Plant, Cell and Environment (2007) 30, 1566–1577
doi: 10.1111/j.1365-3040.2007.01732.x
Aquaporins and unloading of phloem-imported water in
coats of developing bean seeds
YUCHAN ZHOU1, NATHAN SETZ1, CHRISTA NIEMIETZ2, HONGXIA QU3, CHRISTINA E. OFFLER1,
STEPHEN D. TYERMAN2 & JOHN W. PATRICK1
1
School of Environmental and Life Sciences, The University of Newcastle, Callaghan, NSW 2308, Australia, 2School of
Agriculture, Food and Wine, Adelaide University, Waite Campus, Glen Osmond, SA 5064, Australia and 3South China
Botanical Garden, Chinese Academy of Sciences, Guangzhou 510650, China
ABSTRACT
Nutrients are imported into developing legume seeds by
mass flow through the phloem, and reach developing
embryos following secretion from their symplasmically isolated coats. To sustain homeostasis of seed coat water relations, phloem-delivered nutrients and water must exit seed
coats at rates commensurate with those of import through
the phloem. In this context, coats of developing French
bean seeds were screened for expression of aquaporin
genes resulting in cloning PvPIP1;1, PvPIP2;2 and
PvPIP2;3. These genes were differentially expressed in all
vegetative organs, but exhibited their strongest expression
in seed coats. In seed coats, expression was localized to cells
of the nutrient-unloading pathway. Transport properties of
the PvPIPs were characterized by expression in Xenopus
oocytes. Only PvPIP2;3 showed significant water channel
activity (Pos = 150–200 mm s-1) even when the plasma membrane intrinsic proteins (PIPs) were co-expressed in various
combinations. Permeability increases to glycerol, methylamine and urea were not detected in oocytes expressing
PvPIPs. Transport active aquaporins in native plasma membranes of seed coats were demonstrated by measuring rates
of osmotic shrinkage of membrane vesicles in the presence
and absence of mercuric chloride and silver nitrate. The
functional significance of aquaporins in nutrient and water
transport in developing seeds is discussed.
Key-words: Phaseolus vulgaris; functional characterization;
plasma membrane intrinsic proteins; water transport.
INTRODUCTION
Developing seeds are net importers of organic and inorganic nutrients (Zhang et al. 2007). Nutrients unloaded
from the phloem are probably delivered by bulk flow from
sieve elements to their symplasmically connected vascular
parenchyma cells (Fisher & Cash-Clark 2000).Thus, phloem
water, coimported with nutrients, must exit the symplasm to
avoid a build up of hydrostatic pressure in seed coats (cf.
Murphy 1989). In comparison to nutrient efflux to the
Correspondence: J. W. Patrick. Fax: 61 02 49 21 6923; e-mail:
[email protected]
1566
apoplasm of developing grain legume seeds, little is known
about the mechanism(s) of water flows in developing seeds
(Zhang et al. 2007). One possibility is that membrane flows
of phloem-imported water are regulated by aquaporins
embedded in plasma membranes of seed coat cells responsible for water exchange to the seed apoplasm.
Aquaporins belong to a family of major intrinsic proteins
(MIPs). Phylogenetic analyses divide plant aquaporins into
four subfamilies: plasma membrane intrinsic proteins
(PIPs), tonoplast intrinsic proteins (TIPs), NOD26-like
MIPs (NIPs) and small basic intrinsic proteins (SIPs). The
PIPs subfamily can be further separated into two distinct
phylogenetic groups, PIP1 and PIP2 (Johanson et al. 2001).
Functionally, MIP family members fall into two functional
groups: aquaporins and glycerol facilitators (Park & Saier
1996).
Possible models of water flows in developing seeds include
non-selective pores (see, for example, van Dongen et al.
2001) through which nutrient and water fluxes could be
readily coordinated to avoid major shifts in tissue water
relations (Zhang et al. 2007). In this context, PsNIP1,
expressed in coats of developing pea seeds, has been shown
to function as an aquaglyceroporin (Schuurmans et al. 2003).
However, PsNIP1 was found not to transport sugars, amino
acids and ions (Schuurmans et al. 2003), and hence, does not
fulfil the role of a non-selective pore to unload phloemimported nutrients (van Dongen et al. 2001).The presence of
a water-conducting aquaporin, PsPIP2;1 in pea seed coats
(Schuurmans et al. 2003), points to a possible separation of
membrane fluxes of water and nutrients. This is a situation
likely to pertain to nutrient and water release from coats of
developing French bean (Phaseolus vulgaris L.) seeds which
appears to be mediated by a cohort of selective nutrient
transporters. Included among the nutrient transporters so
far detected in French bean seed coats are sucrose proton
antiporters (Walker et al. 1995) and facilitators (Zhou et al.
2007). The sucrose carriers function in series with nonselective (Zhang et al. 2002) and pulsing chloride (Zhang
et al. 2004) channels. Under these conditions, nutrient and
water release across plasma membranes of the unloading
cells must be highly coordinated through regulatory
mechanisms integrating the activities of the various membrane transporters. Therefore, we hypothesized that such
© 2007 The Authors
Journal compilation © 2007 Blackwell Publishing Ltd
Aquaporins unload phloem-imported water 1567
coordination could rely on water exit through aquaporins,
and particularly PIPs, that are subject to regulatory signals
(see Alleva et al. 2006; Tornroth-Horsefield et al. 2006).
Moreover, return flows of phloem-imported water to the
parent plant via the seed coat xylem (Pate et al. 1985) are
likely to be spatially separated from sites of nutrient release.
Such an outcome would account for the observed absence of
any significant nutrient back-flows from legume seed coats
(Bennett, Sweger & Spanswick 1984).
In this context, we searched for PIP genes that may be
expressed in French bean seed coats. Currently, three PIPs
have been cloned from French bean plants, PvPIP1;1,
PvPIP1;2 and PvPIP2;1 (Aroca et al. 2006). In this paper,
the functional significance of any cloned PIPs in nutrient
and water transport in seed coats was deduced from
determining their transport properties when expressed in
Xenopus oocytes, their organ and cellular localization of
expression combined with observed water transport properties of native membranes derived from seed coats. In
addition, any functional disparities between PIPs were
evaluated in relation to their structural characteristics at an
amino acid level.
MATERIALS AND METHODS
subjected to degenerate PCR using primers 5′-GGGGG
ATCCAAYCCNGCNGTBACNTTYGG-3′ and 5′-AACT
GCAGTRCTYCTNGCNGGRTTVAT-3′. The PCR reactions were performed at 35 cycles with annealing temperature at 55 °C. The PCR products were cloned into pGEMT
vector (Promega, Madison, WI, USA) and sequenced. To
isolate full-length PIP genes, total RNA was subjected to 5′
and 3′ SMART RACE RT–PCR followed by full-length
amplification (BD Biosciences Clontech, Palo Alto, CA,
USA). For each 5′ or 3′ RACE PCR product, about 10
clones were sequenced, and their sequence information was
used to design primers for full-length amplification. Five
full-length clones (PIP–pGEMT) from each PIP gene were
sequenced in double strands. The resulting sequences were
compared to those obtained from degenerate PCR and 5′/3′
RACE clones, and analysed by Sequencher (version 4.1)
and Australian National Genomic Information Service
(ANGIS). Three distinct PIP clones were isolated from
bean seed coats: PvPIP2;2, PvPIP2;3 and PvPIP1;1. The
PvPIP1;1 was identical to a putative aquaporin-1 previously found in P. vulgaris (GenBank Acc. U97023). The
sequences of the PvPIP2;2 and PvPIP2;3 reported here are
available in GenBank under accession numbers EF624001
and EF624002, respectively.
Plant growth conditions
Semi-quantitative PCR
French bean plants (P. vulgaris L. cv. Redlands Pioneer)
were cultivated under glasshouse conditions with partial
temperature control (20–26 °C by day and 14–16 °C by
night) with a 14 h photoperiod. Upon shorter day lengths
(<14 h), photoperiod was extended by irradiation from
metal arc lamps providing 120 mmol quanta m-2 s-1 photosynthetically active radiation (PAR) at the surface of the
crop canopy. Potting mix contained 2¥ coarse sand : 1¥
steamed recycled potting mix : 1¥ perlite : 1¥ coconut fibre.
Seedlings were culled to one per 15 cm pot 2 weeks after
germination. Water (100 mL) was supplied twice daily to
bring potting mix to field capacity. Every 14 d, 100 mL of
0.5¥ strength Wuxal fertilizer (AgNova Technologies Pty
Ltd, Eltham, Victoria, Australia) was added to each pot. In
addition, one application of 8 g of 70 d release fertilizer
pellets (Nutricote; Chissi Fertiliser Company Ltd, Tokyo,
Japan) was added to each pot immediately following seedling culling.
For experimental work, pods were removed from plants
(10–12 weeks post-germination) onto ice at seed developmental stage IV (Walbot, Clutter & Sussex 1972) – where
expansion growth is completed but dry weight (DW) gains
remained linear. Seed coats were surgically removed by an
excision around the equatorial plane of each seed.
Total RNA, extracted from various plant tissues, was
reverse transcribed as described earlier. Semi-quantitative
PCR reactions were performed within the linear range of
amplification for each targeted fragment examined at 32
cycles with 65 °C as annealing temperature. Gene-specific
primers for PIP1;1 were 5′-GTTCCTGTTTCAGCGTGT
TGTGGCAAG-3′ and 5′-CAAGCACTGAGCCACCAT
GTACATGATAGC-3′, for PIP2;2 were 5′-GAAGGAAA
AAAACATGGAGGGGAAGGAGC-3′ and 5′-AGGGG
ACCTGAGGACACCCATG-3′ and for PIP2;3 were
5′-CTTCCTCCCACATTCATTCTACCACCAATG-3′ and
5′-ACTCGGTGTTACCGGGAATACTGGT-3′. The bean
actin gene was amplified using primers GGGACGACATG
GAGAAGATCTGGC and TCCAGAACAATACCAGT
TGTACGGCCAC. Actin gene expression levels served as
internal positive controls for quantification of relative
amounts of cDNA.
Cloning of PIP cDNAs
Total RNA was extracted from bean seed coats according to
Heim et al. (1993). Extracted RNA was reverse transcribed
with SuperScript reverse transcriptase and oligo(dT)
(Invitrogen, Carlsbad, CA, USA). The resulting cDNA was
In situ hybridization
Seed coat tissue was fixed for 6 h at 4 °C in 4% (v/v)
paraformaldehyde containing 50 mm piperazine-1,4-bis-2ethanesulphonic acid (PIPES), pH 7.2. Fixed tissue was
then washed, dehydrated through an ethanol series and
embedded in methacrylate (Bulbert, Offler & McCurdy
1998). Sections (1 mm thick) were cut with a Reichert
Ultracut microtome (Jung Co.,Wien,Austria), and mounted
onto 3-aminopropyltriethoxysilane (APES)-coated slides.
Digoxigenin (DIG)-labelled sense and antisense riboprobes were synthesized by in vitro transcription from
PCR products with a T7 promoter sequence incorporated
© 2007 The Authors
Journal compilation © 2007 Blackwell Publishing Ltd, Plant, Cell and Environment, 30, 1566–1577
1568 Y. Zhou et al.
upstream or downstream of the PIP fragments. Probes, 200–
270 bp, were designed to target a specific unique sequence
of each PIP. In situ hybridization was performed as
described by Harrington et al. (1997). Sections were viewed
using a Zeiss (Gottinger, Germany) Axiophot microscope
equipped with standard fluorescein isothiocyanate (FITC)/
rhodamine/4′,6-diamidino-2-phenylindole (DAPI) filter
sets, and photomicrographs were taken using an Olympus
(Hamburg, Germany) digital camera C-5050 Zoom.
Functional analysis in Xenopus oocytes
Oocyte isolation and injection
Xenopus laevis lobes containing oocytes were removed
(stage V or VI) and placed in cold calcium-free modified
Barth’s solution (96 mm NaCl, 2 mm KCl, 5 mm MgCl2, 5 mm
HEPES/KOH pH 7.6). Lobes were cut into clumps (~1 cm
long) into fresh calcium-free modified Barth’s solution containing 3 mg mL-1 collagenase, and rotated gently for
45 min before replacement with new solution for another
45 min to remove the oocyte follicular cell layer. Postcollagenase treatment oocytes were washed three times in
hypotonic buffer (0.5 m KH2PO4, pH 6.5), followed by two
washes in modified Barth’s solution containing 0.5 mm
CaCl2. The oocytes were then placed in modified Barth’s
solution containing 0.5 mm CaCl2, 2.5 mL per 50 mL of
horse serum (H-1270; Sigma, St Louis, MO, USA),
50 mg mL-1 tetracyclin (Sigma) (5 mg mL-1 stock, used 0.5/
50 mL) and 0.5 mL per 50 mL of penicillin–streptomycin
(P4333, 10 000 units penicillin and 10 mg streptomycin per
millilitre; Sigma). Uniform oocytes (size and colouring)
were subsequently partitioned under a dissecting microscope and placed in lots of 100 into Petri dishes containing
modified Barth’s solution with horse serum and antibiotics.
Petri dishes kept at 18 °C overnight enabled easier detection of damaged oocytes as they will be deformed or
mottled in appearance on the animal pole.
cRNA was injected into individual oocytes by a microinjector (Drummond ‘Nanoject II’ automatic nanolitre
injector, Drummond Scientific, Broomall, PA, USA) with a
needle approximately 10–20 mm. The 5′-capped cRNA was
generated by in vitro transcription of the linearized template with T7 RNA polymerase (Stratagene, La Jolla, CA,
USA). Forty-five nanolitres of PvPIP2:2, PvPIP2:3 and
PvPIP1;1 cRNA (~1 mg mL-1) volumes was injected with
45 nL of nuclease-free water, AQP1 and NOD26 providing
positive controls for expression. Oocytes were injected
with one PIP cRNA solely, or in every possible combination. The concentration of the individual constructs was
held constant, but oocytes injected with a combination got
twice as much RNA (about 25 ng per construct) in the
same volume (46 nL per oocyte). All oocytes were injected
at the border between animal and vegetal poles on the
animal side. After injection, the oocytes were kept individually in 200 mL of modified Barth’s solution containing
calcium, horse serum and antibiotics for 3 d before testing,
with solutions changed every day.
Osmotic water permeability (Pos) assay
To measure water permeability, the oocytes were transferred into one-fifth modified Barth’s solution generating
a predicted outward-directed oocyte osmotic gradient
of 150 mOsmol kg-1 (3 d post-cRNA injection). Oocyte
volume change was derived from images captured at 5 s
intervals for 5 min under a dissecting microscope by a
ViCAM color camera CCD (Vista Imaging, San Carlos, CA,
USA) and analysed by Global Lab Image 2 (Data Translation, Inc., Marlborough, MA, USA) software. The Pos was
calculated using the following equation:
Pos =
(
)
Vi ⋅ dVrel t
d i
Ai ⋅Vw ⋅ΔCo
(1)
where Vi and Ai are the initial volume and area, respectively, of the oocyte calculated from the projection image of
the oocyte and assuming a sphere; (dVrel/dt)i is the initial
slope of relative volume as a function of time; Vw is the
partial molar volume of water; and DC is the change in
external osmolarity.
Glycerol urea and methylamine
permeability assays
Five oocytes each were incubated in 200 mL Ca Ringers plus
either: (1) 0.5 mm 3H glycerol (55 cpm pmol-1); (2) 0.1 mm
14
C methylamine (95 cpm pmol-1); and (3) 0.1 mm 14C urea
(51 cpm pmol-1). After 30 min, the incubation was stopped
by putting the sample on ice, sucking off as much of the
incubation medium as possible and washing the oocytes four
times with 5 mL ice-cold Ca Ringers. Individual oocytes
were lysed in 200 mL 10% sodium dodecyl sulphate (SDS).
Scintillant (3 mL) was added, and the samples counted for
5 min. Five replicates per construct were performed.
Transport across native membranes and
determination of Pos
Preparation of plasma membrane vesicles was carried out
at (4 °C) or on ice following procedures described by Alleva
et al. (2006). Essentially, excised French bean seed coats
(40–70 g) were homogenized in ice-cold medium [500 mm
sucrose, 50 mm tris(hydroxymethyl)aminomethane (Tris)
(pH 8.0), 20 mm ethylenediaminetetraacetic acid
(EDTA), 20 mm ethyleneglycoltetraacetic acid (EGTA),
0.6% soluble polyvinylpyrrolidine, 10% glycerol, 1 mm Na
vanadate, 50 mm Na fluoride, 5 mm glycerophosphate, 2 mm
dithiothreitol (DTT), 10 mm ascorbic acid, 0.5 mg mL-1 leupeptin], and centrifuged at 10 000 g for 15 min after filtration through four layers of cheesecloth. The supernatant
was collected and respun for 36 min at 100 000 g to obtain
the microsomal pellet. The resuspended microsomal pellet
(300 mm sucrose, 5 mm KCl, 2 mm DTT, 10 mm Na fluoride,
5 mm KPi, pH 7.8) was partitioned in a two-phase system
[6.6% polyethylene glycol 3350 (PEG3350; Sigma), 6.6%
Dextran T400 (Sigma), 330 mm sucrose, 5 mm KCl, 5 mm
© 2007 The Authors
Journal compilation © 2007 Blackwell Publishing Ltd, Plant, Cell and Environment, 30, 1566–1577
Aquaporins unload phloem-imported water 1569
KPi, pH 7.8] with phase separation aided by centrifugation
at 1000 rpm for 5 min. Upper and lower phases were carefully removed, diluted [in 330 mm sucrose, 10 mm Tris (pH
8.3), 5 mm EDTA, 5 mm EGTA, 50 mm Na fluoride, 10 mm
boric acid] and pelleted at 100 000 g for 36 min. The pellets
were resuspended in 1 mL of the same medium, snap frozen
and kept at -70 °C.
The Pos value of membrane vesicles in the upper phase,
enriched in plasma membranes, was measured using a
stopped-flow fluorimeter (DX.17 MV; Applied Photophysics, Leatherhead, UK) as described by Niemietz & Tyerman
(1997). Membrane vesicles resuspended into 30 mm
sucrose, 15 mm Tris/2-(N-morpholino)ethanesulphonic acid
(MES) (pH 8.3), 0.2–0.7 mg protein mL-1 were injected
against a hyperosmotic solution [650 mm sucrose, 15 mm
Tris/MES (pH 8.3)] to create a final inward sucrose gradient
of 310 mm. All experiments were carried out at 20 °C. The
time-course of vesicle shrinkage was followed by an
increase in light scattering (600 V), detected at a wavelength of 500 nm. No time-dependent light-scattering
change was recorded when vesicles were injected against an
isosmolar solution. Rate constants (kos) of shrinkage were
determined by fitting a double exponential to each relaxation curve for at least four replicate runs. Pos was calculated from the relationship:
Pos = ( V A )(kos VwCo)
(2)
where V/A is measured as one-third times the radius of
spherical vesicles (5.2 ¥ 10-8 m), Vw is the partial molar
volume of water and Co is the external osmolarity.
RESULTS
Isolation of PIPs cDNAs
PIP genes were cloned from bean seed coats by degenerate
PCR using primers corresponding to the conserved regions
surrounding two NPA boxes of all known aquaporins
(Murata et al. 2000). A predicted fragment of 380 bp was
amplified, and 40 degenerate PCR clones were sequenced.
Aquaporin partial sequences were identified by basic local
alignment search tool (BLAST) searches. These included
clones exhibiting sequence homology with PIPs and TIPs.
However, none of the clones exhibited homology with the
NIPs clade. Three PIP partial sequences were isolated in
full length by 3′ and 5′ RACE PCR resulting in two new
PIPs, PvPIP2;2 and PvPIP2;3, and the previously published
PvPIP1;1 (GenBank Acc. U97023).
Sequence analysis showed that the predicted proteins of
the new PIP genes bear all the hallmarks of the aquaporin
protein family, including six transmembrane helices and
two conserved NPA motifs (Fig. 1). Phylogenetic analysis
showed the predicted proteins encoded by these genes are
members of the PIP family. Thus, PvPIP2;2 and PvPIP2;3
cluster to a PIP2 subgroup, and PvPIP1;1 to a PIP1 subgroup (Fig. 2). The deduced PvPIP2;2 amino acid sequence
is 80% identical to that of PvPIP2;3. In addition, PvPIP2;2
and PvPIP2;3 sequences showed 87 and 80% identity,
respectively, to the recently discovered PvPIP2;1 (Acc.
AY995195).
Structure function relationships were deduced by aligning PvPIP2;2, PvPIP2;3 and PvPIP1;1 sequences with that of
SoPIP2;1 (see Fig. 1) whose molecular structure and gating
mechanism were recently defined (Tornroth-Horsefield
et al. 2006). Sequence comparisons detected a number of
gating-associated sites. These were two conserved serine
residues corresponding to two phosphorylation sites, Ser
115 and Ser 274 in SoPIP2;1 (Johansson et al. 1998;
Tornroth-Horsefield et al. 2006). One serine is located in
cytosolic loop B, conserved in PIP1 and PIP2 subfamilies.
The other serine is located in the C-terminal region and is
only conserved in the PIP2 subfamily (Fig. 1). Another
aquaporin gating-associated site, corresponding to His 193
of SoPIP2;1 (Tournaire-Roux et al. 2003) was detected in
loop D of all three PvPIPs (Fig. 1).
PvPIP1;1 and PvPIP2;3 are identical to most PIPs at five
key positions (P1–P5) that are significant for aquaporins or
glycerol permeases (Froger et al. 1998 and see Fig. 1). In
contrast, PvPIP2;2 at P4 has a tyrosine substitution similar
to TIPs (Froger et al. 1998). Both PvPIP2;2 and PvPIP2;3
possess all aromatic/arginine determinants (ar/R region)
identical to PIP family (Wallace & Roberts 2004, and see
Fig. 1). In contrast, while three substitutions in the ar/R
region, helices (H) 2 and H5 and loop E (LE2, see Wallace
& Roberts 2004) of PvPIP1;1 are identical to those of other
PIPs; the LE1 substitution is alanine (Fig. 1).
In addition, while PvPIP1;1 and the new PvPIP2;2
showed high homology to members of the PIP1 and PIP2
subfamilies, respectively (Fig. 2), sequence alignment of all
available aquaporins revealed they have substitutions in
several residues conserved for most plant PIP1 and PIP2
members. The most notable of these substitutions are as
follows. (1) A cysteine in H3, corresponding to a mercurysensitive site in Arabidopsis TIP (Daniels et al. 1996) is
replaced by serine in PvPIP2;2 and by alanine in PvPIP1;1.
The serine substitution also creates a potential glycosaminoglycan attachment site in PvPIP2;2 (double lined in
Fig. 1) detected by ProDom search (Corpet, Gouzy & Kahn
1998). The cysteine substitution in H3 is conserved in most
PIPs except OsPIP2;6 and OsPIP2;7 from rice, but not conserved in NIPs and TIPs. However, all the substitutions are
non-polar residues except AtNIP6;1, which also has a serine
substitution. (2) A non-polar leucine residue in the loop B
(LB) was replaced by a polar serine in PvPIP2;2 (Fig. 1).
This position is conserved in most PIPs except OsPIP2;8,
but is not conserved in most NIPs and TIPs. In contrast to
PvPIP2;2, all these latter substitutions are non-polar residues. (3) A leucine residue in helix 3 (H3) was replaced by
alanine in PvPIP2;2.
Organ and cellular localization of
gene expression
Coats of storage phase seeds, with cotyledon relative water
contents of about 70% (Thomas, Hetherington & Patrick
© 2007 The Authors
Journal compilation © 2007 Blackwell Publishing Ltd, Plant, Cell and Environment, 30, 1566–1577
1570 Y. Zhou et al.
Figure 1. Alignment of SoPIP2;1 and
the deduced amino acid sequences of
PIP1;1, PIP2;2 and PIP2;3 from bean
seed coats. Sequence alignments were
performed by ClustalW algorithm. Black
shading indicates identical residues, and
grey shading conserved residues.
Annotated are: transmembrane helices
H1 to H6; pore helices loop B (LB), loop
E (LE) and loop D (LD); two NPA
motifs (***); two conserved serine
residues (*); and a histidine residue (*)
which correspond to Ser115, Ser274 and
His193 in SpPIP2;1. The ar/R signature
(∧) and residue substitutions (>) are
described in the text. P1–P5 denote the
five key positions used to discriminate
between aquaporins and glycerol
facilitators (Froger et al. 1998). The
potential glycosaminoglycan attachment
site is double lined, and the four residues
associated with ZmPIP1;2 (see
Discussion) are denoted as ‘O’.
2000) were examined for PIP expression, and compared
with their relative levels of expression in other plant organs.
Transcripts of the three PIP genes were detected in all plant
organs including source leaves, sink leaves, flowers, stem,
roots, cotyledons and seed coats (Fig. 3). All three PIPs
exhibited their highest levels of expression in seed coats,
with PvPIP2;3 and PvPIP1;1 expression levels being twice
those found in source leaves. The remaining organs, including developing cotyledons, exhibited relatively low levels of
expression for all three PIPs (Fig. 3).
Cellular localization of each PIP transcript was examined
in bean seed coats by hybridizing tissue sections (1 mm
thick) with DIG-labelled RNA probes specific to each PIP
gene. No significant signals for the three PIPs were detected
in the chlorenchyma and ground parenchyma cells radially
outward from the line of vascular bundles. Rather, expression of all three transcripts was localized to ground
parenchyma cells proximal to, and radially inward from, the
ring of vascular bundles. A signal was also detected in the
vascular bundles as well as the branch parenchyma cells
(Fig. 4b,d,f). Both PvPIP2;3 and PvPIP1;1 exhibited strongest expression in the vascular bundles, while the ground
and branch parenchyma cells appeared to have the same
level of expression (Fig. 4).
Functional analysis in Xenopus oocytes
The PIPs were functionally characterized by expression in
Xenopus oocytes transformed with a vector containing
cRNA of each PIP. Pos values of the transformed oocytes
were determined by measuring initial rates of cell swelling
following their transfer to a hypotonic solution that generated a 150 mOsmol kg-1 osmotic gradient. Only oocytes
transformed with PIP2;3 supported water fluxes above
© 2007 The Authors
Journal compilation © 2007 Blackwell Publishing Ltd, Plant, Cell and Environment, 30, 1566–1577
Aquaporins unload phloem-imported water 1571
Figure 4. Cellular localization of PIP expressed in seed coats
Figure 2. Phylogenetic analysis of PvPIP1;1, PvPIP2;2 and
PvPIP2;3 in relation to other plant aquaporins. The tree was
constructed by the parsimony method in the PHYLIP package
and unrooted by its Drawtree program. All sequences were
obtained from NCBI (http://www.ncbi.nlm.nih.gov). Arabidopsis
thaliana: AtPIP1;1 (At3g61430), AtPIP1;4 (At4g00430), AtPIP2;2
(At2g37170), AtPIP2;5 (At3g54820), AtPIP2;6 (At3g54820),
AtPIP2;7 (At4g35100), AtPIP2;8 (At2g16850), AtNIP1;2
(At4g19030), AtNIP2;1 (At2g34390) and AtTIP2;3 (At5g47450);
Lycopersicon esculentum: LePIP1 (Q08451), MtPIP1;1
AF386739); Nicotiana tabacum: NtPIP1;1 (AAL33585), NtPIP2;1
(AF440272); Spinacia oleracea: SoPIP2;1 (L77969); Pisum
sativum: PsPIP2;1 (CAB45651), PsTIP1;1 (AJ243309); Phaseolus
vulgaris: PvPIP1;1 (U97023), PvPIP1;2 (AY995196) and PvPIP2;1
(AAY22203); and Vitis vinifera: VvTIP (AJ289866). The current
nomenclature was used according to Johanson et al. (2001).
45
PvPIP1
PvPIP2;2
PvPIP2;3
Relative expression (%)
40
35
30
25
20
15
10
5
s
tyle
don
Co
See
dc
oat
s
ot
Ro
Ste
m
er
Flo
w
ave
s
k le
Sin
S ou
rce
le
ave
s
0
Figure 3. Expression of PvPIP1;1, PvPIP2;2 and PvPIP2;3 in
specified bean organs. Relative expression measures were by
semi-quantitative RT–PCR normalized to expression of the actin
gene. Values are the means ⫾ SE from three separate RNA
extractions.
of bean. Light micrographs illustrating in situ hybridization of
digoxigenin (DIG)-labelled PIP sense (a,c,e) and antisense
(b,d,f) riboprobes to coat transverse sections of developing bean
seeds. Seed coat sections treated with riboprobes to PvPIP1;1
(a,b), PvPIP2;2 (c,d) and PvPIP2;3 (e,f). Note that antisense
riboprobe signal is localized from the ground parenchyma inward
of the vascular bundles to and including branch parenchyma
tissues. bp, branch parenchyma; c, chlorenchyma; gp, ground
parenchyma; h, hypodermis; p, palisade; vp, vascular parenchyma;
x, xylem. Bar = 75 mm.
water-injected controls (ca 2 mm s-1, and see Fig. 5a). Calculated Pos values of 200 mm s-1 for PIP2;3 are consistent
with water flux through aquaporins (Fetter et al. 2004).
Some apparently non-functional plant aquaporins, when
expressed with another of the PIP2 class, can interact to
increase water permeability (Fetter et al. 2004). In this
context, we tested affects of co-expressing various combinations of the three aquaporins on Pos values of transformed Xenopus oocytes. No additional water permeability
was detected when PvPIP1;1 or PvPIP2;2 was co-expressed
with PvPIP2;3 (Fig. 5a). Furthermore, the combined expression of PvPIP1;1 with PvPIP2;2 did not yield functional
water channel activity above water-injected control oocytes.
When treated with 1 mm HgCl2, the water permeability of
PvPIP2;3 expressing oocytes was inhibited to 18% of the
control, taking into account the effect of HgCl2 on waterinjected control oocytes (Fig. 5b).
Permeabilities of the PIPs to radiolabelled glycerol, urea
or methylamine were determined from measuring their
rates of uptake by Xenopus oocytes expressing the PIPs
alone or in combination (see previous discussion). Uptake
rates for all solutes tested were found not to differ from the
water-injected controls (data not shown).
Water transport properties of native plasma
membranes isolated from bean seed coats
Plasma membrane vesicles were prepared from intact
seed coats. Their water permeability coefficients were
© 2007 The Authors
Journal compilation © 2007 Blackwell Publishing Ltd, Plant, Cell and Environment, 30, 1566–1577
1572 Y. Zhou et al.
(a)
0.08
250
0.06
200
Volts (V)
Pos (m m s –1)
0.04
150
100
V = –0.0348exp
–0.02
–12.7s
– 0.0400exp
+ 0.05123
P1
2
vP
I
2/
P
PI
P2
;
–0.06
0.00
0.20
0.40
Pv
0.60
0.80
1.00
Time (s)
Pv
PI
P2
;
PI
P2
;
3/
P
Pv
vP
I
vP
I
P2
;2
P1
3
3/
P
PI
P2
;
Pv
PI
P2
;
Pv
Pv
PI
P1
2O
–0.04
H
Pos (m m s –1)
0.00
–56.48s
50
0
(b)
0.02
Figure 6. Time-course of 500 nm light scattering of plasma
150
membrane vesicles (PMVs) isolated from seed coats. Vesicles
in solution containing 30 mm sucrose, 10 mm boric acid,
5 mm ethylenediaminetetraacetic acid (EDTA), 5 mm
ethyleneglycoltetraacetic acid (EGTA), 50 mm Na fluoride and
10 mm tris(hydroxymethyl)aminomethane (Tris) (pH 8.0) were
injected against identical medium containing an additional
650 mm sucrose, creating a final inward 310 mOsmol osmotic
gradient. Protein final concentration in reaction: 100 mg mL-1
(n = 6).
100
50
1+
H
g
1
Pv
PI
P
g
H
+
2;
2
Pv
PI
P
2;
2
Pv
PI
P
Pv
PI
P
2;
3+
H
g
2;
3
Pv
PI
P
g
H
2O
+
H
H
Pv
PI
P
2O
0
Figure 5. Osmotic water permeabilities (Pos) of Xenopus laevis
oocytes expressing PvPIPs. (a) Each PvPIP was expressed alone
or in combination with the other two PvPIPs. There was no
significant difference between the Pos of control (H2O injected)
and PvPIP1;1, PvPIP2;2 and PvPIP2;2/PvPIP1;1 (P > 0.05). There
was also no significant difference between PvPIP2;3 and
PvPIP2;3/PvPIP1;1 and PvPIP2;3/PvPIP2;2 (P > 0.05). (b) The
effect of HgCl2 (1 mm) on the Pos of oocytes expressing each of
the PvPIPs alone or in combination.
The Pos values, for the fast and slow shrinkage phases of
plasma membrane vesicles, were significantly reduced by 72
and 58%, respectively, when the water channel blocker,
HgCl2, was added to the bathing medium (Fig. 7). A similar
response (data not shown) was elicited when vesicles were
treated with AgNO3, an alternative channel blocker to
HgCl2 (Niemietz & Tyerman 2002). Together with the relatively high water permeability of the control vesicles, these
strong levels of inhibition indicate that the native plasma
membranes contain functional aquaporins.
250
Control
Mercury
–1
150
Pos (mm s
determined by measuring rates of vesicle volume change
using stopped-flow fluorimetry (Niemietz & Tyerman 1997),
immediately after exposure to a hypo-osmotic solution with
an initial inward osmotic gradient of 310 mOsmol kg-1. A
typical time-course of bean seed coat plasma membrane
vesicle shrinkage is shown in Fig. 6. The shrinkage trace was
quantified by fitting a double exponential curve, from which
two rate constants for vesicle shrinking were obtained.
These values were used to derive estimates of Pos for the
two kinetic phases of vesicle shrinkage that differed by
an order of magnitude (Fig. 7). Both estimated Pos values
were commensurate with membranes containing waterpermeable aquaporins. The amplitude of the fast phase was
double that of the slow phase (data not shown). Shrinkage
did not occur in the absence of an osmotic gradient (data
not shown) demonstrating that volume change was osmotically driven.
)
200
100
50
0
Fast phase
Slow phase
Figure 7. Effect of mercuric chloride on fast and slow phase
decay constants of seed coat plasma membrane vesicles (PMVs)
exposed to a 310 mOsmol osmotic gradient. PMVs were exposed
to the osmotic gradient after a 5 min pre-incubation in the
absence or presence of 100 mm mercuric chloride. 100 mg
protein mL-1 in the reaction medium (n = 6).
© 2007 The Authors
Journal compilation © 2007 Blackwell Publishing Ltd, Plant, Cell and Environment, 30, 1566–1577
Aquaporins unload phloem-imported water 1573
DISCUSSION
In this paper, we report the isolation of three PIP genes,
PvPIP1;1, PvPIP2;2 and PvPIP2;3, from coats of developing French bean seeds. These findings add two additional
PIPs, PvPIP2;2 and PvPIP2;3, to the known cohort
of P. vulgaris PIPs comprising PvPIP1;1, PvPIP1;2 and
PvPIP2;1 (Aroca et al. 2006). The large size of PIP1 and
PIP2 gene families found in other species such as
Arabidopsis thaliana (5 and 8, respectively; Johanson et al.
2001) and rice (3 and 8, respectively; Sakurai et al. 2005)
suggests that other PvPIP genes remain to be cloned.
However, the profile of three PvPIP genes found to be
expressed in bean seed coats (current study) corresponds
with numbers of PIP isoforms expressed in other major
storage sinks. For example, grape berries express two
VvPIP1s (Picaud et al. 2003), and pea seed coats a PIP1,
PIP2 and a NIP1 (Schuurmans et al. 2003). A structure
function analysis of transport properties of the bean PIPs is
presented. Thereafter, their potential role in water and
nutrient transport in seed coats is discussed.
Structure/Function analysis of PvPIPs
Functional characterization of the bean aquaporins, by
heterologous expression in Xenopus oocytes, demonstrated
that the water permeability of oocytes expressing PvPIP2;3
was between 150 and 200 mm s-1 (Fig. 5). This value compares favourably with water permeabilities reported
for aquaporins expressed in pea seed coats; PsPIP2;1
(52 mm s-1) and PsNIP1;1 (13.6 mm s-1, and see Schuurmans
et al. 2003). Consistent with other PIP2s (Chaumont,
Moshelion & Daniels 2005), the transport activity of
PvPIP2;3 was mercury sensitive (Fig. 5b). In contrast to
the functional observations, predicted protein sequences
encoded by PvPIP1;1, PvPIP2;2 and PvPIP2;3 have typical
aquaporin signatures of other plant PIP members (Fig. 1,
and see text description in Results). However, for PvPIP1;1
and PvPIP2;2, some exceptions occur and these may contribute to their observed transport behaviours when
expressed in Xenopus oocytes (water, Fig. 5; nutrients, data
not shown). These issues are examined as follows.
The lack of water-conducting activity for oocytes expressing PvPIP1;1 (Fig. 5a) was not surprising given similar findings for PIP1 isoforms cloned from a range of plant
species (Chaumont et al. 2005). However, co-expression of
PvPIP1;1 with PvPIP2;3 did not result in enhanced
water conductance (Fig. 5a). This indicated a departure of
PvPIP1;1 from cooperative effects with PIP2s reported for
ZmPIP1;2 (Fetter et al. 2004).
Four residues in loop E of ZmPIP1;2, Ile247, Arg250,
Asp251 and Asn256 were found to be important in the
physical interaction between ZmPIP1;2 and ZmPIP2;5
(Fetter et al. 2004). Sequence comparison indicated
PvPIP1;1 has identical residues corresponding to Ile247 and
Asp251 of ZmPIP1;2. However, the remaining two loop E
residues differ in PvPIP1;1 such that Arg250 (basic) is
replaced by lysine (basic), and the polar neutral residue,
Asn256 is replaced by aspartic acid, a negatively charged
amino acid (Fig. 1). The alteration in the basic amino acid
residues is likely to be without effect, but the shift from a
polar neutral to a negatively charged amino acid could
impact on function. In this context, Asn256 in ZmPIP1;2
was suggested to play a role in pulling the loop E of
ZmPIP1;2 towards the pore vestibule to facilitate protein
oligomerization, and hence, high water conductance
(Chaumont et al. 2005). Whether the aspartic acid substitution at this site in PvPIP1;1 prevents oligomerization with
PvPIP2;3 awaits further investigation.
Despite their differences in water transport abilities
when expressed in Xenopus oocytes (Chaumont et al. 2005,
and see Fig. 5), PIP1 and PIP2 members have been shown
to share two pore selectivity filters required for high water
conductance. These are the dual NPA motif and the
aromatic/arginine region (ar/R; Wallace & Roberts 2004).
In this context, sequence alignment predicted that PvPIP2;2
and PvPIP2;3 possess all ar/R determinants required for
high water conductance (cf. Chaumont et al. 2005, and see
Fig. 1). In contrast, for PvPIP1;1 the alanine substitution in
LE1 (Fig. 1) differs from the vast majority of PIP members,
but is identical to all NIP and some TIP members (Wallace
& Roberts 2004). This suggests that a molecular basis exists
for a disparity in water selectivity between PvPIP1;1 and
PvPIP2;2/PvPIP2;3. However, the effect of this substitution
alone on water transport might not be significant because a
hydrophilic surface could be formed through LE1 carbonyl
of the peptide backbone and the LE2 arginine side chain to
facilitate water transport (Sui et al. 2001;Wallace & Roberts
2004).
Apart from those substitutions referred to earlier, the
remaining unique substitutions in PIP1;1 and PvPIP2;2 are
outside the reported pore lining regions (Fig. 1). These
substitutions may correlate to regulatory sites for these
aquaporins. Firstly, the potential glycosaminoglycan attachment site in PvPIP2;2 (see Results) suggests that posttranslational glycosylation is involved in the proper routing
and membrane insertion of PvPIP2;2. Glycosylated forms
of animal APQ1 and AQP2 have been reported (Baum
et al. 1998; Hendriks et al. 2003). In plants, some evidence
suggests a role for glycosylation in the redistribution of an
ice plant McTIP1;2 between endomembrane sites in
response to osmotic stress (Vera-Estrella et al. 2004). Secondly, the lack of cysteine in H3 of both PvPIP1;1 and
PvPIP2;2 (Fig. 1) suggests that their mercury sensitivity
might be different from that of PvPIP2;3. This mercurysensitive site in Arabidopsis d-TIP is suggested to form a
hydrophilic part of the helix facing the pore (Daniels et al.
1996; Johansson et al. 2000), and therefore, is a potential
gating site.
None of the PvPIPs conducted glycerol, methylamine or
urea when expressed in Xenopus oocytes. Consistent with
this finding is the absence of the appropriate amino acid
substitutions in the PvPIPs at Froger’s five positions (P1–
P5) peculiar to glycerol transporters (Froger et al. 1998, and
see Fig. 1). The substitutions for glycerol transporters are
P1–Tyr or an aromatic amino acid; P2–Asp; P3–Lys or Arg;
© 2007 The Authors
Journal compilation © 2007 Blackwell Publishing Ltd, Plant, Cell and Environment, 30, 1566–1577
1574 Y. Zhou et al.
P4–Pro; P5–Leu or a hydrophobic, non-aromatic amino acid
(Froger et al. 1998).
water conductance (Fig. 5a) while none of the PvPIPs supported transport of glycerol, methylamine or urea (data not
shown).
That PvPIP2;3, and possibly PvPIP1;1 and PvPIP2;2,
function to facilitate water transport in bean seed coats is
supported by estimated Pos values of coat native plasma
membranes exceeding 10 mm s-1 and being sensitive to
mercury (Fig. 7, and cf. Tyerman, Niemietz & Bramley
2002). The two distinct kinetic phases of plasma membrane
vesicle (PMV) shrinkage (Fig. 6) may reflect water loss
from two populations of vesicles with markedly different
densities of aquaporins as illustrated by their derived Pos
values (Fig. 7). This proposition is supported by two observations. Firstly, the apparent expression of PIP2;3 is higher
in vascular compared to ground parenchyma cells of seed
coats (Fig. 4). Secondly, Pos estimates for the slow shrinkage phase (20 mm s-1, and see Fig. 7) are similar to Pos estimates obtained from measuring hydraulic conductivities of
intact seed coat ground parenchyma cells using a cell pressure probe (Lp = 8.2 ¥ 10-8 m s-1 MPa-1, Pos = 11.2 mm s-1;
Zhang et al. 1996). Therefore, ground parenchyma cells of
seed coats may not have high Pos (Lp), and the fast phase of
Physiological role of PvPIPs expressed in
coats of developing bean seeds
Nutrients and accompanying water are imported into bean
seed coats through a reticulate vasculature positioned in
their mid-plane (see Fig. 8a,b). Phloem-imported nutrients
move radially inward to the cotyledons (Fig. 8a) symplasmically from sieve elements to ground parenchyma where
they exit to the seed apoplasm through carriers and channels (Fig. 8b,d, and see Zhang et al. 2007). Symplasmic
movement from sieve elements to vascular parenchyma is
likely to occur as a bulk flow (Murphy 1989; Fisher & CashClark 2000). The high expression of three PvPIPs detected
in coats of developing French bean seeds (Fig. 3) and localized to the nutrient unloading pathway (Fig. 4) indicates
a potential role for aquaporins in water and nutrient
exchange to the seed apoplasm. In this context, functional
characterization of the PvPIP transport properties in
Xenopus oocytes indicated that PvPIP2;3 conferred a high
(b)
(a)
p
(c)
h
sc
se
c
scvb
vp
gp
ft
cot
f
se
vb
vp
Figure 8. Morphology and anatomy of
gp
x
ea
bp
pw
pw
(d)
x
se
vp
gp
To cots
vp
developing Phaseolus vulgaris seeds and
a diagrammatic model illustrating the
cellular pathway for water and nutrient
transport in seed coats. (a) Diagrammatic
representation of the morphology of a
developing seed showing the
interrelationships between the various
organs and associated vascular systems
bounded by the pod wall; not drawn to
scale. (b) Light micrograph illustrating a
cross-section of a coat of a developing
seed. (c) Enlargement of part of the
vascular bundle shown in (b). (d) Model
of cellular pathways for water and
nutrient flows in seed coats. Both
bars = 34 mm. bp, branch parenchyma;
c, chlorenchyma; cot, cotyledon; ea,
embryonic axis; f, funiculus; ft, funicular
trace; gp, ground parenchyma cells;
h, hypodermis; p, palisade; pw, pod wall;
sc, seed coat; scvb, seed coat vascular
bundle; se, sieve element; vb, vascular
bundle; vp, vascular parenchyma cell;
) and
x, xylem elements. Nutrient (
) flows. Aquaporin ( ) and
water (
nutrient transporter ( ).
© 2007 The Authors
Journal compilation © 2007 Blackwell Publishing Ltd, Plant, Cell and Environment, 30, 1566–1577
Aquaporins unload phloem-imported water 1575
Table 1. Estimates of water potential differences (DY) required to account for membrane fluxes of phloem-imported water exiting
across plasma membranes of vascular and ground parenchyma cells of coats of developing French bean seeds
Cell type
Pos (mm s-1)
Derived Lpa
(¥10-8 m s-1 MPa-1)
Potential water fluxb
(¥10-10 m3 m2 s-1)
Estimated DY (¥10-3 MPa)
Vascular parenchyma
Ground parenchyma
181
21
134
15.5
130
5.4
-9.7
-3.4
a
Derived from Pos estimates (this study) from the relationship Lp = (PosVw)/RT
where Vw = 1.8 ¥ 10-5 m3 mol-1 and RT at 20 °C = 2.437 ¥ 10-3 m3 MPa mol-1.
b
Membrane fluxes of phloem-imported water derived from assuming an import of a 10% sucrose solution and that water exit occurs
exclusively across the cell membrane surface area of each cell type (Offler & Patrick 1984).
vesicle shrinkage may be attributable to a higher water
permeability of their vascular parenchyma cells. The slight
disparity between estimates of Pos for the slow shrinkage
phase (21 mm s-1) and pressure probe measurements
(11.2 mm s-1) is that the latter could be underestimated. For
instance, Wan, Steudle & Hartung (2004) showed that large
pressure pulses, when used to evoke a pressure relaxation in
root cells, can inhibit aquaporin activity as judged by a
strong inhibition of Lp. Thus, it is possible that the pressure
probe measurements of Zhang et al. (1996) used pressure
pulses that may have deactivated a proportion of the aquaporin population located in ground parenchyma cells.
The presence of aquaporins in vascular and ground
parenchyma cells raises questions about putative roles of
these two cell types in exchanging phloem-imported water
to the seed apoplasm at rates that match those of phloem
import to prevent pressure build up (Murphy 1989). In
developing wheat seeds, substantial pressure gradients of
1 MPa across sieve element/vascular parenchyma interfaces
are considered sufficient to drive a bulk flow of phloem sap
through their interconnecting plasmodesmata (Fisher &
Cash-Clark 2000). Whether there is an ongoing bulk flow
into ground tissues is less certain (Fisher & Cash-Clark
2000). A similar situation probably pertains to water movement in coats of bean seeds (Murphy 1989) from the
importing sieve elements to vascular and perhaps surrounding ground parenchyma cells (Fig. 8c,d). Thereafter, water
must be released to the seed apoplasm (Fig. 8d), irrespective of whether it moves inward to support cotyledon
expansion growth (see Fig. 8a) or is recycled back to the
parent plant via the xylem (Fig. 8a,c,d; cf. Pate et al. 1985).
The latter predominates as cotyledon expansion slows and
finally ceases while rates of phloem import continue
unabated (Thomas et al. 2000). Hydraulic independence of
developing grain legume seeds from their parent plant
infers their tissues are in quasi-water equilibrium (Zhang
et al. 2007). However, small water potential gradients must
exist across seed coat plasma membranes to drive exit of
phloem-imported water. This water potential gradient is
likely to result from a combination of hydrostatic pressure
transmitted from the phloem and transpiration tensions
transmitted from the parent plant. The latter are markedly
attenuated by low path hydraulic conductivities resulting
from xylem discontinuities located in pedicels supporting
each seed (Patrick & Offler 2001). In this context, it was
considered instructive to estimate water potential differences required to account for phloem water fluxes across
vascular and ground parenchyma plasma membranes.
These values were estimated using Pos values derived from
vesicle shrinkage rates and pressure probe measures of
ground parenchyma hydraulic conductivity (see Table 1).
The predicted water potential differences of -0.3 to
-1.0 ¥ 10-2 MPa are extremely small and represent 1.5% or
less of the seed water potential (cf. Patrick 1994). Thus,
these would have little influence on overall seed water
potential, and hence, phloem import rates. In addition, the
predicted water potential differences indicate that vascular
and ground parenchyma cells are equally capable of supporting the observed membrane fluxes of water. The lesser
plasma membrane surface area generated by vascular
parenchyma cells is offset by their higher water permeability (Table 1). As a consequence, phloem-imported water
could flow symplasmically to the vascular parenchyma cells
where most is expelled to the seed apoplasm (see Fig. 8c,d).
For water recycled back to the parent plant via the xylem,
such an outcome would separate the major release sites for
water (vascular parenchyma cells – this paper) and nutrients (ground parenchyma – Patrick & Offler 2001). In
addition to xylem discontinuities combined with phloem
retrieval (Patrick & Offler 2001), this spatial separation of
water and nutrient flows in seed coats could contribute to
the observed minimal return of seed-imported nutrients to
the parent plant (Bennett et al. 1984).
Aquaporins expressed in bean seed coat did not appear
to be permeable to phloem-imported nutrients (data not
shown, and see also Schuurmans et al. 2003). Thus, coordinating membrane fluxes of phloem-imported water and
nutrients to the seed apoplasm must depend upon a more
sophisticated mechanism than their combined release
through the same channels (van Dongen et al. 2001). One
possible scenario is that aquaporins could play an important
role in the turgor homeostat mechanism envisioned to integrate nutrient demand by the cotyledons with nutrient
release from ground parenchyma cells (Fig. 8b,d), and
phloem import into seed coats (Patrick 1994; Zhang et al.
1996). Aquaporins in ground parenchyma cells could facilitate rapid turgor adjustments in response to altered osmotic
concentrations in the seed apoplasm. As a result, an
increase in cotyledon nutrient demand would be rapidly
compensated by a turgor-induced increase in nutrient
© 2007 The Authors
Journal compilation © 2007 Blackwell Publishing Ltd, Plant, Cell and Environment, 30, 1566–1577
1576 Y. Zhou et al.
release by seed coats. In addition, the apparent nonfunctional PvPIP1;1 and PvPIP2;2 may participate in
regulating the turgor homeostat mechanism by sensing
and transducing the turgor signal (see Hill, Shachar-Hill &
Shachar-Hill 2004). Several putative phosphorylation/
dephosphorylation sites and a histidine residue related to
SoPIP2;1 gating (Johansson et al. 1998; Tornroth-Horsefield
et al. 2006) were detected in both PvPIP2;2 and PvPIP1;1
(Fig. 1). Previous studies with intact seed coats suggested
that a cytoplasmic Ca2+ signalling cascade(s) is involved in
regulating turgor-dependent nutrient efflux (Walker et al.
2000). This might be linked with aquaporin activity as Ca2+
recently has been shown to be an important regulatory
factor involved in aquaporin gating (Alleva et al. 2006).
CONCLUSIONS
Three aquaporin genes belonging to the PIP family,
PvPIP1;1, PvPIP2;2 and PvPIP2;3, were strongly expressed
in seed coats. Their expression was localized to cells located
along the post-phloem nutrient-unloading pathway with the
greatest expression in vascular parenchyma cells. Of these
three genes, only PvPIP2;3 encoded a protein capable of
supporting water transport when expressed in Xenopus
oocytes. Whether the activity of PvPIP1;1 and PvPIP2;2
depends upon additional cofactors requires further study. In
contrast, because none of the aquaporins were capable of
supporting solute transport in Xenopus oocytes, the possibility of a common mechanism for water and nutrient exit
from the seed coat symplasm is not supported. However,
the Pos values measured in native membranes suggest that
PvPIP2;3 might play a key role in discharging phloemimported water to the seed apoplasm and particularly so
from the vascular parenchyma cells. A proposed function
for the non-water conducting aquaporins, PvPIP1;1 and
PvPIP2;2, in turgor sensing and signal transduction deserves
experimental investigation.
ACKNOWLEDGMENTS
We are indebted to Mr Kevin Stokes for supplying healthy
plant material for experimentation, Dr Sunita Remesh for
comments on the manuscript and Wendy Sullivan for expert
technical assistance. N.S. is appreciative of a Newcastle Postgraduate Research Scholarship. Financial support from the
Australian Research Council is gratefully acknowledged.
REFERENCES
Alleva K., Niemietz C.M., Sutka M., Maurel C., Parisi M., Tyerman
S.D. & Amodeo G. (2006) Plasma membrane of Beta vulgaris
storage root shows high water channel activity regulated by cytoplasmic pH and a dual range of calcium concentrations. Journal
of Experimental Botany 57, 609–621.
Aroca R., Ferrante A., Vernieri P. & Chrispeels M.J. (2006)
Drought, abscisic acid and transpiration rate effects on the regulation of PIP aquaporin gene expression and abundance in
Phaseolus vulgaris plants. Annals of Botany 98, 1301–1310.
Baum M.A., Ruddy M.K., Hosselet C.A. & Harris H.W. (1998) The
perinatal expression of aquaporin-2 and aquaporin-3 in developing kidney. Pediatric Research 43, 783–790.
Bennett A.B., Sweger B.L. & Spanswick R.M. (1984) Sink to source
translocation in soybean. Plant Physiology 74, 434–436.
Bulbert M.W., Offler C.E. & McCurdy D.W. (1998) Polarized
microtubule deposition coincides with wall ingrowth formation
in transfer cells of Vicia faba L. cotyledons. Protoplasma 201,
8–16.
Chaumont F., Moshelion M. & Daniels M.J. (2005) Regulation of
plant aquaporin activity. Biology of the Cell 97, 749–764.
Corpet F., Gouzy J. & Kahn D. (1998) The ProDom database of
protein domain families. Nucleic Acids Research 26, 323–326.
Daniels M.J., Chaumont F., Mirkov T.E. & Chrispeels M.J. (1996)
Characterization of a new vacuolar membrane aquaporin
sensitive to mercury at a unique site. The Plant Cell 8, 587–
599.
van Dongen J.T., Laan R.G.W., Wouterlood M. & Borstlap A.
(2001) Electrodiffusional uptake of organic cations by pea seed
coats. Further evidence for poorly selective pores in the plasma
membrane of seed coat parenchyma cells. Plant Physiology 126,
1688–1697.
Fetter K., Van Wilder V., Moshelion M. & Chaumont F. (2004)
Interactions between plasma membrane aquaporins modulate
their water channel activity. The Plant Cell 16, 215–228.
Fisher D.B. & Cash-Clark C.E. (2000) Gradients in water potential
and turgor pressure along the translocation pathway during
grain filling in normally watered and water-stressed wheat
plants. Plant Physiology 123, 139–148.
Froger A., Tallur B., Thomas D. & Delamarche C. (1998) Prediction
of functional residues in water channels and related proteins.
Protein Science 7, 1458–1468.
Harrington G.N., Franceschi V.R., Offler C.E., Patrick J.W., Tegeder
M., Frommer W.B., Harper J.F. & Hitz W.D. (1997) Cell specific
expression of three genes involved in plasma membrane sucrose
transport in developing Vicia faba seed. Protoplasma 197, 160–
173.
Heim U., Weber H., Bäumlein H. & Wobus U. (1993) A sucrosesynthase gene of Vicia faba L. expression pattern in developing
seeds in relation to starch synthesis and metabolic regulation.
Planta 191, 394–401.
Hendriks G., Koudijs M., van Balkom B.W., Oorschot V.,
Klumperman J., Deen P.M. & van der Sluijs P. (2003) Glycosylation is important for cell surface expression of the water
channel aquaporin-2 but is not essential for tetramerization in
the endoplasmic reticulum. The Journal of Biological Chemistry
279, 2975–2983.
Hill A.E., Shachar-Hill B. & Shachar-Hill Y. (2004) What are
aquaporins for? Journal of Membrane Biology 197, 1–32.
Johanson U., Karlsson M., Johansson I., Gustavsson S., Sjövall S.,
Fraysse L., Weig A.R. & Kjellbom P. (2001) The complete set of
genes encoding major intrinsic proteins in Arabidopsis provides
a framework for a new nomenclature for major intrinsic proteins
in plants. Plant Physiology 126, 1358–1369.
Johansson I., Karlsson M., Shukla V.K., Chrispeels M.J., Larsson C.
& Kjellbom P. (1998) Water transport activity of the plasma
membrane aquaporin PM28A is regulated by phosphorylation.
The Plant Cell 10, 451–459.
Johansson I., Karlsson M., Johanson U., Larsson C. & Kjellbom P.
(2000) The role of aquaporins in cellular and whole plant water
balance. Biochimica et Biophysica Acta 1465, 324–342.
Murata K., Mitsuoka K., Hirai T., Walz T., Agre P., Heymann J.B.,
Engel A. & Fujiyoshi Y. (2000) Structural determinants of water
permeation through aquaporin-1. Nature 407, 599–605.
Murphy R. (1989) Water flow across the sieve tube boundary:
estimating turgor and some implications for phloem loading and
© 2007 The Authors
Journal compilation © 2007 Blackwell Publishing Ltd, Plant, Cell and Environment, 30, 1566–1577
Aquaporins unload phloem-imported water 1577
unloading. IV. Root tips and seed coats. Annals of Botany 63,
571–579.
Niemietz C.M. & Tyerman S.D. (1997) Characterization of water
channels in wheat root membrane vesicles. Plant Physiology 115,
561–567.
Niemietz C.M. & Tyerman S.D. (2002) New potent inhibitors of
aquaporins: silver and gold compounds inhibit aquaporins of
plant and human origin. FEBS Letters 531, 443–447.
Offler C.E. & Patrick J.W. (1984) Cellular structures, plasma
membrane surface areas and plasmodesmatal frequencies of
seed coats of Phaseolus vulgaris L. in relation to photosynthate
transfer. Australian Journal of Plant Physiology 11, 79–99.
Park J.H. & Saier M.H. (1996) Phylogenetic characterization of
the MIP family of transmembrane channel proteins. Journal of
Membrane Biology 153, 171–180.
Pate J.S., Peoples M.B., van Bel A.J.E., Kuo J. & Atkins C.A. (1985)
Diurnal water balance of the cowpea fruit. Plant Physiology 77,
148–156.
Patrick J.W. (1994) Turgor-dependent unloading of assimilates
from coats of developing legume seed: assessment of the significance of the phenomenon in the whole plant. Physiologia Plantarum 90, 367–377.
Patrick J.W. & Offler C.E. (2001) Compartmentation of transport
and transfer events in developing seeds. Journal of Experimental
Botany 37, 1006–1019.
Picaud S., Becq F., Dédaldéchamp F., Ageorges A. & Delrot S.
(2003) Cloning and expression of two plasma membrane aquaporins expressed during the ripening of grape berry. Functional
Plant Biology 30, 621–630.
Sakurai J., Ishikawa F., Yamaguchi T., Uemura M. & Maeshima M.
(2005) Identification of 33 rice aquaporin genes and analysis of
their expression and function. Plant & Cell Physiology 46, 1568–
1577.
Schuurmans J.A.M.J., van Dongen J.T., Rutjens B.P.W., Boonman
A., Pieterse C.M.J. & Borstlap A.C. (2003) Members of the
aquaporin family in the developing pea seed coat include representatives of the PIP, TIP, and NIP subfamilies. Plant Molecular
Biology 53, 655–667.
Sui H., Han B.G., Lee J.K., Walian P. & Jap B.K. (2001) Structural
basis of water-specific transport through the AQP1 water
channel. Nature 414, 872–878.
Thomas M., Hetherington L. & Patrick J.W. (2000) Genotypic differences in seed growth rates of Phaseolus vulgaris L. I. General
characteristics, seed coat growth factors and comparative roles of
seed coats and cotyledons. Australian Journal of Plant Physiology 27, 109–118.
Tornroth-Horsefield S., Wang Y., Hedfalk K., Johanson U.,
Karlsson M., Tajkhorshid E., Neutze R. & Kjellbom P. (2006)
Structural mechanism of plant aquaporin gating. Nature 439,
688–694.
Tournaire-Roux C., Sutka M., Javot H., Gout E., Gerbeau P., Luu
D.-T., Bligny R. & Maurel C. (2003) Cytosolic pH regulates root
water transport during anoxic stress through gating of aquaporins. Nature 425, 393–397.
Tyerman S.D., Niemietz C.M. & Bramley H. (2002) Plant aquaporins: multifunctional water and solute channels with expanding
roles. Plant, Cell & Environment 25, 173–194.
Vera-Estrella R., Barkla B.J., Bohnert H.J. & Pantoja O. (2004)
Novel regulation of aquaporins during osmotic stress. Plant
Physiology 135, 2318–2329.
Walbot V., Clutter M. & Sussex I.M. (1972) Reproductive development and embryogeny in Phaseolus. Phytomorphology 22, 59–68.
Walker N.A., Patrick J.W., Zhang W.-H. & Fieuw S. (1995) Efflux of
photosynthate and acid from developing seed coats of Phaseolus
vulgaris L.: a chemiosmotic analysis of pump driven efflux.
Journal of Experimental Botany 46, 539–549.
Walker N.A., Zhang W.-H., Harrington G., Holdaway N. & Patrick
J.W. (2000) Effluxes of solutes from developing seed coats of
Phaseolus vulgaris L. and Vicia faba L.: locating the effect of
turgor in a coupled chemiosmotic system. Journal of Experimental Botany 51, 1047–1055.
Wallace I.S. & Roberts D.M. (2004) Homology modeling of representative subfamilies of Arabidopsis major intrinsic proteins.
Classification based on the aromatic/arginine selectivity filter.
Plant Physiology 135, 1059–1068.
Wan X.C., Steudle E. & Hartung W. (2004) Gating of water channels (aquaporins) in cortical cells of young corn roots by
mechanical stimuli (pressure pulses): effects of ABA and of
HgCl2. Journal of Experimental Botany 55, 411–422.
Zhang W.-H., Atwell B.J., Patrick J.W. & Walker N.A. (1996)
Turgor-dependent efflux of assimilates from coats of developing seed of Phaseolus vulgaris L.: water relations of the cells
involved in efflux. Planta 199, 25–33.
Zhang W.-H., Skerrett M., Walker N.A., Patrick J.W. & Tyerman
S.D. (2002) Non-selective currents and channels in plasma membranes of protoplasts from coats of developing seeds of bean
seeds. Plant Physiology 128, 388–399.
Zhang W.-H., Walker N.A., Tyerman S.D. & Patrick J.W. (2004)
Pulsing Cl- channels linked to hypoosmotically-induced turgor
regulation in coat cells of developing bean seeds. Journal of
Experimental Botany 55, 993–1001.
Zhang W.-H., Zhou Y., Dibley K.E., Tyerman S.D., Furbank R.T. &
Patrick J.W. (2007) Nutrient loading in developing seeds. Functional Plant Biology 34, 314–331.
Zhou Y., Qu H., Dibley K.E., Offler C.E. & Patrick J.W. (2007) A
suite of sucrose transporters expressed in coats of developing
legume seeds includes novel pH-independent facilitators. The
Plant Journal 49, 750–764.
Received 18 June 2007; accepted for publication 20 August 2007
© 2007 The Authors
Journal compilation © 2007 Blackwell Publishing Ltd, Plant, Cell and Environment, 30, 1566–1577