localization and quantification of plasma membrane aquaporin … · 2017. 11. 28. · in a system...

19
Abstract Water movement across root tissues occurs by parallel apoplastic, symplastic, and transcellular pathways that the plant can control to a certain extent. Because water channels or aquaporins (AQPs) play an important role in regulating water flow, studies on AQP mRNA and protein expression in different root tissues are essential. Here, we quantified and localized the expression of Zea mays plasma membrane AQPs (ZmPIPs) in primary root tip using in situ and quan- titative RT-PCR and immunodetection approaches. All ZmPIP genes except ZmPIP2;7 were expressed in primary roots. Expression was found to be dependent on the developmental stage of the root, with, in gen- eral, an increase in expression towards the elongation and mature zones. Two genes, ZmPIP1;5 and ZmPIP2;5, showed the greatest increase in expression (up to 11- and 17-fold, respectively) in the mature zone, where they accounted for 50% of the total ex- pressed ZmPIPs. The immunocytochemical localiza- tion of ZmPIP2;1 and ZmPIP2;5 in the exodermis and endodermis indicated that they are involved in root radial water movement. In addition, we detected a polar localization of ZmPIP2;5 to the external pericli- nal side of epidermal cells in root apices, suggesting an important role in water uptake from the root surface. Finally, protoplast swelling assays showed that root cells display a variable, but globally low, osmotic water permeability coefficient (P f < 10 lm/s). However, the presence of a population of cells with a higher P f (up to 26 lm/s) in mature zone of the root might be corre- lated with the increased expression of several ZmPIP genes. Keywords Aquaporin PIP mRNA and protein expression Root water movement Abbreviations ABA abscisic acid AQP aquaporin Lp r root hydraulic conductivity P f osmotic water permeability coefficient RT-PCR reverse transcription-PCR ZmPIP Zea mays plasma membrane intrinsic protein Introduction To move from the soil into the root vascular tissues, water must flow radially across a series of concentric cell layers, including the epidermis, the exodermis (when present), one or several layers of cortex cells, the endodermis, the pericycle, the xylem parenchyma cells, and, finally, the vessels. This anatomical structure of roots results in a complex pattern of water flow. After reaching the vascular bundles, water moves C. Hachez M. Moshelion E. Zelazny D. Cavez F. Chaumont (&) Unite ´ de Biochimie physiologique, Institut des Sciences de la Vie, Universite ´ catholique de Louvain, Croix du Sud 5-15 B-1348 Louvain-la-Neuve, Belgium e-mail: [email protected] M. Moshelion The Robert H. Smith Institute of Plant Sciences and Genetics in Agriculture, Faculty of Agricultural, Food & Environmental Quality Science, The Hebrew University Rehovot 76100, Israel Plant Mol Biol (2006) 62:305–323 DOI 10.1007/s11103-006-9022-1 123 Localization and quantification of plasma membrane aquaporin expression in maize primary root: a clue to understanding their role as cellular plumbers Charles Hachez Menachem Moshelion Enric Zelazny Damien Cavez Franc ¸ois Chaumont Received: 7 March 2006 / Accepted: 22 May 2006 / Published online: 15 July 2006 Ó Springer Science+Business Media B.V. 2006

Upload: others

Post on 24-Oct-2020

0 views

Category:

Documents


0 download

TRANSCRIPT

  • Abstract Water movement across root tissues occursby parallel apoplastic, symplastic, and transcellular

    pathways that the plant can control to a certain extent.

    Because water channels or aquaporins (AQPs) play an

    important role in regulating water flow, studies on

    AQP mRNA and protein expression in different root

    tissues are essential. Here, we quantified and localized

    the expression of Zea mays plasma membrane AQPs

    (ZmPIPs) in primary root tip using in situ and quan-

    titative RT-PCR and immunodetection approaches.

    All ZmPIP genes except ZmPIP2;7 were expressed in

    primary roots. Expression was found to be dependent

    on the developmental stage of the root, with, in gen-

    eral, an increase in expression towards the elongation

    and mature zones. Two genes, ZmPIP1;5 and

    ZmPIP2;5, showed the greatest increase in expression

    (up to 11- and 17-fold, respectively) in the mature

    zone, where they accounted for 50% of the total ex-

    pressed ZmPIPs. The immunocytochemical localiza-

    tion of ZmPIP2;1 and ZmPIP2;5 in the exodermis and

    endodermis indicated that they are involved in root

    radial water movement. In addition, we detected a

    polar localization of ZmPIP2;5 to the external pericli-

    nal side of epidermal cells in root apices, suggesting an

    important role in water uptake from the root surface.

    Finally, protoplast swelling assays showed that root

    cells display a variable, but globally low, osmotic water

    permeability coefficient (Pf < 10 lm/s). However, thepresence of a population of cells with a higher Pf (up to

    26 lm/s) in mature zone of the root might be corre-lated with the increased expression of several ZmPIP

    genes.

    Keywords Aquaporin Æ PIP Æ mRNA and proteinexpression Æ Root water movement

    Abbreviations

    ABA abscisic acid

    AQP aquaporin

    Lpr root hydraulic conductivity

    Pf osmotic water permeability coefficient

    RT-PCR reverse transcription-PCR

    ZmPIP Zea mays plasma membrane intrinsic

    protein

    Introduction

    To move from the soil into the root vascular tissues,

    water must flow radially across a series of concentric

    cell layers, including the epidermis, the exodermis

    (when present), one or several layers of cortex cells,

    the endodermis, the pericycle, the xylem parenchyma

    cells, and, finally, the vessels. This anatomical structure

    of roots results in a complex pattern of water flow.

    After reaching the vascular bundles, water moves

    C. Hachez Æ M. Moshelion Æ E. Zelazny Æ D. Cavez ÆF. Chaumont (&)Unité de Biochimie physiologique, Institut des Sciences dela Vie, Université catholique de Louvain, Croix du Sud 5-15B-1348 Louvain-la-Neuve, Belgiume-mail: [email protected]

    M. MoshelionThe Robert H. Smith Institute of Plant Sciences andGenetics in Agriculture, Faculty of Agricultural, Food &Environmental Quality Science, The Hebrew UniversityRehovot 76100, Israel

    Plant Mol Biol (2006) 62:305–323

    DOI 10.1007/s11103-006-9022-1

    123

    Localization and quantification of plasma membrane aquaporinexpression in maize primary root: a clue to understanding theirrole as cellular plumbers

    Charles Hachez Æ Menachem Moshelion ÆEnric Zelazny Æ Damien Cavez Æ François Chaumont

    Received: 7 March 2006 / Accepted: 22 May 2006 / Published online: 15 July 2006� Springer Science+Business Media B.V. 2006

  • through the entire plant by the long path (root-stem-

    leaf), which is characterized by a low resistance to

    water movement. The radial transfer through the root

    can be explained by the ‘‘composite transport model’’,

    which includes the contributions of three different

    pathways (Steudle 2000). The first is the apoplastic

    pathway, referring to the flow around protoplasts. The

    second is the symplastic pathway from cell to cell via

    plasmodesmata. The third is called the transcellular

    pathway, in which water and solutes have to cross cell

    membranes. The transcellular pathway is important for

    water, but of minor importance for solutes. The sym-

    plastic and transcellular pathways cannot be experi-

    mentally separated and are considered as a cell-to-cell

    path. This model including a combination of three

    different parallel pathways best describes the radial

    flow of water along the apoplast with its barriers.

    Depending on the conditions, the relative contribution

    of these pathways to overall uptake or hydraulic con-

    ductivity may change substantially. Thus, the radial,

    rather than the axial, component of water transport

    limits water uptake by roots (Steudle 2001).

    Different zones corresponding to different matura-

    tion levels can be distinguished in the primary maize

    root. The first region, at the tip of the root, is the cap,

    which is made up of cells originating from the calyp-

    trogen. Above this is the apical meristem region,

    responsible for the production of new root cells. This

    region shows a high mitotic activity, except in the qui-

    escent center, which lacks such activity (Ishikawa and

    Evans 1995). The region of cell division extends for a

    few millimeters behind the apical meristem and it is here

    that cells start to undergo elongation and enlargement

    (Ishikawa and Evans 1995). In the elongation zone, cells

    enlarge and develop as specific cell types according to

    their position in the root. Cells that leave the meriste-

    matic zone elongate rapidly and their volume can in-

    crease up to 40-fold during their development (Moore

    and Smith 1990). Frensch et al. (1996) reported that,

    during development, the root tissues in the division and

    differentiation zones become less permeable for trans-

    membrane water and solute movement.

    Roots must take up water and necessary solutes from

    the soil into the stele, while avoiding the influx of

    unnecessary solutes. Since roots manage to fulfill these

    contradictory requirements, the internal tissues

    responsible for these selectivity functions must be pre-

    cisely regulated in a manner depending on growth

    conditions (Karahara et al. 2004). The development of

    Casparian strips by the hypodermis and endodermis,

    which cut off apoplastic water and solute flow, occurs

    along the root axis. While the suberization of the

    endodermis surrounding the stele occurs before the

    maturation of the protoxylem and the development of

    root hairs (Raven et al. 1992), formation of the exo-

    dermis, defined as suberized hypodermis, is highly

    variable and highly dependent on growth conditions.

    As an example, hydroponically grown maize seedlings

    usually do not develop an exodermis, whereas aerop-

    onically grown ones develop a suberized layer approx-

    imately 0.5–1 cm after the region in which

    differentiation of the endodermis occurs (Enstone and

    Peterson 2005; Hose et al. 2001). These observations

    reflect a mean of protection against excessive water loss

    due to aeration (Hose et al. 2001). The distance from

    the root tip to the lowest position of the exodermal

    Casparian strip decreases in maize roots under osmotic

    stress (Perumalla 1986). However, the mechanism

    responsible for Casparian strip development closer to

    the root tip remains to be fully characterized (Karahara

    et al. 2004). In addition, the apoplastic barriers may not

    be completely impermeable to water and solutes, as

    bypasses have been demonstrated for both the exo-

    dermis and endodermis (Schraut et al. 2004).

    For a given root size and anatomy, plants also have

    the ability to alter their root hydraulic conductivity (Lpr)

    in less than 3 days (Javot and Maurel 2002). This fine

    and rapid regulation is assumed to be under control by

    internal and external stimuli, such as the day/night cycle,

    nutrient deficiency, water shortage, soil salinity, water

    demand of the transpiring shoot, anoxia, temperature,

    heavy metals, and abscissic acid (ABA) (Lopez et al.

    2003; Steudle 2000). Depending on the species and

    conditions, water conductivity can change rapidly by up

    to three orders of magnitude (Steudle 2001). These

    modifications are due to changes in cell membrane

    permeability triggered by the expression and specific

    regulation of aquaporins (AQPs), which facilitate the

    movement of water or other small solutes through the

    membrane. These proteins are ubiquitous in plants and

    have been detected in high amounts in regions of high

    transcellular water transport, such as the endodermis

    (Otto and Kaldenhoff 2000; Schaffner 1998).

    AQPs are seen as ‘‘cellular plumbers’’ regulating the

    movement of water across membranes. A significant

    proportion of root water transport (20–80%) is under

    rapid metabolic control involving AQPs (Maurel and

    Chrispeels 2001). When present and active, they can

    increase the cell membrane osmotic water permeability

    coefficient (Pf) by up to 20-fold. AQPs are abundantly

    expressed in roots, where they mediate soil water up-

    take (Alexandersson et al. 2005; Boursiac et al. 2005;

    Jang et al. 2004; Javot and Maurel 2002). The contri-

    bution of AQPs to root water transport can be assessed

    by the use of AQP inhibitors, such as mercurial

    compounds. For instance, mercurials reduce the Lpr by

    306 Plant Mol Biol (2006) 62:305–323

    123

  • 32–90% in Buckhorn cholla cactus, tomato, wheat,

    paprika pepper, melon, sugar beet, and barley (Amo-

    deo et al. 1999; Carvajal et al. 2000; 1996; 1999; Maggio

    and Joly 1995; Martre et al. 2001; Tazawa et al. 1997).

    Hukin et al. (2002) also showed a rapid 4-fold decrease

    in maize root elongation rate (from 1.83 mm h–1 to

    under 0.4 mm h–1) after mercurial treatment, suggest-

    ing a central role of AQPs in this process.

    Thirteen plasma membrane AQPs (PIPs), split into

    two groups (PIP1 and PIP2), are expressed in maize

    (Chaumont et al. 2001). Functional assays in Xenopus

    oocytes indicated that all ZmPIP2s tested so far

    (ZmPIP2;1, ZmPIP2;4, ZmPIP2;5) have a high water

    channel activity whereas ZmPIP1s (ZmPIP1;1;

    ZmPIP1;2 and ZmPIP1;5) have no or low activity

    (Chaumont et al. 2000; Fetter et al. 2004; Gaspar

    et al. 2003). However, when co-expressed, ZmPIP1;2

    and ZmPIP2 isoforms physically interact to enhance

    both ZmPIP1;2 targeting to the plasma membrane

    and membrane osmotic water permeability (Fetter

    et al. 2004). Transcripts of all ZmPIP genes except

    ZmPIP2;7 have been detected in the primary maize

    root (Aroca et al. 2005; Zhu et al. 2005). Moreover,

    the expression of several ZmPIPs is regulated by

    chilling and salt stress, ABA and diurnal cycle

    (Aroca et al. 2005; Lopez et al. 2003; Zhu et al. 2005).

    These findings indicate that specific AQP isoform

    can be regulated differently and it is therefore

    essential to have a comprehensive analysis of the

    expression pattern of each AQP gene in maize root

    tissues or cells.

    To obtain a better understanding of PIP function in

    the primary root, we quantified and localized AQP

    mRNA and protein expression using in situ and

    quantitative RT-PCR approaches, Western blots, and

    immunocytochemistry. We found that most ZmPIP

    genes were expressed in the roots and their expression

    was developmentally regulated. The high expression of

    ZmPIP2;1 and ZmPIP2;5 in the exodermis and endo-

    dermis indicates an essential role of these channels in

    facilitating water movement through these critical cell

    layers surrounded by Casparian strip.

    Materials and methods

    Plant material

    Zea mays (B73 line) seeds were disinfected using

    50% sodium hypochloride, rinsed with distilled

    water, placed overnight in oxygenated 1 mM CaCl2solution, and incubated in water for an additional

    24 h at 24�C. Seedlings were then grown aeroponi-

    cally on an 8 h-dark (18�C)/16 h-light (24�C) regimein a system exposing the roots to a humid atmo-

    sphere (de Dorlodot et al. 2005). Humidification of

    the atmosphere was achieved through pulverization

    of nutrient solution (half-strength Hoagland medium,

    Sigma, Steinheim, Germany). The pulverization rate

    was set to create a mist by projection of small

    nutrient medium droplets (1 mm) into the air for

    20 s periods separated by 40 s rest periods. Using

    these conditions, the relative humidity of the atmo-

    sphere was maintained close to 100%. The age of the

    seedlings was calculated from the time they were

    placed in the aeroponic system. For the experiments,

    we chose 7- to 8-day-old roots harvested 3.5 h after

    the start of illumination (200 lmol photons m–2 s–1),which corresponds roughly to the peak of ZmPIP

    mRNA production (Lopez et al. 2003). These growth

    conditions were used in all experiments except the Pfexperiment, in which the system described by Fetter

    (Fetter et al. 2004) was used.

    In situ reverse transcriptase-mediated PCR

    Root segments were fixed, embedded in Paraplast Plus

    (Sherwood Medical, St. Louis, MO, USA), sectioned

    into 10 lm slices, dewaxed, and hydrated as previouslydescribed (Fetter et al. 2004). Slide pre-treatment was

    performed as described by Fetter et al. (2004), except

    that 2% pectinase (Calbiochem, Darmstadt, Germany)

    was used and no proteinase K digestion was carried

    out. RT-PCR and amplicon localization were per-

    formed as described previously (Fetter et al. 2004).

    Signals were recorded 16 h after the beginning of the

    staining procedure using an epifluorescent Leica DMR

    microscope (Wetzlar, Germany).

    Detection of apoplastic barriers

    Detection of apoplastic barriers was performed on

    sections hand-cut from fresh root tissue which were

    stained for 1 h with 0.1% berberine hemisulfate and

    for 45 min with 0.5% toluidine blue O (w/v) (Brundrett

    et al. 1988), mounted in 0.1% FeCl3 in 50% glycerol,

    and immediately examined under a microscope (exci-

    tation at 340–380 nm, emission at 425 nm).

    Quantitative RT-PCR

    Total RNA was extracted from fresh root sections

    using an RNeasy� plant extraction minikit (Qiagen,Maryland, USA). DNAse I digestion was performed

    on the column during RNA extraction according to the

    manufacturer’s recommendations.

    Plant Mol Biol (2006) 62:305–323 307

    123

  • For cDNA synthesis, 3 lg of total RNA and 1 lMoligo dT(18) in a volume of 12.5 ll were incubated at70�C for 10 min and chilled on ice, then 12.5 ll of acocktail containing 500 lM dNTP Mix (Roche,Mannheim, Germany), 20 units of RNase inhibitor

    (Roche, Mannheim, Germany), and 200 units of Mo-

    loney murine leukemia virus reverse transcriptase

    (Promega, Madison, USA) was added and the mixture

    incubated for 60 min at 42�C, followed by 5 min at85�C. cDNA was purified using a DNA purification kit(Roche) and eluted in 50 ll of water at 60�C.

    Real-time PCR analysis was performed using 1 ll of1:10 diluted cDNA samples in 25 ll of reaction med-ium containing 300 nM gene-specific primers (Table 3

    in Supplementary Data) and 12.5 ll of SYBR Green 1PCR Mix (Eurogentec, Liège, Belgium) in an Applied

    Biosystems PE-7000 (Darmstadt, Germany). For each

    RNA sample, a reaction without reverse transcription

    was performed to control for contamination by geno-

    mic DNA. PCR using intron-flanking primers was also

    used to check for genomic DNA contamination. In the

    calibration step, a control experiment without cDNA

    was performed to test for primer–dimer formation,

    primers forming dimers being excluded. The PCR

    program was as follows: 2 min at 50�C, 10 min at 95�Cfor DNA polymerase activation, and 40 cycles of 15 s

    at 95�C and 60 s at 60�C. Data collection was per-formed at 60�C. The melting curve analysis programconsisted of 10 s at 95�C, 30 s at 60�C, and heating to90�C at a rate of 0.2�C s–1, data being continuouslycollected. The data were analyzed by the 2–DDCt

    method (Pfaffl 2001). The calibration step of the

    experiment checked for equivalent PCR efficiency of

    the different genes (to allow comparison and normal-

    ization). Standard curves (log of cDNA dilution vs. Ct)

    using serial 10-fold dilution of cDNA were built for

    each pair of selected primers, a 100% PCR efficiency

    corresponding to a slope of –3.3 (Marino et al. 2003).

    Practically, only pairs of primers yielding a slope of –

    3.3 ± 0.1 were selected. The specificity of the amplifi-

    cation (checked by dissociation curve analysis and gel

    electrophoresis) and the use of appropriate control

    genes were also assessed. Normalization of the results

    was achieved using the glyceraldehyde phosphate

    dehydrogenase (GAPDH) (gi:22237) or alpha tubulin

    (gi:450292) gene as internal control, both yielding the

    same results. Results obtained on the different root

    zones were normalized using three maize internal

    control genes, alpha tubulin, plasma membrane H+-

    ATPase (gi:758354), and polyubiquitin (gi:248338), the

    expression of which was shown to be stable in the

    investigated tissues. This method smoothes normali-

    zation errors due to the small variation of expression of

    a single control gene (Vandesompele et al. 2002). The

    elongation factor 1-alpha (EF1-a) gene (gi:2282583)was used as a control for meristematic activity.

    Root cell protoplast isolation

    Portions of the root tip at known positions from the

    root apex were excised, chopped into small fragments,

    and placed in a multiplate containing 500 ll of300 mOsm isotonic buffer (10 mM KCl, 1 mM CaCl2,

    8 mM MES-KOH, pH 5.7) containing a cocktail of 2%

    cellulase (Worthington, Lakewood, USA), 1% Drisel-

    lase (Fluka, Buchs, Switzerland), 0.1% BSA (Sigma,

    St. Louis, MO, USA), 0.1% Pectolyase (Karlan, Santa

    Rosa, USA), and 1.5% Macerozyme (Yakult Phar-

    maceutical Ind. Co., Tokyo, Japan) to degrade the cell

    wall. Digestion was carried out for 1.5 h at 30�C in thedark on a rotating table (90 rpm), then the protoplasts

    were filtered on a 100 lm mesh and centrifuged at 90gfor 1 min. To remove the enzymes, 80% of the super-

    natant was removed and replaced by the same amount

    of isotonic solution, then the sample was recentrifuged;

    this step was repeated a further twice. The protoplasts

    were kept at room temperature and analyzed as rapidly

    as possible as described in Moshelion et al. (2004).

    GST-ZmPIP-H expression

    Genetic constructs were prepared using a modified

    version of pGex-KG plasmid (pGex-KG’) (Guan and

    Dixon 1991), in which a sequence encoding six Histi-

    dine residues was introduced between the HindIII and

    SmaI restriction sites. The DNA sequences encoding

    the cytoplasmic amino-terminal ends of ZmPIP iso-

    forms were amplified by PCR using specific primers,

    incorporating EcoRI and HindIII sites at the

    ends (Table 4 in Supplementary data) and subcloned

    into EcoRI and HindIII sites of pGex-KG’. These

    constructs were checked by sequence analysis. The

    encoded fusion proteins consist in glutathione

    S-transferase in frame with the amino-terminal part of

    each ZmPIP and a His tag (GST-ZmPIP-H).

    Twelve milliliters of overnight precultures of indi-

    vidual transformed E. coli cells were used to inoculate

    1 liter Luria-Bertani broth medium containing 100 lg/ml of ampicillin. The cells were cultured at 37�C untilmid-log phase (3–4 h, or OD600 = 0.6) before induction

    of GST-ZmPIP-H expression with 1 mM isopropyl b-D-thiogalactopyranoside. After 3 h incubation at 37�Cthe cells were harvested by centrifugation at 5000g,

    resuspended in 10 ml PBS containing 100 lg/ml lyso-zyme and incubated for 15 min on ice. The cells were

    308 Plant Mol Biol (2006) 62:305–323

    123

  • then sonicated 4 · 30 s and the lysate centrifuged for15 min at 20800g. GST-ZmPIP-H proteins were

    extracted from the soluble phase by nickel affinity

    chromatography according to the manufacturer’s rec-

    ommendations (Qiagen, Maryland, USA).

    Subcellular fractionation

    To prepare the microsomal fraction, root sections,

    collected 3.5 h after light onset, were ground in 800 llof a mixture of 250 mM sorbitol, 50 mM Tris–HCl (pH

    8), 2 mM EDTA, and protease inhibitors [1 mM

    phenylmethylsulfonyl fluoride, 1 lg/ml each of leu-peptin, aprotinin, antipain, chymostatin, and pepstatin

    (Sigma, St. Louis, MO, USA)]. All subsequent steps

    were performed at 4�C. The homogenate was centri-fuged at 2,700g for 5 min, and the resulting supernatant

    centrifuged at 10,000g for 10 min. The second super-

    natant was then centrifuged at 100,000g for 30 min and

    the resulting pellet (microsomal fraction) resuspended

    in 30–60 ll of suspension buffer (5 mM KH2PO4,330 mM saccharose, 3 mM KCl, pH 7.8) and sonicated

    twice for 4 s.

    Protein analysis

    Fifteen micrograms of crude microsomal membranes

    or 0.5 lg of purified GST-ZmPIP-H fusion proteinwere solubilized for 15 min at 60�C in solubilizationbuffer (27 mM Tris/HCl, 0.7% SDS, 3.3% glycerol,

    0.0016% bromophenol blue, 1% DTT) and the pro-

    teins separated by SDS-PAGE on a 12% polyacryl-

    amide gel. After electrophoresis, the gel was incubated

    for 15 min in semi-dry buffer (48 mM Tris, 39 mM

    glycine, 20% methanol, 0.0375% SDS) before semi-dry

    transfer to a polyvinylidene fluoride (PVDF; Millipore)

    membrane was performed for 30 min at 22 V, the

    membrane being previously incubated for 1 min in

    pure methanol and 5 min in semi-dry buffer.

    Western blot analysis was performed on the PVDF

    membrane using antisera raised against the amino-

    terminal peptides of ZmPIP1;2, ZmPIP2;1, ZmPIP2;5,

    and ZmPIP2;6 (ZmPIP1;2, AQGAADDKDYKEP;

    ZmPIP2;1, KDDVIESGAGGEFAAKD; ZmPIP2;5,

    AKDIEAAAAHEGKD; ZmPIP2;6, KEVDVSTLEA

    GGVRDRD). The peptides were coupled to KLH

    using an Imject Maleimide Activated mcKLH Kit

    (Pierce, Rockford, USA) and the resulting conjugates

    used to immunize rabbits. The anti-RsPIP1 antiserum

    was raised against radish PIP1 isoforms (Suga

    et al. 2001). The anti-PMA antiserum was raised

    against plasma membrane H+-ATPases (Morsomme

    et al. 1996). The dilutions used were 1/500 for the anti-

    ZmPIP1;2 and anti-ZmPIP2;5 antisera, 1/750 for

    anti-His (Qiagen), 1/1000 for the anti-ZmPIP2;6 and

    anti-RsPIP1 antisera, 1/5000 for the anti-ZmPIP2;1/2;2

    antiserum, and 1/10000 for the anti-PMA antiserum in

    TBS (0.02 M Tris, 0.136 M NaCl). Stripping of the

    membrane was achieved by incubation in 0.5 M NaOH

    for 5 min and 3 washes in TBS-Tween 20 0.1%.

    In situ immunolocalization

    Hand-cut sections of freshly harvested roots were fixed

    in 0.8% paraformaldehyde in PBS (68 mM NaCl,

    1.34 mM KCl, 5.070 mM Na2HPO4, 0.85 mM

    KH2PO4, pH 7.4) for 1 h at room temperature under

    light vacuum, followed by 5 washes with PBS. Perme-

    abilization of the samples was achieved by dipping the

    sections for 30 min in 0.25% Triton X-100 in PBS. The

    sections were incubated for 1 h at 37�C in PBS con-taining 5% BSA (blocking solution), then for 1 h at

    37�C with anti-PIP antiserum diluted 1/100 in blockingsolution. After 3 · 20 min washes in blocking solution,the slides were incubated for 1 h in the dark at 37�Cwith fluorescein-coupled goat anti-rabbit IgG anti-

    bodies (Molecular Probes, Eugen, Oregon, USA;

    dilution 1/100 in blocking solution). To quench auto-

    fluorescence, the sections were incubated for 2 min in

    PBS containing 0.1% toluidine blue. The slides were

    then rinsed for 4 · 5 min in PBS and analyzed underUV (excitation: 460–500 nm, emission: 510–530 nm).

    The controls included immunodetection using pre-im-

    mune serum (primary antibody specificity), detection

    without incubation with primary antibody (secondary

    antibody specificity), and untreated slides (false posi-

    tive signals due to autofluorescence).

    Results

    Expression of ZmPIP mRNAs in aeroponically

    grown maize roots

    The level of expression of ZmPIP genes was first

    determined in entire aeroponically grown primary roots

    using real time PCR on RNA extracted from the first

    10 cm of the root (from the apex). To allow comparison

    between different ZmPIP amplicons in terms of rela-

    tive abundance, we made sure that the PCR efficiency

    was very similar (see Materials and methods). The roots

    were sampled 3.5 h after light onset. The time when the

    roots are collected is of major importance, as some

    maize PIPs display a diurnal expression pattern, with a

    Plant Mol Biol (2006) 62:305–323 309

    123

  • peak of expression at 2–4 h after the beginning of the

    light period (Lopez et al. 2003).

    All ZmPIP genes except ZmPIP2;7 were expressed,

    although to different extents (Fig. 1). ZmPIP1;1,

    ZmPIP1;5, ZmPIP2;1, and ZmPIP2;5, were the most

    highly expressed and, together, accounted for more

    than 75% of all the expressed ZmPIPs, ZmPIP2;5

    alone accounting for 30%. All other genes were ex-

    pressed at a much lower level, each accounting for

    between 0 and 6% of total ZmPIP expression. Abso-

    lute quantification of the expression of the ZmPIP2;4

    and ZmPIP2;6 genes was performed by constructing a

    calibration curve using serial 10-fold dilutions of plas-

    mids bearing ZmPIP2;4 or ZmPIP2;6 cDNA for which

    the initial concentration was known. Assuming an

    equal efficiency of amplification as checked during real

    time PCR experiment set up (see Materials and

    methods) and knowing the relative expression of the

    different ZmPIPs (Fig. 1), we were able to estimate

    ZmPIP expression in terms of transcript copy number

    (Table 1 in Supplementary Material). The four

    most highly expressed genes, ZmPIP1;1, ZmPIP1;5,

    ZmPIP2;1, and ZmPIP2;5, were present at 2500–7000

    copies par nanogram of total RNA. ZmPIP1;2,

    ZmPIP1;4, ZmPIP2;4, and ZmPIP2;6 represented an

    intermediate group with copy numbers in the range

    of 850–1500 copies per nanogram of total RNA.

    Finally, ZmPIP1;3, ZmPIP1;6, ZmPIP2;2, and

    ZmPIP2;3 showed low expression, below 500 copies

    per nanogram of total RNA. These ZmPIP mRNA

    copy numbers are in the range of those obtained in

    Arabidopsis roots for AtPIP genes (500–7,000 copies

    per nanogram of total RNA) (Jang et al. 2004).

    Development of apoplastic barriers

    It is generally believed that AQPs play a role in elon-

    gating tissues. As the root tissues mature, the cells be-

    come more symplastically isolated and apoplastic

    barriers develop in two regions, the endodermis and the

    exodermis. The localization of both cell layers was

    examined to determine whether ZmPIP gene expres-

    sion correlated with regions in which symplastic water

    movement is hindered, i.e. regions requiring a high

    membrane water permeability. We analyzed transverse

    sections of the primary root from aeroponically grown

    plants. Lignin and suberin deposits were examined at

    different positions from the root tip up to 10 cm from the

    cap (Fig. 2 and data no shown). Suberization of the

    endodermis was detected 3 cm from the root tip, the

    suberin and lignin deposits being located in the radial

    cell walls of the endodermis. The suberized hypodermis,

    defined as the exodermis, was detected 0.7–1 cm above

    the first suberized Casparian strip of the endodermis at

    3.7–4 cm from the root tip (Fig. 2). These data agree

    with previous reports concerning suberin deposits in

    maize roots (Enstone and Peterson 2005; Hose et al.

    2001; Karahara et al. 2004).

    Expression of ZmPIPs according to root

    developmental stage

    To investigate ZmPIP gene expression along the lon-

    gitudinal axis of the developing root in aeroponic

    conditions and how it relates to the development of

    apoplastic barriers, we extracted total RNA from dif-

    ferent segments of the root, each corresponding to a

    tissue at a different developmental stage. For practical

    reasons, we studied segments of at least 5 mm along

    the root axis. By visual inspection and measurement of

    cell length, we identified six distinct root zones. The

    first, between 0 and 5 mm, consisted of the cap, apical

    meristem, and distal elongation zone (Ishikawa and

    Evans 1995) and corresponded to a region of fast cell

    division and elongation. The second, between 5 and

    10 mm, corresponded roughly to the elongation zone.

    The third, between 10 and 20 mm, was the zone in

    which the process of cell elongation/differentiation

    stopped and root hair started to appear. The fourth,

    between 30 and 40 mm, was in the root hair region,

    where water absorption is reported to reach a maxi-

    mum (Boyer 1985), the endodermis being suberized,

    while the exodermis is still forming. The fifth, between

    50 and 60 mm, corresponded to what we call ‘‘mature’’

    primary root tissue in this report because both apo-

    plastic barriers are present (Fig. 2). Finally, the sixth,

    0

    1

    2

    3

    4

    5

    6

    ZmPI

    P1;1

    ZmPI

    P1;6

    ZmPI

    P2;1

    ZmPI

    P1;2

    ZmPI

    P1;3

    ZmPI

    P1;4

    ZmPI

    P1;5

    ZmPI

    P2;2

    ZmPI

    P2;3

    ZmPI

    P2;4

    ZmPI

    P2;5

    ZmPI

    P2;6

    ZmPI

    P2;7

    GAP

    DH

    Arb

    itrar

    y un

    its

    Fig. 1 Levels of ZmPIP transcripts in the primary maize root.cDNA was synthesized from total RNA extracted from primarymaize roots and real-time PCR was performed as described inthe Materials and methods. The relative expression level ofZmPIPs was assessed using glyceraldehyde phosphate dehydro-genase (GAPDH) as internal standard. The mean and standarddeviation for three independent experiments are shown

    310 Plant Mol Biol (2006) 62:305–323

    123

  • between 100 and 110 mm, was the zone in which lateral

    roots start to protrude from the root and corresponded

    to the zone with ‘‘secondary’’ mitotic activity.

    Interestingly, with the exceptions of ZmPIP2;7,

    which was not expressed, and ZmPIP2;6, which

    showed a fairly stable level of expression throughout

    the root segments, expression of the ZmPIP genes

    was developmentally regulated (Fig. 3 and Table 2 in

    Supplementary data). ZmPIP expression was low (or

    nonexistent in the case of ZmPIP1;6) in the first and

    second zones and gradually increased to a plateau in

    either zone 3 or zone 4, corresponding to mature

    primary root tissues. ZmPIP1;5 and ZmPIP2;5, two of

    the most highly expressed genes, both showed the

    same dramatic change in their expression pattern

    (Fig. 3). Assuming an equal efficiency of PCR

    amplification (see Materials and methods), we com-

    pared the relative contribution of each ZmPIP within

    one segment. ZmPIP1;5 and ZmPIP2;5 expression

    was low in the first segment (less than 6% of the total

    expressed ZmPIPs), then gradually increased to

    about 20% (ZmPIP1;5) or 30% (ZmPIP2;5) of the

    expressed ZmPIPs in mature primary root tissue,

    accounting for 50% of the expressed ZmPIPs

    (Table 1). ZmPIP1;1 accounted for 27% of the ex-

    pressed ZmPIPs in the first root segment, but its

    contribution gradually decreased to 9% in the fourth

    segment, where root hairs appear, and was about 12%

    in the mature zones. ZmPIP2;1 showed the same

    pattern of a decrease from 28 to 14% (Table 1).

    These decreases in relative contribution were not

    attributable to a decrease in absolute ZmPIP1;1 and

    ZmPIP2;1 expression levels, which actually increased,

    but rather to a huge increase in ZmPIP1;5 and

    ZmPIP2;5 expression. On average, all the other genes

    each accounted for less than 5% of ZmPIP expression

    in the different root segments, with the exczeption of

    ZmPIP2;6, which averaged 8% (Table 1). These data

    show that ZmPIP1;5 and ZmPIP2;5 are major

    isoforms in the mature root.

    The EF1-a (elongation factor) gene was chosen as amarker for meristematic activity. Its expression peaked

    in the meristematic zone, then decreased to a constant

    and much lower value in the subsequent segments, again

    increasing (to a lower level than in zone 1) in segment 6,

    in which secondary root meristems were activated

    4 cm

    3 cm

    (A) (B) (C)

    Suberized hypodermis

    (D)

    (E)

    (F)

    Suberized endodermis

    D

    E

    F

    Fig. 2 Development ofapoplastic barriers in primaryroots cultivated in aeroponicconditions. A, Schematic viewof apoplastic barrierdevelopment in maize rootsgrown in aeroponicconditions. B–F, Transversesections of the roots at 4 (B–D), 3 (E), and 1 (F) cm fromthe root tip stained withberberine hemisulfate andtoluidine blue as described inthe Materials and methods.No suberized cell layer wasdetected in the zone 0–3 cmfrom the root tip (F). At 3 cm,suberization of theendodermis was detected (E,white arrow). At 4 cm, both asuberized endodermis andexodermis were detected (D,white arrows). B and C,Magnifications of theendodermis and exodermis,respectively, at 4 cm from theroot tip. Bars = 25 lm in Band C and 100 lm in D–F

    Plant Mol Biol (2006) 62:305–323 311

    123

  • (Fig. 3). Although it underwent the same normalization,

    the expression pattern of EF1a displayed an oppositetrend to that of the PIP genes, clearly indicating that the

    observed developmental PIP expression results were

    not due to an artifact of the method.

    Expression of ZmPIP genes in the cap, meristem,

    and elongation zones

    In situ RT-mediated PCR is a powerful tool for

    localizing specific gene expression and is claimed to

    be more sensitive than RNA in situ hybridization

    because it includes an amplification step (Nuovo

    1996). As appropriate controls, the full process was

    checked for efficiency and specificity by carrying out

    an in vitro RT-PCR on total root RNA extract using

    the same conditions (primers, PCR cycles, stock

    solutions etc.) as for the in situ reactions. The RT-

    PCR products obtained for all ZmPIP genes were

    checked on a 2% agarose gel. The presence of a

    single band of the expected size was considered as

    resulting from a specific amplification. To distinguish

    between a specific signal and background noise, some

    sections did not undergo the RT step and were con-

    sidered as negative controls; as a result, since genomic

    DNA was destroyed by the DNAse treatment and

    mRNA was not transcribed into cDNA, no signal was

    observed (Fig. 4E, H and J). A true positive signal for

    in situ RT-PCR consists of labeling of both the nu-

    cleus and cytoplasm (Nuovo 1996) (Fig. 4K), while a

    false positive signal generated by incomplete genomic

    DNA digestion results in only nuclear labeling

    (Fig. 4L). A positive signal shows the presence of the

    targeted mRNA, but its intensity cannot be linked to

    0

    10

    20

    30

    40

    50

    60

    70

    800-5 mm

    5-10 mm

    10-20 mm

    30-40 mm

    50-60 mm

    100-110 mm

    ZmPI

    P1;1

    ZmPI

    P1;6

    ZmPI

    P2;1

    ZmPI

    P1;2

    ZmPI

    P1;3

    ZmPI

    P1;4

    ZmPI

    P1;5

    ZmPI

    P2;2

    ZmPI

    P2;3

    ZmPI

    P2;4

    ZmPI

    P2;5

    ZmPI

    P2;6

    EF1-

    Aα-

    tubu

    linUb

    iqui

    tin

    PMA

    Arb

    itrar

    y un

    its

    Fig. 3 Levels of ZmPIP transcripts in different segments of theprimary root. cDNA was synthesized from total RNA extractedfrom six different segments of the primary maize root and real-time PCR performed as in Fig. 1. The geometric mean of theexpression level of three control genes (a-tubulin, ubiquitin, andH+-ATPase) was used to standardize the results. The distance ofthe different segments from the root tip were 0–5 mm (cap,

    meristem, start of the elongation zone), 5–10 mm (elongationzone), 10–20 mm (end of the elongation zone, start of the roothair zone), 30–40 mm (root hair zone), 50–60 mm (matureprimary root tissue), and 100–110 mm appearance of secondaryroots). The mean and standard deviation for three independentexperiments are shown

    Table 1 Expression ofZmPIPs as a percentage ofthe total ZmPIP expression ineach investigated root zone.N.D. stands for ‘‘NotDetected’’

    Percentage contribution to ZmPIP expression in different root segments

    0–5 mm 5–10 mm 10–20 mm 30–40 mm 50–60 mm 100–110 mm

    ZmPIP1;1 26.9 ± 5 25.6 ± 2.8 19.2 ± 3.7 9.3 ± 3.0 11.7 ± 4.7 12.9 ± 4.1ZmPIP1;2 2.1 ± 0.5 6.4 ± 1.9 6.2 ± 1.4 3.5 ± 1.8 2.2 ± 0.4 2.3 ± 0.7ZmPIP1;3 2.7 ± 0.9 3.6 ± 0.6 3.8 ± 1.1 6.0 ± 2.6 4.2 ± 1.2 5.1 ± 1.3ZmPIP1;4 1.1 ± 0.3 1.8 ± 0.8 2.3 ± 1.4 2.6 ± 0.6 2.4 ± 0.3 2.5 ± 0.6ZmPIP1;5 5.8 ± 1.4 12.3 ± 1.8 14.6 ± 3.0 19.9 ± 2.4 21.3 ± 4.1 22.6 ± 3.2ZmPIP1;6 N.D. N.D. 0.2 ± 0.1 1.1 ± 0.1 1.0 ± 0.8 1.1 ± 0.5ZmPIP2;1 27.6 ± 12.6 18.0 ± 13.9 18.3 ± 11.9 19.3 ± 4.7 12.1 ± 5.7 14.1 ± 8.0ZmPIP2;2 8.3 ± 2.5 5.9 ± 1.6 4.9 ± 1.9 5.4 ± 1.3 3.9 ± 0.4 3.3 ± 0.5ZmPIP2;3 1.6 ± 0.5 3.8 ± 0.6 2.6 ± 1.6 2.8 ± 0.8 3.9 ± 0.8 3.0 ± 1.3ZmPIP2;4 0.5 ± 0.3 1.6 ± 0.3 1.5 ± 0.3 1.3 ± 0.1 1.4 ± 0.7 1.4 ± 0.4ZmPIP2;5 5.0 ± 1.4 15.4 ± 2.1 20.0 ± 4.2 22.0 ± 0.5 29.2 ± 3.3 27.2 ± 2.3ZmPIP2;6 18.3 ± 5 5.7 ± 4.5 6.3 ± 2.6 6.7 ± 0.8 6.7 ± 1.2 4.5 ± 0.6

    312 Plant Mol Biol (2006) 62:305–323

    123

  • mRNA levels. Thus, this approach can be seen as a

    complementary to quantitative RT-PCR, which yields

    quantitative results, but no information about the

    location of transcript expression.

    With the exception of the ZmPIP1;6 and ZmPIP2;7

    transcripts, which were not detected, all ZmPIP

    mRNAs were detected in most cell types in the meri-

    stem, elongation, and mature zone (Fig. 4A–F and data

    not shown). The cap constituted an exception to the

    rule. Indeed, with the exception of ZmPIP2;6 and

    ZmPIP2;4 which were found in the root cap (Fig. 4D

    and F), the ZmPIP isoforms yielded a very low signal

    not significantly different from background noise

    (Fig. 4A–C and E). However, the nuclei of the cap

    cells were strongly labeled in the positive controls (no

    DNAse treatment), suggesting that our method is able

    to stain the cap tissue. Transverse sections showed a

    clear ZmPIP1;1 signal in the epidermis and exodermis

    (Fig. 4G) and in that the ZmPIP2;4 signal was stronger

    in the stele than the cortex (Fig. 4I)

    Fig. 4 Localization ofZmPIP mRNAs revealed byin situ RT-PCR. A–F, 10 lmthick longitudinal sectionsthrough the maize root tipshowing the presence ofZmPIP transcripts. AZmPIP2;1, B ZmPIP1;2, CZmPIP1;1, D ZmPIP2;4, Enegative control (no RT), FZmPIP2;6. G–H: Localizationof ZmPIP1;1 mRNA in theepidermis and exodermis (G)and negative control (no RT)(H), I–J: Transverse sectionsat 5 cm from the root tipshowing the localization ofZmPIP2;4 mRNA in all thetissues, with a stronger signalin the stele (I) or the negativecontrol (No RT) (J). K,Typical signal for in situ RT-PCR showing that both thecytoplasm and the nuclei werelabeled for ZmPIP1;1transcripts. L, Positive controlin which no DNAse and noRT treatments wereperformed, leading to a strongnuclear signal (ZmPIP2;2gene). Bars = 200 lm in A–F,I, J and 20 lm in G, H, K, L

    Plant Mol Biol (2006) 62:305–323 313

    123

  • The ubiquitous appearance of AQP mRNAs might

    seem puzzling, but is not abnormal considering the

    fairly high level of ZmPIP gene expression observed

    by quantitative RT-PCR. Moreover, the PCR step

    renders the method ultrasensitive for the detection of

    low transcript copy numbers, a few copies of mRNA

    being theoretically enough to obtain a signal at the

    single cell level (Nuovo 1996).

    ZmPIP protein levels in roots

    Expression at gene level is not always reflected at the

    protein level. Therefore, ZmPIP protein expression in

    maize primary root was investigated using Western

    blots. ZmPIPs were detected using antibodies raised

    against amino-terminal peptides derived from

    ZmPIP2;1, ZmPIP2;5, ZmPIP2;6, ZmPIP1;2, and rad-

    ish PIP1s (Suga et al. 2001) (Fig. 5A–E). To test the

    specificity of the antibodies, fusion proteins consisting

    in the glutathione-S-transferase, amino-terminal end of

    each ZmPIP (21–56 amino acid residues) and six His

    tag were expressed in Escherichia coli, purified on a

    nickel column, separated on SDS-PAGE and immu-

    nodetected using each serum (Fig. 5F). The anti-

    ZmPIP2;1 antibodies recognized ZmPIP2;1 and

    ZmPIP2;2, as expected on the basis of amino acid se-

    quence comparison (Fig. 5A), whereas the antisera

    against ZmPIP2;5, ZmPIP2;6 and ZmPIP1;2 were iso-

    form-specific. The anti-radish RsPIP1 antiserum

    recognized all maize PIP1s but a higher signal was

    obtained for ZmPIP1;6.

    Crude membrane proteins from the different root

    segments were extracted and separated on SDS-PAGE,

    then ZmPIPs were immunodetected. Anti-PMA anti-

    body recognizing plasma membrane H+-ATPases was

    used as a loading control (Morsomme et al. 1996). All

    the immunoblot analyses of the different ZmPIPs re-

    vealed two major bands at approximately 28 and 56 kD,

    corresponding to monomeric and dimeric forms

    (Fig. 6). Interestingly, the variations in protein levels in

    the different root segments correlated quite well with

    the variation in transcript amount. ZmPIP2;1, encoded

    by one of the most highly expressed ZmPIP genes in

    the root, showed low expression in the first segment

    (cap, meristem, and elongation zone), but its expression

    gradually increased to peak in the fourth segment

    (Fig. 6). The same expression pattern was observed at

    the RNA level (Fig. 3). In each of the two major bands,

    Fig. 5 Antibody specificity.A–E, Alignments of peptidesused to raise antisera againstthe N-terminal region ofZmPIP2 or ZmPIP1 proteins.The peptide synthesized tomake the antiserum is shownat the top. Identical aminoacid residues are shown inbold. The numbers indicatethe peptide position withinthe protein. F. Purified GST-ZmPIP-H fusion proteins(0.5 lg) were separated bySDS-PAGE, transferred toPVDF membrane andimmunostained usingantibodies against His-tag,ZmPIP2;1, ZmPIP2;5,ZmPIP2;6, ZmPIP1;2, orRsPIP1 as described in‘‘Materials and methods’’

    314 Plant Mol Biol (2006) 62:305–323

    123

  • this antibody detected two bands with very close

    molecular masses. The upper band could correspond to

    the ZmPIP2;2 isoform, which has a slightly higher

    molecular mass and isoelectric point (30,259 Da, 8.15)

    than ZmPIP2;1 (30,214 Da, 7.81) and was less ex-

    pressed at the RNA level (Figs. 3 and 6), or to a post-

    transcriptional modification of ZmPIP2;1. ZmPIP2;5

    was barely detected in the two first root segments but

    clearly observed in the third one, and its level increased

    up to 5 cm from the root apex (Fig. 6), reflecting the

    changes in mRNA levels. ZmPIP2;6, which had the

    most stable transcript level of the ZmPIP genes

    (Fig. 3), was already detected in the first segment and

    its level did not vary significantly between the different

    regions along the root axis. The ZmPIP1 subgroup

    showed a strong developmental expression pattern, the

    genes being expressed at low levels (or not at all in the

    case of ZmPIP1;6) in the first segment, then at gradu-

    ally higher levels, reaching a fairly steady state in the

    third or fourth segment, corresponding to mature pri-

    mary root tissue (Fig. 3). This trend was faithfully re-

    flected at the protein level, the signal being virtually

    absent in the first zone and gradually increasing to a

    maximum in mature primary root tissues (Fig. 6).

    Together, these data indicate a clear correlation be-

    tween mRNA and protein levels in maize root grown in

    aeroponic conditions. This correlation between AQP

    mRNA and protein levels has not always been

    observed, e.g. in roots exposed to chilling stress (Aroca

    et al. 2005) or in other plant species (Suga et al. 2002);

    (Alexandersson et al. 2005); (Boursiac et al. 2005).

    However, the lack of isoform specificity of the antibody

    and of spatial resolution within the analyzed tissues

    made the interpretation of the data of these studies

    much more difficult.

    In situ immunolocalization of ZmPIP2;1

    and ZmPIP2;5

    ZmPIP2;5 and ZmPIP2;1 are two of the most highly

    expressed ZmPIP mRNAs and proteins in the primary

    maize root tip. To more precisely determine the

    localization of both isoforms in developing roots,

    immunocytochemistry was performed on transverse

    sections taken at 5 mm (young tissue) and 5 cm

    (mature tissue with suberized hypodermis and endo-

    dermis) from the root tip (Figs. 7 and 8). The intensity

    of the green color indicates the abundance of proteins.

    As expected, the signals were observed around the

    cells, indicating a plasma membrane localization of the

    proteins. As shown above, the antibodies against

    ZmPIP2;1 also recognize ZmPIP2;2

    At the start of the elongation zone (Fig. 7A–F),

    ZmPIP2;1/2;2 was mainly detected in the stele tissues,

    the endodermis, and, to a lesser extent, the inner part

    of the cortex (Fig. 7A, D, and E). No signal was

    recorded in the epidermis (Fig. 7D). This labeling

    contrasts with that for H+-ATPases, which were

    uniformly detected in all root tissues, including the

    PMA

    ZmPIP2;1/2;2

    ZmPIP2;5

    ZmPIP2;6

    ZmPIP1;2

    RsPIP1

    100

    55

    24

    55

    24

    55

    24

    55

    24

    24

    kD

    0-5 5-10 50-6030-4010-20 100-110

    Distance from the root apex (mm)

    55

    Fig. 6 ZmPIP protein expression in primary roots. Using 7-day-old primary roots, crude microsomal membranes were preparedfrom the same segments used in Fig. 3 and the proteins (15 lg)subjected to Western blotting using antibodies against PMA (H+-ATPases), ZmPIP2;1/2;2, ZmPIP2;6, ZmPIP1;2, or RsPIP1. ThePMA signal was used to control for gel loading and normalizethe signals. The positions of the molecular mass markers areindicated

    Plant Mol Biol (2006) 62:305–323 315

    123

  • Fig. 7 Immunocytochemical localization of ZmPIP2;1/2;2 inmaize roots. A–M, Transverse sections at 5 mm (A–F) or 5 cm(G–M) from the root tip. A, Detection of ZmPIP2;1/2;2 in thestele, endodermis, and inner cortex. Note the absence of signal inthe epidermis and a strong signal in the stele and the innercortical cell layers. B, Negative control for A. Pre-immune serumwas used instead of anti-ZmPIP2;1/2;2 antibodies. C, Localiza-tion of plasma membrane H+-ATPases (PMA) used as control.A strong signal was recorded throughout the root. D, Localiza-tion of ZmPIP2;1/2;2 in the epidermis and parenchyma cells. Nosignal was detected in the epidermis, whereas cortical cells werelabeled. E, Localization of ZmPIP2;1/2;2 at the junction of thecortex and the stele; a strong signal was seen in the endodermis.F, Localization of PMA in the epidermis and parenchyma cells.

    G, Localization of ZmPIP2;1/2;2 in transverse section 5 cm fromthe root tip. The protein was present in the epidermis, cortex,and endodermis. The signal was much fainter in the stele. H,Negative control for G (pre-immune serum). I, Same as G,except that the signal was stronger in the stele. J, Localization ofZmPIP2;1/2;2 in the epidermis and exodermis; strong labeling ofthese cell layers was seen. K, Localization of ZmPIP2;1/2;2 inroot hairs. L, Localization of ZmPIP2;1/2;2 at the junction of thecortex and the stele; strong labeling of the endodermis was seen.The signal in the cortex was higher than in the stele. M,Localization of PMA in a transverse section 5 cm from the roottip. A strong signal was recorded in the periphery of the cortexand in the stele. Bars = 80 lm in A, B, G, and H, 100 lm in Cand M, 25 lm in D, E, J, K, and L, 40 lm in F, and 110 lm in I

    316 Plant Mol Biol (2006) 62:305–323

    123

  • epidermis (Fig. 7C and F). Interestingly, in the mature

    zone of the primary root, the signal pattern was re-

    versed, with stronger ZmPIP2;1/2;2 labeling in the

    epidermis, exodermis, and cortex than the stele

    (Fig. 7G–L), confirming the results obtained at the

    transcriptional level by Zhu et al. (2005). The intensity

    of the ZmPIP2;1/2;2 signal in the stele varied between

    samples, but was always lower than that in the cortex

    (Fig. 7G and I). The epidermal and endodermal cells

    were strongly labeled, with high ZmPIP2;1/2;2 accu-

    mulation at epidermal cell junctions (Fig. 7J). The root

    hairs also showed labeling of their plasma membrane,

    with some internal spots possibly corresponding to

    fusing vesicles (Fig. 7K).

    ZmPIP2;5 displayed some significant differences in

    its expression pattern compared to ZmPIP2;1/2;2. At

    the start of the elongation zone (Fig. 8A–F), the stele

    tissue was strongly labeled (Fig. 8A and E), as with

    ZmPIP2;1/2;2, but a stronger ZmPIP2;5 signal was seen

    in the cortical cells (Fig. 8A and D). More remarkably,

    contrary to ZmPIP2;1/2.2, a polarized signal was ob-

    served in the epidermal cells of this zone, in which a

    stronger ZmPIP2;5 signal was seen on the plasma

    membrane exposed to the external medium, thus

    exhibiting a characteristic U-shape (Fig. 8A and C). In

    more mature tissue, about 1 cm from the root tip, the

    endodermis was strongly labeled (Fig. 7F). In the ma-

    ture zone, ZmPIP2;5 was highly expressed in the epi-

    dermis/exodermis cell layers and at the endodermis

    boundary (Fig. 7G, J, I, and K), suggesting that this

    isoform is involved in facilitating transcellular water

    movement in these regions. As with ZmPIP2;1/2;2, the

    overall ZmPIP2;5 signal intensity was stronger in the

    cortex than in the stele tissues (Figs. 8G, J, K), con-

    firming the transcriptional data obtained by Zhu et al.

    (2005).

    Osmotic water permeability coefficients of root

    protoplasts

    To determine whether the increase in ZmPIP expres-

    sion according to the root developmental stages cor-

    related with an increase in osmotic water permeability

    of the cell membrane (Pf), we analyzed the swelling of

    the protoplasts in response to hypoosmotic challenge

    as described previously (Moshelion et al. 2004). As

    shown in Fig. 9, all the cells examined (N = 43) in the

    zone at 0–5 mm (cap, meristem, and start of the elon-

    gation zone) had a similar Pf value of < 10 lm/s, while,of the 50 cells examined in the zone at 10–20 mm (root

    hair zone), some (N = 10, 20%) exhibited a Pf higher

    than 10 lm/s, while the remainder (N = 40, 80%)constituted a ‘‘low Pf cell’’ population with Pf

    values < 10 lm/s (Fig. 9B–C). The low Pf cells in the10–20 mm zone had a mean Pf similar to those of the

    low Pf cells in the other zones. Moreover, the same

    division between high and low Pf cells was observed for

    the zone at 40–60 mm (data not shown). AQPs

    appeared to control the water permeability, since the

    Pf of the protoplasts was decreased by 67% after 5 min

    incubation with 250 lM HgCl2, which inhibits AQPs(Fig. 9D).

    Discussion

    ZmPIP genes are widely expressed in the maize

    primary root

    Quantitative and in situ RT-PCR techniques were used

    to monitor PIP gene expression in maize primary

    roots. Since ZmPIP1;1, ZmPIP1;5, ZmPIP2;1, and

    ZmPIP2;5 transcript abundance exhibits a diurnal

    variation, with a peak at 2–4 h after light onset (Lopez

    et al. 2003), all samples analyzed were collected during

    this period of the day. We showed that all except

    ZmPIP2;7 were transcribed at various levels. The four

    most highly expressed genes were ZmPIP1;1,

    ZmPIP1;5, ZmPIP2;1, and ZmPIP2;5. In general,

    these data are in accordance with the number of ex-

    pressed sequence tags obtained from maize root tissues

    (Chaumont et al. 2001) and with DNA array hybrid-

    ization studies recently performed on maize roots (Zhu

    et al. 2005). However, we found ZmPIP1;5 to be less

    abundantly expressed and ZmPIP2;5 more highly ex-

    pressed than in the latter study. These discrepancies

    might be explained by the different maize genotypes

    and growing conditions used in the two studies. Our

    seedlings were grown aeroponically in a system

    exposing the roots to a humid atmosphere, while Zhu

    et al. (2005) grew their plants in aerated hydroponic

    medium.

    The observation that almost all ZmPIP genes were

    widely expressed in the primary root is quite puzzling,

    but is not specific to maize. Similar results have been

    reported in analyses of global AQP expression in

    Arabidopsis, which demonstrated that all AtPIP genes

    were expressed in roots, but showed differential regu-

    lation according to the environmental conditions (Al-

    exandersson et al. 2005; Boursiac et al. 2005; Jang

    et al. 2004). This confirms the general view that AQPs

    play an important physiological role in facilitating

    water movement across cell membranes in roots.

    Nevertheless, the fact that almost all ZmPIP genes

    were transcribed simultaneously in the different root

    tip tissues, as shown by in situ RT-PCR experiments

    Plant Mol Biol (2006) 62:305–323 317

    123

  • (Fig. 4), raises the question whether the ability to

    conduct water is their sole physiological role. Evidence

    is accumulating that plant AQPs exhibit diverse regu-

    lation and transport properties and can behave as

    multifunctional channels, suggesting that these pro-

    teins may have different physiological roles, which

    might explain the isoform diversity (Chaumont et al.

    2005; Luu and Maurel 2005; Tyerman et al. 2002). In

    addition, the observation that most ZmPIP genes from

    both the PIP1 and PIP2 subfamilies were co-expressed

    in the same cells suggests that they might physically

    interact and form heterotetramers in planta. It is

    Fig. 8 Immunocytochemical localization of ZmPIP2;5 in maizeroots. A–K, Transverse sections at 5 mm (A–E), 1 cm (F), or5 cm (G–K) from the root tip. A, Detection of ZmPIP2;5 in thestele, endodermis, cortex, and epidermis. A strong signal wasrecorded in the stele, endodermis and inner cortex. Note thepolarization of the signals in the epidermis. B, Negative controlfor A using pre-immune serum. C, Polarized ZmPIP2;5 signal inthe epidermis D, Detection of ZmPIP2;5 in the inner paren-chyma cells. E, Detection of ZmPIP2;5 in the xylem parenchymacells. F, Detection of ZmPIP2;5 in the stele and endodermis at1 cm from the root tip; strong labeling of the latter was seen. G,Detection of ZmPIP2;5 in a transverse section 5 cm from the

    root tip. A strong signal was recorded in the epidermis, cortexand endodermis. The signal was much fainter in the stele. H,Negative control for G (pre-immune serum). I, Detection ofZmPIP2;5 in the epidermis and exodermis. The signal wasstronger in the exodermis than in the epidermis. J, Same as Gexcept that the signal was concentrated in two ‘‘rings’’, oneconsisting of the epidermis and exodermis and the other the cellsfacing the stele. K, Detection of ZmPIP2;5 at the junction of thecortex and the stele. The signal was stronger in the cortex thanthe stele. Bars = 90 lm in A, B, F, G, and H, 30 lm in C, I, K,18 lm in D and E, and 200 lm in J

    318 Plant Mol Biol (2006) 62:305–323

    123

  • interesting to observe that the ratio between PIP1 and

    PIP2 gene expression in the different root segments is

    relatively constant (Table 1). We previously showed

    that co-expression of ZmPIP1;2 and ZmPIP2s in

    Xenopus oocytes increases both the amount of

    ZmPIP1;2 in the plasma membrane and the membrane

    osmotic water permeability (Fetter et al. 2004). Inter-

    estingly, Martre et al. (2002) pointed out that, when

    comparing the water parameter data obtained from

    PIP1, PIP2, or PIP1 + PIP2 antisense plants, the PIP

    subgroup that was not down-regulated appeared inac-

    tive. This observation further supports the theory that

    PIP heteromerization regulates the cell Pf.

    ZmPIP expression in the root is developmentally

    regulated

    Our data revealed that the content of all ZmPIPs

    mRNAs except ZmPIP2;6 mRNA was strongly

    dependent on the developmental stage of the root, with

    a general increase in expression towards the elongation

    and mature zones (Fig. 3). Two genes, ZmPIP1;5 and

    ZmPIP2;5, showed the greatest increase in expression

    (up to 11- and 17-fold, respectively) in the mature

    zone, where they accounted for 50% of the total

    expressed ZmPIPs, indicating an important role of

    these two AQPs. The use of antibodies recognizing

    ZmPIP2;1/2;2, ZmPIP2;5, ZmPIP2;6, ZmPIP1;2 or

    ZmPIP1s showed that the amounts of protein detected

    in the different root zones correlated well with the

    transcript abundance (Fig. 6). This AQP mRNA-pro-

    tein correlation has not always been observed in plant

    tissues (Aroca et al. 2005; Boursiac et al. 2005; Lopez

    et al. 2003), but this could be partly due to the general

    lack of antibody specificity, which prevented the

    accurate study of a specific isoform. This increase in

    ZmPIP expression in maturing root zones was previ-

    ously reported by Hukin et al. (2002), who showed a

    huge increase in ZmPIP1;2 and ZmPIP2;4 gene

    expression in the cell elongation zone compared to the

    meristematic zone. Growing cells near the root tip are

    symplastically connected to each other and to the

    phloem by plasmodesmata (Bret-Harte and Silk 1994).

    In contrast to the youngest cells, those located in the

    root elongation zone loose symplastic continuity

    (Hukin et al. 2002). The symplastic isolation may

    extend to the more mature regions of the growing

    zone, and may change under different solute avail-

    0

    2

    4

    6

    8

    10

    12

    14

    16

    0

    5

    10

    15

    20

    25

    30

    35

    0 2 4 6 8 10 12 14 16 18 20 22 24 26 28

    zone between 0-5mm

    zone between 10-20 mm

    0-5

    mm

    10-2

    0 m

    m

    (A)

    (C) (D)

    (B)

    95

    100

    105

    110

    115

    120

    125

    130

    135

    140

    0 10 20 30 40 50 60

    High PfLow Pf

    Rel

    ativ

    e vo

    lum

    e

    Time (s)

    Pf at 60 s (µm/s)

    Perc

    enta

    ge

    P f (

    µm/s

    )

    Control

    Mercury

    Fig. 9 Measurement of theosmotic water permeabilitycoefficient of root cells. A,Schematic diagram of the rootshowing the two zones studiedat 0–5 mm and 10–20 mm. B,Distribution (as a percentageof the cell population) of theosmotic water permeabilitycoefficient (Pf) in the twozones. Note the ‘‘high Pf’’population of cells in the zonebetween 10 and 20 mm. C,Comparison of swellingbehavior of sevenrepresentative root protoplastsfrom each of the high and lowPf cell populations from the10–20 mm zone. The meanand standard deviation areshown. D, Pf values of cellsfrom the 0 to 20 mm zonebefore, and after, incubationwith 250 lM HgCl2. The meanand standard deviation for 15controls and 18 mercurial-treated cells are shown

    Plant Mol Biol (2006) 62:305–323 319

    123

  • ability regimes (Patrick and Offler 1996). The increase

    in ZmPIP transcript abundance along the longitudinal

    root axis can be attributed to a decrease in symplastic

    continuity along the growing and maturing zones

    (Hukin et al. 2002) and therefore to a need for facili-

    tated water transport through the cellular plasma

    membrane.

    Another physiological observation that could re-

    quire higher AQP expression in mature regions of the

    roots is that, in this zone, the xylem becomes more

    conductive and water demand driven by the transpi-

    ration stream is higher than in younger regions, where

    water demand is essentially driven by cell elongation.

    In addition, the development of suberized apoplastic

    barriers in the mature zone might require higher AQP

    expression in this zone to facilitate radial water

    movement to the stele. Under aeroponic growth con-

    ditions, the suberized endodermis was detected 3 cm

    from the root apex, whereas the Casparian strips from

    the exodermis were detected 1 cm further from the

    apex (Fig. 2). Interestingly, the zones in which apo-

    plastic barriers were suberized corresponded to regions

    in which ZmPIP expression reached its maximum. The

    development of barriers made of suberin and lignin

    would force the plant to increase its membrane water

    permeability at these crucial places by higher expres-

    sion of AQPs.

    On the other hand, ZmPIP2;6 gene (Fig. 3) and

    protein (Fig. 6) expression did not vary during root

    development, and its mRNA was detected in all root

    cells by in situ RT-PCR (Fig. 4). ZmPIP2;6 expression

    in the outer layers of the roots (epidermis, cortex, and

    endodermis) and in the inner tissue (stele) is also

    reported to be similar (Zhu et al. 2005). Together, these

    data suggest that ZmPIP2;6 isoform is constitutively

    expressed at similar level in maize primary root.

    Functional assay in Xenopus oocytes showed that this

    protein, like the previously characterized ZmPIP2;1,

    ZmPIP2;4, and ZmPIP2;5 (Chaumont et al. 2000; Fet-

    ter et al. 2004; Lopez et al. 2003), forms an active water

    channel (Hachez and Chaumont, unpublished data).

    Differential distribution of ZmPIP2;1/2;2

    and ZmPIP2;5 in the primary root

    Immunocytochemistry experiments allowed us to

    localize ZmPIP2;1 and ZmPIP2;5, two of the most

    highly expressed AQPs. The antibodies against

    ZmPIP2;1 also recognize ZmPIP2;2, while those

    against ZmPIP2;5 are isoform-specific. At 5 mm from

    the root tip, both antisera mainly labeled the stele

    tissues and the inner cell layers of the cortex. These

    expression patterns might be explained by a phloem

    unloading process. Growing root apices function as

    heterotrophic sink organs that are dependent on the

    continuous supply of photoassimilates from the leaves,

    and these are delivered to the subapical region of the

    root tip through the protophloem sieve elements

    (Patrick and Offler 1996). Phloem unloading in the

    root tips is usually believed to occur symplastically, as

    indicated by the plasmodesmata linking the sieve

    elements and meristematic cells (Bret-Harte and Silk

    1994; Oparka et al. 1994). However, the possibility of

    the symplastic and apoplastic pathways operating in

    parallel during phloem unloading has been suggested

    (Bret-Harte and Silk 1994). As sucrose is actively taken

    up, the osmotic potential of the cells decreases, driving

    water entry, and the phloem unloading mechanism

    would then require facilitated water transport to

    rapidly balance the osmotic potential, possibly

    explaining the high PIP expression in the stele. A

    similar function has been attributed to the tonoplast

    AQP, ZmTIP1;1, which is highly expressed in the root

    tip stele (Barrieu et al. 1998).

    Whereas ZmPIP2;1/2;2 was not detected in the

    epidermis of the root apices, ZmPIP2;5 was highly

    expressed in these cells and was specifically localized in

    periclinal plasma membrane domains facing the

    external medium. The polarization of several mammal

    AQPs in apical membranes has been reported (Agre

    and Kozono 2003; Brown 2003) but, to our knowledge,

    this is the first example in plant cells. Even if it is

    estimated that 81% of the growth demand for water in

    the root tip in maize plants growing in vermiculite

    comes from the phloem, the remainder of the water

    might come from inward transport from the root sur-

    face (Bret-Harte and Silk 1994). This directional

    movement of water could be facilitated by a polarized

    localization of ZmPIP2;5 in the epidermis cells.

    ZmPIP2;5 localization underlines the importance of

    the non-apoplastic route in water uptake from the soil

    into the developing tissues. In this zone, root hairs are

    not yet present and a high amount of aquaporins in the

    epidermis apical membrane might compensate for the

    lack of these structures.

    In the mature root zone, 5 cm from the root tip,

    ZmPIP2;1/2;2 and ZmPIP2;5 were detected in the

    epidermis, exodermis, cortex, and endodermis. Due to

    the transpiration stream, water absorption is very

    important in this region (Boyer 1985), and the need for

    facilitated water uptake into the cells is suggested by

    the ZmPIP2;1/2;2 and ZmPIP2;5 expression pattern.

    The development of an exodermis substantially de-

    creases the root hydraulic conductivity in hydrostatic

    experiments (Zimmermann et al. 2000; Zimmermann

    and Steudle 1998) and this effect could therefore be

    320 Plant Mol Biol (2006) 62:305–323

    123

  • partly compensated by the presence of a high level of

    AQPs in the epidermis and exodermis membranes.

    Similarly, the presence of a suberized endodermis

    would require cell water uptake in the cortex, which

    was strongly suggested by the localization of ZmPIP2;5

    (Fig. 8J), which showed increased expression in two

    rings of cells corresponding to the exodermis and cells

    surrounding the endodermis. In the cortex, water could

    move radially in the apoplast, but also use the trans-

    cellular pathway, requiring high expression of ZmPIPs

    (Zimmermann et al. 2000). This is in accordance with

    the composite transport model (Steudle 2000). The

    contribution of AQPs to the overall root hydraulic

    conductivity is reported to account for up to 80% of

    the Lpr in Arabidopsis (Boursiac et al. 2005).

    The anti-ZmPIP2;1/2;2 antibodies also labeled the

    root hair, confirming previous RT-PCR data (Fetter

    et al. 2004). Whereas the whole plasma membrane was

    labeled, cytoplasmic spots were detected in the hair tip,

    possibly reflecting cytoplasmic accumulation of AQP-

    containing vesicles that could, in turn, account for a

    high turn-over of the protein in this region and the

    rapid hair growth.

    Root protoplasts display a low water permeability

    It has been reported that the membrane diffusive water

    permeability is around 10 lm/s, whereas AQP-mediated water transport results in a Pf ranging from

    100 to 1000 lm/s (Tyerman et al. 1999). However,recent findings using different methods indicate that,

    despite the presence of plasma membrane-associated

    AQPs, the plasma membrane can display a low Pf(Chaumont et al. 2005; Moshelion et al. 2004). The low

    water permeability of maize root protoplasts reported

    by Ramahaleo et al. (1999) during the early stages of

    root development are similar to our findings. They also

    reported an increase in the Pf of root cells after 5 days

    and proposed that the larger Pf could be due to

    the developmental expression of AQPs. Such results

    correlate well with our findings. However, protoplasts

    may not ideally reflect the behavior of cells in planta, as

    they are subjected to stressful conditions during their

    preparation. The activity of AQPs in the membrane

    is highly and reversibly regulated by post-translational

    modifications and environmental conditions

    (Chaumont et al. 2005; Luu and Maurel 2005). These

    regulation mechanisms could explain the higher Pfvalues (between 30 and 200 lm/s) measured withpressure probe experiments on intact maize root cells

    (Azaizeh et al. 1992; Zhu and Steudle 1991). A recent

    study comparing the Pf of barley leaf cells measured

    with a pressure probe in intact tissues or using the

    protoplast swelling assay showed that the absolute Pfvalues were higher (up to 5-fold) using the former

    method compared to those determined in protoplasts

    (Volkov V, Hachez C, Moshelion M, Draye X,

    Chaumont F, Fricke W, data not shown). However,

    differences between the Pf of elongating and

    non-elongating cells were similar whatever the method

    used.

    The contribution of AQPs was detectable even in

    low Pf cells, since the Pf was reduced by 67% by incu-

    bation of the protoplasts with 250 lM HgCl2 for 5 min(Fig 9D). Cells with a low water permeability were

    found in the root tip, but also made up the majority of

    the cells in the zones at 10–20 mm and 40–60 mm.

    However, as the root tissues matured, the cells differ-

    entiated and some acquired a significantly higher water

    permeability, a phenomenon that we attributed to

    developmentally regulated expression of PIPs (Fig. 9).

    Given their size (data no shown), these high Pf cells

    might correspond to cells adjacent to apoplastic barri-

    ers with a high membrane permeability to allow cell

    water loading or to cells with a specialized function,

    such as xylem parenchyma cells involved in maintaining

    the transpiration stream and refilling gas-filled xylem

    vessels (Johansson et al. 2000). The existence of some

    high Pf cells in mature parts of the root further supports

    the theory of increased AQP activity in these zones due

    to the symplastic isolation of the cells. This is in

    accordance with the observation that 20 lM mercurycauses higher Lp inhibition in growing and mature cells

    than in cells at the root tip (Hukin et al. 2002).

    Conclusion

    These results show the wide, but developmentally

    regulated, expression of ZmPIPs in maize primary root

    and highlight the importance of two ZmPIP genes,

    ZmPIP1;5 and ZmPIP2;5, which, together, account for

    more than 50% of all expressed ZmPIPs in mature root

    tissue. AQPs provide the plant with the ability to

    rapidly alter its membrane water permeability and this

    is especially important in regions in which cell-to-cell

    water transport acts as a bottleneck at sites of sym-

    plastic isolation or in the vicinity of apoplastic barriers.

    This theory is supported by the increase in AQP

    expression seen during root maturation and the in-

    creased expression of ZmPIP2;1/2;2 and ZmPIP2;5

    proteins in the exodermis and endodermis. Finally the

    polarized positioning of ZmPIP2;5 in the epidermis of

    the root apex suggests a role in facilitating directional

    water movement from the soil. More detailed mapping

    of all the AQPs along the root and an examination of

    Plant Mol Biol (2006) 62:305–323 321

    123

  • the correlation between their localization and cell

    water parameters will clarify their role in root growth

    and plant water supply.

    Acknowledgements This work was supported by grants fromthe Belgian National Fund for Scientific Research (FNRS), theInteruniversity Attraction Poles Programme—Belgian SciencePolicy, and the ‘‘Communauté française de Belgique—Actionsde Recherches Concertées’’. F.C. is a Senior Research Associateand C.H. a Research Fellow at the FNRS; E.Z. is a ResearchFellow at the Fonds pour la Formation à la Recherche dansl’Industrie et dans l’Agriculture. We thank R. Jung (PioneerHi-Bred International) for providing ZmPIP cDNAs,M. Maeshima (Nagoya University, Japan) and M. Boutry (Uni-versité catholique de Louvain) for supplying the anti-RsPIP1 andanti-PMA antibodies, respectively, and T. Trombik andE. Peeters (Université catholique de Louvain) for supplyingpGex-KG’ plasmid. We are very grateful to X. Draye andT. Lavigne for the use of the aeroponics facility and advices, andto M. Boutry and X. Draye for their critical reading of themanuscript.

    References

    Agre P, Kozono D (2003) Aquaporin water channels: molecularmechanisms for human diseases. FEBS Lett 555:72–78

    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

    Amodeo G, Dorr R, Vallejo A, Sutka M, Parisi M (1999) Radialand axial water transport in the sugar beet storage root. JExp Bot 50:509–516

    Aroca R, Amodeo G, Fernandez-Illescas S, Herman EM,Chaumont F, Chrispeels MJ (2005) The role of aquaporinsand membrane damage in chilling and hydrogen peroxideinduced changes in the hydraulic conductance of maizeroots. Plant Physiol 137:341–353

    Azaizeh H, Gunse B, Steudle E (1992) Effects of NaCl andCaCl2 on water transport across root-cells of maize (Zea-mays L.) seedlings. Plant Physiol 99:886–894

    Barrieu F, Chaumont F, Chrispeels MJ (1998) High expression ofthe tonoplast aquaporin ZmTIP1 in epidermal and con-ducting tissues of maize. Plant Physiol 117:1153–1163

    Boursiac Y, Chen S, Luu DT, Sorieul M, van den Dries N,Maurel C (2005) Early effects of salinity on water transportin Arabidopsis roots. Molecular and cellular features ofaquaporin expression. Plant Physiol 139:790–805

    Boyer J (1985) Water transport. Ann Rev Plant Physiol 36:473–516

    Bret-Harte MS, Silk WK (1994) Nonvascular, symplasmic dif-fusion of sucrose cannot satisfy the carbon demands ofgrowth in the primary root tip of Zea mays L. Plant Physiol105:19–33

    Brown D (2003) The ins and outs of aquaporin-2 trafficking. AmJ Physiol Renal Physiol 284:F893–F901

    Brundrett MC, Enstone DE, Peterson CA (1988) A berberine-aniline blue fluorescent staining procedure for suberin, lig-nin, and callose in plant-tissue. Protoplasma 146:133–142

    Carvajal M, Cerda A, Martinez V (2000) Does calcium amelio-rate the negative effect of NaCl on melon root watertransport by regulating aquaporin activity? New Phytol145:439–447

    Carvajal M, Cooke DT, Clarkson DT (1996) Responses of wheatplants to nutrient deprivation may involve the regulation ofwater-channel function. Planta 199:372–381

    Carvajal M, Martinez V, Alcaraz CF (1999) Physiological func-tion of water channels as affected by salinity in roots ofpaprika pepper. Physiol Plant 105:95–101

    Chaumont F, Barrieu F, Jung R, Chrispeels MJ (2000) Plasmamembrane intrinsic proteins from maize cluster in two se-quence subgroups with differential 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

    de Dorlodot S, Bertin P, Baret PV, Draye X (2005) Scaling upquantitative phenotyping of root system architecture using acombination of aeroponics and image analysis. Aspect ApplBiol 73:41–54

    Enstone DE, Peterson CA (2005) Suberin lamella developmentin maize seedling roots grown in aerated and stagnantconditions. Plant Cell Environ 28:444–455

    Fetter K, Van Wilder V, Moshelion M, Chaumont F (2004)Interactions between plasma membrane aquaporins modu-late their water channel activity. Plant Cell 16:215–228

    Frensch J, Hsiao TC, Steudle E (1996) Water and solute trans-port along developing maize roots. Planta 198:348–355

    Gaspar M, Bousser A, Sissoeff I, Roche O, Hoarau J, Mahe A(2003) Cloning and characterization of ZmPIP1-5b, anaquaporin transporting water and urea. Plant Sci 165:21–31

    Guan KL, Dixon JE (1991) Eukaryotic proteins expressed inEscherichia coli: an improved thrombin cleavage and puri-fication procedure of fusion proteins with glutathioneS-transferase. Anal Biochem 192:262–267

    Hose E, Clarkson DT, Steudle E, Schreiber L, Hartung W (2001)The exodermis: a variable apoplastic barrier. J Exp Bot52:2245–2264

    Hukin D, Doering-Saad C, Thomas CR, Pritchard J (2002)Sensitivity of cell hydraulic conductivity to mercury iscoincident with symplasmic isolation and expression ofPlasmalemma aquaporin genes in growing maize roots.Planta 215:1047–1056

    Ishikawa H, Evans ML (1995) Specialized zones of developmentin roots. Plant Physiol 109:725–727

    Jang JY, Kim DG, Kim YO, Kim JS, Kang H (2004) Anexpression analysis of a gene family encoding plasmamembrane aquaporins in response to abiotic stresses inArabidopsis thaliana. Plant Mol Biol 54:713–725

    Javot H, Maurel C (2002) The role of aquaporins in root wateruptake. Ann Bot (Lond) 90:301–313

    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

    Karahara I, Ikeda A, Kondo T, Uetake Y (2004) Developmentof the Casparian strip in primary roots of maize under saltstress. Planta 219:41–47

    Lopez M, Bousser AS, Sissoeff I, Gaspar M, Lachaise B, HoarauJ, Mahe A (2003) Diurnal regulation of water transport andaquaporin gene expression in maize roots: contribution ofPIP2 proteins. Plant Cell Physiol 44:1384–1395

    Luu DT, Maurel C (2005) Aquaporins in a challenging envi-ronment: molecular gears for adjusting plant water status.Plant Cell Environ 28:85–96

    Maggio A, Joly RJ (1995) Effects of mercuric chloride on thehydraulic conductivity of tomato root systems (evidence for

    322 Plant Mol Biol (2006) 62:305–323

    123

  • a channel-mediated water pathway). Plant Physiol 109:331–335

    Marino JH, Cook P, Miller KS (2003) Accurate and statisticallyverified quantification of relative mRNA abundances usingSYBR Green I and real-time RT-PCR. J Immunol Methods283:291–306

    Martre P, Morillon R, Barrieu F, North GB, Nobel PS, Chri-speels MJ (2002) Plasma membrane aquaporins play a sig-nificant role during recovery from water deficit. PlantPhysiol 130:2101–2110

    Martre P, North GB, Nobel PS (2001) Hydraulic conductanceand mercury-sensitive water transport for roots of Opuntiaacanthocarpa in relation to soil drying and rewetting. PlantPhysiol 126:352–362

    Maurel C, Chrispeels MJ (2001) Aquaporins. A molecularentry into plant water relations. Plant Physiol 125:135–138

    Moore R, Smith HS (1990) Morphometric analysis of epidermaldifferentiation in primary roots of Zea-mays. Am J Bot77:727–735

    Morsomme P, de Kerchove d’Exaerde A, De Meester S, ThinesD, Goffeau A, Boutry M (1996) Single point mutations invarious domains of a plant plasma membrane H(+)-ATPaseexpressed in Saccharomyces cerevisiae increase H(+)-pumping and permit yeast growth at low pH. Embo J15:5513–5526

    Moshelion M, Moran N, Chaumont F (2004) Dynamic changes inthe osmotic water permeability of protoplast plasma mem-brane. Plant Physiol 135:2301–2317

    Nuovo GJ (1996) The foundations of successful RT in situ PCR.Front Biosci 1:c4–c15

    Oparka KJ, Duckett CM, Prior DAM, Fisher DB (1994) Real-time imaging of phloem unloading in the root-tip of ara-bidopsis. Plant J 6:759–766

    Otto B, Kaldenhoff R (2000) Cell-specific expression of themercury-insensitive plasma-membrane aquaporin NtAQP1from Nicotiana tabacum. Planta 211:167–172

    Patrick JW, Offler CE (1996) Post-sieve element transport ofphotoassimilates in sink regions. J Exp Bot 47:1165–1177

    Perumalla CJ, Peterson CA (1986) Deposition of Casparianbands and suberin lamellae in the exodermis andendodermis of young corn and onion roots. Can J Bot64:1873–1878

    Pfaffl MW (2001) A new mathematical model for relativequantification in real-time RT-PCR. Nucleic Acids Res29:e45

    Ramahaleo T, Morillon R, Alexandre J, Lassalles JP (1999)Osmotic water permeability of isolated protoplasts. Modi-fications during development. Plant Physiol 119:885–896

    Raven PH, Evert RF, Eichhorn SE (1992) The movement ofwater and solutes in plants. In: Mastalski SAE (ed) Biologyof plants. Worth Publishers, New York

    Schaffner AR (1998) Aquaporin function, structure, andexpression: are there more surprises to surface in waterrelations? Planta 204:131–139

    Schraut D, Ullrich CI, Hartung W (2004) Lateral ABA transportin maize roots (Zea mays): visualization by immunolocali-zation. J Exp Bot 55:1635–1641

    Steudle E (2000) Water uptake by plant roots: an integration ofviews. Plant Soil 226:45–56

    Steudle E (2001) The cohesion-tension mechanism and theacquisition of water by plant roots. Annu Rev Plant PhysiolPlant Mol Biol 52:847–875

    Suga S, Imagawa S, Maeshima M (2001) Specificity of theaccumulation of mRNAs and proteins of the plasma mem-brane and tonoplast aquaporins in radish organs. Planta212:294–304

    Suga S, Komatsu S, Maeshima M (2002) Aquaporin isoformsresponsive to salt and water stresses and phytohormones inradish seedlings. Plant Cell Physiol 43:1229–1237

    Tazawa M, Ohkuma E, Shibasaka M, Nakashima S (1997)Mercurial-sensitive water transport in barley roots. J PlantRes 110:435–442

    Tyerman SD, Bohnert HJ, Maurel C, Steudle E, Smith JAC(1999) Plant aquaporins: their molecular biology, biophysicsand significance for plant water relations. J Exp Bot50:1055–1071

    Tyerman SD, Niemietz CM, Bramley H (2002) Plant aquaporins:multifunctional water and solute channels with expandingroles. Plant Cell Environ 25:173–194

    Vandesompele J, De Preter K, Pattyn F, Poppe B, Van Roy N,De Paepe A, Speleman F (2002) Accurate normalization ofreal-time quantitative RT-PCR data by geometric averagingof multiple internal control genes. Genome Biol 3:RE-SEARCH0034

    Zhu C, Schraut D, Hartung W, Schaffner AR (2005) Differentialresponses of maize MIP genes to salt stress and ABA. J ExpBot 56:2971–2981

    Zhu GL, Steudle E (1991) Water transport across maizeroots—simultaneous measurement of flows at the cell androot level by double pressure probe technique. Plant Physiol95:305–315

    Zimmermann HM, Hartmann K, Schreiber L, Steudle E (2000)Chemical composition of apoplastic transport barriers inrelation to radial hydraulic conductivity of corn roots (Zeamays L.). Planta 210:302–311

    Zimmermann HM, Steudle E (1998) Apoplastic transport acrossyoung maize roots: effect of the exodermis. Planta 206:7–19

    Plant Mol Biol (2006) 62:305–323 323

    123

    Localization and quantification of plasma membrane aquaporin expression in maize primary root: a clue to understanding their role as cellular plumbersAbstractIntroductionMaterials and methodsPlant materialIn situ reverse transcriptase-mediated PCRDetection of apoplastic barriersQuantitative RT-PCRRoot cell protoplast isolationGST-ZmPIP-H expressionSubcellular fractionationProtein analysisIn situ immunolocalizationResultsExpression of ZmPIP mRNAs in aeroponically grown maize rootsDevelopment of apoplastic barriersExpression of ZmPIPs according to root developmental stageFig1Fig2Expression of ZmPIP genes in the cap, meristem, and elongation zonesFig3Tab1Fig4ZmPIP protein levels in rootsFig5In situ immunolocalization of ZmPIP2;1 �and ZmPIP2;5Fig6Fig7Osmotic water permeability coefficients of root protoplastsDiscussionZmPIP genes are widely expressed in the maize primary rootFig8ZmPIP expression in the root is developmentally regulatedFig9Differential distribution of ZmPIP2;1/2;2 �and ZmPIP2;5 in the primary rootRoot protoplasts display a low water permeabilityConclusionAcknowledgementsReferencesCR1CR2CR3CR4CR5CR6CR7CR8CR9CR10CR11CR12CR13CR14CR15CR16CR17CR18CR19CR20CR21CR22CR23CR24CR25CR26CR27CR28CR29CR30CR31CR32CR33CR34CR35CR36CR37CR38CR39CR40CR41CR42CR43CR44CR45CR46CR47CR48CR49CR50CR51CR52CR53CR54CR55CR56CR57CR58CR59CR60CR61CR62

    /ColorImageDict > /JPEG2000ColorACSImageDict > /JPEG2000ColorImageDict > /AntiAliasGrayImages false /DownsampleGrayImages true /GrayImageDownsampleType /Bicubic /GrayImageResolution 150 /GrayImageDepth -1 /GrayImageDownsampleThreshold 1.50000 /EncodeGrayImages true /GrayImageFilter /DCTEncode /AutoFilterGrayImages true /GrayImageAutoFilterStrategy /JPEG /GrayACSImageDict > /GrayImageDict > /JPEG2000GrayACSImageDict > /JPEG2000GrayImageDict > /AntiAliasMonoImages false /DownsampleMonoImages true /MonoImageDownsampleType /Bicubic /MonoImageResolution 600 /MonoImageDepth -1 /MonoImageDownsampleThreshold 1.50000 /EncodeMonoImages true /MonoImageFilter /CCITTFaxEncode /MonoImageDict > /AllowPSXObjects false /PDFX1aCheck false /PDFX3Check false /PDFXCompliantPDFOnly false /PDFXNoTrimBoxError true /PDFXTrimBoxToMediaBoxOffset [ 0.00000 0.00000 0.00000 0.00000 ] /PDFXSetBleedBoxToMediaBox true /PDFXBleedBoxToTrimBoxOffset [ 0.00000 0.00000 0.00000 0.00000 ] /PDFXOutputIntentProfile (None) /PDFXOutputCondition () /PDFXRegistryName (http://www.color.org?) /PDFXTrapped /False

    /Description >>> setdistillerparams> setpagedevice