*For correspondence (fax +49 9318 8861 58;
Significance of plasmalemma aquaporins for water-transport in Arabidopsis thaliana
Article first published online: 5 JAN 2002
The Plant Journal
Volume 14, Issue 1, pages 121–128, April 1998
How to Cite
Kaldenhoff, R., Grote, K., Zhu, J.-J. and Zimmermann, U. (1998), Significance of plasmalemma aquaporins for water-transport in Arabidopsis thaliana. The Plant Journal, 14: 121–128. doi: 10.1046/j.1365-313X.1998.00111.x
This paper is dedicated to Professor G. Richter on the occasion of his retirement.
- Issue published online: 5 JAN 2002
- Article first published online: 5 JAN 2002
- Received 30 June 1997; revised 19 December 1997; accepted 28 January 1998.
The plant plasma membrane intrinsic protein, PIP1b, facilitates water transport. These features were characterized in Xenopus oocytes and it has asked whether aquaporins are relevant for water transport in plants. In order to elucidate this uncertainty Arabidopsis thaliana was transformed with an anti-sense construct targeted to the PIP1b gene. Molecular analysis revealed that the anti-sense lines have reduced steady-state levels of PIP1b and the highly homologous PIP1a mRNA. The cell membrane water permeability was analyzed by swelling of protoplasts, which had been transferred into hypotonic conditions. The results indicate that the reduced expression of the specific aquaporins decreases the cellular osmotic water permeability coefficient approximately three times. The morphology and development of the anti-sense lines resembles that of control plants, with the exception of the root system, which is five times as abundant as that of control plants. Xylem pressure measurement suggests that the increase of root mass compensates the reduced cellular water permeability in order to ensure a sufficient water supply to the plant. The results obtained by this study, therefore, clearly demonstrate that aquaporins are important for plant water transport.
In animals and plants, water-selective channel proteins have been found which are integral constituents of the plasma membrane or the plant vacuolar membrane. Amino acid sequence comparisons of the plant and animal proteins reveal that they are evolutionarily related and belong to the MIP-protein family (Reizer et al. 1993). The functional identification of aquaporins as water channel proteins was established by expression of the human AQP1(CHIP28) -cRNA in Xenopus oocytes and subsequent measurement of water uptake into cells (Preston et al. 1992), as well as by reconstitution of the water transport activity in liposomes after incorporation of the purified protein (Zeidel et al. 1994).
Careful analysis of proteins, which are constituents of the tonoplast, revealed that one class, the tonoplast intrinsic proteins (TIP), belong to the MIP-protein family (Johnson et al. 1990). By means of the Xenopus oocyte swelling assay, which was developed for the functional analysis of animal water channels, it was demonstrated that these proteins are aquaporins, i.e. they are selective for water and facilitate its transport (Maurel et al. 1993).
Similar proteins, characterized as putative constituents of the plasmalemma, were identified by differential cloning procedures which have been initiated to isolate stress-induced genes (Kaldenhoff et al. 1993;Yamada et al. 1995;Yamaguchi-Shinozaki et al. 1992). Electron microscopy and immunological investigations (Kaldenhoff et al. 1995;Kammerloher et al. 1994) provided evidence that a group of aquaporin homologues from Arabidopsis thaliana are located in the plasma membrane and therefore referred to as plasma membrane intrinsic proteins (PIP).
Their molecular function was confirmed by expression of the specific PIP-cRNA in oocytes (Kammerloher et al. 1994). Although initial observations can be made regarding the functional analysis of these plant proteins in an animal model system (such as Xenopus oocytes) concerning their transport properties, the role of the plant water channel proteins as aquaporins in planta is still a matter under discussion. Cellular and long-distance water transport in plants has been explained by models which do not take these proteins into account (Dainty 1985;Dainty and Ginzburg, 1963;Molz and Ikenberry 1974).
Some indirect evidence about the role of aquaporins in plants came from experiments where tomato roots (Maggio and Joly 1995) or algae were treated with chemicals like HgCl2 or ZnCl2 (Henzler and Steudle 1995;Tazawa et al. 1996), which are known to block the pore of water channels if the proteins are expressed in Xenopus oocytes (Preston et al. 1992). Furthermore, studies of transgenic plants expressing a reporter gene, which is under control of an aquaporin-promoter, provided insights into the spatial gene activity. In these investigations an up-regulation of the promoter activity was registered during developmental processes like side root formation or shoot elongation (Kaldenhoff et al. 1995;Ludevid et al. 1992). In addition, a moderate or low activity for the specific promoter was observed in virtually all parts of the plants.
Thus far, the function of aquaporins for plant cellular water permeability has been demonstrated by measurement of bursting rates from protoplasts, which derived from two independent anti-sense lines (Kaldenhoff et al. 1995). The results obtained by these experiments indicated that aquaporins increase the plasma membrane water permeability. However, it was not possible to quantify the effect, i.e. measurement of protoplast swelling rates, in order to determine changes in the plasma membrane water permeability. The importance of aquaporins for vascular water transport has not yet been demonstrated and experiments directed to investigate this uncertainty have been asked for by several authors (Chrispeels and Maurel 1994;Steudle and Henzler 1995).
In order to elucidate these inquiries, we have constructed anti-sense plants with a reduction in the expression of PIP1b (AthH2), an aquaporin which is located in the plant plasma membrane. In comparing these anti-sense plants to control plants we were able to observe differences in cellular water transport and plant morphology, which are induced by a reduced expression of the specific PIPs. In particular, this inhibition of expression has an effect on the water permeability of the plasma membrane and on the morphology of the root system. Since the data obtained from the protoplast swelling experiments of this study demonstrate that aquaporins are substantial to the cellular plant water transport and that this results in an increase in root mass, the significance of these channel proteins with respect to plant growth and development is discussed.
Molecular analysis of the transformed plants
Arabidopsis thaliana plants were transformed with a construct containing the 35S promoter of cauliflower mosaic virus, which is constitutively expressed in virtually all tissues and fused to the complete cDNA of the aquaporin gene PIP1b in an inverted orientation. The integration of the construct into the Arabidopsis thaliana genome was verified by Southern analysis (Fig. 1). For this purpose, genomic DNA was prepared from the T4 progeny of five independently transformed lines and hybridized against probes corresponding to the PIP1b-promoter fragment or the 35S promoter sequence. Assessment of the signals that are related to the 35S promoter sequence revealed that the anti-sense plants were transformed with one or several copies of the construct. Due to the distinct sizes of the fragments corresponding to the native PIP1b gene, it is likely that the insertion site is dissimilar to that of the targeted gene.
In order to analyze effects of the anti-sense construct on the expression level of aquaporin-RNA, total RNA was extracted from anti-sense and control plants which had been transformed with a construct of the 35S promoter fused to the GUS (β-glucoronidase) coding sequence. Evidence for the assumption that the aquaporin gene expression is not affected was given by the determination of PIP1b steady-state levels in non-transformed compared to control plants and revealed identical concentration of the specific mRNA (data not shown).
The consequences of the anti-sense construct with regard to PIP1b expression was determined by measurement of the relative transcript level in correlation to the RNA steady-state concentration of other supposedly unrelated genes, i.e. the small subunit of ribulose-1,5-bisphoshate carboxylase/-oxygenase (SSU).
Total RNA was extracted from leaves of greenhouse grown anti-sense lines (as1, as2, as3, as22 and as23), as well as control plants, and was standardized against SSU-RNA. Subsequently, the filter was stripped and successively re-hybridized with gene-specific probes of the PIP and TIP-families (Fig. 2a). The Northern dot experiments revealed that the steady-state level of the mRNAs for γTIP of the tonoplast as well as those of the plasma membrane aquaporins PIP1c, PIP2a and PIP2b do not change significantly. In contrast, the mRNAs of PIP 1a and PIP 1b are reduced to a level of 20–60%, depending on the anti-sense line being investigated (Fig. 2b).
An antibody, which is targeted for the amino-terminus of PIP1-aquaporins, was employed to microsomal fractions and the Western blot revealed that the as1, as2 and as3 lines showed reduced levels of PIP1-protein. It was not detectable in extracts from as22 and as23 (Fig. 3).
Water-transport at the cellular level
Control as well as anti-sense plants were grown in greenhouse conditions and their leaves used as starting material to generate protoplasts. In order to estimate the water conductivity of the protoplasts’ plasma membrane, the cells were subjected to a swelling assay under hypotonic conditions. From these experiments, osmotic water permeability coefficients (Pf, Zhang et al. 1990) were calculated from the increase of cell diameter at 30–60 sec after shift into hypotonic medium. For protoplasts derived from control plants an average Pf (μm/s) value of 10.8 ± 2.1 was obtained and was decreased approximately three times in the anti-sense lines, i.e. as1: 3.7 ± 1.3, as2: 3.8 ± 1, as3: 3.6 ± 1, as22: 3.1 ± 0.8, as23: 2.9 ± 1 (Fig. 4a).
In order to demonstrate the contribution of those aquaporins, which are targeted by the anti-sense construct on the overall plasma membrane water transport, protoplasts from control plants were incubated for 5 min in a solution containing 1 mm HgCl2. Consequently, Pf was decreased to 1.6 ± 1 μm s–1, which is an additional twofold reduction compared to anti-sense cells. The effect was reversible by the addition of 5 mmβ-mercapto ethanol (data not shown). A time course of protoplast swelling as presented in Fig. 4(b) indicates that anti-sense as well as mercury chloride treated protoplasts show linear swelling rates between 30 and 60 sec. In this respect, those from control protoplasts are possibly already declining and a calculation of Pf based on these data might have been underestimated.
The results indicate that inhibition of plasma membrane aquaporins PIP1a and PIP1b synthesis reduces the water permeability at the cellular level. Therefore, the aquaporins are active as water channel proteins in the plant plasmalemma of leaf cells and increase the membrane water permeability.
Effects of reduced PIP expression on plant physiology and morphology
The significance of aquaporin-facilitated cellular water transport for an individual plant should be deducible from morphological and physiological differences between control and anti-sense plants. However, greenhouse grown anti-sense plants develop normally, without any obvious phenotype of leaves or flowers.
In order to determine whether the water uptake is altered by the expression of the anti-sense construct, 20 individual plants from each transformed line were investigated with a potetometer (Pfeffer 1881), which consists of a water-filled tube that contains the complete root system plus an associated capillary for recording the volume of water absorbed by the roots (Fig. 5). The data were collected over 3 h and revealed that the average rate of water uptake was 1.2 ± 0.2 μl min–1 g fresh weight, irrespective of whether control (n = 25) or anti-sense lines (n = 100) were examined.
According to these data the reduced water permeability, as it was determined by the protoplast swelling assay, does not seem to be relevant for the overall water uptake of the plants. It was expected that those, that lack an aquaporin and therefore possess a less facilitated membrane water transport, show a deficit in growth or development. The aerial parts of the anti-sense plants do not show any difference, although the root system is more abundant compared to the equivalent control plants (Fig. 6). The fresh weight quotient of leaf over root is 3.7 ± 0.5 for the anti-sense (n = 100) and for the control plants is 17.7 ± 1.2 (n = 25), which indicates that individuals showing a reduced aquaporin expression develop about five times the root mass as compared to the control plants.
Osmotic response of roots
In order to determine whether the increase in root surface results in an altered physical parameter of the water transport in the xylem, the root xylem pressure was measured. The xylem pressure probe (Balling and Zimmermann, 1990) was inserted into the xylem of the main roots of plants as described by Zhu et al. (1995) and provided information about the xylem pressure of the entire root system. The absolute xylem pressures were –0.08 ± 0.22 MPa (n = 12) and –0.1 MPa ± 0.22 (n = 15) for the control and anti-sense plants, respectively. If the roots were exposed to a nutrient solution containing 25 mm NaCl, which corresponds to an osmotic pressure of about 0.015 MPa, the xylem pressure dropped within a couple of seconds to more negative values. On average, the ratio of the change in xylem pressure to that of the osmotic pressure in the medium was 0.65 ± 0.23 (control, n = 12) and 0.73 ± 0.27 (anti-sense, n = 15). Despite the five times larger root system, the anti-sense plants showed the same xylem pressure, even after osmotic changes. This indicated that the as-lines are capable of perfectly compensating the reduced plasma membrane water permeability, which might explain the regular development of leaves and flowers.
Transformation of Arabidopsis thaliana with a construct consisting of a 35S promoter and the coding region of PIP1b in an inverted orientation leads to a reduced steady-state level of aquaporin RNAs PIP1b and PIP1a. The restriction of the anti-sense effect on these two mRNA-species is possibly a consequence of nucleotide sequence identity, which is 83% between PIP1a and PIP1b. Further Arabidopsis thaliana aquaporin-mRNA sequences share less identity with PIP1b (78% PIP1c to 66% PIP2b).
The results of the protoplast swelling assay revealed that the water permeability coefficient of the leaf protoplast plasma membrane is decreased from approximately 11 μm s–1 to around 3 μm s–1. This indicates PIPs function as water channels in plants in vivo. Expression of these channel proteins increases the water permeability coefficient of the plant plasmalemma approximately to a factor of three. The further reduction of Pf by incubation with a HgCl2-containing medium can be accounted for in two ways. Mercury might interfere in an unspecific way with additional components of the plasma membrane and restrain the transport of water unrelated to aquaporins. On the other hand, it is likely that facilitated water transport by water channels, which are not affected by the anti-sense construct and are mercury sensitive, is inhibited as well. The mercury insensitive aquaporins (Daniels et al. 1994) are probably still functional. The data observed by these particular experiments would then reflect the water permeability of the lipid membrane plus mercury insensitive aquaporins (Pf = 1.6 μm s–1). The difference (about 9 μm s–1) to that of the control protoplasts denotes the significance of mercury sensitive aquaporins in plant plasma membranes.
Since the P values obtained by the protoplast assay are not based on the initial swelling rates, but on those 30–60 sec after the change in osmolarity, the absolute value might appear underestimated in the instance of a faster swelling in the first 30 sec. It is unlikely, however, that this could be underestimtaed for the anti-sense and mercury chloride treated protoplasts because the swelling rates are almost linear within 90 sec. Furthermore, a difference between anti-sense and control plants can be attributed to the expression of water channels and provides evidence for the function of these proteins.
Experiments which use the pressure probe for cell pressure measurements were performed to provide additional data and a confirmation of the results already obtained by the protoplast assay. However, due to the small size of the Arabidopsis thaliana cells, no reliable data were obtained by this method.
On first view it is surprising that these cellular changes do not have an impact on the overall water transport capacity of the plant (potetometer) or change the xylem pressure of the root system. The observations concerning the increase in root mass by a factor of 5 provides an argument for the interpretation of this unchanged overall water transport in the whole plant. It is conceivable that this morphological change is a feedback reaction of the modifications at the cellular level. It is likely that the root cells of anti-sense plants which express water channels in the controls have a reduced water permeability similar to that of leaf cells because the construct is driven by the 35S promoter. Root protoplast swelling experiments fail to provide conclusive data, probably due to the cell-type specific expression of water channels in adult roots, which is restricted to the cells of the endodermis and elongation zone (Kaldenhoff et al. 1995). Extension of the root surface ensures a sufficient supply of water thereby providing the means for a regular metabolism and development in the aerial parts of the anti-sense plants. It is consistent if plants that lack aquaporins, thus having an increased resistance for the water movement in leaf and root cells, compensate this effect by increasing their root surface.
Future experiments should give information about other physiologically relevant parameters which are difficult to measure with Arabidopsis thaliana. It is planned to transform a larger plant, such as tobacco, with an anti-aquaporin and continue these investigations in the near future. Measurement of altered transpiration rates, which can be achieved easily by using these plants, should give evidence for a possible modified stomatal transpiration rate. It is also intended to use high resolution NMR-spectroscopy in order to show the water distribution in particular organs of these anti-sense plants, as well as the velocity of water movement in distinct tissues.
Arabidopsis thaliana (C24) was used in all experiments. Details of culturing have been described previously (Kaldenhoff et al. 1993).
Construction of the 35S-PIP1b anti-sense construct
The complete coding region of PIP1b (AthH2; EMBL Z17424) was cloned into pGPTV-Kan (Becker et al. 1992) with a 35S-promoter. Using a triparental mating procedure with the helper plasmid pRK 2013 (Ditta et al. 1980), the pGPTV-Kan-35S-anti-PIP1b construct was transferred from Escherichia coli to Agrobacterium tumefaciens LBA 4404 (Bevan 1984). Transconjugates of A. tumefaciens were identified by resistance against streptomycin and kanamycin as well as Southern hybridization.
Transformation of A. thaliana
Arabidopsis thaliana plants (ecotype C24) were transformed with 35S-anti-PIP1b by employing the root transformation method (Valvekens et al. 1988). Transgenic plants were first grown in vitro on kanamycin-containing medium and later transferred to soil for optimal seed production. Homogeneity of transgenic progenies was verified by resistance to kanamycin.
RNA extraction and hybridizations
RNA was extracted according to Kaldenhoff et al. (1993). Protocols for dot blot and Southern hybridizations followed standard procedures (Sambrook et al. 1989). Probes specific for 3′ ends of individual PIPs were kindly provided by A. Schäffner and synthesized as described by Kammerloher et al. (1994). Probes for SSU and γTIP were obtained from cloned RT–PCR products.
Isolation of microsomal fractions and immunodetection
The isolation of microsomes from cell homogenates was performed according to Hodge and Mills (1986). Equal amounts of protein were loaded on each lane of an SDS-PAGE Gel, separated and blotted onto nitrocellulose membrane. Immunodetection of PIP1b proteins was performed as described in Kaldenhoff et al. (1995). Reactions with the secondary antibody followed the protocol given by Blake et al. (1984).
For the construction of the potetometer device, a 1 ml syringe tube was glued to a Falcon tube with a 3 mm hole halfway. The end of the syringe was connected to a 1 ml pipette with a small hose and the whole device was filled with water until the pipette was completely full. A hole (about 5 mm diameter) was punched into the cap of the Falcon tube, and the plant root was fixed in the cap with soft rubber and sealed air tight with Lanolin. Water uptake was measured for 3 h at room temperature.
Plants were grown in continuous white light on 0.8% agar plates supplemented with 0.5 × Murashige-Skoog medium. Leaves were harvested, incubated in a solution of 0.5 m mannitol for 5 min and cut into small pieces with a razor blade. The leaf material was treated overnight with an enzyme solution consisting of 1% w/v cellulase Onozuka R-10, 0.25% w/v Macerozyme R-10, 8 mm CaCl2ı, 0.4 m mannitol (pH 5.5). Protoplasts were concentrated by centrifugation at 60 g and washed with 1 mm CaCl2, 10 mm KCl, 0.4 m mannitol (pH 5.8). 2.5 μl suspension were spotted on a microscopy slide, diluted with 2 Vol. water and photographed at 30 sec intervals. The pictures were digitized and diameters were measured by drawing a circle around individual protoplasts with commercially available computer software (Corel Draw) which indicates the circle size. Pf values were calculated according to Zhang et al. (1990).
Xylem pressure probe
The principle, calibration and possible pitfalls of the xylem pressure probe have been described in detail elsewhere (Balling and Zimmermann 1990;Benkert et al. 1991;Zhu et al. 1995;Zimmermann et al. 1993). For insertion of the probe into a single xylem vessel, a main root of the intact plant was carefully clamped into the holding device of the measuring chamber which was filled with nutrient solution. Because of vibrations, the bath solution was not aerated and root vessels were probed 2–5 mm below the first leaf.
We thank Prof. G. Richter (Institut für Botanik, Universität Hannover, Germany) for fruitful discussions. We are grateful to Dr A. Schäffner (Institut für Biochemie, Ludwig-Maximilians-Universität München, Germany) for supplying clones with PIP specific sequences and Dr D. Becker (Julius-von-Sachs-Institut für Biowissenschaften, Universität Würzburg, Germany) for critical reviewing of the manuscript. We also thank the Deutsche Forschungsgemeinschaft (SFB 251) for providing support.
- 111985 Water transport through the root. Acta Hort., 171, 21 31.
- 141995 Effects of Mercuric Chloride on the Hydraulic Conductivity of Tomato Root Systems. Evidence for a Channel-Mediated Water Pathway. Plant Phys., 109, 331 335.&
- 211881 Wasserbewegung in der Pflanze. In Pflanzenphysiologie ein Handbuch, Leipzig: Verlag von Wilhelm Engelmann, pp. 113–178.
- 11993 The MIP family of intergral membrane channel proteins: sequence comparisons, evolutionary relationships, reconstructed pathway of evolution, and proposed functional differtiation of the two repeated halves of the proteins. Critical Rev. in Biochem. Mol. Biol., 28, 235 257., , &
- 291989 Molecular Cloning: a Laboratory Manual. Cold Spring Harbor Laboratory Press, Cold Spring Harbor., ,
- 161996 Characteristics of Hg- and Zn-sensitive Water Channels in the Plasma Membrane of Chara Cells. Bot. Acta, 5, 388 396., ,
- 81992 Molecular cloning and characterization of 9 cDNAs for genes that are responsive to desiccation in Arabidopsis thaliana: sequence of one cDNA clone that encodes a putative transmembrane channel protein. Plant Cell Physiol., 33, 217 224., , ,