isolation and functional characterization of three
TRANSCRIPT
Planta (2007) 225:381–392
DOI 10.1007/s00425-006-0365-2ORIGINAL ARTICLE
Isolation and functional characterization of three aquaporins from olive (Olea europaea L.)
Francesca Secchi · Claudio Lovisolo · Norbert Uehlein · Ralf KaldenhoV · Andrea Schubert
Received: 15 March 2006 / Accepted: 21 July 2006 / Published online: 19 August 2006© Springer-Verlag 2006
Abstract To study the molecular bases of water trans-port in olive we characterized cDNAs from Olea euro-paea cv “Leccino” related to the aquaporin (AQP)gene family. A phylogenetic analysis of the corre-sponding polypeptides conWrmed that they were partof water channel proteins localized in the plasma mem-brane and in the tonoplast. The full-length sequenceswere obtained by RACE-PCR and were namedOePIP1.1, OePIP2.1 and OeTIP1.1. The OePIP2.1 andOeTIP1.1 encode functional water channel proteins, asindicated by expression assays in Xenopus laevisoocytes. OePIP1.1 and OePIP2.1 expression levels arehigh in roots and twigs and low in leaves. The highesthybridization signal of OeTIP1.1 was detected in twigs,while in roots and leaves the expression was low. Toinvestigate the eVect of abiotic stress on the transcriptlevel of olive AQP genes, olive trees were subjected todrought treatment and the expression levels of thegenes were measured by Northern-blot analysis. Thetranscript levels of each gene diminished strongly inplants submitted to drought stress, when soil moisture,twig water potential and twig hydraulic conductivity
progressively decreased. The downregulation of AQPgenes may result in reduced membrane water perme-ability and may limit loss of cellular water during peri-ods of water stress. A possible role for AQPs on shootembolism repair is discussed.
Keywords Aquaporin · Drought stress · Expression · Olive · Water channel
AbbreviationsAQP AquaporinMIP Major intrinsic proteinPIP Plasma membrane intrinsic proteinTIP Tonoplast intrinsic proteinPf Osmotic permeability coeYcientPCR Polymerase chain reactionKss SpeciWc twig hydraulic conductivity
Introduction
Water Xux across membranes has been shown to occurnot only through the lipid bilayer, but also throughaquaporins (AQPs). They belong to a highly conservedgroup of membrane major intrinsic proteins (MIPs)with molecular masses of 26–30 kDa. AQPs are foundin the cell membranes of all living organisms includingfungi, bacteria, plants and animals, and are tetramersof proteins each containing six membrane-spanning�-helices with both N- and C-termini located on the cyto-plasmic side of the membrane (Tyerman et al. 2002).AQPs can increase the osmotic hydraulic conductivityof the membrane by 10- to 20-fold (Preston et al. 1992).
The role of plant AQPs in water transport acrossmembranes has been proven by their expression in
F. Secchi (&) · C. Lovisolo · A. SchubertDepartment of Arboriculture and Pomology,University of Turin, Via Leonardo da Vinci 44, Grugliasco, 10095 Turin, Italye-mail: [email protected]
N. Uehlein · R. KaldenhoVApplied Plant Sciences, Institute of Botany, Darmstadt University of Technology,Schnittspahnstr 10, 64287 Darmstadt, Germany
123
382 Planta (2007) 225:381–392
Xenopus laevis oocytes. Based on their transport prop-erties, MIPs may be divided into two groups: AQPsand transporters of glycerol and other small neutralmolecules (Johansson et al. 2000). In plants, AQPs areclassiWed in diVerent groups according to theirsequence identity or to their structural features (e.g.sizes or N- and C-termini). The main groups includeplasma membrane intrinsic proteins (PIPs), tonoplastintrinsic proteins (TIPs), NOD26-like intrinsic proteins(NIPs: NOD26, the Wrst identiWed NIP, was found inthe symbiosome membrane surrounding nitrogen-Wxing bacteria of soybean), and small basic intrinsicproteins, a recently formed group which includes vari-ous AQPs whose subcellular localization is generallyunknown (Johanson and Gustavsson 2002). Plantgenomes include a large number of MIP genes; in Ara-bidopsis thaliana 35 diVerent AQP transcripts arefound (Johanson et al. 2001), the genome of Zea maysencodes 33 AQP homologues (Chaumont et al. 2001)and in Oryza sativa 33 like AQP-like genes have beenidentiWed (Sakurai et al. 2005).
Aquaporin activity can be regulated at the transcrip-tional and post-transcriptional levels. Some AQPs areconstitutively expressed (Johansson et al. 1996), whilethe expression of others is regulated by diVerent stim-uli, such as developmental stage, hormones (abscisicacid or gibberellic acid), or by adverse environmentalconditions such as drought and salinity (Vera-Estrellaet al. 2004). Post-transcriptional regulation occurs viaphosphorylation at one or multiple sites on speciWcPIPs, TIPs and NIPs (Baiges et al. 2002); for instance,the water transport activity of the PIP2 memberPM28A in spinach is regulated by phosphorylation oftwo serine residues (Johansson et al. 1998).
Water stress has a strong inXuence on AQP geneexpression and both up- and downregulation havebeen reported (Baiges et al. 2002). AQP upregulationis thought to increase membrane permeability to watertransport when water is less available (Yamada et al.1997). On the contrary, downregulation of AQP geneexpression may encourage cellular water conservationduring periods of water stress (Li et al. 2000; Smartet al. 2001). Probably, in order to maintain a suitablewater status under abiotic stress, both increased watertransport via AQPs in some tissues and reduced watertransport in other tissues are required (Jang et al.2004).
The olive originated in the eastern Mediterraneanarea, and has been cultivated since ancient times. Treesare persistent (up to 1,000 years) evergreens which tol-erate drought and salinity. It is generally cultivated inareas where water is the main limiting factor in agricul-tural production. Olive is an attractive ornamental
plant, produces table fruits and oil of high nutritionalvalue. “Leccino” is one of the most widespread culti-vars for oil production in Italy. Currently more than9,500,000 ha of olive tree groves are grown in the worldand about 2,500,000 t of olive oil are produced peryear.
Despite the adaptation of this plant to dry climates,the molecular bases of water transport in olive treeshave not been studied up to now. In this paper wereport on the isolation, characterization and expressionpattern of three olive AQP-encoding genes and wepresent data showing that in Olea europaea L., a plantthat is well-adapted to arid climates, AQP gene expres-sion is downregulated under water stress.
Materials and methods
Plant materials and drought treatments
Twelve 2-year-old plants of O. europaea cv “Leccino”(Vivai Coppo, Grugliasco, Italy) were grown in thegreenhouse. Each plant grew in a 30-l container Wlledwith a substrate composed of a sandy-loam soil/expanded clay/peat mixture (3:1:2) without fertilizers.
Plants were irrigated every third day to containerpotential (about 3 l water in the period of the experi-ment). A diVerential water treatment was imposed fora period of 4 weeks starting from 3rd of May 2005.During this period, part of the plants were watered asabove described (control plants), while the other (WS)plants were subjected to drought stress by withholdingwater.
Roots (apical 5 cm), adult leaves close to the apexand twigs (3-cm long apical portions of non-ligniWedshoots) were collected from control and WS plantsafter 3 (C1 and WS1 treatments) and 4 (C2 and WS2treatments) weeks from start of the diVerential watertreatment. All tissues were harvested at the same timeof the day (10 a.m.) to avoid eVects of diurnal Xuctua-tions of AQP expression and the samples were frozenimmediately in liquid nitrogen.
Nucleic-acid isolation
RNA was isolated from leaves, roots and twigs, accord-ing to the protocol of Chang et al. (1993). In order toremove contaminant DNA from the RNA samples, thenucleic acid extract was treated with DNaseI (Gibco),according to the manufacturer’s instructions. The con-centration of RNA was quantiWed by measuring theabsorbance at 260 nm and its integrity was checked onagarose gels.
123
Planta (2007) 225:381–392 383
Isolation of sequences encoding partial putative aquaporins
An RT-PCR strategy was used to identify partial AQPtranscripts. cDNA was synthesized from 5 �g totalRNA of roots and leaves using oligo-(dT)12–18 asprimer and Moloney murine leukaemia virus reversetranscriptase (Gibco-BRL).
Two pairs of primers were used for ampliWcation:gene-speciWc primers for PIP-type AQPs (PIP-S andPIP-AS) were designed based on AtPIP1.2 sequence(GenBank Accession No. NM_130159); degenerateprimers for TIP-type AQPs (TIP-S and TIP-AS) weredesigned on conserved regions of AtTIP sequences(Table 1).
PCR was performed using AmpliTaq Gold polymer-ase (Applied Biosystems) and according to the manu-facturer’s instructions. PCR cycles were: 1 cycle of95°C for 9 min, 35 cycles of 94°C for 45 s; 55 and 52°C,respectively, for PIP and TIP, for 45 s; 72°C for 1 min;the Wnal step at 72°C was extended for 7 min. TheampliWcation products were subjected to agarose gelelectrophoresis and stained with ethidium bromide.DNA from bands of the expected product length werecloned by ligation into the pGEM-T Easy vector (Pro-mega) and the product was used to transform Escheri-chia coli JM109 High EYciency Competent Cells(Promega).
The transformed colonies were puriWed using a stan-dard alkaline-lysis method. Three puriWed plasmidswere obtained (pPIP1, pPIP2 and pTIP); they weresequenced and the sequence information was thenused to design the primers for obtaining the full-lengthgenes.
5� and 3� RACE and isolation of full-length genes
Rapid ampliWcation of cDNA ends (RACE-PCR) wasperformed to obtain the 5� and 3� ends of the putativeAQPs. The poly(A) RNA was isolated from totalRNA of leaves and of roots using Genelute mRNAminiprep kit (Sigma).
For the PIP1 sequence, the 5� and 3� Wrst strandcDNAs were obtained and ampliWed using the BD
SMART RACE cDNA ampliWcation kit (BD Bio-sciences). cDNA was synthesized from 1 �g of poly(A)RNA according to the manufacturer’s protocol.AmpliWcation was done in the following conditions: 5cycles at 94°C for 30 s, and at 72°C for 3 min; 5 cycles at94°C for 30 s, 70°C for 30 s and 3 min at 72°C followedby 25 cycles at 94°C for 30 s, 68°C for 30 s and 72°C for3 min. The primers used in addition to the adapter-spe-ciWc primers (Universal Primer Mix, UPM) were PIP1+and PIP1¡. For PIP2 and TIP 5� RACE, the Wrststrand cDNAs were synthesized as above described.The primers used were TIP+ and UPM. The PCR reac-tion of PIP2 cDNA was carried out as follows: 25 cyclesat 94°C for 30 s, 30 s at 62°C, 3 min at 72°C followed byone cycle at 72°C for 5 min, with UPM and PIP2+primers (Table 2).
PIP2 and TIP 3� RACE was performed by RT-PCRfollowing the protocol described by Frohman et al.(1988). 3� end ampliWcations of the two cDNAs weredone with an adapter primer (B25) and a speciWcprimer (TIP¡ or PIP2¡) and with the following PCRprogram: 1 cycle of 95°C for 10 min, 40 cycles of 94°Cfor 40 s, 55 and 65°C, respectively, for PIP2 and TIP,for 2 min, and 72°C for 3 min, followed by a 15 minWnal extension at 72°C.
The 3� and 5� RACE fragments of each gene werecloned into separate pGEM-T Easy plasmid vectors(Promega). The cloned fragments were sequenced andtwo new primers, each carrying a restriction enzymesite at his 5� end, were designed on the ends of the cod-ing region. These primers and the Wrst strand cDNAswere utilized to obtain the full-length sequences bylong-distance PCR (LD-PCR), using Advantage™ 2polymerase mix (BD Biosciences) with the followingprogram: 1 cycle to 94°C for 2 min, then 35 cycles, eachconsisting of 30 s at 94°C, 30 s at 55°C, 60 s at 72°C.The primers used were: OePIP1s and OePIP1as withleaf cDNA as template for the full-length PIP1 gene(OePIP1.1); OePIP2s and OePIP2as with root cDNAas template for the PIP2 gene (OePIP2.1). To obtainthe complete sequence of the TIP gene (OeTIP1.1),the primers OeTIPs and OeTIPas, and root cDNA astemplate were used (Table 2). The ampliWed sequenceswere subcloned into a pGEM-T Easy plasmid vector.
The full-length nucleotide sequences of the threeolive AQPs were determined and submitted to Gen-Bank [accession numbers—DQ202708 (OePIP1.1),DQ202709 (OePIP2.1), DQ202710 (OeTIP1.1)].
Northern-blot analysis
SpeciWc DNA probes corresponding to the full-length olive AQP sequences were generated by PCR
Table 1 Primers used for PCR ampliWcation of putative MIPsequences
Primer code DNA sequence
PIP-S 5�-actacaaagagccaccacctgcgcc-3�PIP-AS 5�-tgaaccaagaacacagcgaatccgatagg-3�TIP-S 5�-acatctcygghcaygtkaaccc-3�TIP-AS 5�-gnccrrcccagtagayccagt-3�
123
384 Planta (2007) 225:381–392
using DIG probe synthesis kit (Roche), with theprimers used for LD-PCR (see above) and cDNA astemplate. Total RNA (10 �g) was separated on 1%formaldehyde-agarose gel electrophoresis and pho-tographed prior to being transferred overnight to apositively charged nylon membrane (Roche) by cap-illary blotting in 20£ SSC. The integrity and theequal amounts of RNA loading were conWrmed byethidium bromide staining. After transfer, nucleicacid was Wxed on the membrane by baking at 120°Cfor 30 min and by UV crosslinking for 4 min (2 minfor each side of the membrane). Northern blots wereprehybridized in the appropriate volume of Dig EasyHyb buVer (Roche) at 50°C for 1 h with gentle agita-tion. Hybridizations were carried out in the samesolution at 50°C overnight with addition of the spe-ciWc probe denatured by boiling for 5 min and rapidlycooling in ice.
After hybridization, the membranes were washedtwice in 2£ SSC, 0.1% SDS for 5 min at room tempera-ture, and twice under high stringency conditions in0.1£ SSC, 0.1% SDS for 15 min at 68°C. Detection ofhybridization signals was performed using the DIGhigh prime DNA labelling and detection starter kit II(Roche). The Wlters were exposed to X-ray Wlms(Kodak) and were developed after an appropriateperiod of time with GBX Wxer and developer solution(Kodak).
Probes were removed from the membranes withthree washings in 0.1% SDS solution for 10 min at100°C. Membranes were then washed brieXy in 2£SSC and reprobed as described above.
Sequence analysis
Database searches were done using the programTBLASTX (National Centre for Biotechnology Infor-mation, USA). Alignment of deduced amino acidsequences of OePIP1.1, OePIP2.1 and OeTIP1.1 withother MIP encoding sequences was obtained using theprogram Multalin. A phylogenetic neighbour-joininganalysis was conducted on deduced amino acidsequences using the program MEGA 3.1 (MolecularEvolutionary Genetics Analysis). The programTHMM (www.cbs.dtu.dk) followed by Wnally manualadjustments was used for the TM prediction, and theprogram ProSite (www.expasy.org) for the identiWca-tion of phosphorylation sites.
Expression analysis in Xenopus laevis
The LD-PCR primers used for ampliWcation ofOePIP1.1 and OePIP2.1 contained BamHI and XbaIrestriction sites in their 5� ends, and those used forampliWcation of OeTIP1.1 contained, respectively,EcoRI and a XbaI restriction sites. The plasmids weredigested with the appropriate restriction enzymes andthe inserts were cloned into the pGEM-HE vector(Liman et al. 1992), an expression vector carrying 5�
and 3� untranslated sequences of the �-globin genefrom X. laevis in order to promote translationeYciency of the plant cRNA. The assay was performedfollowing the method described by Biela et al. (1999).
A possible inhibition of AQP water transport by sul-phydryl reagents was examined by incubation of the
Table 2 List of primers used for RACE-PCR ampliWcation
S sense, AS antisensea Orientation of the oligonucleotide relative to corresponding amino acid sequence
Primer Code
DNA sequence Orientationa Use
P1P1+ 5�-CCA CTACCCTTGGTGTAACCGTGAGC-3� S P1P1 5�RACEP1P1¡ 5�-CTCATGGTCGTTCTACAGAGCTGGCATTG-3� AS P1P1 3�RACEP1P2+ 5�-AGGAAGTGGAGCCAAGACAGGGACAT-3� S P1P2 5�RACEP1P2¡ 5�-CGTTGCCACTTTGCTGTTCCTTTACG-3� AS P1P2 3�RACETIP+ 5�-AAAGACACGGCTGGGTTCATGGAT-3� S TIP 5�RACETIP¡ 5�-GGTGGACTCGAAACATCAGCCTTTG-3� AS TIP 3�RACEOeP1P1s 5�-AAAACGGGATCCCGAAAAGGAGAGCAAAGAGGAAGATGACG-3� S 5�ORF OeP1P1OeP1P1as 5�-AAAAGCTCTAGAGCGCTTATTTTTTGAATGGGATGGCTCG-3� AS 3�ORF OeP1P1OeP1P2s 5�-AAAACGGGATCCCGAGAATGACGAAAGACGTCGAATCTCA-3� S 5�ORF OeP1P2OeP1P2as 5�-AAAAGCTCTAGAGCATTCATGTGCAGTTTTCAAGGTCCTC AS 3�ORF OeP1P2OeTIPs 5�-AAAACGGGATCCCGAAAATGCCTATTTCAAGAGTCGC-3� S 5�ORF OeTIPOeTIPas 5�-AAAAGCTCTAAGAGCTCTTAGTATTCTGCACTGGTAGGGAG-3� AS 3�ORF OeTIPB25 5�-GACTCGAGTCGACATCG-3� Primer 3� FrohmanUPM 5�-CTAATACGACTCACTATAGGGCAAGCAGTGGTATCAACGCAGAGT3�
5�CTAATACGACTCACTATAGGGC-3�Universal Primer Mix
123
Planta (2007) 225:381–392 385
oocytes for 10 min in a modiWed incubation solutioncontaining 0.3 mM HgCl2. The assay was thereafterdone using an incubation solution without HgCl2.
Measurements of physiological parameters
Soil moisture was assessed gravimetrically. Waterpotential measurements were performed on ten repli-cate apical twigs (each having ten leaves) using a Scho-lander-type pressure chamber (Soil MoistureEquipment Corp., Santa Barbara, CA, USA). Hydrau-lic conductivity of the twig was measured on tenreplicate detached twigs through a controlled tension-pressure apparatus according to Lovisolo et al. (2002).About 20-cm-long (20.6 § 0.02) twig portions withleaves were detached from the tree underwater, keep-ing the twig submerged in a plastic basin. The basal endof the portion was transferred into a pressure chamberWlled with tap water, avoiding air contact to the cutbasal surface during transfer. The apical end wasclamped with a rubber sleeve. First, a measurement of“undisturbed” hydraulic conductivity was taken byapplying an apical suction (¡80 kPa) for 5 min throughthe sleeve to the cut twig portion apical end (Kolb et al.1996). Following the measurement under tension, andleaving the same twig portion in the chamber, a basalpressure (+200 kPa) was applied for 5 min to the waterin the chamber to Xush out xylem embolisms (Lo Gulloet al. 2003), and hydraulic conductivity was then mea-sured again upon a ¡80 kPa tension, as describedabove. Before each conductivity measurement, a¡80 kPa tension was applied for 10 min in order to sta-bilize the measurement. SpeciWc conductivity (Kss) wascalculated dividing hydraulic conductivity by the cross-sectional area of the twig measured in the middle of theportion.
Results
cDNA cloning of Olea europaea aquaporins
Putative AQP-encoding cDNAs of O. europaea wereobtained by RT-PCR: three diVerent cDNA bands,respectively, of 580, 598 and 441 bp were isolated andsequenced. The 580-bp sequence was isolated from leafRNA, while the other two bands were obtained fromroot RNA. A comparison of the encoded amino acidsequences of the cDNAs with those of plant AQPs(using the program TBLASTX) showed a high (>80%)identity. The plant PIP group is divided into twogroups (PIP1 and PIP2), which are characterized byspeciWc amino acid residues at the N- and C-terminals
and around the conserved NPA motifs (SchaVner1998). Based on homology analysis the olive 580 bpsequence bears high similarity to PIP1 genes, while the598 bp sequence is intermediate between PIP1 andPIP2 genes. The band of 441 bp shows homology to theTIP group.
Using RACE, the 5� and 3� terminals of the threepartial sequences were cloned and the correspondingfull-length genes were isolated by LD-PCR.
The nucleotide sequence of the Wrst gene, namedOePIP1.1, contains an open reading frame of 858 bp, a5�-UTR of 17 bp and a 3� non-coding region of 277 bp(GenBank Accession No. DQ202708). This cDNAsequence has 96% identity with PIP1.1 of Fraxinusexcelsior (AAT74898).
The other two genes, named OePIP2.1 (GenBankAccession No. DQ202709) and OeTIP1.1 (GenBankAccession No. DQ202710), are, respectively, 1,246 and1,070 bp long. OePIP2.1 contains a 5�-UTR of 67 bpand a 3�-UTR of 315 bp, while OeTIP1.1 has a 5�-UTRof 95 bp and a 3�-UTR of 36 bp. The OePIP2.1sequence shows 90% identity with MipC of Mesembry-anthemum crystallinum (AAB18227) and the OeTIP1.1sequence is most similar to TIP1.2 of Mimosa pudica(82% identity; BAD90703).
A phylogenetic analysis of 48 plant AQPs includingmembers of the PIP1, PIP2 and TIP subfamilies con-Wrmed that two of the full-length sequences isolatedare plasma membrane proteins (PIP); OePIP1.1 clus-ters with the PIP1 subfamily, OePIP2.1 with the PIP2subfamily; the other protein (OeTIP1.1) belongs to theTIP1-group of the TIP subfamily (Fig. 1).
The amino acid sequences of the three full-lengthcDNAs (Fig. 2) included: the two NPA motifs, typicalfeatures of all AQP proteins (Chrispeels and Maurel1994); three amino acids residues involved in the con-striction region of the pore (Sui et al. 2001); two con-served regions speciWc of the PIP subfamily located inloop C and in loop E (Barone et al. 1997); His residuesinvolved in pH sensing (Luu and Maurel 2005) and Serresidues, possible targets for phosphorylation (Johans-son et al. 1998).
Water channel activities of olive aquaporins
The biological activity of putative AQPs was deter-mined using the X. laevis oocyte-swelling assay. cRNAtranscribed from the full-length cDNA clones wasmicroinjected into the cytoplasm of the oocytes. Threedays after injection, the oocytes were transferred fromisotonic to hypotonic medium. The change in volumewas used to calculate the relative water permeabilitycoeYcient (Pf).
123
386 Planta (2007) 225:381–392
Control oocytes (water-injected) showed a Pf of12.8 § 0.78 cm s¡1 £ 10¡4 (n = 8). The Pf of oocytesinjected with OePIP1.1 cRNA had similar values as incontrols (11.0 § 3.31 cm s¡1£ 10¡4; n = 7); whereas thePf of oocytes injected, respectively, with OePIP2.1cRNA and OeTIP1.1 cRNA was, respectively, 40- and24-fold higher than in controls (476 § 36.2 cm s¡1
£ 10¡4; n = 8, and 285 § 17.9 cm s¡1£ 10¡4; n = 10;Fig. 3).
The sensitivity to mercury, a non-speciWc inhibitorof water channels, was also assessed. Incubation ofOePIP2.1-expressing oocytes with HgCl2 reduced thevalues of water permeability by about 2.5-fold. On thecontrary, incubation of OePIP1.1- and OeTIP.1-expressing oocytes with HgCl2 did not inhibit theirosmotic water permeability, suggesting that these twoproteins are not sensitive to mercurials (Fig. 4).
Tissue-speciWc expression of olive aquaporin genes
The accumulation of AQP transcripts in diVerent vege-tative tissues of olive was tested by Northern-blot anal-ysis. An 18S rRNA probe was used to check theuniformity and integrity of mRNA in the diVerentlanes. The expression proWles varied with the diVerentAQP genes.
Transcripts from OePIP1.1 accumulated to high lev-els in roots and twigs with lower levels in leaves.OePIP2.1 mRNAs were very abundant in roots, theirexpression was moderate in twigs, while the mRNAswere detectable at lower levels in leaf extracts. Theaccumulation of OeTIP1.1 mRNA was detected intwigs, while in roots and leaves the expression was low(Fig. 5).
EVect of drought stress on expression of olive aquaporin genes
To investigate the eVect of abiotic stress on the tran-script levels of AQP genes, the olive trees were gradu-ally drought-stressed by withholding irrigation for aperiod of 4 weeks. Then the expression levels of AQPgenes in leaves, roots and twigs of well-watered andstressed plants were measured by Northern-blot analy-sis. 18S RNA was probed in order to ensure equal laneloading in this experiment.
The results showed signiWcant diVerences in tran-script levels under drought stress, with similar expres-sion patterns for the diVerent AQP genes and diVerenttissues. In control plants, transcript abundance of theanalysed AQP-genes conWrmed the diVerential expres-sion in diVerent plant organs. Accumulation ofmRNAs was similar after 3 and 4 weeks from start of
Fig. 1 Phylogenetic analysis of the 48 plant MIPs. Protein se-quences of AQP-encoding genes were aligned using the programMEGA 3.1. The number next to nodes are bootstrap values from2,000 replicates. The length of each branch is proportional to thedivergence of the protein sequence from other members of thefamily. The black dots indicated the olive OeMIP that we identi-Wed. The sequence sources were—M. pudica, MpTIP1.2(BAD90703), MpTIP1.1 (BAD90702), MpPIP2.3 (BAD90699);O. europaea, OeTIP1.1 (DQ202710), OePIP1.1 (DQ202708), Oe-PIP2.1 (DQ202709); Vitis vinifera, VvTIP1.1 (AAW02943),VvPIP2.1 (AAV69744); Vitis berlandieri £ Vitis rupestris,VbxVrTIP3 (AAF78757), VbxVrPIP1.1 (AAF71817); Ricinuscommunis, RcTIP1.1 (CAE53881), RcPIP2.1 (CAE53883); Bras-sica oleracea, BoTIP (AAB51393), BoPIP1.b (AAG23179), Bo-PIP3 (AAG30607); A. thaliana, AtTIP1.1 (P25818), AtTIP1.2(Q41963), AtTIP2.1 (AAM65406), AtTIP4.1 (AAC42249), At-TIP5.1 (CAB51216), AtTIP3.1 (P26587), AtPIP1.3 (AAP13421),AtPIP1.1 (P61837), AtPIP2.2 (AAM63463), AtPIP2.1(AAM65406), AtPIP2.7 (P93004); M. crystallinum, McTIP(AAB17284), McMipc (AAB18227); O. sativa, OsTIP(BAB63833), OsPIP1.a (CAA11896); Z. mays, ZmTIP1.2(AAK26767), ZmPIP1.2 (AAD29676), ZmPIP1.5 (AAK26756),ZmPIP2.3 (AAK26760), ZmPIP2.1 (AAK26758); F. excelsior,FePIP1.1 (AAT74898); Samanea saman, SsAQP1 (AAC17528);Medicago truncatula, MtPIP1.1 (AAK66766); Hordeum vulgare,HvPIP1.5 (BAA23746); Spinacea oleracea (SoPIP2.1); Raphanussativus, RsPIP2.c (BAA92261); Populus tremula £ Populus tre-muloides, PttPIP2.5 (CAH60724), PttPIP2.4 (CAH60723), Ptt-PIP2.1 (CAH60720); Juglans regia, JuPIP2.1 (AAO39007); Pyruscommunis, PcPIP2.1 (BAB40141); Picea abies, PaPIP2.1(T14889), PaPIP1.1 (T144863)
123
Planta (2007) 225:381–392 387
the experiment. In drought-stressed plants, mRNAabundance of the three AQP genes markedlydecreased after 3 weeks and continued decreasing tothe second sampling point. In the experiment, tran-script abundance was very low, while 18S rRNA abun-dance was similar in irrigated and drought-stressedplants (Fig. 6).
EVects of drought stress on water relations and twig water conductivity of olive plants
The twig water potential of control plants after 3 weeks(C1) and 4 weeks (C2) was ¡0.40 § 0.02 and¡0.65 § 0.07 MPa, respectively; the soil moisture was39.4 § 0.08% for C1 plants and 33.9 § 0.03% for C2
plants. Drought stress markedly aVected soil moisturecontent, which dropped to 5.48 § 0.01% after 3 weeks(WS1 plants) and to 3.91 § 0.01% after 4 weeks (WS2plants). In drought-stressed plants, at the moment ofsample collection, leaf turgor was visibly lower than incontrol plants, as twig water potential decreased to¡1.13 § 0.07 MPa in WS1 plants and to¡3.01 § 0.23 MPa in WS2 plants (Fig. 7).
Drought stress, as expected, negatively aVected twighydraulic conductivity. In well-watered, control plants,twig speciWc conductivity was 0.24 § 0.03 kg s¡1
MPa¡1 m¡1 after 3 weeks (C1 plants) and 0.26 § 0.07kg s¡1 MPa¡1 m¡1 after 4 weeks (C2 plants). In plantsexposed to drought stress for 3 weeks (WS1), the twigspeciWc conductivity decreased to 0.17 § 0.03 kg s¡1
Fig. 2 Alignment of the deduced amino acid sequences of Oe-PIP1.1 (DQ202708), OePIP2.1 (DQ202709) and OeTIP1.1(DQ202710) from O. europaea with three A. thaliana AQPs—At-PIP1.1 (P61837), AtPIP2.2 (AAM63463) and AtTIP1.1 (P25818).The sequences were aligned using the ClustalW program. Aminoacid numbers are shown on the right. Bold blue indicates the res-idues that are identical in all sequences analysed, the program Se-quence Date Explorer (MEGA 3.1) was used for searching theconserved amino acids. Underlined residues are conserved in all
MIP sequences (adapted from Zardoya et al. 2005). Boxed re-gions are predicted to form transmembrane domains (TM1–6),the NPA motifs are shaded with yellow and two motifs conservedonly in PIP family are shaded with light grey. Purple boxes indi-cate discriminating amino acids involving in the formation of con-striction region speciWc for the water transport (adapted from Suiet al. 2001). Residues of serine shaded with green represent possi-bly sites of phosphorylation while residues of histidine involved inpH sensing are shaded with light blue
123
388 Planta (2007) 225:381–392
MPa¡1 m¡1. One further week of drought stressreduced twig speciWc conductivity to 0.12 § 0.03 kg s¡1
MPa¡1 m¡1.After embolism removal (Xushing), twig speciWc
conductivity did not diVer among treatments and peri-ods, suggesting that reductions in twig water conductiv-ity were caused by embolism during water stress(Fig. 8).
Discussion
The molecular bases of water transport in olive treeshave not been studied. Many genes encoding AQPswere isolated from diVerent plant species but up tonow no genes have been identiWed from O. europaea.We isolated three AQPs from olive plants, includingone TIP and two PIPs.
Amino acid sequence comparisons with diVerentplant AQPs show that OeTIP1.1 bears sequence simi-larity with cDNAs of the TIP family, while OePIP1.1and OePIP2.1 show sequence similarities, respectively,with members of the PIP1 and PIP2 families. SpeciWcamino acid residues located in the loop C and in theloop E, in the second and in the third apoplastic loops,respectively, allow to distinguish the PIPs from theTIPs family. All plant plasma membrane AQPs,including OePIPs, presented two highly conservedregions, one in the loop C: G-G-G-A-N-X-X-X-X-G-Yand other in loop E: T-G-I/T-N-P-A-R-S-L/F-G-A-A-I/V-I/V-F/Y-N; this motifs are absent in all TIPs familyand also in OeTIP1.1 (Barone et al. 1997).
As described for PIP1 proteins (Chaumont et al.2000), OePIP1.1 showed a longer amino-terminalextension and a shorter carboxy-terminal end com-pared with the PIP2 subfamily proteins, includingOePIP2.1. Both PIP genes have an extended amino-terminal compared with OeTIP1.1 (SchaVner 1998).
In human AQP1 protein four residues (Phe 58, His182, Cys 191, Arg 197) deWned the constriction regionof the channel pore, and three of these are conservedacross the water-speciWc AQPs. The location of theseresidues is essential for deWning the selectivity to wateror to others solutes as glycerol (Sui et al. 2001). TheOeAQPs sequences contain the amino acids speciWcfor the water constriction region: the Wrst residue (Phe)is located in the middle of TM2 domain, an His aminoacid is conserved in the TM5 domain and the loop Econtain a Tyr residue corresponding to AQP1-Cys 191and an Arg amino acid located behind the second NPAmotif. OeTIP1.1 contains an His in the Wrst positionwhile in the other two positions unrelated amino acidsare present.
The activity of plasma membrane AQPs may beaVected by cytosolic pH. A conserved His residue inthe intracellular loop D seems to control this eVect, asshown by the reduced eVects of cytosol acidiWcation
Fig. 3 Assay of water permeability in X. laevis oocytes. Oocyteswere separately injected with 50 ng of three olive AQP cRNAs orwith the same volume of water (control). Relative volume wasplotted against time. OePIP2.1, Wlled squares; OePIP1.1, emptycircles; OeTIP1.1, Wlled triangles; H2O, asterisks. Values aremeans § SE. OePIP2.1, n = 8; OePIP1.1, n = 7; OeTIP1.1, n = 10;H2O, n = 8
Fig. 4 Values of Pf, obtained from volume change measure-ments, of oocytes injected with water or with three olive AQP cRNAs. The assay was performed with a 10-min preincubation in absence (empty columns) or in pres-ence (Wlled columns) of 0.3 mM HgCl2. The Pf values are expressed as means § SE and the number of replicates is indicated above the bars
123
Planta (2007) 225:381–392 389
when this amino acid is substituted by an alanine(Tournaire-Roux et al. 2003). Both olive AQPs iso-lated in this study contain this His residue (OePIP1.1-His207, OePIP2.1-His201), suggesting that also theseolive proteins may sense pH.
Aquaporin phosphorylation is a common post-tran-scriptional modiWcation and probably can enhance thein vivo membrane water permeability (Johansson et al.1996). Experiments with oocytes of X. laevis demon-strated that phosphorylation can increase water chan-nel activity (Chaumont et al. 2005). In the AQPSoPIP2.1 (formerly called PM28A), two Serine resi-dues, Ser 115 located in the Wrst cytoplasmatic loop andSer 274 in the C-terminus can be reversibly phosphory-lated (Johansson et al. 1998). The molecular gatingmechanism is proposed by Törnroth-HorseWeld et al.(2005): the AQP SoPIP2.1 is in the open stage whenphosphorylated on two conserved serine residues andin the close stage when Ser 115 and Ser 274 are dephos-phorylated. The dephosphorylation occurred duringconditions of drought stress.
OePIP1.1 and OePIP2.1 proteins show a serine resi-due in the loop B (OePIP1.1-Ser 129 and OePIP2.1-Ser123), these amino acids are located in domains whichcorrespond to the recognition sequences of proteinkinase A (Arg-Lys-X-Ser), of protein kinase C (Arg-Lys-X-Ser-X-X-Arg), and of calmodulin-like domainprotein kinases (Leu-X-Arg-X-X-Ser; Johansson et al.1998), thus could be in vivo targets for phosphoryla-tion. Another potential target of phosphorylation in
OePIP2.1 is a serine (Ser 282) located in a cAMPdependent protein kinase site (Ser-X-Arg) at the car-boxy end. Interestingly, bean Pv�TIP has diVerent sitesof phosphorylation: two serine residues located in theN-terminus (Ser 7–Ser 23) and one in the loop B (Ser99). Also ZmPIP1.2 has two potential phosphorylationsites: Ser 16 and Ser 131, phosphorylated, respectively,by protein kinases C and A (Chaumont et al. 2000).
The presence of AQPs in a membrane can increasethe osmotic hydraulic conductivity of the membrane(Preston et al. 1992; Siefritz et al. 2002). We tested thebiological activity of olive AQPs using X. laevis oocyte-swelling assay. This proved that OePIP2.1 andOeTIP1.1 function very eYciently as water channelswhen expressed in oocytes, while OePIP1.1 has nowater transport activity in oocytes. Proteins belongingto the plant PIP1 family often show no or very limitedwater channel activity in X. laevis oocytes (Chaumontet al. 2000).
Fig. 5 Expression pattern of three olive AQP genes in diVerenttissues of “Leccino” plants as assessed by Northern hybridization.Each lane contained 10 �g of RNA isolated from roots, leaves andtwigs. The membranes were hybridized with probes speciWc toOeTIP1.1, OePIP1.1 and OePIP2.1 and to a constitutively ex-pressed gene encoding 18S rRNA
Fig. 6 Expression patterns of three olive AQP genes in “Lec-cino” plants under irrigation or drought stress, as assessed byNorthern hybridization. Each lane contained 10 �g of RNA iso-lated from diVerent tissues of plant. Filters were also hybridizedto an 18S rRNA probe as control
123
390 Planta (2007) 225:381–392
Mercury chloride inhibits water movement acrossbiological membranes and is a known inhibitor of AQPactivity (Hukin et al. 2002). However, not all AQPs aresensitive to mercurial compounds (Hukin et al. 2002;Moshelion et al. 2002). Our results show that watertransport through OePIP2.1 is mercury-sensitive andOeTIP1.1 is not sensitive to mercury. The sensitivity tomercury has been traced to speciWc cysteine residues inplant AQPs. In AtTIP2.1 and AtTIP1.1, respectively,Cys 116 and Cys 118, located in transmembranedomain 3, confer the sensitivity to mercury. These resi-dues are conserved in all Arabidopsis TIPs but arefound also in PIPs (Daniels et al. 1996). In domain 3 ofOePIP2.1 no cysteine residues are found in these posi-
tions, while other Cys residues are present in the pro-tein. Moreover, other mechanisms are probablyinvolved in Hg sensitivity: Barone et al. (1997), showedthat the addition of mercury induced a change in pro-tein conformation of beet PMIP31 resulting in theexposure of a proteolytic cleavage site (G-G-G-A-N)located in the loop C.
Aquaporins in plants often show a tissue/organ-spe-ciWc expression. For example �TIP of Phaseolus vulga-ris is seed-speciWc (Johansson et al. 1998), McMipB isroot-speciWc (Yamada et al. 1995), AtTIP3.1 isexpressed in seed storage vacuoles (Ludevid et al. 1992)and AtTIP2.1 is present in developing vascular tissuesof the shoots but not in roots (Daniels et al. 1996). Inolive, OePIP1.1 is expressed at a similar level in all tis-sues examined while OePIP2.1 is predominantlyexpressed in roots. More recent and complete studies ofAtPIPs does support that PIP2s are preferentiallyexpressed in roots compared to, e.g. PIP1s (Jang et al.2004; Alexandersson et al. 2005). On the contrary,higher levels of OeTIP1.1 are detected in shoots than inroots and in leaves. OeTIP1.1 expression in shoot tis-sues is probably related to mechanisms of osmoregula-tion as reviewed by Luu and Maurel (2005). Salinityand drought-induced osmoregulation of olive is knownin leaves and shoots (Brito et al. 2003). Moreover, oliveshoots are known to be very sensitive to drought, show-ing wide trunk diameter Xuctuations (TDF) (Jones2004). TDFs need fast water inXux and eZux from cellto apoplast and inside the cell from cytoplasm to vacu-ole, implying expression and even gating of AQPs inboth plasmalemma and tonoplast.
The regulation of the expression of AQP genes mayplay an important role in a plant highly tolerant todrought like olive. In our experiment we applied asevere water stress which progressively lowered waterpotential to less than ¡3 MPa in apical shoot (twig)portions. This is comparable to the water potentialsmeasured in olive plants under water stress in the Weld(Tognetti et al. 2004). The hydraulic conductivity oftwigs in olive is an important physiological parameter:in a related species (Olea oleaster) it plays dominantrole on the control of total plant water conductivityand it controls stomatal opening in conditions ofdrought stress (Lo Gullo et al. 2003). Application ofdrought stress predictably reduced twig water conduc-tivity by induction of xylem embolization, as previouslyshown in the same plant in similar conditions (LoGullo et al. 2003). This can be considered a negativeeVect of water stress, but can also play the role of pro-tecting the plant from excessive water loss in condi-tions of severe water stress and intense transpirationrequirements.
Fig. 7 Twig water potential (MPa) and soil moisture content(gwater/gwet substrate) were measured on well-watered (C1 and C2)and drought-stressed plants (WS1 and WS2). C1 and WS1, 3weeks after start drought; C2 and WS2, 4 weeks after startdrought. White boxes represent soil moisture content, grey boxesshow twig water potential values. Values are means § SE, n = 10
Fig. 8 Twig hydraulic conductivity. Grey boxes—speciWc con-ductivity (Kss) measured on well-watered plants (C1 and C2) andon stressed plants (WS1 and WS2). White boxes—measurementsperformed on the same plants after imposing a pressure Xushingof +200 kPa to the twig to Xush out xylem embolisms. C1 andWS1, 3 weeks after start drought; C2 and WS2, 4 weeks after startdrought. Values are means § SE, n = 10
123
Planta (2007) 225:381–392 391
The expression of AQP genes in olive drasticallydecreased following water stress, both after 3 weeks(when stress was relatively mild) and after 4 weeks(when stress had progressed to severe). The reductionin expression concerned both PIPs and TIPs. TheeVects of drought treatment on expression of AQPgenes have been studied in various plants and downre-gulation has been frequently observed. PIP and TIPgenes in Nicotiana glauca were downregulated inleaves, shoots and roots under drought stress (Smartet al. 2001).
Downregulation of AQP genes is commonly associ-ated with losses in hydraulic conductance of plantorgans. Such observations were done in roots (Northet al. 2004) and their explanation is straightforward, asAQPs control radial water Xux in roots (Vandeleur et al.2005). In shoots the relationship between MIP expres-sion and hydraulic conductivity is less evident, in partic-ular as drought stress induces a strong increase in vesselembolization. However, embolisms can be repairedeven when vessels are under tension (Holbrook andZwieniecki 1999). ReWlling of embolisms involves theactivity of vessel-surrounding living cells (Salleo et al.2004), and it has been proposed that transmembranetransport of water from these cells to the vessels throughwater channels (AQPs) signiWcantly contributes to theprocess (Sakr et al. 2003). In our experiment, droughtdepressed twig AQP expression, and this may haveimpaired the ability of vessel-surrounding cells to con-tribute to embolism reWlling in xylem vessels.
Acknowledgments This study was Wnanced by the Italian Min-istry of Education and Research—PRIN MIUR “The control ofvegetative growth in olive (O. europaea L.) by the rootstock: eco-physiological, histo-anatomical and molecular aspects.” FS wasfunded by the Vigoni-DAAD program: “CO2 in plants: from theatmosphere towards chloroplasts. A coupled intercellular andtransmembrane pathway, involving aquaporins” for oocyte-swelling assays in Darmstadt. Our particular thanks are due to A.Carra for his advices on molecular biology and to S. Bragagnolofor his help with physiological measurements.
References
Alexandersson E, Fraysse L, Sjovall-Larsen S, Gustavsson S,Fellert M, Karlsson M, Johanson U, Kjellbom P (2005)Whole gene family expression and drought stress regulationof aquaporins. Plant Mol Biol 59:469–484
Baiges I, SchaVner AR, AVenzeller MJ, Mas A (2002) Plant aqu-aporins. Plant Physiol 115:175–182
Barone LM, Shih C, Wasserman P (1997) Mercury-induced con-formational changes and identiWcation of conserved surfaceloops in plasma membrane aquaporins from higher plants.J Biol Chem 272:30672–30677
Biela A, Grote K, Otto B, Hoth S, Hedrich R, KaldenhoV R(1999) The Nicotiana tabacum plasma membrane aquaporin
NtAQP1 is mercury-insensitive and permeable for glycerol.Plant J 18:565–570
Brito G, Costa A, Fonseca HMAC, Santos CV (2003) Responseof Olea europaea ssp. maderensis in vitro shoots exposed toosmotic stress. Sci Hortic 97:411–417
Chang S, Puryear J, Cairney J (1993) A simple and eYcient meth-od for isolating RNA from pine tree. Plant Mol Biol Rep11:113–116
Chaumont F, Barrieu F, Jung R, Chrispeels MJ (2000) Plasmamembrane intrinsic proteins from maize cluster in two se-quence subgroups with diVerential aquaporin activity. PlantPhysiol 122:1025–1034
Chaumont F, Barrieu F, Wojcik E, Chrispeels MJ, Jung R (2001)Aquaporins constitute a large and highly divergent proteinfamily in maize. Plant Physiol 125:1206–1215
Chaumont F, Moshelion M, Daniels MJ (2005) Regulation ofplant aquaporin activity. Biol Cell 97:749–764
Chrispeels MJ, Maurel C (1994) Aquaporins: the molecular basisof facilitated water movement through living plant cells?Plant Physiol 105:9–13
Daniels MJ, Chaumont F, Mirkov TE, Chrispeels J (1996) Char-acterization of a new vacuolar membrane aquaporin sensi-tive to mercury at a unique site. Plant Cell 8:587–599
Frohman MA, Dush MK, Martin GR (1988) Rapid production offull-length cDNAs from rare transcripts: ampliWcation usinga single gene-speciWc oligonucleotide primer. Biochemistry85:8998–9002
Holbrook NM, Zwieniecki MA (1999) Embolism repair and xy-lem tension: do we need a miracle? Plant Physiol 120:7–10
Hukin D, Doering-Saad C, Thomas CR, Pritchard J (2002) Sensi-tivity of cell hydraulic conductivity to mercury is coincidentwith symplasmic isolation and expression of plasmalemmaaquaporin genes in growing maize roots. Planta 215:1047–1056
Jang JK, Kim DG, Kim YO, Kim JS, Kang H (2004) An expres-sion analysis of a gene family encoding plasma membraneaquaporins in response to abiotic stresses in Arabidopsis tha-liana. Plant Mol Biol 54:713–725
Johanson U, Gustavsson S (2002) A new subfamily of majorintrinsic proteins in plants. Mol Biol Evol 19:456–461
Johanson U, Karlsson M, Johansson I, Gustavsson S, Sjovall S,Fraysse L, Weig AR, Kjellbom P (2001) The complete set ofgenes encoding major intrinsic proteins in Arabidopsis pro-vides a framework for a new nomenclature for major intrin-sic proteins in plants. Plant Physiol 126:1358–1369
Johansson I, Larsson C, Ek B, Kjellbom P (1996) The major inte-gral proteins of spinach leaf plasma membranes are putativeaquaporins and are phosphorylated in response to Ca2+ andapoplastic water potential. Plant Cell 8:1181–1191
Johansson I, Karlsson M, Shukla VK, Chrispeels MJ, Larsson C,Kjellbom P (1998) Water transport activity of the plasmamembrane aquaporin PM28A is regulated by phosphoryla-tion. Plant Cell 10:451–459
Johansson I, Karlsson M, Johanson U, Larsson C, Kjellbom P(2000) The role of aquaporins in cellular and whole plantwater balance. Biochim Biophys Acta 1465:324–342
Jones HG (2004) Irrigation scheduling: advantages and pitfalls ofplant-based methods. J Exp Bot 55:2427–2436
Kolb KJ, Sperry JS, Lamont BB (1996) A method for measuringxylem hydraulic conductance and embolism in entire rootand shoot systems. J Exp Bot 47:1805–1810
Li L, Li S, Tao Y, Kitagawa Y (2000) Molecular cloning of a novelwater channel from rice: its products expression in Xenopusoocytes and involvement in chilling tolerance. Plant Sci154:43–51
123
392 Planta (2007) 225:381–392
Liman ER, Tytgat J, Hess P (1992) Subunit stoichiometry of amammalian K+ channel determined by construction of mul-timeric cDNAs. Neuron 9:861–871
Lo Gullo MA, Salleo S, Rosso R, TriWlò P (2003) Drought resis-tance of 2-year-old saplings of Mediterranean forest trees inthe Weld: relations between water relations, hydraulics andproductivity. Plant Soil 250:259–272
Lovisolo C, Hartung W, Schubert A (2002) Whole-plant hydrau-lic conductance and root-to-shoot Xow of abscisic acid areindependently aVected by water stress in grapevines. FunctPlant Biol 29:1349–1356
Ludevid D, Hofte H, Himelblau E, Chrispeels MJ (1992) Theexpression pattern of the tonoplast intrinsic protein �TIP inArabidopsis thaliana is correlated with cell enlargement.Plant Physiol 100:1633–1639
Luu DT, Maurel C (2005) Aquaporins in a challenging environ-ment: molecular gears for adjusting plant water status. PlantCell Environ 28:85–96
Moshelion M, Becker D, Biela A, Uehlein N, Hedrich R, Otto B,Levi H, Moran N, KaldenhoV R (2002) Plasma membraneaquaporins in the motor cells of Samanea saman: diurnal andcircadian regulation. Plant Cell 14:727–739
North GB, Martre P, Nobel PS (2004) Aquaporins account forvariations in hydraulic conductance for metabolically activeroot regions of Agave deserti in wet, dry, and rewetted soil.Plant Cell Environ 27:219–228
Preston GM, Carroll TP, Guggino WB, Agree P (1992) Appear-ance of water channels in Xenopus oocytes expressing redcell CHIP28 protein. Science 256:385–387
Sakr S, Alves G, Morillon R, Maurel K, Decourteix M, GuilliotA, Fleurat-Lessard P, Julien JL, Chrispeels MJ (2003)Plasma membrane aquaporins are involved in winterembolism recovery in walnut tree. Plant Physiol 133:630–641
Sakurai J, Ishikawa F, Yamaguchi T, Uemura M, Maeshima M(2005) IdentiWcation of 33 rice aquaporin genes and analysisof their expression and function. Plant Cell Physiol l46:1568–1577
Salleo S, Lo Gullo MA, TriWo’ P, Nardini A (2004) New evidencefor a role of vessel-associated cells and phloem in the rapidxylem reWlling of cavitated stems of Laurus nobilis L. PlantCell Environ 27:1065–1076
SchaVner AR (1998) Aquaporin function, structure, and expres-sion: are there more surprises to surface in water relations?Planta 204:131–139
Siefritz F, Tyree MT, Lovisolo C, Schubert A, KaldenhoV R(2002) PIP1 plasma membrane aquaporins in tobacco: fromcellular eVects to function in plants. Plant Cell 14:869–876
Smart LB, Moskal WA, Cameron KD, Bennett AB (2001) MIPgenes are down-regulated under drought stress in Nicotianaglauca. Plant Cell Physiol 42:686–693
Sui H, Han BG, Lee JK, Wallan P, Jap B (2001) Structural basisof water-speciWc transport through the AQP1 water channel.Nature 414:872–878
Tognetti R, D’Andria R, Morelli G, Calandrelli D, Fragnito F(2004) Irrigation eVects on daily and seasonal variations oftrunk sap Xow and leaf water relations in olive trees. PlantSoil 263:249–264
Törnroth-HorseWeld S, Wang Y, Hedfalk K, Johanson U, KarlssonM, Tajkhorshid E, Neutze R, Kjellbom P (2005) Structuralmechanism of plant aquaporin gating. Nature 439:688–694
Tournaire-Roux C, Sutka M, Javot H, Gout E, Gerbeau P, LuuDT, Bligny R, Maurel C (2003) Cytosolic pH regulates rootwater transport during anoxic stress through gating of aqu-aporins. Nature 425:393–397
Tyerman SD, Niemietz CM, Bramley H (2002) Plant aquaporins:multifunctional water and solute channels with expandingroles. Plant Cell Environ 25:173–194
Vandeleur R, Niemietz C, Tilbrook J, Tyerman SD (2005) Rolesof aquaporins in root responses to irrigation. Plant Soil274:141–161
Vera-Estrella R, Barkla BJ, Bohnert HJ, Pantoja O (2004) Novelregulation of aquaporins during osmotic stress. Plant Physiol135:2318–2329
Yamada S, Katsuhara M, Kelly WB, Michalowski CB, BohnertHJ (1995) A family of transcripts encoding water channelproteins: tissue-speciWc expression in the common ice plant.Plant Cell 7:1129–1142
Yamada S, Komori T, Myers PN, Kuwata S, Kubo T, Imaseki H(1997) Expression of plasma membrane water channel genesunder water stress in Nicotiana excelsior. Plant Cell Physiol38:1226–1231
Zardoya R (2005) Phylogeny and evolution of the major intrinsicprotein family. Biol Cell 97:397–414
123