Phloem unloading follows an extensive apoplasmic pathway in cucumber (Cucumis sativus L.) fruit from anthesis to marketable maturing stage


Z. Zhang. e-mail:


The phloem unloading pathway remains unclear in fruits of Cucurbitaceae, a classical stachyose-transporting species with bicollateral phloem. Using a combination of electron microscopy, transport of phloem-mobile symplasmic tracer carboxyfluorescein, assays of acid invertase and sucrose transporter, and [14C]sugar uptake, the phloem unloading pathway was studied in cucumber (Cucumis sativus) fruit from anthesis to the marketable maturing stage. Structural investigations showed that the sieve element–companion cell (SE–CC) complex of the vascular bundles feeding fruit flesh is apparently symplasmically restricted. Imaging of carboxyfluorescein unloading showed that the dye remained confined to the phloem strands of the vascular bundles in the whole fruit throughout the stages examined. A 37 kDa acid invertase was located predominantly in the cell walls of SE–CC complexes and parenchyma cells. Studies of [14C]sugar uptake suggested that energy-driven transporters may be functional in sugar trans-membrane transport within symplasmically restricted SE–CC complex, which was further confirmed by the existence of a functional plasma membrane sucrose transporter (CsSUT4) in cucumber fruit. These data provide a clear evidence for an apoplasmic phloem unloading pathway in cucumber fruit. A presumption that putative raffinose or stachyose transporters may be involved in soluble sugars unloading was discussed.


Phloem unloading has been defined as assimilates moving from sieve element–companion cell (SE–CC) complexes to sites of utilization/storage in the recipient sink cells (Oparka 1990; Patrick 1997). Thus, phloem unloading includes transfer across the SE–CC complex boundary (SE unloading) and subsequent transport through a diverse range of sink parenchyma cells (post-phloem transport) (Oparka 1990; Patrick 1997). It is now well accepted that phloem unloading plays an important role in the partitioning of photoassimilates, so determining to a large extent crop output and quality (Patrick 1997). Evidence from several plant species suggests that unloading at the terminal end of the phloem is mostly symplasmic via plasmodesmata, although an apoplasmic step may occur at some point along the post-phloem transport (Patrick 1997; Roberts et al. 1997; Imlau, Truernit & Sauer 1999; Haupt et al. 2001). Apoplasmic phloem unloading is barely theoretically possible in strong sinks such as immature leaves and seeds (Wang et al. 1995; Haupt et al. 2001) because of the probable limitations caused by trans-membrane solute movement involving a transporter-mediated and energy-coupled process (Patrick 1997).

The phenomenon of phloem unloading has been investigated over the past 20 years (Turgeon & Wolf 2009), but the pathway of phloem unloading in economically important sink organs such as fruits has only been elucidated in a relatively small number of species. For example, in grapefruit (Citrus paradisi), disruption of the symplasmic continuity within the stalks of juice vesicles and [14C]assimilates tracing suggest an apoplasmic post-phloem transport (Koch & Avigne 1990). In apple (Malus domestica) fruit, phloem unloading follows an extensive apoplasmic pathway (Zhang et al. 2004). In walnut (Juglans regia) fruit, phloem unloading is symplasmic in the seed pericarp and apoplasmic in the fleshy pericarp (Wu et al. 2004). In addition, several studies revealed that the phloem unloading routes are changeable, and may shift in response to sink development and function. In tomato (Lycopersicon esculentum) fruit and grape (Vitis vinifera) berry, a symplasmic pathway operates at the early stages, but an apoplasmic pathway occurs at the late stages of fruit development (Ruan & Patrick 1995; Zhang et al. 2006). In jujube (Zizyphus jujuba) fruit, the phloem unloading pathway is predominantly apoplasmic, but it is interrupted by a symplasmic pathway at the middle stage (Nie et al. 2010). In potato (Solanum tuberosum) tuber, the symplasmic pathway is required for bud outgrowth following dormancy, but growth is initially prevented by substrate limitation mediated by symplasmic isolation; however, the unloading pathway switches to symplasmic again upon the swelling of potato stolons (Viola et al. 2001, 2007). It seems likely that the patterns of phloem unloading pathway in fruits are complicated, as it varies greatly in different kinds of fruits on the base of sink type and developmental stages.

Sucrose is the most prevalent sugar that is translocated in many plant species, such as in Arabidopsis (Arabidopsis thaliana), maize (Zea mays) or tobacco (Nicotiana benthamiana). However, in numerous other plants, stachyose is the primary sugar that is translocated between organs in the phloem (Bachmann, Matile & Keller 1994; Hu et al. 2009). To date, there are many unanswered questions about the transport and metabolism of stachyose, which are far more complicated than that of sucrose. It is believed that elucidation of the cellular pathways of phloem loading and unloading in stachyose-transporting species will be helpful to understand these questions. Although a symplasmic phloem loading pathway in source leaves, known as polymer trapping (Rennie & Turgeon 2009), has been extensively investigated, there is no available report about phloem unloading in sink organs of stachyose-transporting species to date.

Cucumber (Cucumis sativus), a typical stachyose-transporting plant (Hu et al. 2009), is an important horticultural crop, and its fruit is one of the economically momentous sink organs with the whole body as the edible portion. However, the phloem unloading pathway in cucumber fruit is not investigated. In the present study we examined the pathway of phloem unloading in cucumber fruit from anthesis to the marketable maturing stage using a combination of electron microscopy, transport of the phloem-mobile symplasmic tracer carboxyfluorescein (CF), assays of acid invertase and a putative sucrose transporter (CsSUT4), and studies of [14C]sugar uptake.


Plant materials and growth curve

Cucumber (Cucumis sativus cv. Guonong 25) plants were grown in a growth chamber with a 10 h light period (600 µmol m−2 s−1) at 25 °C and a 14 h dark period at 18 °C, and relative humidity of 70%. Plants were watered once daily and fertilized weekly with Hoagland nutrient solution. Fruit age was determined by tagging each flower at anthesis. Based on the growth curve (Supporting Information Fig. S1), cucumber fruits were collected at the following developmental stages: beginning of development [0 d after anthesis (DAA); stage I]; early developmental stage (4 DAA; stage II); middle developmental stage (7 DAA; stage III); and marketable maturing stage (10 DAA; stage IV).

14CO2 labelling and autoradiography

The 14CO2 labelling and autoradiography were performed as described previously (Zhang et al. 2004). Leaves located at the same nodes of fruits were labelled with 14CO2 to follow the subsequent movement of radioactivity into fruits. The leaves were enclosed in plastic bags, and the radiolabel was injected into a vial inside the bag. Each leaf received 1.85 MBq of 14CO2 released from [14C]bicarbonate by the addition of excess 3 m lactic acid. The leaves were exposed to 14CO2 for 1 h before injecting excess 3 m KOH to neutralize the acid and to absorb any remaining 14CO2. After labelling, the leaves were left for 24 h to translocate 14C to fruits.

After 14CO2 labelling, fruit tissue for autoradiography was selected and sectioned before rapid freezing in liquid nitrogen between sheets of paper. The frozen sample was gently compressed between aluminum plates and freeze-dried. After drying, the tissue was pressed flat and autoradiographed using Kodak BioMax MR-1 film (Kodak, Rochester, NY, USA) at −80 °C for 5 d.

CF diacetate (CFDA) and Texas red labelling

The CFDA (Sigma-Aldrich, solution was introduced into cucumber fruit from the pedicel (Fig. 1a) as described previously (Zhang et al. 2004). A few cortical cell layers of the pedicel were removed, avoiding any damages to the phloem. Afterwards, the pedicel was enlaced by a cotton thread at one end and the other end of the cotton thread was immersed in a tube with 200 µL 1 mg mL-1CFDA aqueous solution (prepared from a stock solution in acetone). Plants were allowed to translocate the CF for 24 h, and the fruit tissues were subsequently sectioned and examined using confocal laser scanning microscopy (CLSM).

Figure 1.

Structure of cucumber fruit and [14C]labelling of vascular bundles. (a) An intact cucumber fruit, showing the loading site of carboxyffuorescein diacetate (CFDA) into the pedicel (indicated by red arrow). Unless otherwise noted, segments taken for analysis are the central area of fruit. (b) A schematic diagram of the transverse section of a cucumber fruit, showing the distribution of the vascular bundles and the sampling site (indicated by blue arrow). (c, d) Transverse (c) and longitudinal (d) anatomical sections of the close-up of boxed region in (b). Arrows indicate the vascular bundles that are bicollateral. (e, f) Transverse sections of a cucumber fruit at stage IV, showing the vascular bundles before (e) or after (f) [14C]autoradiography. [14C]labelled assimilates were transported in the vascular bundles.

To distinguish xylem from phloem of the vascular system in fruit under CLSM, the CFDA-treated fruits were sliced into columns containing vascular bundle zones (Fig. 1b), about 2 cm in length. The column was immersed, to a depth of 3 mm, in a solution containing 1 mg mL−1 3 kDa Texas Red dextran (Invitrogen, for 30 min before preparing the sections for examination under CLSM. Furthermore, columns were hand sectioned from the loaded cut surface upwards at 1 cm intervals to ensure that only xylem transport (and not parenchyma transport) was responsible for dye delivery to the apical region.

Tissue sectioning and microscopy

Prior to CLSM, the phloem zones of the tissues (Fig. 1b) were manually cut into transverse or longitudinal sections that were immersed immediately into 80% (v/v) glycerol to prevent dye loss.

A D-ECLIPSE C1 CLSM (Nikon, Tokyo, Japan) was used to image CF and Texas Red dextran transport. CF was excited by the 488 nm beam produced by an argon laser. On occasions where detached fruits were allowed to transpire Texas Red dextran after CF import, the sections were imaged first for CF by using blue (488 nm) light and immediately afterward for Texas Red dextran by using green excitation (543 nm) light (Roberts et al. 1997). Images taken at the different wavelengths were then superimposed to reveal phloem and xylem transport on a single image of the section.

Tissue preparation for ultrastructural observation

The method described by Zhang et al. (2004) was used for the ultrastructural observations. Briefly, the glutaraldehyde and OsO4 double-fixed tissues (Fig. 1b for the sampling site) were embedded with Spurr for preparation of ultrathin sections. These sections were stained with lead citrate and uranyl acetate, and viewed with a JEM-100S transmission electron microscope (Jeol Ltd., Tokyo, Japan).

Measurement of plasmodesmal density

The method for measuring plasmodesmal density was adapted from Kempers, Ammerlaan & van Bel (1998). Five group series of transverse ultrathin sections were prepared from the Spurr-infiltrated flesh samples, in which each group was cut at a distance of approximately 20 µm from the previous one. From each group, six pieces of ultrathin sections were picked at random and put on the copper grids of 100-mesh. Five scopes (each consisting of phloem and its surrounding parenchyma cells) were observed from each ultrathin section. Plasmodesmata were counted at all cell interfaces, that is, the interfaces between SE/CC, SE/phloem parenchyma cell (PP), CC/PP, PP/PP, PP/flesh parenchyma cell (FP) and FP/FP in each selected field. In the case of branched plasmodesmata, each branch was counted as one plasmodesma. Where branching was unequal in a complex, the numbers of branches on the two sides were averaged. The results of the plasmodesmal counting were given as the number of plasmodesmata per micron of specific cell/cell interface length on transversal section, which is referred to as plasmodesmal density (no. plasmodesmata µm−1).

Enzyme assay and immunoblotting

Soluble acid invertase (SAI) and cell wall acid invertase (CWI) extracts of the mesocarp tissues enclosing the phloem (located in the vascular bundle) (Fig. 1b) were prepared essentially as described previously (Pan et al. 2005a). Acid invertase activities were assayed in the soluble and insoluble fractions according to Schaffer et al. (1987).

Sodium dodecyl sulfate polyacrylamide gel electrophoresis (SDS-PAGE) and immunoblotting of the fractions of SAI and CWI were carried out according to Zhang et al. (2001) with modifications. Briefly, after electrophoretic transfer from the polyacrylamide gels, the polyvinylidene fluoride membranes were blocked overnight at 4 °C and incubated for 2 h at 37 °C in the antiserum directed against SAI of the apple fruit (Pan et al. 2005b). Following extensive washes, the membranes were incubated with goat anti-rabbit IgG-alkaline phosphatase conjugate (Sigma-Aldrich). The membranes were stained with 5-bromo-4-chloro-3-indolyl phosphate and nitro blue tetrazolium.

Protein concentrations were determined by the method of Bradford (1976) with bovine serum albumin as a standard.

Immunogold labelling

Specimen preparation and immunogold labelling were conducted essentially according to Zhang et al. (2001). Briefly, the ultrathin sections prepared as described above for ultrastructural observation from Spurr-infiltrated fruit tissues in the vascular bundle zones (Fig. 1b) were initially incubated with rabbit antiserum directed against SAI of the apple fruit, and then with secondary antibody (goat anti-rabbit IgG antibody conjugated with 10 nm gold, Sigma-Aldrich). The sections were finally double-stained with uranyl acetate-lead citrate and examined with a JEM-100S electron microscope.

Studies of [14C]sugar uptake

Columns containing vascular bundle zone (Fig. 1b) or just common flesh parenchyma tissues were sampled from freshly harvested fruit at stage IV using a cork borer, 5.0 mm in diameter. Each column was cut transversely into discs, 0.1 cm thick. Cellular debris was removed by 3 × 1 min washes in ice-cold carrier solution (20 mm MES, pH 5.8, 0.5 mm CaCl2, 0.5 mm K2SO4, 175 mm mannitol, 198 mOsmol). Discs were preequilibrated in the carrier solution with or without inhibitors [3 mm p-chloromercuribenzenesulfonic acid (PCMBS), Toronto Research Chemicals, or 0.05 mm carbonylcyanide m-chlorophenylhydrazone (CCCP), Sigma-Aldrich)] for 20 min at 25 °C with gentle shaking. After removing excessive PCMBS and CCCP by 3 × 1 min washes in carrier solution, pretreated discs were transferred to the same carrier solution plus 1 mm[U−14C]sucrose or [U−14C]glucose (3.70 × 10−2 MBq mL−1) (PerkinElmer, for 15 min at 25 °C with gentle shaking. Fluxes of sucrose or glucose were estimated as previously described (Ruan & Patrick 1995).

Molecular cloning of a CsSUT4 cDNA

Total RNA was extracted from cucumber ovaries (about 2 d before anthesis) using Column Plant RNAout kit (Tiandz, according to the manufacturer's instructions, treated with DNAase and reverse transcribed (SuperScript II; Invitrogen) with an oligo(dT)15 primer. The synthesized single-stand cDNA was subjected to polymerase chain reaction (PCR) amplification. Two primers (forward 5′-ATGGTGATGCCGGAGTCGTCT-3′ and reverse 5′-TCATGTGAGGTTTCTGGGGTTC-3′) were designed based on the published consensus sequence of sucrose transporters and the complete genome sequence of cucumber (Huang et al. 2009). The full-length CsSUT4 cDNA sequences were amplified using the following cycling conditions: 95 °C, 5 min, followed by 30 cycles of 94 °C for 0.5 min; 57 °C for 0.5 min; 72 °C for 2 min. PCR products were cloned into pGEM-T Easy vector (Promega, and sequenced.

Quantitative real-time PCR analysis

The CsSUT4 gene expression was quantified using the iQTM SYBR® Green Supermix kit following manufacturer's instructions (Bio-Rad,, and performed in an iCycler iQTM Real-Time PCR Detection System (Bio-Rad). The first-strand cDNAs used for real-time PCR analysis were obtained from source leaves, sink leaves, stems and mesocarp tissues enclosing the phloem (located in the vascular bundle) (Fig. 1b) with the same method as CsSUT4 cloning. Gene specific primers (forward 5′- GGTTTCTGGGGTTCTGAGCA and reverse 5′- CTTAGCAATAGTTTTCCCACAGGTT) for CsSUT4 were used in real-time PCR analysis. The thermal cycling conditions used were 5 min at 95 °C and then 40 cycles consisting of 0.5 min at 94 °C, 0.5 min at 57 °C, and 0.5 min at 68 °C. The specificity of the individual PCR amplification was confirmed using a heat dissociation protocol from 79 °C to 90 °C following the final cycle of the PCR. All quantifications were normalized to actin cDNA fragments amplified in the same conditions by the following primers: forward 5′-CCACGAAACTACTTACAACTCCATC and reverse 5′-GGGCTGTGATTTCCTTGCTC. Moreover, a non-RT control was conducted to detect any possible DNA contamination in all real-time PCRs. Real-time PCR experiments were repeated three times, with the threshold cycle (CT) determined in triplicate. The average for the triplicate of one representative experiment was used in all subsequent analyses. The relative levels of CsSUT4 transcription were calculated by using the 2−▵▵CT method (Livak & Schmittgen 2001).

Transient expression of CsSUT4-green fluorescent protein (GFP) in onion epidermis cells

For observation of the subcellular localization of CsSUT4, the open reading frame of its cDNA was PCR amplified by using primers 5′-CCGCTCGAGATGGTGATGCCGGAGTCGTCTGAAGGT-3′ (forward primer) and 5′-CGCGGATCCCGTGTGAGGTTTCTGGGGTTCTGAGCA-3′ (reverse primer). The PCR product was then fused to the upstream of the enhanced GFP (Cormack, Valdivia & Falkow 1996) at the XhoI (5′ end)/BamHI (3′ end) sites in pEZS-NL vector ( The strips of onion bulb epidermis were bombarded with gold particles coated with plasmids using a Bio-Rad PDS-1000/He particle delivery system (Scott et al. 1999). Fluorescence of GFP was observed by a CLSM after incubation at 22 °C in dark for 24 h.


CF is confined to functional phloem strands

An intact cucumber fruit is a typical pseudocarpous fruit with its ectocarp, mesocarp and endocarp originating from the ovary and receptacle (Fig. 1a,b). Anatomical observations showed that the vascular bundles of cucumber fruit were bicollateral (Fig. 1b–d). Autoradiography of the cucumber fruit showed that the vascular bundles in the mesocarp were more labelled than surrounding parenchyma tissues (Fig. 1e,f). This demonstrated that the [14C]assimilates were unloaded from the vascular bundles, revealing that the phloem in these bundles was functional for unloading and the mesocarp, one of the main edible fleshy portions, is fed with photoassimilates through the vascular bundles embedded in itself (Fig. 1b). Therefore, the vascular bundles are the main sampling sits in our subsequent study (Fig. 1b).

When it is loaded into cells, the membrane-permeable and non-fluorescent CFDA is degraded into CF, a membrane-impermeable fluorescent dye. CF is often used as a fluorescent marker of phloem transport and symplastic phloem unloading (Roberts et al. 1997; Haupt et al. 2001), and its behaviour is similar to the pattern of photoassimilates unloading determined by autoradiography (Viola et al. 2001). Our preliminary experiments showed that CF supplied to the pedicel of the fruit reached the flesh 2–3 h after CF loading. Furthermore, the unloading route of CF remained the same afterwards, even when the samples were collected 48 h after CF application (data not shown). CLSM images of CF movement in the fruits sampled 24 h after CF supply to the pedicel were presented in Fig. 2 and Supporting Information Fig. S2. A series of experiments conducted at different stages showed that CF was always confined to the phloem strands along the phloem pathway in the vascular bundles without apparent diffusion to the surrounding tissues at stage I (Fig. 2a,b,a′,b′), II (Fig. 2c–f,c′,e′), III (Fig. 2g,h,g′,h′) and IV (Fig. 2i–k, i′–k′). A survey of CF unloading by serial sections along the whole fruit confirmed the above observations. CF was restricted to the phloem strands along the vascular bundles throughout the fruit from the basal (pedicel side) to the apical region (Supporting Information Fig. S2).

Figure 2.

Confocal laser scanning microscopy (CLSM) imaging of carboxyffuorescein (CF) unloading during cucumber fruit development. CF reached the fruit flesh phloem 2–3 h after loading at the pedicel. The treated fruit was sampled 24 h after CF loading, and the hand-sections were prepared from vascular bundle zones in the central area (indicated in Fig. 1a,b) of the developing fruit. All the fluorescence images showed that CF was restricted to the phloem strands along the phloem pathway in the vascular bundle without apparent diffusion into the surrounding tissues. The red areas in (d) and (f) indicate the xylem zone labelled by Texas red dextran. (a, b, a′, b′) Stage I. (c–f, c′, e′) Stage II. (g, h, g′, h′) Stage III. (i–k, i′–k′) Stage IV. (a–k) Fluorescence images. (a′–c′, e′, g′–k′) Bright-field images. (a, c, d, g, i, a′, c′, g′, i′) Transverse sections. (b, e, f, h, j, k, b′, e′, h′, j′, k′) Longitudinal sections. EP, external phloem; IP, internal phloem; X, xylem. Bars = 100 µm.

In order to distinguish the phloem and xylem within the fruit, we successfully introduced 3 kDa Texas Red dextran, a widely used xylem vessel tracer (Viola et al. 2001; Zhang et al. 2006), into cucumber fruit. Texas Red fluorescence was seen in the xylem zone, and was not superposed with CF fluorescence (Fig. 2d,f). This indicated that CF was really delivered by phloem, and not by xylem. Thus, the method of introducing CF into cucumber fruit was effective, and the results of CF unloading were reliable.

Structurally, the SE–CC complex is symplasmically restricted with surrounding parenchyma cells

To give a more reliable assessment of the plasmodesmal connection between SE–CC complexes and their surrounding cells, the plasmodesmal densities (Kempers et al. 1998) at the interfaces between these cells were counted at selected stages (Fig. 3, Supporting Information Fig. S1, Table 1). The results showed that the CC had an electron-dense cytoplasm and enriched with mitochondria (Fig. 3a), which indicated active physiological function in the cells. Plasmodesmata were rarely observed between the SE–CC complex and its surrounding PPs (Table 1), which resulted in a symplasmic restriction of SE–CC complexes. This symplasmic restriction from the surrounding PPs remained unchanged throughout the selected developmental process (Table 1). However, a number of plasmodesmata that branched at the CC side and converged at the SE side into a single pore were found between SE and CC (Fig. 3b, Table 1). Plasmodesmata were also numerous between parenchyma cells (including PPs and FPs) (Fig. 3c–f, Table 1). There were more plasmodesmata at the interface between adjacent PPs than between PP and FP or between adjacent FPs (Fig. 3c–f, Table I). This held true throughout the selected developmental stages (Table 1). Furthermore, we observed that at stage IV a small portion (about 5%) of plasmodesmata channels are branched at the interfaces between two PPs and also between the adjacent FPs (Fig. 3g,h), and some plasmodesmata channels (approximately 20–30%; Table 1) were apparently blocked by electron-opaque material (Fig. 3i,j, Table 1). These phenomena were not observed at the earlier stages.

Figure 3.

The ultrastructure of the sieve element-companion cell (SE–CC) complex and its surrounding parenchyma cells in developing cucumber fruit. All sections were cut transversely. (a, b) The samples were picked at stage II. (a) The phloem of the vascular bundle, showing the structure of SE–CC complex and its surrounding PPs. Dashed arrows indicate mitochondria. (b) An amplification of a transverse section of the SE–CC complex showing the existence of plasmodesmata that branched at the CC side and converged at the SE side into a single pore. (c–f) The samples were picked at stage III. Note numerous single pore plasmodesmata at the interface between PP and FP (c), between FP and FP (d), and between PP and PP (e, f). (g–j) The samples were picked at stage IV. (g, h) Branched plasmodesmata (indicated by white stars) at the interface between PP and PP (g), and between FP and FP (h). (i, j) Electron-opaque material-blocked plasmodesmata (indicated by white triangles) between SE and PP (i), between CC and PP (i), and between FP and FP (j). CC, companion cell; Ch, chloroplast; FP, flesh parenchyma cell; G, golgi apparatus; M, mitochondrion; N, nucleus; Pd, plasmodesmata; PP, phloem parenchyma cell; RER, rough endoplasmic reticulum; SE, sieve element; V, vacuole; Ve, vesicle. Bars = 1 µm.

Table 1.  Plasmodesmal densities between different cells in the phloem of developing cucumber fruit
Developmental stagesSE/CCSE/PPCC/PPPP/PPPP/FPFP/FP
  1. Unit of plasmodesmal densities, number of plasmodesmata µm−1. Each value is the mean ± standard deviation of 30 replicates. Values in parentheses represent the density of plasmodesmata blocked by electron-opaque material.

  2. CC, companion cell; FP, flesh parenchyma cell; PP, phloem parenchyma cell; SE, sieve element.

I1.55 ± 0.510.08 ± 0.030.12 ± 0.061.72 ± 0.931.13 ± 0.520.75 ± 0.41
II1.63 ± 0.780.06 ± 0.020.08 ± 0.041.78 ± 0.691.23 ± 0.720.68 ± 0.38
III1.67 ± 0.820.04 ± 0.020.05 ± 0.022.13 ± 0.911.45 ± 0.680.73 ± 0.53
IV1.59 (0.44) ± 0.670.03 (0.01) ± 0.010.04 (0.01) ± 0.021.62 (0.48) ± 0.541.34 (0.41) ± 0.350.62 (0.25) ± 0.47

Acid invertase is localized in cell walls of SE–CC complex and parenchyma cells

Stachyose, raffinose and sucrose accounted for 45.84, 11.27 and 29.42% of the translocated sugars in cucumber phloem sap, respectively (Hu et al. 2009). Clearly, besides stachyose and raffinose, sucrose is another main soluble sugar that is unloaded from the phloem of cucumber fruit. It is well accepted that sucrose unloading is closely related with acid invertase (β-fructosidase, EC, which converts sucrose into glucose and fructose, and is present in the cell wall in an insoluble form or in the cytoplasm and vacuole in a soluble form (Quick & Schaffer 1996; Patrick 1997). Figure 4a showed that, whatever the developmental stage, the activity of CWI was significantly higher than that of SAI. In addition, the activities of SAI and CWI both increased from stage I to III when the maximal level was achieved, and decreased at stage IV (Fig. 4a). Furthermore, anti-apple SAI antibodies (Pan et al. 2005b) were used to analyse the amounts and subcellular localization of acid invertase in cucumber fruit. The anti-SAI serum recognized a single polypeptide with a molecular weight of approximately 3-kDa in the fractions of both SAI and CWI extracted from the mesocarp tissues enclosing the phloem (Fig. 4b). The immunoblot-estimated amounts of both the SAI and CWI changed substantially in parallel with their activities (Fig. 4a,b). The immunogold labelling assay showed clearly that acid invertases (visualized by immunogold particles) were localized predominantly in the cell walls of SE–CC complexes and their surrounding parenchyma cells (including PPs and FPs), and only little signals were found in the cytoplasm and vacuoles of CCs, PPs and FPs, or in the lumens of SEs (Fig. 5a–e). Importantly, no gold particles were substantially found in any of the controls without the antiserum or with the pre-immune serum controls (Fig. 5f,g), indicating that the antiserum was specific and the unspecific labelling was negligible. Additionally, SAI and CWI exhibit a high degree of antigenic homology, which allows for the interchange of antibodies from one to the other (Quick & Schaffer 1996; Zhang et al. 2001). Therefore, the anti-apple SAI antibodies should recognize a similar specificity of CWI and SAI, and the results of acid invertse immunoblotting and immunogold localization in the present study were reliable.

Figure 4.

Changes in activities and amounts of acid invertase in cucumber fruit. (a) Activities of SAI and CWI in developing cucumber fruit. Each value is the mean ± standard deviation of three replicates. (b) Immunoblotting with apple SAI polyantibodies. Equal amounts (30 µg) of protein were loaded. A 37 kDa polypeptide was detected in the soluble and cell wall extracts of the mesocarp tissues enclosing the phloem.

Figure 5.

Immunogold localization of acid invertase in the flesh of cucumber fruit. (a–e) Immunogold localization of acid invertase in cucumber fruit at stage I. The protein was reacted with anti-apple soluble acid invertase (SAI) serum. The immunogold particles (invertase) are intensively localized in the cell walls between SE and PPs (a), between SE and CC (b), between CC and PP (c), between PP and PP (d), and between PP and FP (e). (f, g) Controls without antiserum (f) or with preimmune serum (g). No substantial signal was detected. CC, companion cell; FP, flesh parenchyma cell; PP, phloem parenchyma cell; SE, sieve element; V, vacuole; W, cell wall. Bars = 1 µm.

[U−14C]Sucrose and [U−14C]glucose uptake are inhibited by PCMBS and CCCP

Table 2 showed that PCMBS, a sulfhydryl modifier (Riesmeier, Willmitzer & Frommer 1992; Sauer & Stolz 1994), inhibited [U−14C]sucrose uptake into the vascular-bundle-enriched and common flesh-parenchyma-enriched discs by 81% and 69%, respectively. Similarly, [U−14C]glucose uptake was also largely reduced by PCMBS (Table 2). Additionally, [U−14C]sucrose and [U−14C]glucose uptake into the vascular-bundle-enriched and common flesh-parenchyma-enriched discs were largely reduced by CCCP, a general uncoupler of trans-membrane proton gradients (Riesmeier et al. 1992; Sauer & Stolz 1994). These data suggested that an energy-dependent trans-membrane step is involved in sugar transport in cucumber fruits.

Table 2.  Effect of p-chloromercuribenzenesulfonic acid (PCMBS) and carbonylcyanide m-chlorophenylhydrazone (CCCP) on [U−14C]sucrose and [U−14C]glucose uptake by discs of cucumber fruit at stage IV
Tissue typeTreatmentUptake rate [µmol (g DW)−1 h−1]a
  • a

    Concentrations of sucrose, glucose, PCMBS and CCCP were 1.0, 1.0, 3 and 0.05 mm, respectively. Values in parentheses are percentages of the control. Each value is the mean ± standard deviation of three replicates.

  • DW, dry weight; FP, common flesh-parenchyma-enriched discs; V, vascular bundle-enriched discs (indicated in Fig. 1b).

VControl6.70 ± 0.17 (100)5.83 ± 0.33 (100)
PCMBS1.27 ± 0.11 (19)1.79 ± 0.18 (31)
CCCP2.18 ± 0.22 (32)3.25 ± 0.07 (56)
FPControl5.23 ± 0.33 (100)6.01 ± 0.21 (100)
PCMBS1.56 ± 0.13 (30)1.52 ± 0.14 (25)
CCCP2.55 ± 0.19 (49)2.60 ± 0.28 (43)

A sucrose transporter, expressed mainly in mesocarp tissues enclosing the phloem, is localized to plasma membrane

Further experiments were conducted to determine the presence of sugar transporters and whether they are involved in trans-membrane transport in cucumber fruit. To achieve this, a full-length (1512 bp) cDNA encoding a sucrose transporter (CsSUT4) (GenBank accession number GQ903726) was cloned from cucumber ovaries using reverse transcription-PCR (Fig. 6a). The spatiotemporal expression analysis showed that transcript levels of CsSUT4 were much higher in mesocarp tissues enclosing the phloem than that in source leaves, sink leaves and stems (Fig. 6b). In mesocarp tissues enclosing the phloem, the transcript level of CsSUT4 increased from stage I to II when the highest expression was achieved, followed by a rapid decline to the lowest expression at stage III, and then returned quickly at stage IV (Fig. 6b). The localization of CsSUT4 was determined by expressing CsSUT4:GFP fusion protein in chloroplast-free epidermis cells of onion (Allium cepa) (Fig. 6c–f). Non-targeted GFP was found to be distributed in both the cytoplasm and nucleus of the bombarded epidermal onion cells (Fig. 6c, d). In contrast, CsSUT4:GFP was localized exclusively to the periphery of bombarded cells (Fig. 6e, f). Careful analysis of serial optical sections through a number of CsSUT4:GFP-expressing cells clearly showed that fluorescence was only ever associated with the cell periphery and was not observed around the inner edge (tonoplast side) of the nucleus in these cells (Fig. 6e, f). These results indicated that CsSUT4 is not associated with the tonoplast membrane and suggested that it is targeted to the plasma membrane in plants. Therefore, CsSUT4 may play an important role in the sucrose transport across the plasma membrane in cucumber fruit, which is consistent with the results of [U−14C]sucrose uptake assay.

Figure 6.

Cloning, expression and subcellular localization of CsSUT4. (a) PCR amplification of full-length CsSUT4. Lane 1, DNA marker; Lanes 2 and 3, PCR product of full-length CsSUT4. (b) Quantitative real-time PCR analysis of CsSUT4 expression level in different organs of cucumber plant. Each value is the mean ± standard deviation of three replicates. (c–f) Subcellular localization of CsSUT4 in onion epidermal cells. (c, d) Non-targeted GFP was found to be distributed in both the cytoplasm and nucleus. (e, f) CsSUT4:GFP fusion protein is present in the periphery of bombarded cells. Arrows indicate the nucleus. (c, e) Fluorescence images. (d, f) Overlay of fluorescence and bright-field images. Bars = 50 µm.


Phloem unloading follows an apoplasmic pathway in cucumber fruit

The present cytological studies reveal that the SE–CC complexes in the phloem of vascular bundles that feed the fruit flesh (mesocarp) with assimilates are symplasmically restricted. This was shown by the in vivo functional investigation of CF movement by CLSM. It was clearly showed that the CF, which could only be transported via plasmodesmata and not transmembranously (Nie et al. 2010), is confined strictly to the phloem strands in the fruit fleshy mesocarp (Fig. 2, Supporting Information Fig. S2). Furthermore, there was almost no plasmodesma at the interface between the SE–CC complexes and their surrounding PPs (Table 1). Autoradiograph (Fig. 1e,f) indicated that the phloem strands are functional for assimilate unloading. These data provide a clear evidence for an extensive apoplasmic phloem unloading pathway in the whole fruit of cucumber throughout the selected developmental process (Figs 1–3, Supporting Information Figs S1 & S2, Table 1).

The present study proved the predominant occurrence of CWI in cucumber flesh tissue by biochemical analysis of enzyme activity, immunoblotting and immunogold localization (Figs 4 & 5). It is believed that the continuous hydrolysis of sucrose by acid invertase at the apoplasmic unloading site increases the steepness of the sucrose gradient, leading to a faster unloading of sucrose from the SE–CC complex, and so the presence of acid invertase in cell wall favors apoplasmic sucrose unloading (Quick & Schaffer 1996; Patrick 1997; Wu et al. 2004; Zhang et al. 2006). The sucrose unloaded to the apoplasm from the SE–CC complex may more or less be reloaded into the parenchyma cells of the cucumber fruit at the vascular interface or continue to move apoplasmically to each sink parenchyma cell (Wu et al. 2004). The acid invertase in the cell walls of parenchyma cells (Fig. 5d, e) provides the possible mechanism to degrade these two types of sucrose. In contrast to apoplasmic unloading, the symplasmic unloading seems to depend more on SAI than on CWI (Zhang et al. 2006; Nie et al. 2010). Another indirect approach to distinguish between apoplastic and symplastic unloading is to analyse PCMBS sensitivity (Ruan & Patrick 1995; Ruan, Llewellyn & Furbank 2001; Zhang et al. 2004). PCMBS blocks carrier-mediated uptake of sugars from the apoplast (Ruan & Patrick 1995; Ruan et al. 2001; Zhang et al. 2004), but does not inhibit symplastic transport (Turgeon 1996). Our study found that PCMBS largely inhibited [U−14C]sucrose and [U−14C]glucose uptake into fruit discs (Table 2). These data supported the apoplasmic phloem unloading pathway in cucumber. These data supported the apoplasmic phloem unloading pathway in cucumber. In addition, post-phloem transport may occur by simultaneous apo- and symplasmic pathways, as indicated by the acid invertase in the cell walls of parenchyma cells (Fig. 5d, e), the numerous plasmodesmata connecting these cells (Fig. 3, Table 1), and large inhibition of [14C]sugar uptake into fruit discs by PCMBS and CCCP (Table 2).

It must be pointed out that in the study we focused on the main vascular bundles (Fig. 1b) of which the external and internal phloem belong to fascicular phloem (Crafts 1932; Turgeon & Oparka 2010). Besides, there still are some smaller vascular bundles of which phloem is called extrafascicular phloem (Crafts 1932; Turgeon & Oparka 2010) distributed as scattered elements throughout the mesocarp (Supporting Information Fig. S3a). Recently, Zhang et al. (2010) have found that the fascicular phloem is largely responsible for sugar transport, whereas the extrafascicular phloem may function in signalling, defence and transport of other metabolites. In this study, we found that both fascicular and extrafascicular phloem can transport CF, and that CF remains largely contained within the two kinds of phloem (Supporting Information Fig. S3), which suggested that the metabolites may be also apoplasmically unloaded from the extrafascicular phloem. Furthermore, the CF fluorescence in external and internal phloem indeed distributed in the similar pattern (being restricted), although in some pictures CF fluorescence was just found in either external or internal phloem (not in both) for the reason that the two phloem on a hand-section were not in the same focal plane (Fig. 2, Supporting Information Fig. S2). It is suggested that the photoassimilates may be unloaded in the same way in external and internal phloem. Accordingly, we randomly sampled the external or internal phloem in the subsequent experiments of ultrastructural observation and acid invertase immunogold labelling, and sampled the whole main vascular bundles in other experiments.

Possible mechanism of apoplasmic unloading in fruits

To recapitulate, symplasmic unloading is the common path because of its lower resistance and greater transport capacity compared with apoplasmic unloading (Patrick 1997). However, several studies revealed the interchanging pattern between apo- and symplastic phloem unloading pathways as fruit development processed (Ruan & Patrick 1995; Zhang et al. 2006; Nie et al. 2010), and even an extensive apoplasmic pathway during fruit development (Zhang et al. 2004). This leads to a consideration why fruits choose a more difficult trans-membrane apoplasmic pathway for the mode of phloem unloading during fruit development, and how sugars transport across the membranes.

In fruits, sugar content is a major determinant of yield and quality. As a matter of fact, some studies have suggested that if sugar content is the lowest in pedicel phloem, higher in the apoplasmic space and the highest in fruit cytosol, the phloem unloading pathway is apoplasmic, otherwise it is symplasmic (Yamaki & Ino 1992; Beruter, Feusi & Ruedi 1997; Zhang et al. 2004, 2006). It is a widely accepted fact that this concentration gradient from the phloem to the flesh cells could lead to reversal of turgor gradients between SEs and the sink cells. A reversed turgor gradient with an open symplasmic pathway would counter unloading by diffusion and therefore reduces the turgor gradient between the import phloem in source leaves and the terminal release phloem in the sink fruits, and hence breaks long-distance phloem transport from source leaves to the developing fruits (Patrick 1997). The negative effects of a reversed turgor gradient would be compensated by symplasmically isolating SEs from sink cells with unloading following an apoplasmic route (Patrick 1997). In addition, hydraulic resistance resulting from xylem discontinuity contributes to prevent apoplasmic sugar flow from moving out of the fruits (Choat et al. 2009). This anatomical barrier to solute movement has been found in grape (Choat et al. 2009) and tomato (Lee 1989). Similar anatomical barrier may be also present in cucumber.

In fruits, assimilates unloading into the apoplast from the SE–CC complexes is a transporter-mediated active process coupled with protons returning down the electrochemical gradient generated by a plasma membrane H+-ATPase (Patrick 1997; Sondergaard, Schulz & Palmgren 2004). In apple fruits, abundant H+-ATPase and a 52-kDa putative monosaccharide transporter are found in the plasma membrane of SE–CC complexes, consisting with the extensive apoplasmic unloading pathway (Zhang et al. 2004). In grape berries, the expression of two grape sucrose transporters (VvSUC11 and VvSUC12) and three hexose transporters (VvHT1, VvHT2 and VvHT3) increased at the onset of ripening and remained until fruit maturity (Davies, Wolf & Robinson 1999; Fillion et al. 1999; Hayes, Davies & Dry 2007). Plasma membrane H+-ATPase, which drives sugar transporters by generating the proton motive force across plasma membranes (Sondergaard et al. 2004), was shown to be activated by an abscisic acid-stimulated protein kinase that was up-regulated from the onset of grape berry ripening (Yu et al. 2006). These data supported the occurrence of an enhanced capability of trans-membrane sugar transport from the onset of ripening, and are consistent with the developmental switch in unloading mechanisms in grape berry (Zhang et al. 2006). In the present study, [U−14C]sucrose and [U−14C]glucose uptake into the discs were largely reduced by PCMBS and CCCP (Table 1). Furthermore, a plasma membrane sucrose transporter (CsSUT4), highly expressed in the mesocarp tissues enclosing the phloem, was identified in cucumber fruits (Fig. 6). These results potentially provide the machinery to mediate the trans-membrane transport pathway of soluble sugars.

Hypothetical phloem loading and unloading strategies for soluble sugars in cucumber

Based on the results obtained, an integrated model of soluble sugars phloem loading and unloading in cucumber plant is proposed (Fig. 7). According to the polymer trapping hypothesis in cucurbitaceous species (Rennie & Turgeon 2009), stachyose, raffinose and sucrose are symplasmically loaded into the phloem of source leaves and move towards fruits by hydrostatic pressure (Fig. 7a). In the case of the symplasmic restriction of SE–CC complex with surrounding parenchyma cells in the phloem of cucumber fruit (Fig. 2, Supporting Information Fig. S2, Table 1), the first step of exporting soluble sugars from SE–CC complex must be trans-membrane via specific transporters (Fig. 7b). The primary sugar accumulated in the mesocarp tissues of cucumber fruit is hexose, and stachyose is found only in negligible quantities (Handley, Pharr & Mcfeeters 1983; Hu et al. 2009). These data indicate that stachyose may be quickly broken down to sucrose by α-galactosidase (EC (Smart & Pharr 1980), and the resulting sucrose could be further hydrolysed to hexose after reaching the vascular bundles of fruits. However, it still remains poorly understood how these happen, and we are doing further experiments to probe these questions. Besides hexose, sucrose levels in the mesocarp tissue of cucumber fruit cannot be ignored (Handley et al. 1983; Hu et al. 2009). In addition, the acid invertase is also located in the cell walls of parenchyma cells (Fig. 5d,e). These data suggest that most of the sucrose present in the apoplasmic space may be converted into hexose by CWI before being taken into PP via a putative hexose transporter (Fig. 7b), and a few sucrose molecules may be transported into PP without hydrolysis (Fig. 7b).

Figure 7.

An integrated model of soluble sugars phloem loading and unloading in cucumber plant. (a) Phloem loading in cucumber source leaves (Rennie & Turgeon 2009). (b) Phloem unloading in cucumber fruits. CC, companion cell; CWI, cell wall acid invertase; MC, mesophyll cell; PP, phloem parenchyma cell; R, raffinose synthase (EC; S, stachyose synthase (EC; SE, sieve element.

According to the model, putative transporters of hexose, sucrose, raffinose and stachyose are likely involved in phloem unloading in cucumber fruit (Fig. 7b). A plasma membrane sucrose transporter (CsSUT4) (Fig. 6), as well as a hexose transporter (unpublished data in our laboratory), has been identified in cucumber fruits. However, neither a putative raffinose transporter nor stachyose transporter has been identified at the molecular level to date in any of the higher plants. Our results validated proceeding with the next step, to systematically characterize the putative phloem raffinose or stachyose transporter at the biochemical level and to search for putative raffinose or stachyose transporter genes in cucumber plant using the complete genome sequence released recently (Huang et al. 2009). Once candidate sugar transporter genes involved in phloem unloading in cucumber fruit are identified, a reverse genetic approach will facilitate the determination of the importance of these genes, which will allow the soluble sugars unloading mechanism to be fully characterized.

Whatever the exact mechanism of phloem unloading is determined to be in cucumber fruit, the data in our study indicate that the unloading pathway of soluble sugars from SE–CC complex may be predominantly apoplasmic in cucumber fruit from anthesis to the marketable maturing stage, and the unloaded sucrose may be hydrolysed by functional acid invertase localized in the cell walls. The next challenge is to determine how environmental such as cold or low light influences phloem unloading, so infecting the output and quality of cucumber fruit.


We thank Dr. Dapeng Zhang (Tsinghua University, Beijing, China) for kindly providing the antiserum against apple SAI, and Dr Junping Gao (Agricultural University, Beijing, China) for his generous gift of pEZS-NL vector. This project was supported by the National Natural Science Foundation of China (30972004), and the earmarked fund for Modern Agro-industry Technology Research System (Nycytx-35-gw22).