Down-Regulating CsHT1, a Cucumber Pollen-Specific Hexose Transporter, Inhibits Pollen Germination, Tube Growth, and Seed Development^,

A hexose transporter affects seed number and seed size in cucumber by controlling pollen tube growth.

Abstract

Efficient sugar transport is needed to support the high metabolic activity of pollen tubes as they grow through the pistil. Failure of transport results in male sterility. Although sucrose transporters have been shown to play a role in pollen tube development, the role of hexoses and hexose transporters is not as well established. The pollen of some species can grow in vitro on hexose as well as on sucrose, but knockouts of individual hexose transporters have not been shown to impair fertilization, possibly due to transporter redundancy. Here, the functions of CsHT1, a hexose transporter from cucumber (Cucumis sativus), are studied using a combination of heterologous expression in yeast (Saccharomyces cerevisiae), histochemical and immunohistochemical localization, and reverse genetics. The results indicate that CsHT1 is a plasma membrane-localized hexose transporter with high affinity for glucose, exclusively transcribed in pollen development and expressed both at the levels of transcription and translation during pollen grain germination and pollen tube growth. Overexpression of CsHT1 in cucumber pollen results in a higher pollen germination ratio and longer pollen tube growth than wild-type pollen in glucose- or galactose-containing medium. By contrast, antisense suppression of CsHT1 leads to inhibition of pollen germination and pollen tube elongation in the same medium and results in a decrease of seed number per fruit and seed size when antisense transgenic pollen is used to fertilize wild-type or transgenic cucumber plants. The important role of CsHT1 in pollen germination, pollen tube growth, and seed development is discussed.

In the early stages of pollen development, large amounts of sugar are constantly supplied to anthers for pollen development. Anthers are regarded as having the greatest sink strength in floral development due to their high metabolic activity (Goetz et al., 2001; Zhang et al., 2010; Slewinski, 2011). After germination of mature pollen grains, the pollen tube must grow rapidly because the distance from the stigma to the embryo sac is often thousands of times the diameter of the pollen grain (Cheung, 1996). The extraordinary speed of pollen tube growth and the extreme length of this single expanding cell require that large amounts of carbohydrate be imported from the pistil for energy consumption and de novo cell wall biosynthesis (Garrido et al., 2006). Any disturbance of either carbohydrate supply or transport capability can significantly impair pollen development or pollen tube growth and cause male sterility (Goetz et al., 2001; Datta et al., 2002; Oliver et al., 2005, 2007b; Mamun et al., 2006; Ji et al., 2010).

Because the pollen grain and pollen tube are both symplastically isolated (Lemoine et al., 1999; Schneidereit et al., 2003, 2005), sugar must be delivered to them through the apoplast. As the main transport form of photoassimilates, Suc can theoretically be loaded into the pollen grain and pollen tube directly by Suc transporters (SUTs) or after cleavage into hexoses by hexose transporters (Schneidereit et al., 2003, 2005).

To date, there have been a number of studies conducted on the roles of SUTs on pollen tube germination and growth (Ylstra et al., 1998; Schneidereit et al., 2003, 2005; Scholz-Starke et al., 2003; Sivitz et al., 2008; Hirose et al., 2010). Several SUTs have been detected in pollen or pollen tubes (Lemoine et al., 1999; Stadler et al., 1999; Hackel et al., 2006; Sivitz et al., 2008; Hirose et al., 2010). In tomato (Solanum lycopersicum), antisense suppression of LeSUT2 reduces germination rate as well as pollen tube length, leading to less efficient pollination and small fruit size (Hackel et al., 2006). Also, disruption of AtSUC1 in Arabidopsis (Arabidopsis thaliana; Sivitz et al., 2008) and OsSUT1 in rice (Oryza sativa; Hirose et al., 2010) impairs pollen germination.

The role of hexose transporters in pollen development is not as well understood. This is unfortunate because, as Büttner and Sauer (2000) point out, with so many identified monosaccharide transporters in plants, their localization and knowledge of their individual roles is required to understand many poorly understood physiological processes, including pollen development.

The literature on hexoses and hexose transporters in pollen development is complex and species specific. The pollen of Japanese pear (Pyrus serotina) germinates more readily on Glc than Suc, while germination is inhibited by Fru (Okusaka and Hiratsuka, 2009). By contrast, pollen tubes of petunia (Petunia hybrida) grow equally well on medium containing Glc, Fru, or Suc (Ylstra et al., 1998). Although several hexose transporter genes are highly expressed in pollen grains or pollen tubes of Arabidopsis (Büttner, 2010) and rice (Oliver et al., 2007a; Wang et al., 2007, 2008), single knockouts of these genes do not result in a phenotype, and in vitro experiments demonstrate that the germination of Arabidopsis pollen seems to be strictly dependent on Suc in the medium. Addition of Glc causes instant bursting of Arabidopsis pollen tubes (Schneidereit et al., 2003, 2005; Scholz-Starke et al., 2003). Similarly, there is no phenotype associated with an insertion mutant of Petunia Monosaccharide Transporter1 (PMT1), a putative pollen-specific hexose transporter, of petunia (Ylstra et al., 1998; Garrido et al., 2006).

Due to these complex results, there is as yet no evidence that any specific hexose transporter is necessary for pollen tube growth. In this study, we used antisense suppression and overexpression to study the physiological properties of CsHT1, a pollen-specific hexose transporter of cucumber (Cucumis sativus) plants. The results demonstrate that CsHT1 is a hexose/H⁺ symporter with high affinity for Glc, that overexpression of CsHT1 in cucumber pollen promotes pollen tube growth on Glc or Gal medium, and that down-regulation inhibits growth on the same medium. Also, there are fewer seeds in the fruits when pollen from antisense transgenic plants is used to fertilize flowers of either transgenic or wild-type plants. Taken together, the results indicate that CsHT1 plays a leading role in pollen tube growth in the cucumber ovary/fruit in which the soluble sugar composition is mainly Glc and Fru. This is, to our knowledge, the first identification of a hexose transporter that is required for efficient pollen germination and pollen tube growth.

RESULTS

Isolation and Homology Analysis of CsHT1 cDNA

A 1,557-bp open reading frame (ORF) of a putative cucumber hexose transporter gene CsHT1 (GenBank accession no. HQ202746) was isolated by reverse transcription (RT)-PCR from total RNA of cucumber male flowers. The encoded CsHT1 protein has 519 amino acids, with a calculated molecular mass of 57.3 kD. Hydrophobicity profile analysis indicated that CsHT1 contains two sets of six putative transmembrane domains separated by a long central hydrophilic region with both the N- and C-terminal domains located on the intracellular side of the plasma membrane (Supplemental Fig. S1; Supplemental Text S1). This strongly suggests that CsHT1 is a transmembrane protein with 12 transmembrane domains, which is consistent with sugar transporters in microbes, mammals, and plants (Büttner and Sauer, 2000).

A phylogenetic tree analysis (Fig. 1; Supplemental Table S1) indicated that CsHT1 is highly homologous to plasma membrane-localized monosaccharide transporters from higher plants, especially to Arabidopsis monosaccharide transporter9 (AtSTP9), AtSTP10, AtSTP11, and Pmt1, which are pollen specific (Ylstra et al., 1998; Schneidereit et al., 2003, 2005). CsHT1 is clearly separated from the three plastid-localized monosaccharide transporters (Olea europea plastidic glucose translocator [OepGlcT], Mesembryanthemum crystallinum chloroplast glucose transporter [McpGlcT], and spinach [Spinacia oleracea] plastidic Glc translocator [SocTP1]) and the five tonoplast-localized monosaccharide transporters (AtTMT1, AtTMT2, AtTMT3, OsTMT1, and OsTMT2). Furthermore, multisequence alignment analysis indicated that the deduced amino acid sequence of CsHT1 shares a high degree of homology with pollen-specific hexose transporters AtSTP9, AtSTP11, and Pmt1 (Supplemental Fig. S2), and all of them have 12 hydrophobic transmembrane domains located at similar positions. These results indicate that CsHT1 is probably a plasma membrane-localized hexose transporter that is mainly expressed in pollen.

Figure 1.

Phylogenetic analysis of CsHT1. The first identified cucumber monosaccharide transporter CsHT1 shows high homology with plasma membrane monosaccharide transporters and is clearly separated from the three plastid localized monosaccharide transporters OepGlcT, McpGlcT, and SocTP1 and from the five tonoplast-localized monosaccharide transporters AtTMT1, AtTMT2, AtTMT3, OsTMT1, and OsTMT2. For the GenBank accession numbers of the sequences used for the analyses, see Supplemental Table S1. CkHUP3, Chlorella monosaccharide transporter3; MtST1, Medicago truncatula hexose transporter1; OsMST4, Oryza sativa monosaccharide transporter4; RCSTC, Ricinus communis sugar carrier protein; SspSGT2, sugar cane monosaccharide transporter2.

Functional Characterization of CsHT1 by Heterologous Expression in Yeast

Functional characterization of CsHT1 was performed by expressing the CsHT1 gene in the hexose transporter-deficient yeast strain EBY.VW4000, which cannot grow on monosaccharides but can grow on maltose (Schneidereit et al., 2003). A drop test revealed that both transformant lines of EBY.VW4000 with pDR196/CsHT1 and empty yeast expression vector (pDR196) can grow on maltose, but only the yeast with CsHT1 expression can grow well on 2% (w/v) Glc (Fig. 2A). This result demonstrates that CsHT1 encodes a functional Glc transporter. [¹⁴C]Glc uptake analysis revealed that yeast with pDR196/CsHT1 was able to accumulate [¹⁴C]Glc rapidly, while yeast with the empty vector accumulated a negligible amount of [¹⁴C]Glc (Fig. 2B). The optimum pH of [¹⁴C]Glc uptake by CsHT1 is approximately 5.5 (Fig. 2C). Kinetic analysis of [¹⁴C]Glc uptake by EBY.VW4000 expressing CsHT1 revealed a K_m for Glc of 107.3 μm at pH 5.5 (Fig. 2D). To determine the substrate specificities of CsHT1, transport of [¹⁴C]Glc was studied in the presence of other sugars supplied at 10 times higher concentrations. The results revealed that the uptake of [¹⁴C]Glc in yeast was only inhibited strongly by Gal. Furthermore, low concentrations of the proton uncoupler carbonyl cyanide m-chlorophenylhydrazone (CCCP) significantly decreased Glc uptake (Fig. 2E), suggesting that sugar uptake via CsHT1 was driven by a proton gradient across the plasma membrane. The data from heterologous expression in yeast suggests that CsHT1 is an energy-dependent hexose/H⁺ symporter with high affinity for Glc and may also transport Gal but apparently with a lower affinity than for Glc.

Figure 2.

Functional and kinetic characterization analysis of CsHT1 in yeast (Saccharomyces cerevisiae). A, Growth recovery on Glc of the hexose transporter-deficient yeast cells by CsHT1 expression. B, Time course of [¹⁴C]Glc uptake by yeast mutant EBY.VW4000 transformed with pDR196/CsHT1 (black circle) or pDR196 alone (white circle). Carrier [¹⁴C]Glc concentration was 100 μm. C, pH dependence of [¹⁴C]Glc uptake by EBY.VW4000 transformed with pDR196/CsHT1. Carrier [¹⁴C]Glc concentration was 100 μm. Time interval for [¹⁴C]Glc uptake was 4 min. D, Concentration-dependent [¹⁴C]Glc uptake. Eadie-Hofstee transformation of the data used to estimate K_m. The estimated K_m is 107.3 ± 3.78 μm. V_max = 35.7 ± 0.53 pmol min^–1 mg^–1 cells. E, Substrate specificity and effects of metabolic inhibitors on the activity of the CsHT1 expressed in the EBY.VW4000. Substrate specificity was determined by competitive inhibition of [¹⁴C]Glc uptake. Carrier [¹⁴C]Glc concentration was 100 μm (without competitor as one internal control, indicated by Control). Competing sugars were added 30 s prior to the addition of labeled [¹⁴C]Glc at a concentration of 1 mm (1 mm Glc as another control, indicated by Glc). The final concentration of metabolic inhibitors was 50 μm. The time interval for [¹⁴C]Glc uptake was 4 min. The values are presented as units relative to the values from the internal control taken as 100%. Values in B to E are means ± se (n = 3). FW, Fresh weight; OD₆₂₃, optical density at 623 nm.

Transcript Level of CsHT1 Increases with Pollen Development, But Translation Does Not Start until Pollen Germinates

The spatial expression pattern of CsHT1 transcripts was analyzed using RT-PCR with 18S ribosomal RNA as a control. The results indicate that CsHT1 mRNA is transcribed only in male flowers (Fig. 3A). It is found at high levels in anthers, pollen grains, and pollen tubes but not in petals or sepals (Fig. 3B). Analysis based on RT-PCR (Fig. 3C) and RNA in situ hybridization (Supplemental Fig. S3; Supplemental Text S1) revealed that CsHT1 transcription increases gradually during the final stages of cucumber male flower development as the pollen matures (stages 9–12; Bai et al., 2004). Similar results were obtained in cucumber (Fig. 3, D and E) and Arabidopsis (Supplemental Fig. S4, B–E) transformed with the pCsHT1-GUS construct (Supplemental Fig. S4A).

Figure 3.

Expression analysis of the CsHT1 gene. A and B, Spatial expression of CsHT1 by RT-PCR. C to E, Temporal expression of CsHT1 in the male flower by RT-PCR (C) and histochemical analysis of pCsHT1:GUS plants (D and E) at different development stages of the male flower from 9 to 12. Inset of the picture of stage 12 in E shows the GUS expression is also in the germinated pollen and pollen tube. F to L, Immunolocalization of CsHT1 protein in pollen grains or pollen tubes of cucumber wild-type plants using affinity-purified anti-CsHT1 antiserum/IgG-alkaline phosphatase. F, No signal can be detected in the male flowers from stages 9 to 12 before the pollen germinated. G, I, and K, Strong signal can be detected in the germinating pollen and pollen tube in vitro (G) on the stigma (I) and in the ovule (K). H, J, and L, Same sections as in G, I, and K incubated with preimmune serum. The results in F to L represent at least five experimental replicates. R, Root; St, stem; L, leaf; mFl, male flower; fFl, female flower; Fr, fruit; C, carpopodium; Pe, petal; Se, sepal; An, anther; Po, pollen; PT, pollen tube. Bars = 4 mm (D), 100 μm (E), 20 μm (inset), 50 μm (G and H), and 100 μm (F and I–L).

To detect CsHT1 protein expression in pollen, an immunolocalization experiment was performed. As shown in Figure 3, F to L, and Supplemental Figure S5, CsHT1 protein expression was only detected in germinating pollen and in the growing pollen tube. Whether the pollen tube grew in vitro (Fig. 3G; Supplemental Fig. S5), inside the stigma (Fig. 3I), or deep in the ovule (Fig. 3K), the CsHT1 protein could be detected by immunolocalization. However, no CsHT1 protein was detectable in other organs (data not shown) or in ungerminated pollen from male flowers at developing stages 9 to 12 (later stages of pollen development; Fig. 3F), when CsHT1 was transcribed strongly (Fig. 3, C–E). These results indicate that pollen grains are preloaded with CsHT1 mRNA during maturation, but translation does not start until the pollen germinates.

CsHT1 Is a Plasma Membrane Protein

The subcellular location of CsHT1 was determined by expressing CsHT1-GFP in chloroplast-free epidermal cells of onion (Allium cepa) and cucumber protoplasts with chloroplasts, using the vector pCAMBIA 1302 containing GFP as a control. Transient expression of CsHT1-GFP demonstrated that the fusion protein was targeted to the plasma membrane of epidermal cells of onion (Fig. 4A) and cucumber protoplasts (Fig. 4B).

Figure 4.

Localization of CsHT1 in cell plasma membrane. Transient expression of GFP (as control) and CsHT1:GFP under the control of 35S promoter in onion epidermal cells (A) and cucumber protoplast (B). Laser-scanning confocal microscopy images show fluorescence (GFP-) and merged images. Chlorophyll autofluorescence (Auto-), and the bright-field images are also presented. Arrows indicate the nucleus. Bars = 50 μm (A) and 10 μm (B).

Sugar Specificity and Energy Dependence

Several sugars (Suc, Glc, Fru, Gal, and Man) were tested as energy sources for pollen tube growth in vitro. Effective sugars were Suc (the conventional carbon source), Glc, and Gal (Fig. 5A). The latter two are substrates of CsHT1 (Fig. 2E). However, the optimal concentration for pollen germination and tube growth varied with the type of sugar: Suc, Glc, and Gal optima are 5% to 20% (w/v), 5% to 15% (w/v), and 5% to 10% (w/v), respectively (Fig. 5A). Interestingly, no pollen germination or tube growth occurred in medium containing Fru or Man at any concentration (Fig. 5A). In addition, pollen germination was seriously inhibited by the proton uncoupler CCCP and apoplast transport inhibiter p-chloromercuribenzene sulfonate (PCMBS; Fig. 5B), indicating that pollen germination and tube growth need an energy-dependent H⁺ symporter.

Figure 5.

Effect of sugar composition and metabolic inhibitors on cucumber pollen germination in vitro. A, Light-microscopic photographs of wild-type cucumber pollen tubes grown in vitro with different carbon sources in different concentration after 4 h. B, The inhibition of pollen germination and tube growth by CCCP or PCMBS. Pollen germination 4 h after treatment of different concentrations of CCCP or PCMBS in the pollen germination medium with 15% (w/v) Suc as carbon source. Bars = 200 μm (A) and 500 μm (B).

CsHT1 Is Required for Pollen Germination and Tube Growth in Low Levels of Glc or Gal

To establish the function of CsHT1 in cucumber pollen development and germination, CsHT1 was expressed in sense and antisense directions under the control of the constitutive Cauliflower mosaic virus (CaMV) 35S promoter (Fig. 6A). The pollen of seven antisense lines and four overexpression lines from more than 20 independent transgenic plants were screened (Fig. 6B). Only two of 13 putative hexose transporter genes that demonstrated homology with CsHT1 in cucumber were expressed in the pollen of wild-type plants, with CsHT1, and neither of these was affected in CsHT1-antisense lines (Supplemental Fig. S6). No obvious differences were observed in pollen size, pollen number, pollen appearance (Fig. 6C), or pollen viability (Fig. 6D) in any of the transgenic lines compared with the wild type. However, when cultivated in 3% (w/v) Glc-containing medium or 3% (w/v) Gal-containing medium, compared with wild-type pollen, few pollen grains of CsHT1 antisense plants germinated and the tubes of those that did germinate grew slowly (Fig. 6, E–H). Meanwhile, the germination ratios of CsHT1 overexpression pollen in the same medium was higher than in the wild type, and the pollen tubes of CsHT1 overexpression lines grew more rapidly than the wild type (Fig. 6, E–H). These results indicate that CsHT1 plays an important role in cucumber pollen germination and pollen tube growth when provided with 3% (w/v) Glc or 3% (w/v) Gal.

Figure 6.

Characterization of pollen of CsHT1 sense and antisense transgenic T1 generation cucumber lines. A, The schematic structure of the sense expression (top) and antisense expression (bottom) vector carrying CsHT1. B, qRT-PCR analysis of the relative RNA transcription in the stage 12 male flowers of CsHT1-antisense expression lines (A) and sense expression lines (S). C, Raster electron microscopy images of pollen of wild-type (WT), A3, and S4 lines. D, No significant difference of pollen viability between the wild type, A3, and S4. E and F, Pollen grains of the wild type, A3, and S4 germinated in vitro in medium consisting of 3% (w/v) Glc (E) or 3% (w/v) Gal (F). G and H, Pollen germination ratio (G) and average pollen tube length (H) of wild-type, A3, and S4 lines after germination for 4 h. Only the pollen grains that germinated were calculated in H. Means of at least 1,000 pollen grains from three independent experiments. LB and RB indicate transfer DNA left and right borders, respectively. Npt II (Kan) indicates resistance genes encoding kanamycin. NOS-pro and NOS-ter indicate a promoter and terminator sequence, respectively. Bars = 200 μm.

To determine whether the same inhibition can be observed in planta, pollen tubes in vivo were stained with aniline blue. The results indicated that both wild-type pollen and antisense pollen geminated, and tubes grew well, on the stigma (Fig. 7). However, 24 h after pollination, most of the wild-type pollen tubes reached the bottom of the transmitting tracts and grew toward ovules, whereas few pollen tubes of the antisense plants reached the bottom of the transmitting tracts (Fig. 7). These results demonstrated that down-regulating CsHT1 inhibited the elongation of pollen tubes in planta.

Figure 7.

Pollen germination and pollen tube growth in vivo of cucumber wild-type (WT) and CsHT1 antisense transgenic line A3. A, A cucumber ovary/fruit (cv Xintaimici) shows the observation sketch. The ovaries/fruits of the wild type and A3 were cut into longitudinal slices, and the slices were fixed and stained with aniline blue. The red box indicated the approximated location that was photographed. B, The pollen of the wild type and A3 geminated, and pollen tubes grew in the stigma (top row) or in the ovaries/fruits (bottom row; arrow indicates the pollen tubes) 24 h after pollination. Bars = 500 μm.

Antisense Suppression of CsHT1 Affects Pollen Tube Growth and Seed Development

To determine if CsHT1 expression in pollen affects fertilization and seed development, two representative lines of antisense transformants were selected for study. The results indicated that CsHT1 transcript (Fig. 8A) and protein levels (Fig. 8B) in the pollen of most of the subsequent T2 generation lines were significantly reduced. After artificial self-pollination, we obtained a number of cucumber fruits from these CsHT1-antisense transgenic lines. Many abnormal phenotypes were found. In particular, fruit placenta development and seed development were inhibited significantly. In longitudinal sections, the placentas of wild-type fruits were almost fully filled with viable seeds. By contrast, the transgenic line placentas were only half full of viable seed, closest to the flowers, the rest being empty cavities (Fig. 8, C and F). Seed counts indicated 36% and 34% reductions in the two antisense transgenic lines compared with the wild type (Fig. 8G). In addition, the seed size, including length (Fig. 8, D and H) and width (Fig. 8, E and I), of T2 was less than that of the wild type. Accordingly, the 1,000-seed weight of the transgenic lines was significant lower than that of the wild-type plants (Fig. 8J).

Figure 8.

Phenotype analysis of CsHT1-antisense transgenic lines (T2 generation). A, Transcript levels of CsHT1 by qRT-PCR and semiquantitative RT-PCR in mature pollen of transgenic lines (T2). B, Immunolocalization of CsHT1 protein in pollen tubes. C to E, Phenotypic analysis of fruits and seeds of CsHT1 antisense transgenic lines. Longitudinal section of cucumber fruits 45 d after fertilization (C) and the figure of head-to-end (D) or side-to-side (E) setup of 20 seeds. F, The average fertile length (FL) of fruit (i.e. the region of the fruit with viable seeds) as shown in C. G to J, Total viable seed number per fruit (G), 10-seed length (H), 10-seed width (I), and 1,000-seed weight (J) of wild-type (WT) and CsHT1 antisense transgenic lines. X indicates the seeds of eight to 10 independent fruits from relative transgenic lines were counted. Bars = 100 μm.

To rule out the possibility that results obtained were caused by interference with a homologous gene with a broad expression pattern, reciprocal cross experiments between wild-type plants and CsHT1-antisense plants were carried out. As showed in Table I, when pistils of transgenic plants were hand pollinated with wild-type pollen, the seed number per fruit and 1,000-seed weight were not significantly different from those of the wild-type self-fertilized plants. However, when pistils of wild-type or transgenic plants were pollinated with pollen from antisense transgenic plants, the seed number per fruit and 1,000-seed weight were significantly lower than those of wild-type self-fertilized plants (Table I). These results demonstrate that antisense expression of CsHT1 impedes the function of pollen and affects seed development but does not impede the function of female tissues and organs.

Table I. Antisense expression of CsHT1 only impeded the function of pollen but not the female.

The seeds of eight to 10 independent fruits from each line were counted. The means in each column followed by the same letter are not significantly different at the 5% level according to Student's t test (P ≤ 0.05).

Male Parent × Female Parent	Seed No. per Fruit	1,000-Seed Weight
		g
Wild type × wild type (self)	388.3 ± 25.46 a	17.2 ± 0.69 a
Wild type × A3	378.0 ± 15.56 a	16.3 ± 0.79 a
A3 × wild type	296.5 ± 15.32 b	14.6 ± 0.21 b
A3 × A3 (self)	283.2 ± 11.10 b	14.6 ± 0.61 b

DISCUSSION

CsHT1 Is a Plasma Membrane Protein Expressed Specifically in Pollen

The cucumber monosaccharide transporter (CsHT1) we report in this paper is a plasma membrane protein (Fig. 4) expressed in pollen (Fig. 3) possessing 12 transmembrane domains, a structure common to plant monosaccharide transporter proteins, and sharing high homology with earlier reported plasma membrane H⁺-monosaccharide symporters (Fig. 1; Supplemental Figs. S1 and S2; Büttner and Sauer, 2000; Wang et al., 2007). It shares especially high homology with the Arabidopsis genes AtSTP4, AtSTP9, and AtSTP11 and petunia gene Pmt1, which have been reported to be expressed mainly in pollen grain and the pollen tube (Fig. 1; Supplemental Fig. S2). Spatiotemporal expression analysis indicated that CsHT1 transcripts were exclusively found in the developing pollen grain and germinating pollen tube. The CsHT1 protein was not detectable in pollen grains before germination but was abundant in germinating pollen and in the growing pollen tube. This result is similar to a monosaccharide transporter in Arabidopsis, AtSTP9 (Schneidereit et al., 2003), and there are many other pollen-specific genes having similar translational control (Mascarenhas, 1989). It has been reported that the upstream open reading frame (uORF), located in the 5′-untranslated region of a mature mRNA, may act as a trans-factor to regulate the expression of a major ORF (Hu et al., 2005). Functional uORFs are between four and six codons in length, occur between 50 and 150 nucleotides from the start of the main ORF, and do not overlap with, and are clearly separated from, neighboring uORFs (Cvijović et al., 2007). Here, the predicted transcript of CsHT1 has one uORF six codons in length and 119 nucleotides distant from the start of the main ORF (Supplemental Fig. S7). These features suggested that this uORF may regulate the translation of CsHT1. This translation regulation is the trend in evolution, because the environment of a pollen grain from the time of its release from the anther to deposition on the stigma surface is quite inhospitable, so the pollen grain has to contain all that is necessary to germinate rapidly and enable the tube to penetrate the stigma, where conditions are favorable for its further development (Mascarenhas, 1975). A calcium-dependent calmodulin-independent protein kinase that has the similar translational control has been proven to be required for normal germination and pollen tube growth in maize (Zea mays; Estruch et al., 1994). The preloading of CsHT1 mRNA in pollen here suggests that CsHT1 may also have an important function during pollen germination and tube growth.

CsHT1 Prefers Glc and Gal and Uses an Energy-Dependent H⁺ Symporter

Pollen germination and tube growth in vitro require sugar uptake from the medium. In conventional pollen germination medium, Suc is the carbon source (Jahnen et al., 1989; Rodriguez-Enriquez et al., 2013). Previous studies have shown that tobacco (Nicotiana tabacum) and Arabidopsis pollen prefer Suc as a carbon source. Addition of Glc to in vitro-germinated pollen led frequently to an immediate bursting of the tubes for unknown reasons (Lemoine et al., 1999; Scholz-Starke et al., 2003). However, petunia pollen can germinate and grow using not only Suc, but also Glc or Fru (Ylstra et al., 1998). Cucumber pollen tubes are different; they can grow on Suc, Glc, and Gal but not on Fru or Man (Fig. 5A). Moreover, high concentration of Glc (greater than 15% [w/v]) and Gal (greater than 10% [w/v]) inhibit cucumber pollen germination (Fig. 5A).

In petunia, almost all of the Suc is converted to Glc and Fru before absorption by the pollen tube, suggesting that pollen tubes of this species normally absorb carbohydrates only as monosaccharides (Ylstra et al., 1998). In addition, competition by addition of excess Glc reduces the uptake rate of ¹⁴C-labeled Suc by tomato pollen tubes, indicating that at least part of the Suc is cleaved extracellularly and taken up in the form of hexoses (Hackel et al., 2006). Also, many hexose transporters (Ylstra et al., 1998; Schneidereit et al., 2003, 2005) and cell wall invertases (Singh and Knox, 1984; Miller and Ranwala, 1994; Maddison et al., 1999) have been identified in pollen and pollen tubes. While these data suggest that monosaccharides may be the preferred form for carbohydrate uptake in many plants, there is as yet no definitive evidence for the role of a specific hexose transporter in pollen germination and growth.

Unlike many other plant monosaccharide transporters reported so far that transport a broad spectrum of monosaccharides (Gear et al., 2000; Ngampanya et al., 2003; Schneidereit et al., 2005), CsHT1 shows a high specificity for Glc and Gal. That is to say, the preference of cucumber pollen for Glc and Gal, but not for Fru or Man (Fig. 5A), is consistent with the substrate specificities of CsHT1 (Fig. 2E). Moreover, inhibition by the proton uncoupler CCCP and apoplast transport inhibiter PCMBS (Fig. 5B) is consistent with the necessary function of an energy-dependent H⁺ symporter. All of these results are consistent with CsHT1 playing a leading role in cucumber pollen germination and tube growth.

CsHT1 Is a Key Gene for Seed Development Due to Its Control over Pollen Tube Growth

In the race to deliver its sperm nucleus to the ovule, a pollen tube lives on nutrients from female tissues of the pistil (Konar and Linskens, 1966). Because germinating pollen grains and growing pollen tubes have no symplastic connection to surrounding tissue, nutrients, including hexoses, must be transferred into the pollen tube by transporters. Although the functional characterization of hexose transporters is fundamental to our understanding of reproductive biology, studies on this subject are mainly limited to Arabidopsis (Schneidereit et al., 2003, 2005; Scholz-Starke et al., 2003) and petunia (Ylstra et al., 1998; Garrido et al., 2006). Furthermore, it is difficult to assign functions to those transporters because single knockouts in these species do not result in phenotypic change. Moreover, the identity of available sugars and the roles of hexose transporters after pollination in developing fruits are unknown.

In studies reported here, we demonstrate that cucumber pollen tubes with down-regulated hexose transporter CsHT1 grow through the stigma and style as well as wild-type pollen but are severely limited in growth after arrival in the ovary/fruit (Fig. 7B). These data are consistent with data showing that Suc is the primary carbon source in cucumber stigma/styles (Supplemental Fig. S8), while Glc and Fru are the most abundant sugars in the cucumber fruit (Handley et al., 1983; Hu et al., 2009). Therefore, although pollen with down-regulated CsHT1 can grow on Suc in vitro (Supplemental Fig. S9), in vivo, the pollen tubes are deprived of their major source of carbon skeletons and energy as they attempt to reach the ovules. Given the extreme length of the cucumber ovary on the day of flowering (generally, 4–6 cm; Fig. 7A), this results in reduced 1,000-seed weight, seed number per fruit, and length of fruit with viable seeds in antisense transgenic plants (Fig. 8). Some seeds in the region of the fruit closest to the flower are produced, but they are small, and no seeds are produced in the more remote fruit tissues. The reason for smaller seed size in transgenics is unclear. Seed size is influenced by a complex regulatory network (Li and Li, 2015). Factors associated with seed development and seed size, and that could be causally linked to CsHT1, include coat-associated invertases and the carbohydrate status of the developing seed (Weber et al., 1996) and Suc synthase activity in developing cotyledons (Turner et al., 2009). Resolving the interactions in this network will require additional study.

In summary, CsHT1 is a plasma membrane hexose transporter expressed only in pollen and translated only during pollen germination and tube growth. CsHT1 specifically transports Glc and Gal and affects seed development by controlling pollen tube growth.

MATERIALS AND METHODS

Plant Material and Bacterial/Yeast Strains

Wild-type cucumber (Cucumis sativus ‘Xintaimici’) plants and transgenic cucumber lines were grown under greenhouse conditions from the end of February to July or from the end of August to December. For RT-PCR, different organs from 2-month-old plants, different tissues from mature male flowers, and anthers from male flowers at developing stages 9 to 12 (Bai et al., 2004) were used. For in situ hybridization, male flowers at developing stages 9 to 12 were used.

Escherichia coli strain DH5a (purchased from Tiangen Biotech) was used for cloning. Heterologous expression was performed in yeast (Saccharomyces cerevisiae) strain EBY.VW4000.

Cloning of Cucumber CsHT1 and Sequencing

A 1,557-bp PCR fragment containing the complete CsHT1 coding sequence was amplified from male flower complementary DNA (cDNA) using primers CsHT1f (5′-ATG GCG GGA GGT GGA TTT GTT TCT-3′) and CsHT1r (5′-TCA GAC ACC TTT GCC ATA CGG CTC CAT ACT-3′), cloned into pGEM-T Easy (Promega) vector, and subsequently sequenced. The primers were designed according to the cucumber genome that was sequenced completely in 2009 (Huang et al., 2009).

RT-PCR and Real-Time PCR

Total RNA was extracted from specified tissues and treated with DNase (Promega) and reverse transcribed using an oligo(dT) primer according to the supplier’s instructions (Promega). The cDNA was then used as a template for RT-PCR or quantitative RT (qRT)-PCR analysis. For the RT-PCR analysis, the clone primers of CsHT1 (CsHT1f and CsHT1r) were used, and the 18S ribosomal RNA was used as a control. Quantitative RT-PCR analysis was performed with the ABI 7500 system (Bio-Rad) using the SYBR green detection protocol (TaKaRa).

Tubulin mRNA was used as an internal control, and relative amounts of mRNA were calculated using the comparative threshold cycle method. Primers specific to CsHT1 for qRT-PCR analyses were as follows: 5′-CGGGAGGTGGATTTGTTTCT-3′ and 5′-GCTGCCTTGGCTTGTTGTT-3′.

Functional Characterization of CsHT1 by Heterologous Expression in Yeast

To test the functionality of CsHT1, the ORF of CsHT1 was cloned into the yeast expression vector pDR196 (Fan et al., 2009). Appropriate restriction sites and a stop codon were introduced with designed primers (forward primer, 5′-CCGGAATTCATGGCGGGAGGTGGATTTGTTTCT-3′; and reverse primer, 5′-ACGCGTCGACTCAGACACCTTTGCCATACGGCTCCATACT-3′), with the EcoRI and SalI sites being underlined. The PCR products were digested either with EcoRI and SalI to yield CsHT1. The fragment was subcloned into the corresponding restriction sites of pDR196, yielding pDR196/CsHT1, and the construct was confirmed by sequencing. The hexose transporter-deficient yeast strain EBY.VW4000 (Wieczorke et al., 1999) was transformed with pDR196/CsHT1, according to the method of Morita and Takegawa (2004), and transformation with empty vector pDR196 was used as a control.

The drop test for yeast growth was performed according to Loqué et al. (2007). Cells were grown in liquid synthetically defined (lack ura) medium supplemented with 2% (w/v) maltose as sole carbon source to an optical density at 600 nm of 0.6. Different dilutions of this cell suspension were dropped on solid synthetically defined (lack ura) medium (pH 5.5) supplemented with 2% (w/v) maltose or Glc as sole carbon source. Cells were incubated 3 to 5 d at 30°C prior to photography.

For Glc uptake assays, yeast strain EBY.VW4000, carrying either pDR196/CsHT1 or empty pDR196, was grown in liquid minimal medium containing maltose at 30°C to an optical density at 623 nm of approximately 0.8. Cells were harvested by centrifugation and washed twice with 25 mm sodium phosphate buffer, pH 5.5, and suspended in the same buffer to an optical density at 623 nm of 20. Uptake assays were initiated by adding ¹⁴C-Glc to a final specified concentration and incubation at 30°C in a shaking water bath. Samples (200 μL) were collected at specified time intervals by vacuum filtration onto 0.8-μm glass fiber filters (Whatman) and washed three times with 4 mL of ice-cold distilled water before radioactivity was determined by liquid scintillation counting.

Generation of Transgenic Cucumber Plants

To obtain the overexpression and antisense plant lines, the ORF of CsHT1 was cloned into the expression vector pBI121 in sense and antisense orientations. The XbaI and BamHI restriction sites were used, and PCR product was linked between the CaMV 35S promoter and the octopine synthase terminator. Both sense and antisense vectors were transformed into Agrobacterium tumefaciens strain LBA4404. The sense and antisense vectors were transformed into the cucumber pure line ‘Xintaimici’ using the fresh expanding cotyledon disk transformation method (Liu et al., 2010; Sui et al., 2012). Briefly, 1 d after the seeds germinated, the growing points and the upper halves of cotyledons were removed, and the other halves of cotyledons were dipped in the A. tumefaciens-diluted suspension containing one-half-strength Murashige and Skoog liquid medium (optical density at 600 nm = 0.3–0.5) for 15 min. These explants were placed upside down on shoot induction medium containing 0.5 mg L^–1 6-benzylaminopurine and 1 mg L^–1 abscisic acid and cocultured for 2 d at 28°C in the dark. After that, the tissues were transferred to germination medium containing 0.5 mg L^–1 6-benzylaminopurine, 1 mg L^–1 abscisic acid, 25 mg L^–1 kanamycin, and 500 mg L^–1 carbenicillin (Sigma) and then cultured for 14 to 20 d at 28°C, 100 μM m^–2 s^–1 photosynthetic photon flux density. The shoots with kanamycin resistance were cut out and cultured in Murashige and Skoog medium with 50 mg L^–1 kanamycin and 200 mg L^–1 carbenicillin for development into whole plants. Regenerated plants were screened by PCR for integration of the construct.

Localization of CsHT1 Expression by GUS Reporter Gene Analyses

The putative promoter region of CsHT1, a 1,620-bp PCR fragment upstream of the start codon ATG was amplified from cucumber genomic DNA using primers HT1-Pf (5′-CCCAAGCTTACTTAGTAACAGTTTCGGATTGAC-3′) and HT1-Pr (5′-CGCGGATCCCTTTTTGTTTTTCTTTGAATTTTCT-3′), with the HindIII and BamHI sites being underlined. The PCR product was digested with HindIII and BamHI and was cloned in front of the β-glucuronidase enzyme gene (GUS) in vector pBI121, yielding construct pCsHT1-GUS. The pCsHT1-GUS plasmid was transferred into A. tumefaciens strain LBA4404. A. tumefaciens-mediated transformation of cucumber was performed as described above. Arabidopsis (Arabidopsis thaliana) plants were transformed with A. tumefaciens harboring the constructs pCsHT1-GUS using the floral dip method according to the protocol of Clough and Bent (1998). Transgenic Arabidopsis and cucumber samples were incubated with GUS staining solution overnight at 37°C as described by Jefferson et al. (1987). After being stained, the tissues were rinsed and dehydrated through an ethanol series before photographed. Some of them were hand sectioned or made into paraffin sections for better observation.

Subcellular Localization of CsHT1-GFP Fusion Protein

To examine the subcellular localization of CsHT1, the ORF of CsHT1 cDNA was amplified with gene-specific primers HT1-Gf (5′-GGAAGATCTGATGGCTACAGTTGGAAGAGTC-3′) and HT1-Gr (5′-CGGACTAGTATATGTTCTTATCTCATAAATCACATCC-3′), with the BglII and SpeI sites being underlined. The PCR product was digested with BglII and SpeI and was cloned in front of the GFP reporter gene in vector pCAMBIA1302 to produce the CsHT1-GFP fusion protein under the control of the CaMV 35S promoter. Transient expression of the CsHT1-GFP fusion protein in onion (Allium cepa) epidermal cells and cucumber protoplast was conducted as described by Hayes et al. (2007) and Huang et al. (2013), respectively. GFP fluorescence was visualized by an Olympus Confocal Laser Scanning Microscope.

Generation of Anti-CsHT1 Antisera and Immunolocalization of CsHT1 Protein

Three specific peptide fragments (CDAVIGHHVSMEPYGKGV, FPSVYEQQAKAAGGNQYC, and CERGDMEKARKMLKKI) derived from the CsHT1 protein sequence were chosen to synthetize polypeptides and used for the immunization of two rabbits (specific pathogen free) by Beijing B&M Biotech.

For immunohistochemical analyses, samples were collected at the same time, from the same plants, and processed under the same conditions. Samples were taken from male flowers (stages 9–12), anthers with in vitro germinated pollen (6 h after cultivation in liquid pollen germination medium), and pollinated stigmas and ovaries (24 h after pollination). The immunocytochemical protocol used on 16- to 20-µm-thick paraffin sections has been described (Wang et al., 2014). Specimens were viewed with an Olympus BX53 microscope.

In Vitro Pollen Germination Tests

For the pollen in vitro germination assay, mature pollen grains were collected from male flowers at anthesis and spread onto cucumber pollen medium, pH 6.0, consisting of 15% (w/v) Suc, 0.01% (w/v) boric acid, 1 mm Ca(NO₃)₂, and 0.3% (w/v) Phytagel (Sigma) in a humid chamber at approximately 25°C. Single images of the cultures were obtained after 4 h with the Olympus BX53 microscope.

To detect the effect of carbohydrates on pollen germination and pollen tube elongation in vitro, Suc in the germination medium was replaced with D(+)-Glc, D(–)-Fru, Gal, and Man. Pollen germination rate was determined with a Fuchs-Rosenthal chamber. Pollen was taken from two to four individual transformed lines and from 10 individual wild-type plants and scored as geminated when the pollen tube was longer than the pollen grain diameter. For each experiment, between 200 and 500 pollen grains were counted. Pollen germination rate reflects the mean of three independent flowers with at least 100 pollen grains each.

Scanning Electron Microscopy

Cucumber pollen was collected in Eppendorf tubes and attached to double-sided adhesive tape (Plano). Scanning electron microscopy aluminum specimen holders were covered with conducting adhesive tape, and the pollen was stuck on the tape. The specimens were coated with 20 nm of gold using an SCD 005 sputter coater (BAL-TEC). Scanning electron microscopy was carried out with a LEO 1430 scanning electron microscope (Zeiss).

Pollen Viability Assay

Triphenyltetrazolium chloride (TTC) test procedures were used in determining pollen viability. Petri dishes containing 1% (w/v) TTC solution (0.2 g of TTC and 12 g of Suc dissolved in 20 mL of distilled water) were dusted with fresh pollen and kept at room temperature for 2 h under daylight, and then pollen grains were examined using a light microscope. Viable pollen grains were dyed red and light red; dead pollen was not dyed in TTC.

In Vivo Pollen Tube Staining

Pollen tubes in pistils were stained with aniline blue as described by Zhu et al. (2013) with slight modification. Briefly, preclamped mature female flowers of CsHT1 antisense lines and wild-type lines were self-pollinated separately. After 24 h, pollinated pistils were cut into longitudinal slices and incubated in fixative solution (ethanol:acetic acid [v/v] = 9:1) for at least 3 h at room temperature and then in 8 m NaOH softening solution overnight. After being washed in phosphate-buffered saline three times for 5 min each, samples were then stained with aniline blue solution in the dark. Samples were observed using an Olympus Confocal Laser Scanning Microscope. The UV excitation wavelength maximum was 405 nm.

Bioinformatic Analysis

Nucleotide sequences and amino acid sequences were primarily analyzed using the primer premier 5.0. Sequence homology searches in GenBank were carried out with the BLAST program (http://www.ncbi.nlm.nih.gov/BLAST/; Madden et al., 1996).

The phylogenetic analysis was conducted using MEGA version 5.2 (Kumar et al., 2004) adopting Poisson correction distance and was presented using traditional rectangular TreeView. Support for the tree was assessed using the bootstrap method with 1,000 replicates.

Statistical Analysis

The Student’s t tests were performed using the algorithm embedded into Microsoft Excel, and significance was determined by the Student’s t test (P < 0.05).

Sequence data from this article can be found in the GenBank/EMBL data libraries under accession number CsHT1 (HQ202746).

Supplemental Data

The following supplemental materials are available.

• Supplemental Figure S1. Membrane-spanning model of CsHT1.

• Supplemental Figure S2. Comparison of amino acid sequence of monosaccharide transporter from cucumber and other organisms.

• Supplemental Figure S3. Vertical sections of male flowers at stages 9 to 12 were used for RNA in situ hybridization analysis.

• Supplemental Figure S4. Expression of GUS reporter genes in Arabidopsis under the control of the CsHT1 promoter.

• Supplemental Figure S5. Immunolocalization of CsHT1 protein in pollen tubes of cucumber wild-type plants using affinity-purified anti-CsHT1 antiserum.

• Supplemental Figure S6. RT-PCR analyses the expression of CsHT1 homologous genes in pollens of wild-type and CsHT1 antisense lines.

• Supplemental Figure S7. The 5′-untranslated region sequence of CsHT1.

• Supplemental Figure S8. Sugar levels in the stigma of wild-type cucumber.

• Supplemental Figure S9. No significant difference of pollen germination between wild-type plant and CsHT1 antisense lines in Suc medium.

• Supplemental Table S1. GenBank accession numbers of the sequences used for the analysis in Figure 1.

• Supplemental Text S1. The methods of in situ hybridization and bioinformatical analysis.

Glossary

SUT

sucrose transporter

RT

reverse transcription

CCCP

carbonyl cyanide m-chlorophenylhydrazone

uORF

upstream open reading frame

ORF

open reading frame

qRT

quantitative reverse transcription

CaMV

Cauliflower mosaic virus

TTC

triphenyltetrazolium chloride

cDNA

complementary DNA