The Plasma Membrane of Arabidopsis thaliana Contains a Mercury

Plant Physiol. (1994) 106: 1325-1333

The Plasma Membrane of Arabidopsis thaliana Contains a Mercury-Insensitive Aquaporin That I s a Homolog of the

Tonoplast Water Channel Protein TIP’

Mark J. Daniels, T. Erik Mirkov, and Maarten J. Chrispeels*

Department of Biology, University of California, San Diego, 9500 Gilman Drive, La Jolla, California 92093-01 16

Plant cells contain proteins that are members of the major intrinsic protein (MIP) family, an ancient family of membrane channel proteins characterized by six membrane-spanning domains and two asparagine-proline-alanine (NPA) amino acid motifs in the two halves of the protein. We recently demonstrated that y- TIP, one of the MIP homologs found in the vacuolar membrane of plant cells, i s an aquaporin or water channel protein (C. Maurel, J. Reizer, 1.1. Schroeder, M.J. Chrispeels [1993] EMBO J 12: 2241- 2247). RD28, another MIP homolog in Arabidopsis thaliana, was first identified as being encoded by a turgor-responsive transcript. To find out i f RD28 i s a water channel protein, rd28 cRNA was injected into Xenopus laevis oocytes. Expression of RD28 caused a 10- to 15-fold increase in the osmotic water permeability of the oocytes, indicating that the protein creates water channels in the plasma membrane of the oocytes and i s an aquaporin just like i t s homolog ?-TIP. Although RD28 has several cysteine residues, i ts activity is not inhibited by mercury, and in this respect it differs from y-TIP and all but one of the mammalian water channels that have been described. Introduction of a cysteine residue next to the second conserved NPA motif creates a mercury-sensitive water channel, suggesting that this conserved loop i s critical to the activity of the protein. Antibodies directed at the C terminus of RD28 were used in combination with a two-phase partitioning method to demonstrate that RD28 is located in the plasma membrane. The protein is present in leaves and roots of well-watered plants, suggesting that i t s presence in plants does not require a specific desiccation regime. These results demonstrate that plant cells contain constitutively expressed aquaporins in their plasma membranes (RD28), as well as in their tonoplasts (?-TIP).

The vacuolar membranes of plant cells and the plasma membranes of mammalian cells contain aquaporins, proteins that form water-selective channels (see Chrispeels and Maurel, 1994, for review). These 27-kD integral membrane proteins belong to a family of proteins that has cognates in mammals, yeasts, and bacteria; they are part of the larger MIP family (see Reizer et al., 1993, for a recent review). In plants, more than half a dozen derived amino acid sequences and/or proteins have been recently identified. Some of them are expressed in a tissue-specific manner, whereas others are induced by specific physiological conditions. For example, a- TIP is a seed-specific protein found in the tonoplasts of

Supported by a grant from the US. Department of Agriculture (Competitive Research Grants Organization).

* Corresponding author; fax 1-619-534-4052.

protein storage vacuoles (Johnson et al., 1990), tobRB7 mRNA is found in the roots of tobacco (Nicotiana tabacum) (Yamamoto et al., 1991; Opperman et al., 1994), NOD26 is specific to the peribacteroid membranes of soybean (Glycine max) nodules (Sandal and Marcker, 1988), and trg-31 (Guer- rero et al., 1990; Guerrero and Crossland, 1993) and rd28 (Yamaguchi-Shinozaki et al., 1992) transcripts are induced by desiccation of pea (Pisum sativum) and Arabidopsis thaliana, respectively. Dip is expressed in mature seeds and dark- grown seedlings of Antirrhinum majus (Culianex-Macia and Martin, 1993) and MIP cognates tmp-A and tmp-B (Shagan and Bar-Zvi, 1993; Shagan et al., 1993) are highly expressed in etiolated A. thaliana plants.

Amino acid sequence comparisons of these plant proteins and of all members of the MIP family show that they are evolutionarily related. All proteins have six membrane- spanning domains and a small number of amino acids that are absolutely conserved, including the signature sequence NPAXT. This signature sequence is partially repeated as NPA in the second half of the protein. Maurel et al. (1993) recently showed that y-TIP, a protein that is present in the tonoplast of cells in shoots, roots, and flowers of Arabidopsis, is an aquaporin. When y t i p cRNA was injected into Xenopus laevis oocytes and the eggs moved to a hypotonic solution 2 to 3 d later, we observed a 5- to 10-fold increase in the osmotic water permeability of the eggs. The presence of y-TIP in the plasma membrane of the oocyte causes them to rapidly swell and burst when bathed in a hypotonic solution. Similar results have been reported for three mammalian proteins in the MIP family, the aquaporins CHIP28 (Preston et al., 1992), WCH-CD (Fushimi et al., 1993), and MIWC (Hasegawa et al., 1994). The injection of cRNAs for other membrane proteins-including homologs of MIP that are not aquaporins, such as the bacterial glycerol channel GlpF-do not have this effect (Maurel et al., 1994). We now report that the desiccation-induced MIP homolog rd28 of A. thaliana also encodes an aquaporin that creates water channels in oocyte membranes. Furthermore, we show that the RD28 protein is a plasma membrane protein expressed throughout the plant, except in seeds. Therefore, we refer to it as RD28-PIP (plasma membrane intrinsic protein).

Abbreviations: MIP, major intrinsic protein; mosmol, milliosmolal; Pr, osmotic water permeability; PIP, plasma membrane intrinsic protein; TIP, tonoplast intrinsic protein.

1325

1326 Daniels et al. Plant Physiol. Vol. 106, 1994

MATERIALS AND METHODS

Growing Conditions for Arabidopsis Plants

Arabidopsis thaliana (Columbia ecotype) was grown from seed under continuous bright illumination at 22 to 24OC after 2 d of vemalization at 4OC. Plants were grown in 4-inch pots filled with Terra-Lite RediEarth (W.R. Grace and Co., Ajax, Ontario, Canada) and subirrigated with 0.05% Peter's peat- lite fertilizer (J.M. McConkey, Sumner, WA). Plants for plasma membrane isolation were grown using Terra-Lite Vermiculite (W.R. Grace and Co., Cambridge, MA) instead of RediEarth.

Plasmid Constructions and In Vitro RNA Synthesis

DNA fragments encoding 7-TIP (Hofte et al., 1992) and RD28-PIP (Yamaguchi-Shinozaki et al., 1992) were cloned into the BglII site of a pSP64T-derived Bluescript vector carrying 5' and 3' untranslated sequences of a &globin gene of Xenopus (Preston et al., 1992). pSP64T-derived vectors provide high mRNA stability and translation efficiency in Xenopus oocytes (Krieg and Melton, 1984). The cloned DNA fragments were (a) a BamHI-XbaI fragment containing the entire 7-TIP coding sequence plus 7 bp and 1 bp at the 5' and 3' termini, respectively, and with the XbaI end recessed and ligated to an 8-bp BamHI linker (New England Biolabs, Beverly, MA); and (b) a NdeI-BamHI fragment containing the entire RD28-PIP coding sequence plus 1 bp and 217 bp at the 5' and 3' termini, respectively. Capped RNA encoding ?.-TIP and RD28-PIP were synthesized in vitro using T3 RNA polymerase and purified as described (Preston et al., 1992).

DNA Sequence Analysis

Aquaporin protein sequences were aligned using NEWAT87 software within the DNASYSTEM (Smith, 1988) package with adjustments made by eye where appropriate. Amino acid similarities and consensus sequence were deter- mined by the LINEUP and PRETTY programs of the Genetics Computer Group (Devereux et al., 1984) sequence analysis package.

Site-Directed Mutagenesis to Create a Mercury-Sensitive RD28-PIP

The pSP64T-derived Bluescript vector containing the RD28-PIP coding sequence was used for site-directed mutagenesis by recombinant PCR as described by Higuchi (1990). The sequence of the mutagenic oligonucleotide primer (sense orientation) is given in Figure 1. The sense oligonucleotide was used in combination with an oligonucleotide complementary to the T7 promoter in the pSP64T-derived Bluescript vector, whereas the antisense oligonucleotide was used in combination with an oligonucleotide complementary to the T3 promoter in the pS64T-derived Bluescript vector. The products of these two separate PCR reactions were gel purified and used as templates for a secondary PCR using only the T7 and T3 primers. The resulting PCR product was digested with Hind111 and PsfI and cloned into the corresponding sites of pBS (Stratagene). The resulting clone was

218 219 220 221 222 223 224 225 226 221 228

I P I T G T G I N P A

5' - CCG ATC ACC GGA TGC GGT ATC AAC CC - 3'

rd28 cDNA ... A n CCG ATC ACC GGA ACC GGT ATC AAC' CCG GCT '.'

* * primer T223C (sense)

C

Figure 1. Site-directed mutagenesis of rd28 cDNA to create rd28 T22317. The sequence of the sense orientation mutagenic primer is shown. Substituted nucleotides are marked with asterisks. Amino acids are given by their one-letter codes and are numilered from the putative amino terminus of the protein.

sequenced to verify that the desired mutation of 1'223C had been introduced.

Oocyte Preparation and Injection

Fully grown oocytes (stage V and VI) were isolated from Xenopus llievis and incubated in Barths solution (88 m NaCl, 1 m KCl, 2.4 IT~M NaHC03, 10 mM Hepes-NaOH:, 0.33 m Ca[N03],, 0.41 m CaC12, 0.82 m MgS04, pH 7.4) as described (Cao et al., 1992). In vitro transcripts (0.5-1 mg/ mL) or nuclease-free water were injected in a volume of 50 nL.

Osmotic Water Permeability Assay

Oocytes were transferred 3 d after cRNA injection from Barth's solution (200 mosmol) at room temperature to the same solution diluted to 40 mosmol with distilled water. Changes in cell volume were followed with a microscope by taking photographs at 30- to 90-s intervals. Oocyte diameters were measured four times along two sets of perpendicular axes. The volume V was estimated as the mean of two ellipsoid volumes. The osmotic permeability coefficient was calculated from Pr = V,[d(V/V,)/df]/[S x V , (Osm, - Osm,,,)], where initial oocyte volume V , = 9 x cm3, initial oocyte surface area S = 0.045 cm', and mohr volume of water V , = 18 cm3/mol (Zhang and Verkman, 1991).

Antiserum to RD28-PIP

A peptide representing the carboxy terminus of the derived amino acid sequence of rd28 was chemically synthesized and coupled to BSA with glutaraldehyde using the procedures described by Harlow and Lane (1988). The sequence of the synthetic peptide is H2N-KSLGSFRSAANV-COOH. The coupled peptide was injected into a New Zealand White rabbit and the peptide was boosted at 4-week intervals. After three boosts, we obtained a highly specific serum that w , ~ used in a 1:250 dilution for the preparation of immunoblols.

lmmunodetection

For immunoblotting, appropriate quantities of protein were fractionated by SDS-PAGE and transferred to nitrocellulose and the proteins were detected using the rabbit anti-RD28- PIP antiserum. Goat anti-rabbit IgG coupled to horseradish peroxidase (Bio-Rad) was used as the secondary antibody. Extracts were incubated in denaturing buffer for 'LO min at room temperature or for 5 min at 55OC before loading the gel.

Plasma Membrane Aquaporin 1327

Chimeric Gene Constructs and Tobacco Transformation

rd28 was cloned as an EcoRV to SstI fragment between the cauliflower mosaic virus 35s promoter and the transcriptional terminator from the 3’ end of the nopaline synthase gene contained in pBIl2l (Clonetech, Palo Alto, CA) after removal of the @-glucuronidase gene as a SmaI to SstI fragment. The pBI121-derived constructs were transformed into Agrobacter- ium tumefaciens strain LBA4404 as described by Hofgen and Willmitzer (1988). Transformants were selected on medium containing 50 pg/mL kanamycin.

Leaf discs of Nicotiana tabacum (cv Xanthi) were transformed with A. tumefaciens harboring the chimeric rd28 gene cloned in pBI121 as described by Voelker et al. (1989). Transformants were selected by their resistance to kanamycin, and transformed plants were regenerated from shootlets by transfer to a root-inducing, kanamycin-containing agar medium as described by Voelker et al. (1989). The kanamycin-resistant plants were transferred to soil and grown in the greenhouse. Young leaves were collected for immunoblot analysis. The highest expressors were used for further analysis.

Protein Extraction and Preparation of Microsomes

Protein extracts were obtained by homogenizing plants in grinding buffer (100 mM Tris [pH 7.51, 1 mM EDTA, 12% [w/v] SUC, 0.2 mM aminoethylbenzenesulfonylfluoride [Cal- biochem-Novabiochem, La Jolla, CAI, 2 pg/mL aprotinin, l pg/mL leupeptin, 1 &mL pepstatin A) and collecting the supernatant after centrifugation at 15,OOOg for 10 min. After centrifugation, the supernatant was filtered through a single layer of cheesecloth.

Microsomes were prepared by centrifuging protein extracts at 100,OOOg for 2 h and resuspending the resulting pellet in grinding buffer.

Purification of Plasma Membranes by Two-Phase Separation

Plasma membranes from 3-week-old A. thaliana plants or leaves of transgenic tobacco expressing RD28-PIP were separated from other cellular membranes according to Kjellbom and Larsson (1984) with some modifications. The U3 upper phase containing the plasma membranes was repartitioned an additional time with fresh lower phase to give U4. The L1 lower phase was repartitioned three additional times with fresh upper phase to give L4 containing the intracellular membranes. The upper phases produced by the lower phase repartitioning were discarded. Purification of plasma membranes from tobacco transgenic for rd28 was as above, except that the last repartitioning was omitted so as to produce a U3 phase and an L3 phase. The @-glucan synthase I1 assay was performed as reported by Kauss and Jeblick (1987).

Water-Stress Time Course

Four-week-old rosette A. thaliana plants were harvested and washed gently to remove soil from the roots. Plants were then placed into a chamber consisting of a magenta box with a I-mm nylon mesh insert upon which the plants were laid,

raised above the box floor by sections of 2-cm-tall plastic tubing. Three different regimes of desiccation were applied by varying the amount of water present in the box. Under the first (low-stress) treatment, plants dried to 96% of their original weight; under the second (medium-stress) treatment, plants dried to approximately 60% of their original weight; and under the third (high-stress) treatment, plants dried to approximately 10% of their original weight. During treatment, plants were left under continuous dim illumination and collected after 24 and 48 h, at which time they were weighed and frozen at -7OOC.

RESULTS

RD28-PIP Forms Water Channels in Xenopus Ooctyes

In vitro-transcribed cRNAs encoding ?-TIP or RD28-PIP, each with the 5’ and 3’ untranslated regions of a P-globin gene from X . laevis, were injected into Xenopus oocytes and osmotic water transport was investigated 3 d after injection. For this, oocytes were exposed to hypoosmotic conditions (160 mosmol gradient) and initial changes in cell volume were measured. Control oocytes that were injected with water instead of cRNA swelled very slowly after the sudden de- crease in osmotic strength (Fig. 2A) and burst only after approximately 40 min (data not shown). This result illustrates a limited osmotic water flux through the oocyte plasma membrane, whose permeability can be accounted for mainly by lipid-mediated water transport (Zhang and Verkman, 1991). In contrast, oocytes injected with rd28 cRNA or y-tip cRNA rapidly increased their volume by up to 40% (Fig. 2A) and ruptured in 2 to 3 min, displaying the appearance of a new pathway for facilitated diffusion of water into the hy- pertonic cell. With RD28-PIP and y-TIP, the Pr increased 13- to 16-fold and 11- to 14-fold, respectively, over the control value (Fig. 2B). Recent evidence indicates that the 7-TIP protein is responsible for the formation of pores highly permeable to water (Maurel et al., 1993), similar to the mammalian proteins CHIP28 (Preston et al., 1992; van Hoek and Verkman, 1992; Zeidel et al., 1992), WCH-CD (Fushimi et al., 1993), and MIWC (Hasegawa et al., 1994). Such proteins are referred to as aquaporins.

It has been thought that a character of water channels was their inhibition by mercurial sulfhydryl reagents. Mercury readily complexes with protein sulfhydryls (Webb, 1966; Vallee and Ulmer, 1972) and it is presumed that this complex physically blocks water from passing through the aquaporin channel (Preston et al., 1993). Three water channels known so far (CHIP28, WCH-CD, and y-TIP) appear to be sensitive to mercury, but along with the recently discovered MIWC (Hasegawa et al., 1994), aquaporin RD28-PIP is refractory to mercury inhibition (Fig. 2B) in spite of the presence of Cys residues at positions 73, 95, 131, and 136 of the protein.

Introduction of a Cys Residue Creates a Mercury-Sensitive Channel

A comparison of the amino acid sequences of the four known aquaporins and RD28-PIP is shown in Figure 3. The NPAXT signature sequence of these MIP family proteins is shown in boldface, as is the corresponding NPA in the second


A

---o- - - water

__._.._.o........ YTIP

A R D 2 8

---* ------o- - - - - I . I . I . , .

O 1 2 3 4 5 6 7 8 time (min)

h

N 3.0 , I

water Y TIP RD28

Figure 2. Pf of rd28- and y-tip-injected oocytes. A, Time course of osmotic swelling of individual oocytes injected with water or in vitro-synthesized cRNA encoding RD28 or y-TIP. Oocytes in Barth's buffer were perfused at t = O with a 5-fold dilution of Barth's buffer with distilled water. Measurements of oocytes injected with rd28 and y-tip cRNA stopped at time of cell rupture. B, Pf values of individual oocytes derived from volume change measurements made over three independent preparations of oocytes. When indicated, the assay was performed in the presence of mercury ions (Hg, 1 m M HgCI2, 10 min preincubation). Data are expressed as the mean * SE, with the number of replicates above.

half of the protein. This sequence is present in nearly all members of the MIP family and its repetition is an indication of the internal duplication of the gene (Reizer et al., 1993). Both sequences are present in loops on opposite sides of the membrane: NPAXT on the cytoplasmic side and NPA on the extracellular or vacuolar side. By using site-directed mutagenesis, Preston et al. (1993) were able to show that Cysla9, located just in front of the second NPA sequence, is responsible for the mercury sensitivity of CHIP28, indicating that this is a functionally important region of the protein. Replace- ment of this Cys residue by Ser created a mercury-insensitive water channel. RD28-PIP and MIWC lack a Cys residue at the position homologous to that of CHIP28 where a Cys was found to cause mercury sensitivity. Through site-directed mutagenesis, an RD28 T223C mutant was made in which the Thr residue present was changed to Cys (Fig. 1). The resulting mutant RD28 T223C retains the original water channel activity but is also sensitive to mercury inhibition, showing an inhibition of Pf of approximately 80% (Fig. 4) in

the presence of mercury ions (1 mM HgC12). -/-TIP also lacks a Cys residue in this position; however, it is likely that a Cys at a different position is responsible for the mercury sensitivity of this aquaporin.

RD28-PIP Protein I s Present in All Organs But Not in Seeds

To determine the tissue and subcellular location 13f RD28- PIP, we prepared a monospecific antiserum. The antiserum was raised against a peptide consisting of the amino-terminal 12 amino acids of RD28-PIP coupled to BSA. This region of the protein has no sequence identity with other MII' proteins of A. fhaliana. Fractionation of whole organ extracts by SDS- PAGE followed by immunoblotting showed that the serum detected a single band around 27 kD in all organs examined except seeds. Loading the lanes with equal amounts of protein, we found the strongest signal in the floral bolt (Fig. 5). In samples of total protein, RD28-PIP appears in monomeric form (Fig. 5) , whereas in microsome samples (Figs. 6, 7, and 8) RD28-PIP was observed as a dimer or as a rruxture of monomeric and dimeric forms. Additional reducing agent in the samples partially solved the aggregates. The highly hy- drophobic nature of the protein, the presence of inaccessible disulfide bonds between monomers, or a combination of these two forces may cause this aggregation. From what is known, expression patterns of plant MIPS vary among the family members. Immunodetection methods show that a-TIP is exclusive to seeds, whereas y-TIP appears in most other plant tissues (Hijfte et al., 1992). The tobRB7 promoier is root specific, driving expression in meristematic and immature vascular cylinder tissue normally, and is induced by nema- tode infection in the giant cells from which the parasite feeds (Opperman et al., 1994). The trg-31 promoter of pea caused P-glucuronidase expression in leaves but not in roots of dehydrated transgenic tobacco (Guerrero and Crossland, 1993).

Abundance of RD28-PIP Protein I s Not Altered by Water Stress

Previous results have found induction of rd28 mRNA transcription beginning 10 h after A. thaliana plants w12re dehydrated to 10% of their initial fresh weight (Yamaguchi- Shinozaki et al., 1992). rd28 is very closely relakd to the trg-31 (clone 7A) gene from Pisum sativum (Rei2:er et al., 1993). RNA transcript levels of trg-31 also respond to water stress and increase shortly after dehydration (Guenero et al., 1990). Microsomal fractions of well-watered Arabidopsis plants contain RD28-PIP protein, and dehydration of the plants using three different dehydration regimes did not result in an increase in the level of RD28-PIP over the course of 48 h (Fig. 6). Low levels of rd28 mRNA were detected in Arabidopsis under normal growing conditions (K. Shinozaki, personal communication) and these may sustain the appar- ently constitutive level of RD28-PIP abundance in well- watered control plants. It is likely that this semi-quantitative method to detect protein did not detect subtle changes in RD28-PIP abundance resulting from water stress.


CHIP28MIWC

WCH-CDgTIPRD28

consensus

1 70H. ........ .......... ......asEf kkklFwrAVv AEFlATtlFV FisiGSalgF kypv....GnM. ........ .......... ....vafkgv wtqaFwkAVt AEFlAmLiFV LlsvGStinW g.......GsM. ........ .......... .......wEl rsiaFsrAVl AEF1ATL1FV FfglGSalqw ........asM. ........ .pirniaigr p......dEa trpdalkAal AEFisTLiFV vagsGSgmaF nklte..nGaMakdvegpdg fqtrdyedpp ptpffdaeEl tkwsLyrAVi AEFvATLlFl YvtvltvigY kiqsdtkaGgM- AEF-

71 140nqtavqdnvk VslAFGLsIa tLaqsvghlS GaHlNPAVTl gILlscqISi fRalMYiiAQ cvGAIvatalenplpvdmvl IslcFGLsIa tMVqcfghlS GgHINPAVTv amvctrkisi aksvFYitAQ cLGAIigaglsppsv...lq lavAFGLglg iLVqalghvs GaHINPAVTv acLvgchVSf IRaaFYvaAQ ILGAVagaalttpsglvaaa VahAFGL..f vaVsvganIS GgHVNPAVTf gaFiggnltl IRgiLYwiAQ ILGsVvaclIvdcggvgilg lawAFGgmlf iLVyccaglS GgHINPAVTf glFlarkVSl iRavLYmvAQ cLGAtcgvgf

CHIP2BMIWC

WCH-CDgTIPRD28

consensus ————————— ———FG——— ————————S G-H-NPAVT- ———————— ————Y —AQ --G——————

CHIP28MIWC

WCH-CDgTIPRD28

consensus

CHIP28MIWC

WCH-CDgTIPRD28

consensus

CHIP2BMIWC

WCH-CDgTIPRD28

consensus

141LsgiTssl.t gnslgrndLa dgvnsGqgLg iEilgTLqLV LcVLAtLylvTpps.v vgglgvttvh gnltaGhgLl vEHiTFqLV FtlFAsLheiTpve.i rgdlavnaLh nnataGqavt vElflTMqLV LcIFAsLkfaTg.... glavpafgLs agvgvlnaFv fEiVmTFgLV YtVYAt.vkafqsshyv nyggganfLa dgyntGtgLa aEilgTFvLV YtVF.s

209.Dr rRrdlgGS.. ..apLalGLs.Ds kRtdvtGS.. ..vaLalGFs.De rRgdnlGS.. ..paLsIGFsiDp k....nGSlg tiapialGFitop kRnardshvp vlapLpIGFa

-LV -IG-

210 T 279ValghLlald YTGcgiNPAR SFGsAVI... thnFsnHWIF WVGPFIGgal AvLiYdFiLa prssdltdrvValgpLfaln YTGasmHPAP SFGpAVI... mgnWenHWIY WVGPilGavl AgalYeYvFc p.dvelkrrLVtlghLlgly FTGcsmSPAR SLapAW... tgkFddHWVF WIGPLVGaii gsLlYnYlLf psakslqerLVganiLagga FsGasmNPAv aFGpAW. . . swtWtnHWVY WaGPLVGggi AgLiYevfF. ..inttheqLvfmvhLatlp iTGtgiNPAR SFGaAVIfnk skpWddHWIF WVGPFIGati AaFyhqFvLr asgskslgsFV———L——— --G——NPA- ———AV——— —————HW-- W-GP--G—— ———————— ————————

285k. ........ .......... vwtsgqveey dldad..... dinsrvemkp k......... .keafskaaqq tkgsymeved nrsqvetedl ilkpgvvhvi didrgdekkg kdssgevlss va......... ......vlkg lepdtdweer evrrrqsvel hspqslprgs ka........ .pttdy..... .......... .......... .......... .......... .......... .rsaanv.... .......... .......... .......... .......... .......... .

Figure 3. Amino acid comparison of five aqua-porins, CHIP28 (Preston and Agre, 1991),MIWC (Hasegawa et al., 1994), WCH-CD(Fushimi et al., 1993), 7-TIP (gTIP) (Hofte et al.,1992), and RD28 (Yamaguchi-Shinozaki et al.,1992). The single-letter amino acid code isused. Gaps added for aligning sequences aredenoted by periods (.). Numbers indicate thedistance from the RD28 protein amino-terminalend. Similar and conserved amino acids are inuppercase letters, others are in lowercase. Con-sensus line shows those residues conserved inall sequences. The MIP family signature NPAXTand NPA residues contained in extramembraneloops (Reizer et al., 1993) are shown in boldfacetype. Cys189, the mercury-sensitive Cys residueof CHIP28, is marked with an arrowhead.

RD28-PIP Is Localized to the Plasma Membrane

To determine the subcellular location of RD28-PIP and toinvestigate the possibility that it may be a plasma membraneaquaporin, we purified plasma membranes with the two-phase partition method of Kjellbom and Larsson (1984). Thismethod is based on the observation that vesicles derivedfrom different membranes are separated into different poly-mer phases as a result of their unique surface properties (seeSandelius and Morre, 1990, for a review). The purified mem-branes were assayed for the presence of the plasma mem-brane marker /3-glucan synthase II. Ninety-five percent ofthe glucan synthase activity was found in the plasma mem-

mea

bility

(cm/

s x

102)

o

in

c

u '

osm

otic

wat

r> o

=

Int .

. .

. l ,

Q - Hg++

m + Hg++

109 r̂ |

r— — 1..;.;..;, -v/.j

13

14

ill

10

,0

water RD28 RD28 T223C

Figure 4. PI values of re/28- and re/28 T223C-injected oocytes de-rived from volume-change measurements made over four inde-pendent preparations of oocytes. When indicated, the assay wasperformed in the presence of mercury ions (Hg, 1 mM HgCI2, 10min preincubation). Data are expressed as the mean ± SE, with thenumber of replicates above.

brane fraction when compared to the intracellular membranefraction. This purity of plasma membranes is typical with thetwo-phase partitioning method (Sandelius and Morre, 1990).As seen in Figure 7, the plasma membrane fraction is greatlyenriched in RD28-PIP compared to the internal membranes.The slight amount of RD28-PIP in the intracellular mem-branes may be attributed to newly synthesized protein thatis in transit through the secretory pathway and/or residualplasma membrane in the intracellular fraction. The samemembrane fractions were assayed for pyrophosphatase, atonoplast marker (Rea and Sanders, 1987), and as an immu-noblot using pyrophosphatase-specific antibodies (kindlysupplied by P.A. Rea). The results showed that the concen-tration of pyrophosphatase was 2 to 4 times greater in the

-46.3

- 27.6

-18.6

Figure 5. Immunodetection of RD28-PIP in tissues of bolting A.thaliana. Protein extracts were fractionated by SDS-PAGE and trans-ferred to nitrocellulose and the blot was developed with anti-RD28antibodies. Samples were loaded for equal protein in each lane.Molecular mass markers are indicated on the right.


95% 60% 10%Oh 24h 48h 24h 48h 24h 48h kD

-46.3-27.6

-18.6

Figure 6. Immunodetection of RD28-PIP in a water-stress timecourse of A. thaliana. Plants were dehydrated to and maintained atthe percentage of their original fresh weight as indicated. Theywere harvested at the time after desiccation shown. Microsomeswere fractionated by SDS-PACE and transferred to nitrocelluloseand the blot was developed with anti-RD28-PIP antibodies. Sampleswere loaded for equal protein in each lane. Molecular mass markersare indicated on the right. Both the 27- and the 54-kD bandsrepresent authentic RD28-PIP, which aggregates under theseconditions.

74.7 kD

46.3 kD

27.6 kD

18.6 kD

Untr»n»-f o r m e d

RD28Transformed

MIC ICM PM MIC ICM PM

1 2 3 4 5 6

Figure 8. Immunodetection of RD28-PIP expressed in transgenictobacco. Plasma membranes (PM), intracellular membranes (ICM),or total microsomes (MIC) purified from nontransgenic tobacco ortobacco transgenic for rd28 were separated by SDS-PAGE andtransferred to nitrocellulose. Blots were probed with anti-RD28-PIPantibodies. Samples were loaded for equal protein in each lane.

intracellular membranes compared to the plasma membrane-enriched fraction (data not shown).

N. tabacum transgenic for RD28 was grown and membraneswere purified from the leaves by two-phase partitioning.RD28-PIP in transgenic tobacco accumulates in the plasmamembrane, with very little protein in evidence in the intra-cellular membranes. No RD28-PIP was detected in any mem-branes of nontransgenic tobacco (Fig. 8). The lack of anycross-reacting species in nontransgenic tobacco immuno-detection, and the poor amino acid conservation of C-termi-nal residues in MIP family proteins (Reizer et al., 1993),indicate that the anti-RD28 antiserum is highly specific todetect the presence of RD28-PIP. Tobacco cells probably alsohave a plasma membrane aquaporin, but there is no cross-reactivity between the tobacco and A. thaliana proteins withthis antiserum.

-74.7

-46.3

-27.6

Figure 7. Immunodetection of RD28-PIP in membranes of A. thal-iana. Plasma membrane and intracellular membrane samples weretaken from fractions U4 and L4, respectively, of a two-phase sepa-ration purification of plasma membranes. These fractions and amicrosome sample were fractionated by SDS-PAGE and transferredto nitrocellulose and the blot was developed with anti-RD28-PIPantibodies. Samples were loaded for equal protein in each lane.Molecular mass markers are indicated on the right.

DISCUSSION

In this report we show RD28-PIP to be a mercury-insensitive aquaporin of the plasma membrane that is ex-pressed in all organs but not in seeds. rd28 was first discov-ered as an mRNA induced by water stress in A. thaliana(Yamaguchi-Shinozaki et al., 1992). The protein encoded bythis gene is a member of the MIP family of proteins whosemembers are functionally related in facilitating diffusion ofsmall molecules across a membrane (Reizer et al., 1993).Recently, two other cDNAs were identified in etiolated seed-lings of A. thaliana (Shagan and Bar-Zvi, 1993; Shagan et al.,1993) that are members of the MIP family and are closelyrelated to this gene, as well as to the turgor-responsive genefound in pea by Guerrero et al. (1990).

RD28-PIP Is a Mercury-Resistant Aquaporin

Based on the amino acid sequence identity between RD28-PIP and the aquaporin 7-TIP, we examined RD28-PIP for itsactivity as a water channel. rd28 cRNA injected into X. laevisoocytes resulted in the oocyte plasma membrane becomingapproximately 10-fold more permeable to water than can beaccounted for by intrinsic water permeability alone. Oocytesinjected with water swell very slowly after being shifted intohypotonic conditions, bursting only after 40 min, whereasoocytes injected with RD28-PIP cRNA swell rapidly whenintroduced into a hypotonic solution, bursting within 3 to 4min. We assume that this increase in P, is caused by theappearance of RD28-PIP protein in the Xenopus oocyteplasma membrane. This increase in Pf is not the result of aforeign protein that disrupts membrane integrity; oocytesinjected with glycerol facilitator GlpF cRNA or potassiumchannel Xsha2 cRNA have plasma membranes with novelproperties but there is no increase in membrane water perme-ability (Maurel et al., 1993, 1994).

In the presence of mercury ions RD28-PIP water channelactivity was not inhibited. It was thought that a pharmaco-logic character of water channels was their sensitivity toinhibition by mercury (Zhang and Verkman, 1991; Maurel etal., 1993; Preston et al., 1993); however, RD28-PIP and the


recently discovered MIWC (Hasegawa et al., 1994) lack this property. There is as yet no evident biological significance for mercury insensitivity. The mercury-sensitive site of ./-TIP is likely to be the single Cys residue (CYS"~) on the cytoplasmic side of the membrane. Our finding that the T223C mutant is mercury sensitive indicates that the second NPA region, in an extracellular loop of the protein, is equally important for the correct functioning of plant aquaporins as it is for mammalian aquaporins. Characterization of additional PIP aquaporins will be necessary to confirm this idea. It is interesting to note in this respect that TMP-A, TMP-B, and TRG-31, the three plant MIP sequences most closely related to RD28-PIP, also lack a Cys residue in this position (Guerrero and Crossland, 1993; Shagan and Bar-Zvi, 1993; Shagan et al., 1993). However, it is not yet known whether they are plasma membrane proteins and whether they are aquaporins.

RD28-PIP Is a Plasma Membrane Aquaporin

./-TIP, the only plant aquaporin characterized to date (Maurel et al., 1993), is a tonoplast protein. In mammalian cells, aquaporins are present in the plasma membrane (see Chrispeels and Maurel, 1994, for review). Preliminary evidence presented by Kammerloher and Schaffner (1993) indicated that the plant plasma membrane probably contains MIP proteins. To find out where RD28 is located, we used the two-phase separation technique of Kjellbom and Larsson (1984) to separate plasma membranes from intemal membranes. The technique was applied to leaves of A. thaliana and to leaves of transgenic tobacco plants that expressed RD28 from the cauliflower mosaic virus 35s promoter. Im- munoblots of the subcellular fractions showed clearly that the protein was located in the plasma membrane fraction, with relatively little cross-reactivity in the internal membranes. The plasma membrane fraction was identified not only by its partitioning characteristics in the two-phase system, but also by the presence of glucan synthase 11, a known plasma membrane marker. Because the intracellular membrane fraction reacted positively with RD28 antibodies even after several partitioning steps, we cannot rule out the possibility that RD28 does not also occur in the tonoplast. However, it seems very unlikely that the same membrane protein would be targeted to these very divergent cellular destinations. It is more likely that cells contain distinct tonoplast and plasma membrane aquaporins.

The antiserum used in these studies detects a single 27-kD protein in the whole tissue extracts, but in membrane preparations it also detects aggregates of this protein. This con- forms to similar findings with other MIP proteins: they readily form aggregates even when treated with SDS (see Johnson et al., 1989, for a discussion). The related protein CHIP28 is known to form a homotetramer in membranes (Verbavatz et al., 1993). Although we cannot exclude the possibility that the serum reacts with more than the 27-kD RD28-PIP protein, the serum was made to a carboxy-terminal peptide, and the carboxy termini of the MIP proteins are generally poorly conserved (Reizer et al., 1993). There is little amino acid sequence conservation at the carboxy terminus between RD28-PIP and TmpA and TmpB, two other closely

related sequences present in A. thaliana (Shagan and Bar-Zvi, 1993; Shagan et al., 1993).

Our finding that RD28-PIP is a plasma membrane protein has two important implications. First, cells contain aquaporins in both membranes that delimit the cytoplasm: ./-TIP in the tonoplast and RD28-PIP in the plasma membrane. This confirms the view that transcellular water movement is an important component of water movement through living tissues. It is likely, but this remains to be investigated, that water transport through aquaporins can be regulated, since it is known that water transport through the plant can be regulated (see Steudle, 1992, for a recent review). Second, plant cells can target two homologous proteins with consid- erable amino acid sequence identity to two different locations in the cell. At this time we do not know which destination requires information and which one requires only bulk flow (Gomez and Chrispeels, 1993). However, domain swap ex- periments and site-directed mutagenesis should reveal which domains and specific residues are responsible for correct targeting.

Abundance of RD28-PIP in Different Organs

Our results show that RD28-PIP is expressed throughout the plant and is present in all organs except seeds. The protein is most abundant in floral bolts. In this respect, its distribution matches that of the tonoplast aquaporin ./-TIP, which is also present everywhere except in seeds (Hofte et al., 1992). We do not know whether RD28-PIP and -/-TIP are expressed in the same tissues or cell types. Seeds contain a-TIP, but this protein has not yet been shown to be an aquaporin. The mRNA for RD28-PIP was first identified as a water stress- induced sequence (Yamaguchi-Shinozaki et al., 1992), and it is homologous to trg-31 from peas (Guerrero et al., 1990), which is also turgor responsive. Although trg-31 is markedly and rapidly induced in leaves when pea plants are subjected to water stress (Guerrero et al., 1990), RD28 is only weakly induced and only after prolonged stress (Yamaguchi- Shinozaki et al., 1992). Our immunoblot showed no change in the total abundance of RD28-PIP protein after 24 h of water stress of the A. thaliana plants, indicating that any change in protein may be restricted to cells that make up a relatively small fraction of the total tissue mass.

What Is the Role of Aquaporins, Such as RD28-PIP and y-TIP?

In animal cells, aquaporins are found wherever water transport through the membrane is high and/or regulated for physiological reasons. In plants, water flows through living tissues because cells are losing water as a result of the transpiration stream or because expanding cells need to increase their volume. Such water flow can go through the apoplast or through the symplast. Measurements with pres- sure probes indicate that the preferred route appears to depend on the plant organ or tissue, the plant species, and the physiological condition of the plant (see Steudle, 1992). Symplastic water movement is thought to be transcellular, mezning that it passes through the plasma membrane, as well as through the tonoplast. We speculate that the presence


of aquaporins allows water transport through living tissues to be regulated in plants, as it is in animals. Regulation could be achieved by altering the expression level of the aquaporins (more or less protein in each of the two membranes) or their activity (more or less water passing through a single channel). Plant MIP proteins are now being discovered in many labo- ratories (Fray et al., 1994) and a number of these will un- doubtedly be found to be aquaporins. Further research will reveal the physiological role of this important new class of proteins.

NOTE ADDED IN PROOF

While this paper was in press, Kammerloher et al. (Plant J 6: 187- 199, 1994) presented evidence that the plasma membrane of A. thaliana contains aquaporins (called PIPS). The C termini of PlP2a and PIP2b are identical to the C terminus of RD28-PIP, and the antibodies used in our study will therefore in all likelihood cross- react with those PIP proteins. Our results therefore confirm the subcellular location of PIP2a and PIP2b but do not rule out the possibility that RD28-PIP is a turgor-responsive protein, as originally suggested by the results of Yamaguchi-Shinozaki et al. (1992).

ACKNOWLEDGMENTS

The authors would like to thank Elad Gil and Audrey Ichida for sharing their preparations of Xenopus oocytes and Julian Schroeder for use of the equipment to carry out the oocyte swelling assays. We thank K. Shinozaki and K. Yamaguchi-Shinozaki for providing us with the RD28 cDNA clone and Lyn Alkan for typing the manu- script. We thank Phil Rea for providing antibodies to tonoplast pyrophosphatase.

Received May 27, 1994; accepted August 23, 1994. Copyright Clearance Center: 0032-0889/94/106/1325/09.

LITERATURE CITED

Cao Y, Anderova M, Crawford NM, Schroeder JI (1992) Expression of an outward-rectifying potassium channel from maize mRNA and complementary RNA in Xenopus oocytes. Plant Cell 4

Chrispeels MJ, Maurel C (1994) Aquaporins: the molecular basis of facilitated water movement through living plant cells. Plant Phys- io1 105 9-15

Culianex-Macia FA, Martin C (1993) DIP: a member of the MIP family of membrane proteins that is expressed in mature seeds and dark-grown seedlings of Antirrhinum majus. Plant J 4

Devereux J, Haeberli P, Smithies O (1984) A comprehensive set of sequence analysis programs for the VAX. Nucleic Acids Res 12

Fray RG, Wallace A, Grierson D, Lycett GW (1994) Nucleotide sequence and expression of a ripening and water stress-related cDNA from tomato with homology to the MIP class of membrane channel proteins. Plant Mol Biol24 539-543

Fushimi K, Uchida S, Hara Y, Hirata Y, Marumo F, Sasaki S (1993) Cloning and expression of apical membrane water channel of rat kidney collecting tubule. Nature 361: 549-552

Gomez L, Chrispeels MJ (1993) Tonoplast and soluble vacuolar proteins are targeted by different mechanisms. Plant Cell 5:

Guerrero FD, Crossland L (1993) Tissue-specific expression of a plant turgor-responsive gene with amino acid sequence homology to transport-facilitating proteins. Plant Mol Biol 21: 929-935

Guerrero FD, Jones JT, Mullet JE (1990) Turgor-responsive gene transcription and RNA levels increase rapidly when pea shoots are

96 1-969

717-725

387-395

1109-1 118

wilted. Sequence and expression of three inducible genes. Plant Mol Biol 1 5 11-26

Harlow E, Lane D (1988) Antibodies: A Laboratory Miinual. Cold Spring Harbor Laboratory Press, Cold Spring Harbor, NY

Hasegawa H, Ma T, Skach W, Matthay MA, Verkmani AS (1994) Molecular cloning of a mercurial-insensitive water channel expressed in selected water-transporting tissues. J Biol Chem 269

Higuchi R (1990) Recombinant PCR. In MA Innis, DH Gelfand, JJ Snmsky, TJ White, eds, PCR Protocols: A Guide to Methods and Applications. Academic Press, San Diego, CA, pp 177--183

Hofgen R, Willmitzer L (1988) Storage of competent cells for Agrobacterium transformation. Nucleic Acids Res 1 6 91177

Hofte H, Hubbard L, Reizer J, Ludevid D, Herman EM, Chrispeels MJ (1992) Vegetative and seed-specific isoforms of a putative solute transporter in the tonoplast of Arabidopsis thaliana. Plant Physiol99 561-570

Johnson KD, Herman EM, Chrispeels MJ (1989) An abundant, highly conserved tonoplast protein in seeds. Plant Physiol 91:

Johnson KD, Hofte H, Chrispeels MJ (1990) An intrinsic tonoplast protein of protein storage vacuoles in seeds is structui~ally related to a bacterial solute transporter (GlpF). Plant Cell 2 5215-532

Kammerloher W, Schaffner AR (1993) PIP-an A. thal,:ana plasma membrane MIP homologue cloned by expression in :nammalian cells (abstract No. 185). In Fifth Intemational Cor.ference on Arubidopsis Research. Ohio State University, Columbus, OH, p 185

Kauss H, Jeblick W (1987) Solubilization, affinity chromatography, and Caz+/polyamine activation of the plasma membrane-located 1,3-@-~-glucan synthase. Plant Sci 48: 63-69

Kjellbom P, Larsson C (1984) Preparation and polypepiide compo- sition of chlorophyll-free plasma membranes from leaves of light- grown spinach and barley. Physiol Plant 62: 501-509

Krieg PA, Melton DA (1984) Functional messenger RNAs are produced by SP6 in vitro transcription of cloned cDNAs. Nucleic Acids Res 1 2 7057-7070

Maurel C, Reizer J, Schroeder JI, Chrispeels MJ (19931 The vacuolar membrane protein y-TIP creates water specific channels in Xenopus oocytes. EMBO J 12: 2241-2247

Maurel C, Reizer J, Schroeder TI, Chrispeels MJ, Sa.ier MH Jr (1994) Functional characterization of the Escherichia coli glycerol facilitator GlpF in Xenopus oocytes. J Biol Chem 269 11869-11872

Opperman CH, Taylor CG, Conkling MA (1994) Root-knot ne- matode-directed expression of a plant root-specific gene. Science

Preston GM, Agre P (1991) Isolation of the cDNA for qthrocyte integral membrane protein of 28 kilodaltons: member 01. an ancient channel family. Proc Natl Acad Sci USA 88: 11 110-1 1 114

Preston GM, Carroll TP, Guggino WB, Agre P (1992) fippearance of water channels in Xenopus oocytes expressing red cell CHIP28 protein. Science 256: 385-387

Preston GM, Jung JS, Guggino WB, Agre P (1993) The mercury- sensitive residue at cysteine 189 in the CHIP28 water channel. J Biol Chem 268: 17-20

Rea PA, Sanders D (1987) Tonoplast energization: two pumps, one membrane. Plant Physiol71: 131-141

Reizer J, Reizer A, Saier MH Jr (1993) The MIP family of integral membrane channel proteins: sequence comparisons, evolutionary relationships, reconstructed pathway of evolution, and proposed functional differentiation of the two repeated halves ‘3f the proteins. Crit Rev Biochem Mol Biol 28: 235-257

Sandal NN, Marcker KA (1988) Soybean nodulin 26 is homologous to the major intrinsic protein of the bovine lens fiber inembrane. Nucleic Acids Res 16: 9347

Sandelius AS, Morre DJ (1990) Plasma membrane isolation, based on vesicle membrane surface properties. In C Larsson, IM Mbller, eds, Plant Plasma Membrane: Structure, Function and Molecular Biology. Springer-Verlag, Berlin, pp 52-65

Shagan T, Bar-Zvi D (1993) Nucleotide sequence of an .4rabidopsis thaliana turgor-responsive TMP-A, a transmembrane protein containing the major intrinsic protein motif. Plant Physiol 101: 1397-1398

Shagan T, Meraro D, Bar-Zvi D (1993) Nucleotide sequence of an

5497-5500

1006-1013

263 221-223


Arabidopsis thaliana turgor-responsive TMP-B cDNA clone encoding transmembrane protein with a major intrinsic protein motif. Plant Physiol102 689-690

Smith DW (1988) A complete, yet flexible, system for DNA/protein sequence analysis using VAX/VMS computers. CABIOS 4: 212

Steudle E (1992) The biophysics of plant water: compartmentation, coupling with metabolic processes, and water flow in plant roots. In GN Somero, CB Osmond, CL Bok, eds, Water and Life: A Comparative Analysis of Water Relationships at the Organismic, Cellular and Molecular Levels. Springer-Verlag, Berlin, pp

Vallee BL, Ulmer DD (1972) Biochemical effects of mercury, cad- mium and lead. Annu Rev Biochem 41: 91-128

van Hoek AN, Verkman AS (1992) Functional reconstitution of the isolated erythrocyte water channel CHIP28. J Biol Chem 267:

Verbavatz JM, Brown D, Sabolíc I, Valenti G, Ausiello DA, Van Hoek AN, Ma T, Verkman AS (1993) Tetrameric assembly of CHIP28 water channels in liposomes and cell membranes: a freeze- fracture study. J Cell Biol 123: 605-618

173-204

18267-18269

Voelker TA, Herman EM, Chrispeels MJ (1989) In vitro mutated phytohemagglutinin genes expressed in tobacco seeds: role of glycans in protein targeting and stability. Plant Cell 1: 95-104

Webb JL (1966) Enzyme and Metabolic Inhibitors. Academic Press, New York

Yamaguchi-Shinozaki K, Koizumi M, Urao S, Shinozaki K (1992) Molecular cloning and characterization of 9 cDNAs for genes that are responsive to desiccation in Arubidopsis thaliana: sequence analysis of one cDNA clone that encodes a putative transmembrane channel protein. Plant Cell Physiol33 217-224

Yamamoto YT, Taylor CG, Acedo GN, Cheng C-L, Conkling MA (1991) Characterization of cis-acting sequences regulating root- specific gene expression in tobacco. Plant Cell 3 371-382

Zeidel ML, Ambudkar SV, Smith BL, Agre P (1992) Reconstitution of functional water channels in liposomes containing purified red cell CHIP28 protein. Biochemistry 31: 7436-7440

Zhang R, Verkman AS (1991) Water and urea permeability properties of Xenopus oocytes: expression of mRNA from toad urinary bladder. Am J PhysiolZ60 C26-C34

Documents

The Plasma Membrane of Arabidopsis thaliana Contains a Mercury