The anatomy of developing pea seeds is characterized by transfer cells present in both coats and cotyledons at the maternal/filial interface. To determine the nature and cellular localization of sucrose transporters in pea seeds, a full-length clone of a sucrose/H+ symporter ( PsSUT1) was isolated from a cotyledon cDNA library. Northern blot analyses of different organs showed that PsSUT1 is expressed in non-seed tissues, including sucrose sinks and sources. Within developing seeds, transcripts of PsSUT1 and PsAHA1 genes were detected in all tissues, while transcripts of a sucrose binding protein ( GmSBP) were confined to cotyledon epidermal transfer cells. Signal intensities of PsSUT1 and PsAHA1 transcripts and protein products were most pronounced in the thin-walled parenchyma cells of seed coats and epidermal transfer cells of cotyledons. For cotyledons, the highest transporter densities were localized to those portions of plasma membranes lining the wall ingrowth regions of epidermal transfer cells. Responses of [14C]sucrose influx to metabolic inhibitors indicated that proton-coupled sucrose transport was operative in both seed coats and cotyledons. Cotyledon epidermal transfer cells were shown to support the highest sucrose flux. Maximal transport activity was found to account for the sucrose flux differences between seed tissues. Intercellular movement of the symplasmic tracer, 5-(6)-carboxyfluorescein (CF), demonstrated that symplasmic pathways interconnect the vascular tissues to thin-walled parenchyma transfer cells of seed coats and, for cotyledons, epidermal transfer cells to storage parenchyma cells.
Sieve element unloading and post-sieve element transport processes are considered to play a central role in regulating photoassimilate transport to, and partitioning between, sinks ( Patrick 1997). For most sink regions, post-sieve element transport occurs deep within their tissues, where vascular and sink cells share an intimate anatomical relationship. As a consequence, direct and unambiguous experimental investigation of post-sieve element transport encounters significant technical difficulties. This problem is somewhat circumvented in developing seeds of grain legumes. In particular, the lack of anatomical interconnection between maternal and filial tissues of these seeds provides access to plasma membrane transport events responsible for photoassimilate movement to and from the seed apoplasm ( Patrick & Offler 1995; Weber et al. 1997b ).
Considerable evidence demonstrates that both sucrose release from seed coats and uptake by cotyledons are facilitated membrane transport events ( Patrick & Offler 1995; Patrick 1997). With the recent development of suitable expression systems ( Frommer & Ninnemann 1995), opportunities now exist to characterise putative transporters in developing seeds at a molecular level. Indeed, sucrose/H+ symporters have been cloned from cotyledons of developing Vicia faba seeds and functionally characterized in a yeast complementation system ( Weber et al. 1997b ). Once cloned, the purported regulatory significance of sucrose transporters as modulators of post-sieve element transport of photoassimilates in developing seeds ( Harrington et al. 1997a ; Harrington et al. 1997b ; McDonald et al. 1996b ; Weber et al. 1997b ) may be evaluated by manipulating their expression levels using transgenic plants ( Lemoine et al. 1996 ). The best understood legume seed systems are those of Glycine max, Phaseolus vulgaris, Pisum sativum and Vicia faba ( Patrick & Offler 1995). However, to date, stable transformation has only been achieved for Pisum sativum L. ( Christou 1997).
Assessment of altering expression of sucrose transporter genes on post-sieve element movement of photoassimilates in developing seeds is predicated upon knowing which cells are responsible for membrane transport of sucrose to and from the seed apoplasm. Such information is not available for developing pea seeds ( Patrick & Offler 1995). To this end, the present paper describes an attempt to identify the cellular localization of genes and their functional protein products considered responsible for membrane transport of sucrose. The anatomical context of our approach draws upon, and extends, the findings of Hardham (1976). For completeness, symplasmic continuity to ( Grusak & Minchin 1988), and from sites of, membrane exchange is examined using the symplasmic tracer, 5-(6)-carboxyfluoroscein ( McDonald et al. 1995 ).
Seed anatomy provides a structural framework to evaluate cellular pathways of post-sieve element transport of photoassimilate and cellular locations of potential sucrose transporters facilitating exchange between maternal and filial tissues. In this context, there is a dearth of relevant anatomical information available for developing pea seeds. Hence a study was undertaken at the light and electron microscope levels to redress this situation.
Hardham (1976) established that the seed coat vasculature arises from a single funicular vein that enters the seed coat at the chalazal end of the hilum. Here it branches to form a chalazal vein located along the integumentary fusion line ( Fig. 1a) and two lateral veins; the latter veins are composed entirely of phloem and extend either side of the hilum toward the site of radicle insertion ( Hardham 1976). This vascular organization is identical to that found in the seed coats of Vicia faba ( Offler et al. 1989 ).
The vascular system is embedded in parenchymatous tissue underlying the protective single-cell layers of palisade and hypodermis ( Fig. 1a). The parenchymatous tissue is composed of circumferential bands of chlorenchyma, ground parenchyma and thin-walled parenchyma arranged sequentially inward from the hypodermis ( Fig. 1a). In the vascular region, chlorenchyma and ground parenchyma overarch the chalazal vein which separates these layers from the thin-walled parenchyma; the latter forms the innermost cell layers of the seed coat ( Fig. 1a). Similar to Vicia seed coats ( Offler et al. 1989 ), papillate secondary wall ingrowths ( Fig. 1c) were observed in the innermost two layers of rectangular-shaped thin-walled parenchyma cells ( Fig. 1b). The wall ingrowths are polarized to the inner periclinal walls of the thin-walled parenchyma cells ( Fig. 1b) and decrease in density with distance from the chalazal vein.
Elliptically shaped and closely packed storage parenchyma cells comprise the bulk of the cotyledon volume which is encased in a single-layered epidermis ( Fig. 1d). The abaxial epidermal cells, which abut the seed coat, exhibit an organelle-enriched cytoplasm bounded by polarized papillate secondary wall ingrowths in their outer periclinal walls ( Fig. 1e,f). Similar wall ingrowths are observed in abaxial epidermal cells of Vicia cotyledons ( Offler et al. 1989 ). However, wall labyrinth development in the pea is much less developed ( Fig. 1e). In addition, a subepidermal layer is not present ( Fig. 1a; McDonald et al. 1995 ). The storage parenchyma cells undergo considerable ultrastructural change during the filling phase of seed development. Their vacuoles vesiculate to form protein bodies ( Craig et al. 1979 ), while proplastids differentiate into dense clusters of starch-storing amyloplasts ( Bain & Mercer 1966).
Molecular cloning of a sucrose/H+ symporter
A full-length sucrose H+/symporter gene (PsSUT1) was isolated from a P. sativum cDNA library. The cDNA (2034 bp) encodes an open-reading frame of 1575 bp with a calculated protein of 55 kDa. The protein sequence contains 12 putative membrane spanning domains (data not shown). The sequence shares 96.3% identical amino acids with the sucrose transporter VfSUT1 from Vicia faba ( Weber et al. 1997a ). Sequence comparison between sucrose transporters ( Fig. 2) shows that PsSUT1 falls in the same group with VfSUT1 and RcSUT1 (Ricinus communis), indicating a closer relationship to each other than to the other SUTs. Northern blot analysis showed that PsSUT1 is expressed in all tissues, with the strongest expression in flowers and the weakest expression in sink leaves ( Fig. 3a). This was confirmed by rehybridization with the 3′ untranslated region of PsSUT1.
Expression analysis of sucrose-transport related genes in developing pea seeds
Total RNA was prepared from seed coats, paradermal slices of cotyledons enriched in epidermal transfer cells ( Fig. 1d) and cotyledon tissue segments containing storage parenchyma cells alone. Full-length (GmSBP) or partial (PsSUT1;PsAHA1) probes hybridized to single bands of approximately 1.7, 1.8 and 3.4 kb that correspond to reported mRNA lengths of these genes (e.g. Ewing & Bennett 1994; Grimes et al. 1992 ; Ward et al. 1997 ). Transcripts of PsSUT1 and PsAHA1 genes were detected in all seed tissues, while SBP transcripts were confined to cotyledon tissues enriched in epidermal transfer cells ( Fig. 3b). Signal strength of PsSUT1 and PsAHA1 transcripts were greatest in cotyledon epidermal transfer cells; progressively lesser signal strengths were detected in cotyledon storage parenchyma cells and seed coat tissues, respectively ( Fig. 3b). This pattern was particularly notable for PsSUT1 expression.
The cellular localization of PsSUT1 gene expression in seed coats and cotyledons was examined by exposing tissue sections to DIG-labelled riboprobes prepared from identical cDNA fragments used for Northern blot analyses. Antisense riboprobe of PsSUT1 hybridized to thin-walled parenchyma cells of seed coats with weaker and dispersed signals being observed in the ground parenchyma and vascular bundles (data not shown). In contrast, exposure of cotyledon sections to antisense riboprobe caused strong labelling of epidermal transfer cells ( Fig. 4d versus 4b) with notably decreasing radial gradient of expression through consecutive rows of storage parenchyma cells ( Fig. 4c versus 4a).
Localization of sucrose transport-related proteins in developing pea seeds
Western blot analyses of microsomal fractions from seed tissues demonstrated that specific proteins cross-react with antibodies raised against a plasma membrane H+-ATPase, SUT1 and SBP. These corresponded to molecular masses of 100 kDa ( Serrano 1989), 47 kDa ( Ward et al. 1997 ) and 62 kDa ( Grimes et al. 1992 ), respectively ( Fig. 3c). Only SBP showed weak cross-reaction with additional polypeptides (40 and 42 kDa). These proteins were exclusively observed in samples containing SBP, thus indicating that they represent degradation products (data not shown). Plasma membrane H+-ATPase and sucrose/H+ symporter signals were greatest in the epidermal transfer cell enriched fraction of cotyledons. As anticipated from Northern blot analyses ( Fig. 3b), SBP was restricted to microsomal fractions isolated from cotyledon tissues enriched in epidermal transfer cells ( Fig. 3c).
Cellular localization of transporter proteins was examined by immunohistochemistry. Labelling patterns of cotyledon sections, exposed to H+-ATPase and sucrose/H+ symporter antibodies, demonstrated that epitopes to these antibodies were present in plasma membranes of epidermal transfer cells and storage parenchyma cells ( Fig. 5b,d). However, stronger signals were detected in epidermal transfer cells ( Fig. 5b,d). Within transfer cells, the greatest densities of transporters were localized mainly to regions lining the outer periclinal walls, probably the plasma membrane ( Fig. 5b,d). This labelling pattern correlates with the position of secondary wall ingrowths ( Fig. 1e,f). In contrast, signals for SBP were detected only in epidermal transfer cells and were restricted to their outer periclinal walls ( Fig. 5c).
Immunofluorescent labelling patterns of seed coat sections showed that H+-ATPases and sucrose/H+ symporters localized to the periphery of thin-walled parenchyma cells. Weaker signals for an PsAHA1 were also detected in other parenchymatous tissues and the chalazal vein (data not shown).
Sucrose transporter activity
To assess whether the detected sucrose transport-related proteins were functional, [14C]sucrose transport activity and its sensitivity to the membrane impermeant sulfhydryl reagent PCMBS were determined ( M’Batchi et al. 1986 ). In addition, proton coupling of sucrose influx was tested using EB, an inhibitor of proton pumps ( Beffagna & Romani 1988).
Sucrose uptake into seed coats was PCMBS insensitive at sucrose concentrations above 1 m m, whereas at 0.1 m m, sucrose influx was inhibited by both PCMBS and EB ( Table 1). In cotyledons, [14C]sucrose influx was monitored either for epidermal transfer cells plus storage parenchyma (intact cotyledons) or storage parenchyma alone (abraded cotyledons). Sucrose transport activity of intact cotyledons is approximately fourfold greater than that found for seed coats ( Table 1). Sucrose influx into both epidermal transfer cells and storage parenchyma cells was sensitive to PCMBS and EB ( Table 1) suggesting that it is mediated by H+/sucrose symport. However, comparing PCMBS-sensitive sucrose fluxes indicated that sucrose/H+ symporters, located in epidermal transfer cells, are sevenfold more active than those in storage parenchyma cells ( Table 1).
Table 1. Effect of PCMBS and EB on [14C]sucrose influx into coats and cotyledons (± abaxial surface abraded) of developing pea seeds
[14C] sucrose influx (n mol.cm–2.h–1)
Carrier sucrose conc.(mM)
Concentrations of PCMBS and EB were 10–3 and 10–4 M, respectively. The data in parentheses are the differences in the sucrose fluxes between the control and PCMBS-treated cotyledons. Each value is the mean ± SE of four replicates per treatment. ND, no data available.
133 ± 6
133 ± 12
127 ± 5
16.5 ± 1.1
15.3 ± 0.9
1.35 ± 0.04
0.88 ± 0.01
0.91 ± 0.04
471 ± 23
73 ± 10 (398)
119 ± 8
149 ± 8
95 ± 7 (54)
95 ± 4
Kinetic properties of the cotyledon sucrose/H+ symporters were evaluated in an attempt to account for the observed flux differences. Sucrose influx for both tissue systems exhibited saturation kinetics with a linear, apparently non-saturating component dominating at concentrations above 50 and 10 m m for the unabraded and abraded cotyledons, respectively (data not shown). Fitted slopes of the non-saturable components provide estimates for first order rate constants for this exchange and were found to be identical for the two tissues ( Table 2). Non-saturable components were subtracted from the total sucrose flux ( Maynard & Lucas 1982) and fluxes of the saturable component were fitted, using linear regression analyses, to Eddie–Hofstee transformations. The fitted transformed flux data provide estimates of apparent Km’s and φmax’s for sucrose influx which showed that marked differences exist between the two cotyledon cell types. The Km and φmax of sucrose influx into storage parenchyma cells (abraded cotyledons) were, respectively, 2.6- and 9.4-fold less than the corresponding values for the abaxial epidermal transfer cells (intact cotyledons and see Table 2).
Table 2. The kinetic parameters for [14C]sucrose influx into epidermal transfer cells (intact cotyledons) and storage parenchyma cells (abraded cotyledons) of developing pea cotyledons
The first-order rate constants (k’) for the non-saturable component of sucrose influx were estimated from the slopes of the linear portion of concentration-dependent fluxes. The Michaelis–Menten constants (Km) and maximal velocities (φmax) of the saturable components of the sucrose fluxes were derived from Eadie–Hofstee transformation of the flux data fitted by linear regressions. Each value is the mean ± SE of 10 replicates per treatment.
k′ (cm h–1 × 103)
8.82 ± 0.87
8.30 ± 0.69
7.72 ± 1.07
2.84 ± 0.55
φmax (n mol cm–2 h–1)
853 ± 72
91 ± 18
Symplasmic continuity assessed using 5-(6)-carboxyfluorescein
The anatomical study identified pit-fields of plasmodesmata in seed coats and cotyledons (data not shown). These may function to support symplasmic transport of photoassimilates to, and away from, the thin-walled parenchyma transfer cells of seed coats ( Figs 1a–c and 6) and cotyledon epidermal transfer cells ( Figs 1d–f, 4 and 5), respectively. The membrane impermeant fluorochrome, CF, was used to assess symplasmic continuity in seed coats and cotyledons.
Phloem-imported CF preferentially moved radially inward from the chalazal vein and then circumferentially around the seed coat in the thin-walled parenchyma transfer and chlorenchyma cells ( Fig. 6a versus 6b). For cotyledons, CF pulse-loaded into epidermal transfer cells ( Fig. 6c versus 6f) was found to move to, and through, the storage parenchyma cells ( Fig. 6d,e versus 6f).
Gene isolation and characterization
A full-length clone (PsSUT1), isolated from a pea cotyledon cDNA library, clustered with known sucrose/H+ symporters ( Fig. 2) and, not surprisingly, exhibited the greatest amino acid sequence similarity with a sucrose/H+ symporter isolated from developing seeds of Vicia faba (VfSUT1;Weber et al. 1997a ; see Fig. 2). The distinct clustering of PsSUT1 and VfSUT1 may indicate that they might play an important role in seed development. However, similar to VfSUT1 ( Weber et al. 1997b ) and RcSCR1 ( Weig & Komor 1996), PsSUT1 is expressed in non-seed tissues ( Fig. 3a). This suggests a more general role in sucrose transport for these genes. Indeed, significant expression of PsSUT1 in source leaves ( Fig. 3a) is likely to be vascular and function in phloem loading ( Ward et al. 1997 ). Dual expression patterns between vascular and non-vascular tissue has been reported for PmSUC1 which, similar to PsSUT1 ( Fig. 3a), exhibits the highest expression in developing flowers ( Ghartz et al. 1996 ). Moreover, concurrence of the signal intensities for Northern blot analyses performed with coding or 3′-untranslated cDNA probes suggests that PsSUT1 is the most highly expressed sucrose transporter throughout the plant.
Sucrose transport in seed coats
Transcripts of an H+-ATPase and sucrose/H+ symporter were detected in seed coats ( Fig. 3b), along with their protein products ( Fig. 3c). Existence of a functional sucrose/H+ symporter in pea seed coats is supported by the finding that in vitro sucrose influx was inhibited, at low external sucrose concentrations, by PCMBS and EB ( Table 1). Similar to other grain legume species (e.g. soybean, Bennett et al. 1984 ; Vicia,Patrick 1993), significant capacity for photoassimilate retrieval from the seed apoplasm by coats is absent from pea ( Table 1; Minchin & Thorpe 1990). The PCMBS- and EB-insensitive sucrose transport activities found at higher sucrose concentrations ( Table 1) probably represent facilitated diffusion through a membrane channel through which sucrose release to the seed apoplasm is envisaged to occur ( De Jong et al. 1996 ).
PsAHA1 and PsSUT1 transcripts were spread throughout seed coat tissues with the greatest densities co-localizing to thin-walled parenchyma cells (data not shown). This diffuse transporter distribution, combined with a depressed transporter activity ( Table 1), is consistent with a low capacity retrieval mechanism. This could function to re-load sucrose leaked to the coat apoplasm during passage along a symplasmic pathway of post-phloem transport delineated by the movement of phloem-imported CF ( Fig. 6; Grusak & Minchin 1988).
Sucrose transport in cotyledons
PsSUT1 and plasma membrane H+-ATPase gene (PsAHA1) are expressed in abaxial epidermal transfer cells and storage parenchyma cells of developing pea cotyledons ( Figs 3b and 4). The saturable components of sucrose influx across plasma membranes of epidermal transfer cells and storage parenchyma cells was slowed by PCMBS and EB ( Table 1). This observation provides circumstantial evidence that PsSUT1 and H+-ATPase are functionally active in plasma membranes of both these cotyledon cell types and probably mediate sucrose influx by proton symport (see also Lanfermeijer et al. 1991 ).
Epidermal transfer cells were found to support a threefold higher sucrose flux than storage parenchyma cells when supplied with an external sucrose concentration of 10 m m ( Table 1). This disparity in sucrose fluxes can be ascribed to differing kinetic properties of the respective sucrose/H+ symporters as first-order rate constants for non-saturable influxes are identical ( Table 2). The 2.6× lower Km exhibited for sucrose/H+ symport by the storage parenchyma cells ( Table 1) may result from expression of another PsSUT or modification of PsSUT1. The 9.4× difference in φmax ( Table 2) can partially be accounted for by relative levels of gene expression of PsSUT1 and H+-ATPase PsAHA1 ( Figs 3b and 4). Our observations do not support the proposition that SBP makes a substantial contribution to the non-saturable sucrose flux ( Overvoorde et al. 1996 ). Here we find that the first-order rate constants for this flux into epidermal transfer cells and storage parenchyma cells ( Table 2) are unaffected by exclusive localization of SBP to the epidermal transfer cells ( Figs 3b,c and 5c).
The presence of functional sucrose symporters in storage parenchyma cells of pea cotyledons contrasts with their absence from these cotyledon tissues in Phaseolus and Vicia seeds ( Harrington et al. 1997a ; McDonald et al. 1995 ; McDonald et al. 1996b ; Weber et al. 1997a ). The putative role of these symporters in pea cotyledons may be deduced from a comparison of predicted Michaelis–Menten sucrose influxes derived from kinetic constants of the symporter ( Table 2) and an endogenous apoplasmic sucrose concentration of 140 m m (J.W. Patrick, unpublished results). Flux estimates of 808 and 89 nmol cm–1 h–1 for epidermal transfer cells and storage parenchyma cells, respectively, demonstrate that storage parenchyma sucrose/H+ symporters make a minor contribution to the net gain of sucrose by developing pea cotyledons. Thus, their likely function is the retrieval of sucrose leaked to the cotyledon apoplasm from storage parenchyma cells. This is consistent with the relatively low Km exhibited by these sucrose/H+ symporters compared to those located in epidermal transfer cells ( Table 2).
Sucrose accumulation into pea cotyledons occurs against a substantive outward-directed transmembrane concentration difference (approximately 54 m m sucrose, J.W. Patrick, unpublished results). This suggests that the sucrose/H+ symporters, and particularly those located in epidermal transfer cells, play a central role in the cotyledon’s carbon economy. Similar conclusions have been reached for developing cotyledons of Phaseolus ( McDonald et al. 1995 ) and Vicia ( Harrington et al. 1997b ; McDonald et al. 1995 ; McDonald et al. 1996a ). Use of transgenic peas in which sucrose symporter expression has been manipulated will provide a definitive test for this hypothesis. As found for Phaseolus and Vicia cotyledons ( McDonald et al. 1995 ), subsequent transport of accumulated sucrose to storage parenchyma cells is likely to occur through a symplasmic route, as suggested by the observed mobility of the membrane impermeant fluorochrome CF ( Fig. 6d,e versus f).
A full-length clone of PsSUT1, isolated from a pea cotyledon cDNA library, clustered with other seed-specific sucrose symporters. However, the gene is expressed throughout the plant body suggesting a general role in sucrose transport. For developing pea seeds, PsSUT1 is expressed in coats and cotyledons.
Phloem unloading is symplasmic with subsequent circumferential symplasmic movement through thin-walled parenchyma cells. PsSUT1, located along this post-sieve element transport pathway, operates to recover sucrose leaked to the seed apoplasm during symplasmic passage. Photoassimilate exchange to the seed apoplasm is mediated by facilitated diffusion through plasma membrane channels ( De Jong et al. 1996 ) of the thin-walled parenchyma transfer cells.
Net gain of sucrose from the seed apoplasm by the cotyledons principally occurs through the epidermal transfer cells mediated by PsSUT1 energized by H+-ATPases. The highest densities of both transporters are co-localized to portions of the plasma membranes lining the outer periclinal wall ingrowth regions of the epidermal transfer cells. Accumulated sucrose moves to the storage parenchyma cells through a symplasmic route and sucrose leaked to the cotyledonary apoplasm is retrieved by symporters localized to the storage parenchyma cells.
Plants of Pisum sativum (cv. Greenfeast) were raised in 1.5 l pots under glasshouse conditions (partial temperature control of 20–26°C by day, 15–17°C by night; supplementary lighting with metal halide lamps to ensure a minimum photosynthetically active radiation (PAR) on the upper most leaves of 200 μmol m–2 s–1 and a 14 h photoperiod) in a potting mix of coarse sand, peat and perlite (3:1:1) with the addition of lime (4 gl–1) and slow release fertiliser (6 g l–1; Nutricote, Chuso Asaki Fertiliser, Tokyo, Japan). Mineral nutrition of plants was supplemented with full-strength Hoaglands no. 1. Developing seeds were harvested for observation during their linear phase of cotyledon dry weight gain. The water content of seed coats and cotyledons was 80 and 68–75%, respectively.
Light and transmission electron microscopy
Tissue samples of seed coat and cotyledon were fixed on ice in 3% gluteraldehyde and 3% paraformaldehyde in 25 m m cacodylate buffer (pH 7.6) containing 10 m m sucrose, 2 m m CaCl2 and 0.5% caffeine. For electron microscopy, tissue was post-fixed overnight in 1% osmium tetroxide. After dehydration the tissue was embedded in L.R. White ( Harrington et al. 1997b ). For light microscopy 1 μm thick sections were stained with toluidine blue, and for electron microscopy, 50 nm thick sections were stained with uranyl acetate and lead citrate.
Isolation of sucrose symporter by cDNA library screening
A pea cotyledon cDNA library in UniTMXR-ZAP (Stratagene) was kindly provided by Trevor Wang (John Innes Institute, Norwich, UK). The library was screened using potato sucrose/H+ symporter (SoSUT1), tobacco (StSUT1) and tomato (LeSUT1) as probes (for a review see Ward et al. 1997 ). Hybridization was performed for 16 h at 42°C in 5× SSC, 5× Denhardt’s, 0.5% SDS, 0.1 mgl–1 tRNA, 25% formamide. Subsequently, filters were washed three times for 20 min in 2× SSC (0.15 m NaCl, 0.015 m sodium citrate) and 0.1% SDS. After two rounds of screening, plaque-purified phages were converted to pBluescript SK-derivatives by in vivo excision and inserts were sequenced.
Isolation of a partial H+-ATPase cDNA by degenerate PCR
Reverse transcriptase polymerase chain reaction (RT-PCR) was performed to isolate cDNA encoding a P-type H+-ATPase from cotyledons of P. sativum. Total RNA (3 μg), extracted from transfer cell-enriched tissue of P. sativum cotyledons, was taken as template for first-strand cDNA synthesis with superscript reverse transcriptase (GIBCO BRL) according to the manufacture’s protocol. Subsequently, cDNA was amplified using degenerate primers derived from peptide sequences conserved in eukaryotic H+-ATPases (EMBL data library; Ewing & Bennett 1994). Amplified cDNAs were subcloned into pGEM-T (Promega Corporation, Madison, WI, USA) and sequenced.
Northern blot analysis
Total RNA of different plant organs was extracted from a pool of 10 plants according to Frommer et al. (1994) . Total RNA (20 μg) was electrophoresed and then transferred to Hybond N (Amersham) following standard procedures. Filters were hybridized with full-length GmSBP from soybean ( Grimes et al. 1992 ) and partial sequences of PsSUT1 (200 bp of 3′ untranslated region or 623 bp of coding region) and PsAHA1 as 32P-labelled cDNA probes for 16 h at 42°C in a hybridization solution containing 5× SSC, 2× Denhardt’s, 0.5% SDS, 0.1 mg ml–1 tRNA, 50% formamide pH 6.4. Post-hybridization filters were washed in 2× in 0.5× and 0.1× SSC containing 0.1% SDS each for 20 min and at either 50°C (GmSBP), 45°C (PsAHA1) or 65°C (PsSUT1).
Antibody preparation and purification
Polyclonal antibodies to the 62 kDa sucrose binding protein were prepared in rabbit ( Ripp et al. 1988 ). The antibodies were further purified by affinity chromatography with the highly purified 62 kDa protein coupled to CNBr-activated Sepharose 4 ( Grimes et al. 1992 ). The anti-H+-ATPase polyclonal antibody was raised against a fusion protein containing the C-terminal sequence of the seed-specific (AHA10 isoform) plasma membrane-bound H+-ATPase of Arabidopsis thaliana. For SUT, polyclonal antibodies were raised in rabbit against a fusion peptide of loop VII of StSUT1 ( Lemoine et al. 1996 ). Compared to the predicted sequence of PsSUT1 (this paper), nine of the 15 amino acids were identical to the corresponding amino acid sequence of StSUT1.
Microsomal protein extraction and Western blot analysis
Microsomal protein fractions were extracted from seed coats, transfer cell-enriched tissue was surgically removed from cotyledons and cotyledon storage parenchyma cells according to the method described by Grimes et al. (1992) . Protein samples (15–30 μg lane–1) were electrophoresed and electroblotted. The SBP blots were blocked for 1.5 h in TBS (10 m m Tris, 0.5 m NaCl), 0.1% Tween 20, 1% BSA, 10% powdered milk (nonfat), 5% glycerol. For SUT and H+-ATPase blots, blocking was for 1 h in TBS, 1% BSA. Blots were treated with primary antibody diluted in block solution. Membranes were incubated with SBP antibodies (1:1000) for 4 h at room temperature or with SUT (1:5000) and H+-ATPase (1:150) antibodies overnight (12 h) at 4°C. Blots were washed with alkaline phosphatase-conjugated secondary antibody diluted 1:1000 in TBST + 1% BSA. Washed blots were incubated in alkaline phosphatase substrate medium (Western Blue, Promega).
Immunocytochemistry and in situ hybridization
For immunolocalization studies, pieces of seed coat and cotyledon tissue (10 × 4 mm) were embedded in L.R. White resin ( Harrington et al. 1997a ). Tissue sections (1 μm) were probed for 4 h with SUT (1:50), SBP (1:50) or H+-ATPase (1:75) primary antibodies in TBST-BSA. Primary antibody binding was detected by 2 h incubation in FITC-conjugated anti-rabbit IgG (1:50, v/v in TBST-BSA; Harrington et al. 1997a ). Control sections were incubated in non-immune serum (H+-ATPase, SUT) or IgG affinity purified from non-immune serum (SBP) at the same dilution as the primary antibody. After overnight incubation (4°C), sections were mounted in Moviol containing 0.1% w/v phenylenediamine solution and viewed with a Zeiss Axiophot fluorescent microscope using epifluorescence.
Digoxigenin (DIG)-labelled PsSUT1 sense and antisense riboprobes were synthesized by in vitro transcription. The plasmid was linearized and DIG-11-UTP was incorporated with SP6 or T7 RNA polymerase according to the manufacturer’s instructions (Boehringer Mannheim). In situ hybridization was performed as described previously ( Harrington et al. 1997b ).
Uptake of [14C]sucrose
All solutions used for sucrose uptake contained [14C]sucrose (0.12 MBq ml–1) at a specified carrier concentration, 1 m m CaCl2 buffered at pH 5.5 with 5 m m Mes/KOH. Final solution osmolalities were adjusted to 300 mOsmol kg–1 ( Lanfermeijer et al. 1991 ) with sorbitol.
Seed coats were separated into halves around integumentary fusion lines. Inner surfaces of excised cotyledons were covered with Apiezon grease to ensure that sugar uptake followed the in vivo path ( McDonald et al. 1995 ). The sucrose transport capacity of storage parenchyma cells was assessed using cotyledons with their abaxial epidermis removed ( McDonald et al. 1996a ). Cellular debris was removed by 3 × 3 sec washes in ice-cold carrier sucrose solution. Seed tissues were pre-equilibrated in carrier sucrose solution for 20 min with or without metabolic inhibitors [1 m m p-chloromercuribenzenesulphonic acid (PCMBS) or 10–4 M erythrosin B (EB)]. Pre-treated seed tissues were transferred to [14C]sucrose solution of identical composition. Plasma membrane fluxes of sucrose were estimated by measuring rates of [14C]sucrose accumulation across a known tissue surface area exposed for 10 min to [14C] sucrose solutions ( McDonald et al. 1996a ).
Application and visualization of 5-(6)-carboxyfluorescein
For phloem import of 5-(6)-carboxyfluorescein (CF) into seed coats, 5-(6)-carboxyfluorescein diacetate (CFDA) solution (100 μg ml–1 in 300 m m sorbitol, pH 5.5) was delivered to wells located on the adaxial surfaces of pod walls. Cotyledons were prepared as detailed for sucrose uptake studies (see above) and exposed to control or CFDA solutions for 3 min to selectively pulse load epidermal transfer cells followed by a chase of 2 h in control solution. Hand sections cut transversely to the longitudinal axis of each seed were mounted in 50% aqueous glycerol on microscope slides. Sections were observed under epifluorescence using a Zeiss Axiophot Photomicroscope fitted with the filter set: E 450–490 blue excitation filter, F 510 chromatic beam splitter and LP 420 barrier filter.
We gratefully acknowledge the supply of a pea cotyledon cDNA library by Dr Trevor Wang (John Innes), SBP cDNA and polyclonal antibody by Dr Bill Hitz (Monsanto), and A.t. H+-ATPase antibody by Dr Jeff Harper (Scripps Research Institute). The expert technical expertise of Louise Hetherington and the provision of healthy plant material by Kevin Stokes is much appreciated. Research was supported by an Australian Research Council Large Grant (A19530955) awarded to J.W.P. and C.E.O.