tzounopoulos, Maylie & Adelman ﬁed in the literature (for a review, see Davenport 2002) 1995). Ba2+ can permeate many Ca2+-permeable channels were also applied to oocytes, but no novel currents were but not the native hyperpolarization-activated Ca2+ chan- detected in AtGLR cRNA injected oocytes over controls nels (Kuruma, Hirayama & Hartzell 2000). Substitution [abscisic acid (ABA); b-methylamino L-alanine (BMAA), of Ca2+ with Ba2+ signiﬁcantly reduced the magnitude 6,7-Dinitroquinoxaline-2,3-dione (DNQX), g-aminobutyric of AtGLR3.7-induced currents (n = 9, 12, respectively) acid (GABA), NMDA and glutamine (data not shown)]. compared with H2O-injected control oocytes (n = 10, 11, However, constitutive currents (greater than those of respectively) (Fig. 5c). Currents were further reduced at all © 2008 The Authors Journal compilation © 2008 Blackwell Publishing Ltd, Plant, Cell and Environment, 31, 861–871 868 S. J. Roy et al.

potentials and significantly at depolarized potentials (0–60 mV) when Na+ was replaced by N-methyl-D- glucamine (NMDG+)(n = 9), a larger cation impermeable to most ion channels. DNQX, a blocker of AMPA (alpha- amino-3-hydroxy-5-methyl-4-isoxazolepropionic acid) and kainate-type iGLuR in animals (see Dingledine et al. 1999), which is often used as an indication of GLR activity in plants (see Davenport 2002), did not give any consistent reduction of AtGLR3.7 cRNA-induced current (n = 7). All treatments listed above shifted the reversal potential of the currents to more negative values and closer to that of the controls, indicating we were abolishing permeable ions from the bathing solution through the induced conductance. Because of the presence of permeable cations in the interior of the oocyte, outward currents would still be present unless a specific blocker could be applied. To test for clade specificity of possible channel activity, AtGLR 3.4 and 2.1 were expressed singly.While AtGLR 3.4 expression resulted in a very similar conductance to that of AtGLR 3.7 (n = 46), expression of AtGLR2.1 did not change plasma membrane conductances when assessed against the water- injected control (n = 18; data not shown). Co-injection of AtGLR3.4 and 3.7 made no difference to the currents recorded (n = 25; data not shown). These results are consistent with AtGLR3.7 and 3.4 expression inducing a Ba2+- permeable cation conductance in Xenopus oocytes in addition to endogenous currents known to be induced by foreign protein expression.

DISCUSSION Tissue-specific GLR expression Analysis of the expression patterns of all 20 AtGLRs revealed that they were all expressed in root tissue, albeit very low for AtGLR2.7, while AtGLRs 2.2, 2.3, 2.4 and 2.6 were not expressed in either leaves or stems and petioles. Interestingly, in our study we detected the expression of AtGLRs 2.1 and 2.9 in our 4-week-old leaf tissue, which was Figure 5. Electrical response across the plasma membrane not detected in 8-week-old plants (Chiu et al. 2002), suggest- of Xenopus laevis oocytes expressing AtGLR3.7. (a) Fifty ing that the expression of these genes are downregulated micromolar l-glutamate elicits no current response, Vm =-80 mV during development. Microarray results gathered through (n = 39). (b) Instantaneous current-voltage plots of Xenopus Genevestigator suggest that all AtGLRs are expressed in laevis oocytes injected with 4 mm BAPTA and cRNA of mature leaf tissue at some stage (Zimmermann et al. 2004), AtGLR3.7 (solid) or H O (hollow). Instantaneous currents at a 2 with the exceptions of AtGLR2.3 (which has negligible range of voltages measured in oocytes bathed with (in mm)96 expression) and GLR2.2 (on which there are no data). NaCl, 1.8 CaCl2, 2 MgCl2, 2 KCl (n = 12). (c) Instantaneous currents at a range of voltages observed when CaCl2 was replaced with BaCl2 (circles) (n = 11); or when CaCl2 and NaCl Improved single-cell sampling enables were replaced with BaCl2 and NM-D-G-Cl, respectively characterization of the expression patterns of (squares) (n = 9), Vhold =-80 mV. Solid symbols are oocytes multiple genes from a single cell injected with cRNA of atGLR3.7 and hollow are H2O controls; error bars represent standard errors of the mean. Currents below Currently, it is possible to carry out RT-PCRs on a small x-axis indicate net inwardly directed cation flux (or outwardly number of genes (usually a control and two to three experi- directed anion flux), whereas positive currents indicate the reverse. BAPTA, 1,2-bis(o-aminophenoxy)ethane- mental genes; Brandt et al. 1999; Laval 2002) or create a N,N,N′,N′-tetraacetic acid. cDNA library,through a lengthy process (Karrer et al. 1995), on sap extracted by a microcapillary from a single cell. The methods detailed within this paper allow for the quick amplification of at least 21 genes from a single cell with, theoreti- cally, the possibility of increasing this number to 100. © 2008 The Authors Journal compilation © 2008 Blackwell Publishing Ltd, Plant, Cell and Environment, 31, 861–871 Glutamate receptor-like gene co-expression in Arabidopsis thaliana 869

Although micropipette-based single-cell sampling pro- more similarity between the GLRs expressed within epider- vides significant advantages over using laser dissection, such mal and mesophyll cells in the same plant leaf than between as the use of fresh tissue, the elimination of lengthy fixation cells of the same type in different plants. This perhaps sug- steps and the ability to sample repeatedly from the same gests that the functional roles of GLRs in both mesophyll tissue to build up a time course of responses, there are and epidermal cells are similar under non-stressed condi- inevitably problems with this approach. Most of these, tions, although whether these change under stress remains however, can be controlled for using certain experimental to be seen. In the future, it would be interesting to produce manipulations. A common criticism of this technique a map of AtGLR expression in all cell types of the leaf and regards the potential contamination with RNA from neigh- to investigate whether there are gradients in expression as bouring cells and an uncertainty in sample origin and iden- one moves closer to the extremities of the leaf or towards tity (Kehr 2003). This is particularly a problem for the veins. Certainly, there must be other cell types that show subsurface cells, such as leaf mesophyll and bundle sheath a different AtGLR expression pattern than mesophyll and cells, where the sampling takes place behind the field of epidermal cells if 16 of these genes are found in whole leaf view and there is the danger of sampling from surface cells tissue, but on average the same five to six genes are found in when the capillary is being guided to its target.The problem mesophyll and epidermal cells. The variation observed in is further exacerbated when the sampled sap is to be used in GLR expression between different plants also suggests a sensitive molecular biological techniques involving gene redundancy or functional overlap within the GLR family. amplification, where even the slightest contaminant could The lack of a consistent complement of GLRs in single be detectable after amplification. Previously, it has been cells does not exclude the possibility that heteromer forma- difficult to address this problem because of the small tion does take place. Further investigation is warranted number of cell-type specific markers that could be used to using AtGLR cRNA co-injection in oocytes, yeast 2-hybrid examine for any contamination. However, the creation of or split ubiquitin studies and bifluorescence complementa- enhancer trap Arabidopsis lines, with cell-specific expres- tion of AtGLRs commonly seen expressed together. sion of GFP provides a visual and robust marker. The In total, 17 AtGLRs were detected in different individual results shown here demonstrate convincingly that it is pos- leaf cells with AtGLRs 2.2 and 2.3 detected in single cells sible to sample from subsurface cell types without any con- but not at the leaf tissue level, suggesting that their tran- tamination taking place when suitable precautions have scripts are found at very low numbers in specific cells in the been taken and the sampling capillary is coated with salmon leaf. The detection of these low-abundance transcripts in testes DNA blocking any remaining binding sites (left by single-cell samples also suggests that the observed distri- silanization) for nucleic acids (Fig. 2). bution of transcripts is not influenced by the sampling For researchers wishing to use micropipette sampling in approach, that is, an inability to extract the majority of the quantitative gene-expression studies, a consideration to the cytoplasmic content. Importantly, however, there was no varying amounts of cytoplasm and RNA from different expression of AtGLR2.8 in mesophyll cells, even though cell types must be given. For instance, we have noted an expression of this gene was recorded in whole leaves. Inter- approximately five to ten times greater amount of RNA can estingly, AtGLR2.8 has been found in the mesophyll cells, be extracted from a single Arabidopsis mesophyll cell com- predominately those around the vascular bundle, in senesc- pared with a single epidermal cell (data not shown). A ing leaves, as shown by promoter::uidA fusions (Meyerhoff related source of error is the limited ability to extract all of and Becker, unpublished results, as cited in Gilliham et al. the RNA from a single cell, although this is equally true of 2006). As sampling of mesophyll cells in this present study RNA samples collected from laser micro-dissected cells. took place away from leaf veins and in a different growth Therefore, it is possible that the final amplified product may stage from those used in the previous experiment, it is pos- be misrepresentative of the complete RNA profile of the sible that cells expressing this gene were never sampled in cell, making it vital that a large enough sample size is our study. Single-cell expression of a selection of AtGLRs considered when using either technique. It should be and actin2 were confirmed using methods outlined by Laval emphasized that the results presented in this study are non- et al. (2002) using nested primers (data not shown). A very quantitative because of the number of PCR cycles used to clear result from both the single-cell sampling and the pro- amplify products from the MEX amplification. However, moter::uidA fusions is that AtGLR3.7 is expressed in most use of the MEX technique does result in cDNA products of cells of the plant throughout development (Figs 3 & 4), similar size, therefore reducing the bias in subsequent PCR suggesting that it has an important function in transport or reactions of the disproportionate amplification of smaller signalling. PCR fragments over larger fragments. AtGLR3.7 forms a glutamate-insensitive Cell-specific AtGLR expression cation-permeable conductance Although there was no cell-specific pattern of GLR expres- Results using Xenopus oocytes were consistent with sion in the different leaf cell types, it was clear that within AtGLR3.7 cRNA inducing a constitutively active plasma each individual cell there was an average of five to six GLRs membrane voltage-independent cation conductance that expressed at one time (Fig. 3a,b). Interestingly, there is is permeable to Na+,Ca2+ and Ba2+ (Fig. 5; Gilliham et al. © 2008 The Authors Journal compilation © 2008 Blackwell Publishing Ltd, Plant, Cell and Environment, 31, 861–871 870 S. J. Roy et al.

2006). These characteristics distinguish the AtGLR3.7 con- Brenner E.D., Martinez-Barboza N., Clark A.P., Liang Q.S., ductance from the endogenous hyperpolarization-activated Stevenson D.W. & Coruzzi G.M. (2000) Arabidopsis mutants Ca2+ conductance, which is neither Ba2+ permeable nor resistant to S(+)-beta-methyl-alpha,beta-diaminopropionic acid, a cycad-derived glutamate receptor agonist. Plant Physiology voltage independent (Kuruma et al. 2000). Moreover, the 124, 1615–1624. 2+ Ca -permeable but glutamate-insensitive conductance Cheffings C.M. (2001) Calcium channel activity of a plant formed by AtGLR3.7 (and AtGLR3.4) is redolent of the glutamate receptor homologue. Paper presented at the 12th constitutive Ca2+-permeable NSCC conductance of Arabi- International Workshop on Plant Membrane Biology, Madison, dopsis root epidermis (Demidchik et al. 2002). It is interest- WI, USA. ing to note that no currents were observed with AtGLR2.1 Chen G-Q, Cui C., Mayer M.L. & Gouax E. (1999) Functional expression despite significant similarity between these characterisation of a potassium selective prokaryotic glutamate receptor. Nature 403, 817–821. three gene products. Expression of AtGLR2.8 cRNA in Chiu J.C., Brenner E.D., DeSalle R., Nitabach M.N., Holmes T.C. & Xenopus oocytes has also been shown to be ineffective in Coruzzi G.M. (2002) Phylogenetic and expression analysis of the stimulating currents (Lacombe et al. 2001), suggesting this glutamate-receptor-like gene family in Arabidopsis thaliana. might be a feature associated with clade 2 AtGLRs. The Molecular Biology and Evolution 19, 1066–1082. ubiquity of AtGLR3.7 expression in planta may reflect an Chomczynski P. (1993) A reagent for the single-step simultaneous ability to form ligand-independent cation channels and/or isolation of RNA, DNA and proteins from cell and tissue perhaps act as a ‘platform’ for heteromeric structures samples. BioTechniques 15, 532–537. Clough S.J. & Bent A.F. (1998) Floral dip: a simplified method including AtGLRs that do respond to ligands. for Agrobacterium-mediated transformation of Arabidopsis thaliana. The Plant Journal 16, 735–743. CONCLUSIONS Colquhoun D. & Sivilotti L.G. (2004) Function and structure in glycine receptors and some of their relatives. Trends in Neuro- We have shown that, by taking certain precautions, it is sciences 27, 337–344. possible to sample mRNA from single cells with no con- Davenport R. (2002) Glutamate receptors in plants. Annals of tamination from neighbouring cells. Using mRNA amplifi- Botany 90, 549–557. Demidchik V., Davenport R.J. & Tester M. (2002) Nonselective cation, we found, on average, five to six AtGLRs expressed cation channels. Annual Reviews of Plant Biology 53, 67–107. in three different cell types of Arabidopsis leaf tissue. Dennison K.L. & Spalding E.P. (2000) Glutamate-gated calcium However, there were no cell-specific expression patterns fluxes in Arabidopsis. Plant Physiology 124, 1511–1514. discernable, nor were there any apparent patterns of Dingledine R., Borges K., Bowie D. & Traynelis S.F. (1999) The co-expression. This lack of a consistent distribution of gene glutamate receptor ion channels. Pharmacological Reviews 51, transcripts within a particular tissue type suggests there are 7–61. Dixon A. & Skynner M. (2006) Nucleic acid amplification method. other determinants of transcript heterogeneity controlling US2006240420. GLR expression, such as cell position or other unknown Dreyer I., Poree F., Schneider A., Mittelstadt J., Bertl A., Sentenac factors. It is interesting that plants grown and sampled H., Thibaud J.B. & Mueller-Roeber B. (2004) Assembly of plant under identical conditions show completely different pat- Shaker-like K(out) channels requires two distinct sites of the terns of overall GLR expression in mesophyll and epider- channel a-subunit. Biophysical Journal 2, 858–872. mal cells. This points to some inherent biological variability Dubos C., Huggins D., Grant G.H., Knight M.R. & Campbell M.M. within those plants that we do not fully understand at this (2003) A role for glycine in the gating of plant NMDA-like receptors. The Plant Journal 35, 800–810. stage. However, we did find that AtGLR3.7 was expressed Dubos C., Willment J., Huggins D., Grant G.H. & Campbell M.M. in every cell sampled and appeared ubiquitous throughout (2005) Kanamycin reveals the role played by glutamate recep- plant development when monitored using plants expressing tors in shaping plant resource allocation. The Plant Journal 43, uidA driven by a putative AtGLR3.7 promoter. AtGLR3.7 348–355. expression in oocytes induces a Ba2+-, Ca2+- and Na+- Filleur S., Walch-Liu P., Gan Y. & Forde B.G. (2005) Nitrate and permeable plasma membrane conductance. glutamate sensing by plant roots. Biochemical Society Transac- tions 33, 283–286. Gilliham M., Campbell M., Dubos C., Becker D. & Davenport R. ACKNOWLEDGMENTS (2006) The Arabidopsis thaliana glutamate-like receptor family (AtGLR). In Communication in Plants (eds S. Baluška, S. The authors would like to thank Prof.JohnWood,UCL,UK, Mancuso & D. Volkmann), pp. 187–204. Springer-Verlag, Berlin, for the gift of the xenopus expression vector and the follow- Germany. ing for funding: Biotechnology and Biological Science Kang J. & Turano F.J. (2003) The putative glutamate receptor 1.1 Research Council (UK), Broodbank Research Fund (Uni- (AtGLR1.1) functions as a regulator of carbon and nitrogen versity of Cambridge) and the Australian Research Council. metabolism in Arabidopsis thaliana. Proceedings of the National Academy of Science of the United States of America 100, 6872– 6877. REFERENCES Karrer E.E., Lincoln J.E., Hogenhout S., Bennett A.B., Bostock R.M., Martineau B., Lucas W.J., Gilchrist D.G. & Alexander D. Brandt S., Kehr J., Walz C., Imlau A., Willmitzer L. & Fisahn J. (1995) In situ isolation of mRNA from individual plant cells: (1999) A rapid method for detection of plant gene transcripts creation of cell-specific cDNA libraries. Proceedings of the from single epidermal, mesophyll and companion cells of intact National Academy of Sciences of the United States of America 92, leaves. The Plant Journal 20, 245–250. 3814–3818.

Kehr J. (2003) Single cell technology. Current Opinion in Plant Richardson P.J., Dixon A.K., Lee K., Bell M.I., Cox P.J., Williams Biology 6, 617–621. R., Pinnock R.D. & Freeman T.C. (2000) Correlating physiology Kim S.A., Kwak J.M., Jae S.-K., Wang M.-H. & Nam H.G. (2001) with gene expression in striatal cholinergic neurones. Journal of Overexpression of the AtGluR2 gene encoding an Arabidopsis Neurochemistry 74, 839–846. homolog of mammalian glutamate receptors impairs calcium Rodrigues R.J., Almeida T. & Richardson P.J. (2005) Dual presyn- utilisation and sensitivity to ionic stress in transgenic plants. aptic control by ATP of glutamate release via facilitatory Plant and Cell Physiology 42, 74–84. P2X(1), P2X(2/3), and P2X(3) and inhibitory P2Y(1), P2Y(2), Kozak M. (1996) Interpreting cDNA sequences: some insights from and/or P2Y(4) receptors in the rat hippocampus. Journal of studies on translation. Mammalian Genome 7, 563–574. Neuroscience 25, 6286–6295. Kuruma A., Hirayama Y. & Hartzell H.C. (2000) A Soloviev M.M. & Barnard E.A. (1997) Xenopus oocytes express a hyperpolarisation- and acid-activated nonselective cation unitary glutamate receptor endogenously. Journal of Molecular current in Xenopus oocytes. American Journal of Physiology – Biology 273, 14–18. Cell Physiology 279, C1401–C1413. Tomos A.D. & Leigh R.A. (1999) The pressure probe: a versatile Lacombe B., Becker D., Hedrich R., et al. (2001) The identity of tool in plant cell physiology. Annual Review of Plant Physiology plant glutamate receptors. Science 292, 1486–1487. and Plant Molecular Biology 50, 447–474. Lam H.M., Chiu J., Hsieh M.H., Meisel L., Oliveira I.C., Shin M. & Turano F.J.,Muhitch M.J., Felker F.C. & McMahon M.B. (2002) The Coruzzi G. (1998) Glutamate receptor genes in plants. Nature putative glutamate receptor 3.2 from Arabidopsis thaliana 396, 125–126. (AtGLR3.2) is an integral membrane peptide that accumulates Laval V., Koroleva O.A., Murphy E., Lu C., Milner J.J., Hooks M.A. in rapidly growing tissues and persists in vascular-associated & Tomos A.D. (2002) Distribution of actin gene isoforms in the tissues. Plant Science 163, 43–51. Arabidopsis leaf measured in microsamples from intact indi- Tzounopoulos T., Maylie J. & Adelman J.P. (1995) Induction of vidual cells. Planta 215, 287–292. endogenous channels by high levels of heterologous membrane Li L., Kim B.-G., Cheong Y.H., Pandey G.K. & Luan S. (2006) A proteins in Xenopus oocytes. Biophysical Journal 69, 904– Ca2+ signaling pathway regulates a K+ channel for low-K 908. response in Arabidopsis. Proceedings of the National Academy of Zimmermann P., Hirsch-Hoffmann M., Hennig L. & Gruissem W. Sciences of the United States of America 103, 12625–12630. (2004) GENEVESTIGATOR. Arabidopsis microarray database Ludewig U., von Wiren N. & Frommer W.B. (2002) Uniport of and analysis toolbox. Plant Physiology 136, 2621–2632. ammonium by the root hair plasma membrane ammonium transporter LeAMT1;1. Journal of Biological Chemistry 277, Received 29 November 2007; received in revised form 3 February 13548–13555. 2008; accepted for publication 5 February 2008 Maathuis F.J.M., Filatov V., Herzyk P., et al. (2003) Transcriptome analysis of root transporters reveals participation of multiple gene families in the response to cation stress. The Plant Journal SUPPLEMENTARY MATERIAL 35, 675–692. Meyerhoff O., Müller K., Roelfsema M.R.G., Latz A., Lacombe B., The following supplementary material is available for this Hedrich R., Dietrich P. & Becker D. (2004) AtGLR3.4 represents article: an Arabidopsis glutamate-, touch-, and cold-sensitive receptor. Book of abstracts, 13th International Workshop on Plant Mem- Table S1. Nucleotide sequences of the used primers. brane Biology, 6–10 July, Montpellier, France. Sequences are given form 5′ to 3′ end. Meyerhoff O., Müller K., Roelfsema M.R.G., Latz A., Lacombe B., Hedrich R., Dietrich P. & Becker D. (2005) AtGLR3.4 repre- This material is available as part of the online article sents an Arabidopsis glutamate-, touch-, and cold-sensitive from http://www.blackwell-synergy.com/doi/abs/10.1111/ receptor gene. Planta 222, 418–427. j.1365-3040.2008.01801.x Qi Z., Stephens N.R. & Spalding E.P. (2006) Calcium entry mediated by GLR3.3, an Arabidopsis glutamate receptor with a broad (This link will take you to the article abstract) agonist proﬁle. Plant Physiology 142, 963–971. Please note: Blackwell Publishing is not responsible for the Reintanz B., Szyroki A., Ivashikina N., Ache P., Godde M., Becker content or functionality of any supplementary materials D., Palme K. & Hedrich R. (2002) AtKC1, a silent Arabidopsis potassium channel alpha-subunit modulates root hair K+ inﬂux. supplied by the authors. Any queries (other than missing Proceedings of the National Academy of Sciences of the United material) should be directed to the corresponding author States of America 99, 4079–4084. for the article.