Alkaline α-galactosidase 2 (CsAGA2) plays a pivotal role in mediating source–sink communication in cucumber

Abstract

Sugars are necessary for plant growth and fruit development. Cucumber (Cucumis sativus L.) transports sugars, mainly raffinose family oligosaccharides (RFOs), in the vascular bundle. As the dominant sugars in cucumber fruit, glucose and fructose are derived from sucrose, which is the product of RFO hydrolysis by α-galactosidase (α-Gal). Here, we characterized the cucumber alkaline α-galactosidase 2 (CsAGA2) gene and found that CsAGA2 has undergone human selection during cucumber domestication. Further experiments showed that the expression of CsAGA2 increases gradually during fruit development, especially in fruit vasculature. In CsAGA2-RNA interference (RNAi) lines, fruit growth was delayed because of lower hexose production in the peduncle and fruit main vascular bundle (MVB). In contrast, CsAGA2-overexpressing (OE) plants displayed bigger fruits. Functional enrichment analysis of transcriptional data indicated that genes related to sugar metabolism, cell wall metabolism, and hormone signaling were significantly downregulated in the peduncle and fruit MVBs of CsAGA2-RNAi plants. Moreover, downregulation of CsAGA2 also caused negative feedback regulation on source leaves, which was shown by reduced photosynthetic efficiency, fewer plasmodesmata at the surface between mesophyll cell and intermediary cell (IC) or between IC and sieve element, and downregulated gene expression and enzyme activities related to phloem loading, as well as decreased sugar production and exportation from leaves and petioles. The opposite trend was observed in CsAGA2-OE lines. Overall, we conclude that CsAGA2 is essential for cucumber fruit set and development through mediation of sugar communication between sink strength and source activity.

Alkaline α-galactosidase 2 hydrolyzes raffinose family oligosaccharides in companion cells of cucumber fruit to initiate phloem unloading of sugars for fruit growth and source–sink communication.

Introduction

Source-to-sink transport of sugar is one of the major determinants of plant growth and implements a fine-tuned regulation of carbon allocation across plant organs (Lemoine et al., 2013). In most plants, the major sugar exported from source leaves is sucrose, which is then distributed to heterotrophic sink tissues through the phloem (Lemoine, 2000). The partitioning of sugars in plant sinks depends on the rate of long-distance translocation and requires the specific activity of sugar metabolism enzymes and membrane sugar transporters (Durand et al., 2018; Fan et al., 2019b). During grape (Vitis vinifera) berry development, VvSWEETs (Sugars Will Eventually be Exported Transporters) are important players in sugar mobilization (Breia et al., 2019). PuSWEET15 could also affect sucrose accumulation in pear (Pyrus ussuriensis) fruit (Li et al., 2020). In addition to the sugar transporters, certain other enzymes are related to sugar allocation and partitioning within the plant, such as sucrose synthase (SUS; EC 2.4.1.13) for sucrose catabolism, which is involved in import capacity to sink organs and enhancing sink strength. Inhibition of SUSs decreases fruit setting and sucrose unloading capacity in young tomato (Lycopersicon esculentum) fruit (Aoust et al., 1999), causes a decrease in total tuber dry weight of potato (Solanum tuberosum; Zrenner et al., 1995), and retards the growth of cotton (Gossypium hirsutum) fiber cell and the development of seed (Ruan et al., 2003). In contrast, overexpression (OE) of SUSs induces fast growth and soluble sugar accumulation in tobacco (Nicotiana tabacum; Coleman et al., 2006), as well as increases starch accumulation and tuber yield in potato (Baroja-Fernández et al., 2009).

Unlike sucrose-translocating plants, the Cucurbitaceae, Lamiaceae, Oleaceae, Scrophulariaceae, and several other families translocate sugars as raffinose family oligosaccharides (RFOs; Keller and Pharr, 1996; ElSayed et al., 2014). The RFOs are galactosyl-sucrose oligosaccharides containing an α-type galactosidic linkage, including the trisaccharide raffinose and the tetrasaccharide stachyose, etc. These oligosaccharides are ubiquitous in land plants. Numerous economically important horticultural crops are typical RFOs-translocating plants, such as cucumber (Cucumis sativus) and watermelon (Citrullus lanatus). Cucumber accumulates sugars as the main nutrients in its fleshy fruit. However, the dominant sugars in cucumber fruit are not RFOs but hexoses (Hu et al., 2009). RFOs are hydrolyzed into sucrose, with further cleavage into hexoses by SUS and/or invertase (INV; EC 3.2.1.26; Li et al., 2021). The initial enzyme in RFO catabolism is an α-galactosidase (α-Gal; α-D-galactoside galactohydrolase; EC 3.2.1.22), which hydrolytically remove the terminal galactose moiety from RFOs to produce sucrose and galactose. In watermelon, the α-Gal ClAGA2 functions in the hydrolysis of RFOs during fruit development (Ren et al., 2021). The α-Gals most likely play a vital role in fruit development of RFOs-translocating plants, similar to the role of SUSs in sucrose-transporting plants.

The α-Gal family is classified into acid and alkaline hydrolases, according to the optimal pH for their activities (Keller and Pharr, 1996). Some of the eukaryotic acidic α-Gals with broad pH optima in the acidic range have high affinities mainly toward raffinose (Keller and Pharr, 1996). An alkaline α-Gal with an optimal pH of 7–7.5 was initially discovered in Cucurbita leaves (Gaudreault and Webb, 1983), and subsequently found in the stachyose-translocating species bugle (Ajuga reptans, Lamiaceae family; Bachmann et al., 1994). In the next few years, an alkaline α-Gal, which can hydrolyze both raffinose and stachyose, was purified and cloned from melon fruit (Gao and Schaffer, 1999; Carmi et al., 2003). In cucumber, it was reported that the alkaline α-Gal preferred stachyose to raffinose, while the acid α-Gal activity was higher with the raffinose as the substrate (Gu et al., 2018).

An increasing number of studies have shown that the α-Gals are involved in multiple physiological processes of plant development. During phloem unloading in cucumber fruit, RFO hydrolysis by α-Gals begins at the peduncle vasculature (Ohkawa et al., 2010). In tobacco, the acidic α-Gal (NTα-Gal) in floral nectar functions in reproductive development (Zha et al., 2012). Silencing the rice (Oryza sativa) alkaline α-Gal (OsAkαGal) resulted in delayed leaf senescence (Lee et al., 2009). In addition, α-Gal has been considered part of the response to abiotic stress in plants. Downregulation of α-Gal could lead to an increase in cold tolerance in nonacclimated and cold-acclimated petunia (Petunia × hybrida cv Mitchell) plants (Pennycooke et al., 2003). A similar function of α-Gal under cold stress and during a subsequent recovery period has been observed in cucumber leaves (Gu et al., 2018).

In watermelon, alkaline α-Gal contributes to sugar accumulation in fruit flesh by facilitating the conversion of RFOs into sucrose (Ren et al., 2021), but there is so far no detailed report on the function of this enzyme in cucumber fruit. As our previous study found that alkaline α-galactosidase 2 (CsAGA2) was involved in phloem unloading together with the hexose transporter CsSWEET7a in cucumber fruit (Li et al., 2021), we decided to pursue an in-depth study on CsAGA2. Here, we report that CsAGA2 was expressed during different development stages of fruit in both the peduncle and fruit main vascular bundles (MVBs). Further research showed that knockdown and knockout of CsAGA2 led to weak sink strength and impaired fruit development compared to the wild type (WT) plants. Meanwhile, expression levels of genes related to sugar metabolism, cell wall biosynthesis, and sugar and auxin signaling were significantly changed by the downregulation of CsAGA2 in both the peduncle and fruit MVB in transgenic plants. Photosynthesis and phloem loading in source leaves were significantly inhibited or promoted in CsAGA2-RNA interference (RNAi) or CsAGA2-OE plants, respectively, indicating a feedback regulation mechanism centered around CsAGA2 between sink and source tissues in cucumber.

Results

CsAGA2 of cultivated cucumber has undergone domestication selection

In our previous research, the sugar transporter CsSWEET7a in cucumber, which could transport hexoses including glucose, Gal, fructose, and mannose, was demonstrated to be expressed in sink tissues especially in fruit MVB (Li et al., 2021). Further findings indicated that CsSWEET7a functioned in fruit apoplasmic phloem unloading in companion cells (CCs), and CsAGA2 was speculated to play a synergistic role in this process (Li et al., 2021). Here, we first compared the expression of CsAGA2 and CsSWEET7a in the fruits of eight cucumber lines from different geographic groups (Qi et al., 2013). The accessions included the wild cucumber [CG0002 (C. sativus var. hardwickii), CG8099], which originate from India, and six cultivated cucumber lines from three regionally selected groups of C. sativus, the Eurasian group (CG4182 and CG6601), the Xishuangbanna group (CG9198 and CG9203), and the East-Asian group (XTMC and CG3079) (Figure 1A). In the cultivated lines, the expression levels of CsAGA2 (Figure 1B) and CsSWEET7a (Figure 1C) were higher than that in the wild accessions. Furthermore, analysis of sugar content indicated that the fruits from the cultivated accessions, especially in the East-Asian group, contained much more hexose (fructose and glucose) than those in the wild group (Figure 1D). Notably, the RFOs can be detected only in the fruit of the wild accessions, indicating that the oligosaccharides transported to the fruits were not entirely hydrolyzed by α-Gals (and subsequent sucrose hydrolases) into hexoses (Figure 1D). We also noted that the CsAGA2 but not CsSWEET7a gene showed a strong positive correlation between mRNA level and sugar content among the different geographic groups (Figure 1, E and F), suggesting that the expression of CsAGA2 gene is closely related to sugar accumulation in cucumber fruit.

Figure 1.

Selective domestication sweeps at the CsAGA2 and CsSWEET7a region. A, Fruit morphology of different cucumber geographic groups. The wild group includes Indian (CG0002 and CG8099), the cultivated groups included the Eurasian (CG4182 and CG6601), Xishuangbanna (CG9198 and CG9203) and East-Asian (XTMC, CG3079) groups. The fruits were digitally extracted for comparison. Scale bar: 1 cm. B and C, Relative expression of CsAGA2 (B) and CsSWEET7a (C) in the MVB of fruits at 9 days after pollination (DAP) from the four groups of cucumbers. D,) Sugar content in the main vascular bundle (MVB) of fruits at 9 DAP from the four groups. Data are presented as the means of three biological replicates (±se) in (B–D). E and F, Sugar content versus expression levels of CsAGA2 (E) and CsSWEET7a (F) in different geographic groups. Data are presented as the means of three biological replicates. Significance was determined by Student’s t test in (E and F). P = 0.001 in (E), P = 0.728 in (F). G–I, Detected domestication sweeps on the fourth chromosome (H) of CsAGA2 (G) and CsSWEET7a (I). Gold bars represent the 50-kb windows defining the ratio of genetic diversity (π) between the cultivated groups (π_c) and the wild group (π_w). The horizontal dashed line indicates the threshold of 15.4, defining the top 5% of π_w/π_c values. CsAGA2 and CsSWEET7a show selective sweeps between the Indian and East-Asian groups. Abbreviations: FW, fresh weight; nd, none detectable.

The CsAGA2 (Csa4G167980) and CsSWEET7a (Csa4G054810) genes are located on chromosome 4 (Figure 1H). To assess whether the CsAGA2 and CsSWEET7a genes had undergone human selection during cucumber breeding, we checked the signature on chromosome 4 using genome resequencing data from a total of 115 cucumber varieties (Qi et al., 2013). The genomic region harboring CsAGA2 and CsSWEET7a showed lower genetic diversity in the varieties of the cultivated groups than those in the Indian group (Figure 1, G–I). This indicated that the cultivated alleles of CsAGA2 and CsSWEET7a had undergone selection in fruit development during cucumber domestication.

Spatiotemporal expression of CsAGA2 in cucumber

There are eight genes predicted to encode α-Gals in the cucumber genome, including four alkaline forms and four acidic forms (Sui et al., 2018), here named CsAGA1-4 and CsGAL1-4, respectively. The four alkaline CsAGAs shared 58.9% amino acid identity, while the four acidic CsGALs showed 38.2% (Supplemental Figure S1). Two known conserved motifs in AGA, DDxW and KxD (Li et al., 2017), were also present in the CsAGAs (Supplemental Figure S1). The deduced amino acid sequences of cucumber α-Gals were homologous to genes from other Cucurbitaceae species, including melon (Cucumis melo, average of 95.1% identity), watermelon (Citrullus lanatus, average of 85.2% identity), and pumpkin (Cucurbita maxima, average of 86.3% identity; Supplemental Figure S2). Both alkaline and acidic α-Gals have also been found in the sucrose-translocating plants such as Arabidopsis (Arabidopsis thaliana), Solanum lycopersicum, and O. sativa (Supplemental Figure S2), indicating that α-Gals are highly conserved across plant species. In our previous transcriptomic analysis, the α-Gal genes were ubiquitously expressed in the phloem tissues of cucumber fruit (Sui et al., 2018). This result was confirmed by reverse transcription quantitative polymerase chain reaction (RT-qPCR) analysis in fruit, with CsAGA2 showing the highest expression compared to the other α-Gal genes (Figure 2A). This indicated that CsAGA2 may play a major role in cucumber fruit sugar partitioning.

Figure 2.

Expression analysis of CsAGA2 in cucumber fruit. A, The expression of α-Gal genes (α-Gals) in fruit. B, Schematic showing the detailed sampling of cucumber fruit. C, The expression of CsAGA2 at different fruit developmental stages and in different parts of fruit at 9 DAA. Data are presented as the means of three biological replicates (±se) in (A and C). D–I, In situ hybridization of CsAGA2 in peduncle vascular bundle (D–F) and fruit MVB (G–I). E and H show a magnification of (D) and (G), respectively. The CCs are indicated by the arrows. The sense probe was hybridized as a control in (F) and (I). Scale bar: 100 μm. J–M, Immunogold localization of the CsAGA2 protein in the MVB of fruit. K and L, the signals in chloroplast and cytoplasm, respectively, through a magnified view of (J). The preimmunized serum was used as a control in (M). The arrowheads highlight the signals. Scale bar: 1 μm. DAA, days after anthesis; Ped, peduncle; Gy, gynophore; Ep, epicarp; MVB, main vascular bundle; Me, Mesocarp; Pl, placenta; EP, external phloem; IP, internal phloem; Xy, xylem; SE, seive element; CC, companion cell; PPC, phloem parenchyma cell.

A more detailed spatiotemporal expression pattern of CsAGA2 (Figure 2, B and C) showed that the mRNA levels of CsAGA2 increased sharply during fruit development from −2 d after anthesis (DAA) to 9 DAA, the marketable mature stage of fruit (Figure 2C). Among the different tissues of the fruit at 9 DAA (Figure 2B), CsAGA2 was highly expressed in the peduncle and fruit MVB, followed by mesocarp, placenta, gynophore, and epicarp (Figure 2C). Using in situ hybridization technology, the CsAGA2 transcript was found to be localized in the phloem region of the peduncle (Figure 2, D and E) and fruit MVB (Figure 2, G and H). Sense probes detected no signals (Figure 2, F and I).

To detect CsAGA2 protein expression, immunogold localization was performed. An antibody was generated using specific peptide fragments derived from CsAGA2, and antibody specificity was determined by western blot (Supplemental Figure S3). After labeling with the CsAGA2 antibody and gold-conjugated secondary antibody, the signal could be detected in the phloem region of the fruit MVB (Figure 2J), especially in the cytoplasm and mitochondria of CCs, but not of the sieve elements (SEs; Figure 2, K and L). The preimmune serum showed no detectable signals (Figure 2M). These results were consistent with the expression of CsAGA2 at the mRNA level (Figure 2), suggesting that CsAGA2 might function in sugar catabolism and phloem unloading in cucumber fruit.

Functional characterization of CsAGA2 by heterologous expression in Escherichia coli

The CsAGA2 protein was produced by expressing the CsAGA2 gene in E. coli., with CsGAL1 as a control to distinguish alkaline or acidic activities (Supplemental Figure S4). Enzymatic assays showed that CsGAL1 preferred raffinose as the substrate and that its activity was highest at pH 5.8, while CsAGA2 preferred stachyose and its activity was highest at pH 7.6 (Supplemental Figure S4, A and B), which was similar to the purified melon α-Gals (Gao and Schaffer, 1999; Carmi et al., 2003). The temperature response curve revealed that CsAGA2 and CsGAL1 attained maximal activity between 30°C and 35°C, which declined sharply at 40°C (Supplemental Figure S4C), also similar to the melon proteins (Gao and Schaffer, 1999). CsAGA2 and CsGAL1 activities increased with the extension of reaction time and reached the maximum level at 1 h (Supplemental Figure S4D). Using stachyose as the substrate, kinetic analysis revealed that the K_m value of CsAGA2 (Supplemental Figure S4, E and F) is quite different from melon (Gao and Schaffer, 1999) and watermelon AGA (Ren et al., 2021). These results demonstrated that cucumber CsAGA2 encodes a functional alkaline α-Gal.

Manipulating the expression of CsAGA2 alters the sugar metabolism in transgenic cucumber fruits

To investigate the function of CsAGA2, we generated multiple stable RNAi and OE transgenic lines. Two independent CsAGA2-RNAi (#11 and #29) and two CsAGA2-OE lines (#16 and #35) were selected for further analysis (Figure 3A). The results of RT-qPCR analysis and western blot indicated that the transcription and protein expression levels of CsAGA2 were substantially reduced in the CsAGA2-RNAi lines but increased in the CsAGA2-OE lines (Figure 3A). Compared with the WT plants, the RNAi lines had a smaller plant phenotype, were developmentally delayed, and produced smaller and lighter fruits, while the CsAGA2-OE lines yielded bigger and weightier fruits (Figure 3, B–E). Furthermore, we generated CsAGA2-edited plants using clustered regularly interspaced short palindromic repeat (CRISPR)–associated protein 9 (Cas9) technology (Supplemental Figure S5). The knockout mutant of CsAGA2 with a 3-bp deletion displayed slow growth with curly and shrunken leaves on the top of the plant, as well as abnormal development of female flowers (Supplemental Figure S5). This mutant died in about 30 d after planting. Overall, these results indicated that CsAGA2 is most likely indispensable and plays a pivotal role in plant growth and fruit development in cucumber.

Figure 3.

Sugar metabolism in the fruit of CsAGA2 transgenic lines. A, The relative levels of CsAGA2 mRNA (top) and CsAGA2 protein (bottom) in fruit from WT plants, RNAi lines (RNAi-11 and RNAi-29), and OE lines (OE-16 and OE-35) as determined by RT-qPCR and western blot, respectively. B–E, Phenotypic analysis of CsAGA2 transgenic lines. B, Plant height 30 d after sowing. C, Fruit phenotype at 9 DAA. The whole plants and fruits were digitally extracted for comparison. D and E, Length, perimeter (D), and weight (E) of fruit at 9 DAA. Scale bars: 20 cm in (B); and 5 cm in (C). F–J, The ¹⁴C-labeled total sugar content (F), single sugar content (G and H) and distribution percentage (I and J) in peduncle (F, G, and I) and fruit MVB (F, H, and J) at 9 DAA from the WT and CsAGA2 transgenic lines. K–N, The enzyme activities of AGA (K), SUS (L), and INV (M and N) in fruit MVB at 9 DAA. O–R, Immunohistochemical localization of CsSUS4 in the peduncle vascular bundle (O and P) and fruit MVB of WT plants (Q and R). Longitudinal sections with FITC as secondary antibody. The SP was stained with aniline blue in (O–R) and indicated by arrows. Scale bar: 100 μm. Three biological replicates are used for each experiment. Data are presented as the means of three biological replicates (±se). Significance compared to WT was determined by Student’s t test in (D–N); *P < 0.05; **P < 0.01. DAA, days after anthesis; MVB, main vascular bundle; Sta, stachyose; Raf, raffinose; Suc, sucrose; Glu, glucose; Fru, fructose; AGA, alkaline α-galactosidase; SUS, sucrose synthase; CWIN, cell wall invertase; INV, invertase; SE, seive element; SP, seive plate; CC, companion cell.

To further dissect CsAGA2-mediated RFO hydrolysis, ¹⁴CO₂ tracer experiments were carried out. In CsAGA2-RNAi lines, the total ¹⁴C sugar content in the peduncle and fruit MVB was significantly decreased at 24 h after ¹⁴CO₂ labeling when compared to the WT plants (Figure 3F). However, the RNAi lines retained a larger proportion of ¹⁴C-labeled RFOs (including stachyose and raffinose, up to 60.0% of total labeled carbohydrate) in the peduncle and fruit MVB (Figure 3, G–J), indicating that downregulation of CsAGA2 significantly reduced RFO hydrolysis along the phloem unloading pathway from the peduncle to the fruit MVB in cucumber. In the CsAGA2-OE lines, the plants accumulated much more [¹⁴C]sucrose and [¹⁴C]hexose (over 83.6% of total carbohydrate) in both the peduncle and fruit MVB (Figure 3, G–J), suggesting the OE of CsAGA2 could result in a substantial accumulation of ¹⁴C-sugars in fruit, especially monosaccharides, such as glucose and fructose.

As monosaccharides are the dominant sugar in cucumber fruit (Hu et al., 2009), the activities of sugar/RFO hydrolases, including AGAs, as well as two sucrose hydrolases, SUS and INV, were assayed (Figure 3, K–N). The results showed that AGA (Figure 3K) and SUS activities (Figure 3L) were significantly decreased in CsAGA2-RNAi lines and increased in CsAGA2-OE lines, respectively; while there was almost no significant difference between the transgenic and WT plants either in the acid (cell wall invertase) or neutral/alkaline INV (cytoplasmic invertase) activities (Figure 3, M and N). Our previous studies indicated that cucumber sucrose synthase 4 (CsSUS4) was highly expressed in most sink organs, but abundantly with fruit development at early stage (Fan et al., 2019b), especially in fruit phloem (Sui et al., 2018). Moreover, the CsSUS4 protein was localized in the CCs of cucumber fruit MVB by immune-gold labeling (Fan et al., 2019b). In this study, CsSUS4, like CsAGA2 and CsSWEET7a proteins (Li et al., 2021), was detected clearly in the CCs of peduncle vascular bundle and fruit MVB (Figure 3, O–R) by using a fluorescein isothiocyanate (FITC)-labeled secondary antibody and aniline blue counterstaining to the sieve plates of SEs adjacent to CCs. In summary, these results obtained here are consistent with those of our previous studies showing, on one hand, that AGA is a key enzyme for the initiation of the breakdown of RFOs along the phloem unloading pathway from the peduncle to the fruit (Li et al., 2021) and, on the other hand, that it is SUS, rather than INV, that plays the main role in sugar accumulation and fruit development of cucumber (Fan et al., 2019b; Li et al., 2021).

The feedback regulation on source leaves by CsAGA2

To determine if the source organs were affected after manipulating CsAGA2 gene expression, we observed the leaf morphology (Figure 4, A–C). Leaves of CsAGA2-RNAi plants were smaller and had a folded shape with fewer chloroplast granum lamellas than those in WT plants (Figure 4, A and B; Supplemental Figure S6, A–D), which probably reduced the efficiency of photosynthesis. Although the CsAGA2-OE lines had no obvious changes in leaf morphology compared to WT plants (Figure 4C), the size of chloroplast grana stacks was much larger than that in WT leaves (Supplemental Figure S6, E and F). The measurement of net photosynthetic rate in the leaves of WT and transgenic plants further indicated that the function and structure of leaves were compatible (Figure 4D). These results implied that the synthesis of assimilate in the source was most likely affected on account of the changes in CsAGA2 expression.

Figure 4.

CsAGA2 affects sugar phloem loading in cucumber leaf through feedback regulation. A–C, Morphology of leaves from WT (A), CsAGA2-RNAi (B) and CsAGA2-OE lines (C). Scale bar: 2 cm. D, Net photosynthesis of source leaves from WT and CsAGA2 transgenic plants. Samples from two RNAi lines (RNAi-11 and RNAi-29) and two OE lines (OE-16 and OE-35) and WT plants as control. CO₂ concentration, relative humidity, air temperature, and air flow rate inside the leaf chamber were maintained automatically by the equipment at 360 ± 10 μmol mol⁻¹, 50%–60%, 27°C ± 1°C, and 800 μmol s⁻¹, respectively. E–M, Transmission electron micrographs of minor veins (E–G) and the PD between MC and IC (H–J), as well as between IC and SE (K–M) from the WT and CsAGA2 transgenic plants. Asterisks in (E–G) and (H–M) indicated SEs and PD, respectively. Scale bars: 5 μm in (E–G), and 0.2 μm in (H–M). N and O, Number of PD at each cell interface between MC and IC (N), as well as between IC and SE (O) in the WT and CsAGA2 transgenic plants was counted respectively from high-resolution transmission electron images. P and Q, The ¹⁴C-labeled sugar content in leaves (P) and petiole (Q) from the WT and CsAGA2 transgenic plants. R–T, The enzyme activities of GolS (R), RS (S) and STS (T) in source leaves from the WT and CsAGA2 transgenic plants. U, Starch content in leaves from the WT and CsAGA2 transgenic plants. Data are presented as the means of three biological replicates (±se). Significance compared to WT was determined by Student’s t test in (D, N, O, and P–U); *P < 0.05; **P < 0.01; ***P < 0.001.MC, mesophyll cell; IC, intermediary cell; CC, companion cell; SE, seive element; Tr, tracheids cell; PPC, phloem parenchyma cell; Suc, sucrose; Raf, raffinose; Sta, stachyose; GolS, galactinol synthase; RS, raffinose synthase; STS, stachyose synthase; FW, fresh weight.

As cucumber mainly follows the polymer-trapping model of phloem loading of RFOs through plasmodesmata (PD; symplasmic pathway; Turgeon et al., 1993; Zhang and Turgeon, 2018; Ma et al., 2019), the minor vein structure and the density of PD in CsAGA2 transgenic cucumber leaves were further studied (Figure 4, E–O). We found to a certain extent that, there were more intermediary cells (ICs, a unique type of CC), CCs, SEs and tracheid cells (Tr) in the minor veins of CsAGA2-OE lines when compared to WT plants (Figure 4, E and G). Moreover, the frequency of PD between mesophyll cells (MCs) and ICs (Figure 4, H–J and N) as well as between ICs and SEs (Figure 4, K–M and O) were significantly increased in CsAGA2-OE lines but reduced in RNAi lines compared to WT plants, suggesting that perhaps the phloem loading of carbohydrates has been influenced by the manipulation of CsAGA2.

The ¹⁴C labeling experiments showed that the content of both [¹⁴C]RFOs, that is, [¹⁴C]stachyose and [¹⁴C]raffinose, and [¹⁴C]sucrose significantly decreased in leaves and petioles of CsAGA2-RNAi lines and increased in CsAGA2-OE lines when compared to WT plants (Figure 4, P and Q). Moreover, RFO synthase enzyme activities, such as the galactinol synthase (CsGolS), raffinose synthase (CsRS), and stachyose synthase (CsSTS; Sui et al., 2012; Lü et al., 2017; Ma et al., 2019), were reduced dramatically in CsAGA2-RNAi lines and increased in CsAGA2-OE lines (Figure 4, R–T). These results indicated that RFO synthesis and sugar export from the leaves of CsAGA2-RNAi lines were reduced to a large extent. Meanwhile, a large amount of starch accumulated in the leaves of RNAi plants, in sharp contrast to lower starch content in leaves of WT or CsAGA2-OE plants (Figure 4U;Supplemental Figure S7). Collectively, we speculated that the changes of fruit sink strength induced by CsAGA2 expression could lead to feedback regulation on source leaves in cucumber.

CsAGA2 mediates the dynamic balance of gene expression at the source and sink in cucumber

To analyze the disruption of sugar metabolism and transport across the whole plant and the changes in sink and source development, transcriptomic analysis was conducted for the WT, RNAi, and OE plants. Total RNA from mature leaves, peduncle, and dissected fruit MVB at 9 DAA stage was extracted and subjected to RNA-sequencing. Principal component analysis (PCA) divided the data from WT, CsAGA2-RNAi, and CsAGA2-OE samples into three groups (Figure 5A). Within the different tissues of the transgenic plants, 255–1,668 differentially expressed genes (DEGs) were upregulated, and 571–1,997 DEGs (P < 0.05, fold change ≥1.5) were downregulated compared to the WT tissues (Figure 5B). In the RNAi plants, there were 1,755 downregulated DEGs in the fruit MVB, the major region of phloem unloading (Sui et al., 2018), and the significantly enriched processes included sugar metabolism, as represented by categories like “carbohydrate metabolic process” (P = 3.7E-5) and “polysaccharide biosynthetic process” (P = 2.0E-4; Figure 5C). Within these metabolic processes, the CsAGA2 gene and the sucrose cleavage enzyme CsSUS4 (Csa5G322500; Figure 6), as well as cell wall biosynthesis- and regulation-related genes, including cellulose synthase genes (Polko and Kieber, 2019), COBRA-like genes (Li et al., 2019), and fasciclin-like arabinogalactan genes (Wang et al., 2015; Supplemental Table S1) showed reduced expression in the CsAGA2-RNAi lines. However, the INV genes were not differentially regulated (Figure 6). It has been reported that, with the continuous extension of the vascular bundle structure from the cucumber peduncle to gynophore and then fruit (Sui et al., 2018), the breakdown of RFOs begins at peduncle and extends down the fruit MVB (Ohkawa et al., 2010; Sui et al., 2018; Fan et al., 2019b). Consistently, co-downregulated DEGs (in total 578) in both the peduncle and the fruit MVB in RNAi plants (Figure 5D) were also enriched in sugar metabolism, especially polysaccharide metabolism (Figure 5E). On the one hand, these analyses showed that RFO catabolism in the peduncle and fruit in CsAGA2-RNAi plants was seriously hindered; on the other hand, polysaccharide biosynthesis for (primary/secondary) cell wall biogenesis, represented by hemicellulose, xylan, glucuronoxylan, cellulose, lignin, and pectin metabolism, was decreased synchronously and significantly.

Figure 5.

Transcriptome analysis in leaf, peduncle and fruit MVB of CsAGA2 transgenic plants. A, Principal component analysis (PCA) showing transcriptional relationships between different tissues in CsAGA2 transgenic cucumber plants. B, Number of differentially expressed genes (DEGs) in leaf, peduncle, and fruit MVB of the CsAGA2 transgenic plants compared to the WT. C, GO terms that were significantly enriched (P < 0.05) based on P-value among the significantly downregulated DEGs in fruit MVB from CsAGA2-RNAi lines (total number of genes is 1,755) compared to WT. I, II, and III indicate the processes of carbohydrate metabolism, cell wall-related sugar metabolism, and sugar signaling, respectively. D, Venn diagrams showing the number of significantly downregulated DEGs in both CsAGA2-RNAi peduncle and fruit MVB compared to WT. E, GO terms that were significantly enriched (P < 0.05) in downregulated DEGs in both CsAGA2-RNAi peduncle and fruit MVB (total number of genes is 578) compared to WT. I, II, and III indicate the processes of carbohydrate/polysaccharide metabolism, cell wall organization or biogenesis, and hormone signaling, respectively. The percentage of genes in a certain biological process (GO term) to total genome genes in the cucumber is “genome frequency,” represented by upper bars in (C and E). The number of significantly downregulated DEGs assigned a certain functional category divided by the total number of downregulated DEGs [1,755 genes in (C) or 578 genes in (E)] in each comparison between the tested tissue samples in CsAGA2-RNAi plants and WT is the “observed frequency,” represented by lower bars, respectively. Values on the right side of the bars are the number of genome genes or downregulated DEGs in CsAGA2-RNAi plants for each GO term. *P-value < 0.05; **P-value < 0.01. MVB, main vascular bundle; F_MVB, fruit MVB.

Figure 6.

Expression profiles of genes involved in the RFO biosynthesis and catabolism pathways in cucumber source and sink phloem from WT and CsAGA2 transgenic plants. Two transgenic lines (RNAi-11 and OE-16) were selected for study and WT plants as control. Numbers below the small squares indicate three biological replicates for each plant line (1–3 for WT, 4–6 for RNAi line, and 7–9 for OE line). Gene IDs or gene names are listed on the left or right of a row of small squares. Solid lines indicate metabolic flux, and dotted lines indicate transport stream. In source leaves, the TP is synthesized after light reaction and Calvin cycle, and then can be further catalyzed to F6P after a series of reactions. SPS catalyzes F6P and UDPG to synthesize S6P, and the presence of SPP leads to the formation of Suc in MCs. Suc is then transferred to ICs to synthesize Raf and Sta with Gala by RS and STS, respectively. Gala is synthesized in ICs by GolS. The Sta, Suc, and Raf are transported to sink tissues. After arrival in the fruit, the Sta and Raf are first hydrolyzed by α-Gal, like CsAGA2, in the peduncle and MVB to produce Suc and Gal. Suc is further hydrolyzed by SUS and/or INV into Fru and UDPG and/or Fru and Glu, respectively. UDPG and Gal are substrates of cell wall biosynthesis. Gene IDs: CsAGA2 (Csa4G167980), cucumber ALKALINE α-GALACTOSIDASE 2; CsSUS4 (Csa5G322500), cucumber SUCROSE SYNTHASE 4. MVB, main vascular bundle; LHCPs, light-harvesting chlorophyll a/b-binding proteins; Rubisco, Ribulose-1,5-bisphosphate carboxylase/oxygenase; rbcS, Rubisco small subunit; rbcL, Rubisco large subunit; ATP, adenosine triphosphate; NADPH, reduced nicotinamide adenine dinucleotide phosphate; ADP, adenosine diphosphate; Pi, orthophosphate; NADP⁺, nicotinamide adenine dinucleotide phosphate; TP, triose phosphate; F6P, fructose-6-phosphate; UDP-Gal, UDP-galactose (uridine diphosphate galactose); UDPG, UDP-glucose (uridine diphosphate glucose); Gala, galactinol; S6P, sucrose-6-phosphate; Suc, sucrose; Raf, raffinose; Sta, stachyose; Gal, galactose; Fru, fructose; Glu, glucose; GolS, galactinol synthase; SPS, sucrose phosphate synthase; SPP, sucrose-phosphatase; RS, raffinose synthase; STS, stachyose synthase; α-Gal, α-galactosidase; SUS, sucrose synthase; INV, invertase; FPKM, Fragments Per Kilobase of exon model per Million mapped fragments.

Remarkably, within these biological processes mentioned above, the overrepresented terms, including response to various sugars (Figure 5C) and response to hormone/auxin (Figure 5E), were significantly enriched amongst the downregulated DEGs in the peduncle and fruit MVB of the RNAi plants, suggesting that the sugar and auxin signaling pathways were most likely to be disrupted at the site of phloem unloading in fruit. Moreover, ultra-performance liquid chromatography (UPLC) followed by tandem mass spectrometry (MS/MS) assay further proved that the indoleacetic acid (IAA) level was reduced in RNAi peduncle and fruit MVB (Supplemental Figure S8A). Consistent with this result, the expression levels of the auxin biosynthetic gene YUCCA (CsYUC6), auxin efflux carriers PIN-FORMED (CsPIN1a/2) and some early auxin response genes, including 15 AUX/IAAs and 3 SAURs (SMALL AUXIN UP RNAs), were sharply decreased (Supplemental Figure S8, B–D; Supplemental Table S2).

In source leaves of CsAGA2-RNAi plants, gene ontology (GO) enrichment analysis showed that a large number of downregulated transcripts were associated with photosynthesis, sugar metabolism, and sugar signaling (Supplemental Figure S9), such as light-harvesting chlorophyll a/b-binding proteins (lhcps, Csa1G445860, Csa5G646700, and Csa6G522690) and ATP synthase genes (Csa1G011450 and Csa6G507210), and ribulose-1,5-bisphosphate carboxylase/oxygenase (Rubisco) large subunit (rbcL, CsGy7G003420) and Rubisco small subunit (rbcS, Csa5G609710), which are related to photosynthetic light and dark reactions, respectively, as well as CsSPS4 (Csa2G401440), CsGolS1 (Csa6G000080), CsRS (Csa3G838720), and CsSTS (Csa7G407800), which are involved in sucrose and RFO biosynthesis (Figure 6). These results suggested that the downregulation of CsAGA2 in the sink could induce negative feedback regulation of expression of genes related to photosynthesis, RFO synthesis and phloem loading in the source tissues.

Conversely, in the CsAGA2-OE lines, with the upregulation of sugar catabolism genes such as CsAGA2 and CsSUS4 in fruits, the mRNA levels of CsSPS4, CsGolS1, CsRS, and CsSTS in source leaves increased substantially (Figure 6). This suggested that photoassimilate synthesis in leaf tissue and phloem loading may be positively regulated through feedback about the sugar content and signals in the sink fruits of the CsAGA2-OE lines. Collectively, the above results indicated that there might be a feedback regulatory network for sugar/RFO metabolism with CsAGA2 playing a central role in cucumber.

Discussion

Most plants load and transport fixed carbon predominantly as sucrose through the phloem, but those in the Cucurbitaceae family such as cucumber mainly load and transport RFOs which are then hydrolyzed by α-Gal and sucrose hydrolase into hexoses in the sink organs (Ohkawa et al., 2010; Li et al., 2021, 2022). Accordingly, α-Gal is indispensable for RFO catabolism and phloem unloading and may play a key role in sink organ development. However, up to now, the exact function of α-Gal in cucumber fruits, an economically and nutritionally important sink organ was unknown. In this research, we revealed the central role of CsAGA2 during fruit development and for source–sink communication in cucumber.

CsAGA2 encodes an RFO hydrolysis enzyme that is active during cucumber fruit development

In most species, carbon fixed during the day is exported from the source organs in the form of sucrose and will be transported into sink organs such as young leaves, roots, flowers, fruits, and seeds. Generally, as a fleshy fruit develops, there is a continuous increase in sugar content, which requires the tissue to maintain its sink strength (Hu et al., 2009; Durán-Soria et al., 2020). In sink organs, the cleavage of sucrose can be reversibly converted by SUS or irreversibly catalyzed by INV (Koch, 2004). In those sucrose-translocating plants, such as Arabidopsis and rice, sugar export from the source leaves and carbohydrate accumulation/partitioning in the sink tissues are mediated to varying degrees by several sugar transporters and SUSs (Eom et al., 2011; Durand et al., 2018; Fan et al., 2019a). In tomato, antisense expression of sucrose synthase SuSy and sucrose transporter LeSUT2 caused the failure of fruit set or growth inhibition (Aoust et al., 1999; Hackel et al., 2006). In addition, the tomato cell wall invertase, Lin5, which was under severe selection during domestication, is involved in fruit development and contributes to the accumulation of sugar in fruit (Tieman et al., 2017). Therefore, the utilization of photoassimilates in various metabolic pathways depends on the cleavage of sucrose for most plants (Koch, 2004).

For the RFOs-translocating plants, the involvement of sucrose transporters and SUS in fruit development has been reported (Cheng et al., 2018; Ren et al., 2018; Fan et al., 2019b). In cucumber, the major sugars accumulated in the fleshy fruit are fructose and glucose, and these carbohydrates most likely come mainly from RFOs transported from source leaves through the phloem (Hu et al., 2009, 2011). However, it has remained unknown how the sink tissues hydrolyze RFOs and regulate subsequent sugar distribution. In our previous study, RFOs were predicated to be firstly hydrolyzed by α-Gals into sucrose and then further hydrolyzed by SUS and/or INV in CCs to maintain sink strength together with the sugar transporter CsSWEET7a (Fan et al., 2019b; Li et al., 2021). In this research, we identified an alkaline α-Gal highly expressed in cucumber fruit, which is homologous to ClAGA2 in watermelon (Ren et al., 2021), and named it CsAGA2. We further provide the molecular and physiological evidence that CsAGA2 plays a role in cucumber fruit development through a nutritional pathway. The CsAGA2 knockdown plants produced aborted fruits with backlogged RFOs and decreased hexoses in the peduncle and fruit MVB, while CsAGA2-OE lines yielded larger and heavier fruits with elevated glucose and fructose levels (Figure 3). Accordingly, the enzyme activity of SUS and the transcript levels of CsSUS4 were significantly changed in CsAGA2 transgenic fruits. The invertases did not show different expression or activity levels, further proving that SUS is more essential than invertase in fruit sugar accumulation of RFOs-transporting plants as previous research in sucrose-transporting plants (Katz et al., 2011). In addition, the transcript levels of CsSWEET7a were significantly decreased in the fruits of CsAGA2-RNAi lines, but increased in CsAGA2-OE lines, as compared to WT plants (Supplemental Figure S10). Overall, we propose that the rapid expansion in size and weight of cucumber fruits is most likely through rapid hydrolysis of RFOs by CsAGA2 and subsequent hydrolysis of sucrose by CsSUS4, which along with the phloem unloading of hexoses mediated by the sugar transporter CsSWEET7a in CCs (Fan et al., 2019b; Li et al., 2021).

The evolutionary history of sugar hydrolysis and sugar partitioning can provide important ways to improve the sugar content of fruit crops. In soybean (Glycine max), the domestication of the SWEET family of sugar transporters leads to dramatic changes of sugar content, seed size and oil content (Wang et al., 2020). Recently, the revelation of the evolutionary changes for the ClAGA2 and ClSWEET3 shows how these two genes have influenced fruit development in watermelon (Ren et al., 2021). In our previous study, it was clearly observed that the CsAGA2 protein co-localized with CsSWEET7a in CCs, showing a close cooperative relationship between them in sugar phloem unloading and fruit development (Li et al., 2021). Here, we found that the CsAGA2 and CsSWEET7a genes were located in two selective sweep regions (Figure 1). The Indian cucumber (wild species) contained less hexoses than an East-Asian cultivar (cultivated species), while the RFOs was only detectable in the fruit of the Indian cucumber, and the hexose contents were consistent with their respective mRNA levels of CsAGA2 and CsSWEET7a. Moreover, ¹⁴C-labelling experiment further showed that ¹⁴C-stachyose and ¹⁴C-raffinose accumulated in the peduncle and fruit MVB of the CsAGA2-RNAi lines but not in the WT and CsAGA2-OE fruits (Figure 3). These results supported the possibility that the evolutionary changes of CsAGA2 and CsSWEET7a led to increases in RFO hydrolysis and subsequent sugar unloading in the release phloem of fruits in domesticated cucumber cultivars.

CsAGA2 may participate in sugar and auxin signaling for fruit sink strength

Chemical components of the plant cell wall include the cellulose and hemicelluloses, the most high-molecular-weight polysaccharides, and pectic polysaccharides, polyphenolic lignin, and minor amounts of structural proteins (Burton et al., 2010). Here, some genes involved in cell wall biogenesis and organization were significantly downregulated in the fruit MVB of CsAGA2-RNAi plants, including the WALLS ARE THIN1 (WAT1) gene, cellulose synthase genes, and COBRA family genes (Supplemental Table S1). Mutation of WAT1 leads to a defect in cell elongation, a dwarfed habit and little or no secondary cell walls in fibers of Arabidopsis (Ranocha et al., 2010). Loss of cellulose synthase-like D (ZmCSLD1) function in maize (Zea mays) causes decreased cell division and smaller organs with fewer cell numbers (Li et al., 2018). Meanwhile, the mutant of the COBRA family gene Brittle Stalk2-Like3 exhibits a dwarf phenotype and decreased sucrose export along with significant cellulose deficiency (Julius et al., 2021). In Arabidopsis, Gal and uridine diphosphate galactose (UDP)-glucose (UDPG) were reported to function in the formation of cell wall polysaccharides and cell expansion via cellulose biosynthesis (Haughn and Chaudhury, 2005; Lee et al., 2007; Poschet et al., 2010). UDPG, the substrate for cellulose synthesis, may come from the reactions catalyzed by either SUS or both INV and UDPG pyrophosphorylase (McFarlane et al., 2014). Accordingly, overexpression of OsSUS3 in rice could enhance cellulose biosynthesis in the primary cell wall (Fan et al., 2019a). In this study, the downregulation of genes related to cell wall biogenesis in CsAGA2-RNAi cucumber is likely due to the decreased availability of Gal and UDPG for polysaccharide formation, since the enzyme activities of AGA and SUS were significantly reduced. All these results indicated that CsAGA2 is most likely a major player in delivering sugars as nutrients for cell wall biosynthesis, which also then affects sink strength.

Interestingly, GO analysis showed significant enrichment of genes in the auxin signaling pathway among the downregulated DEGs in both peduncle and fruit MVB of CsAGA2-RNAi plants (Figure 5). And changes of sugar metabolism triggered the reduction in IAA level, together with downregulated expression of CsYUC6, CsPIN1a/2, CsAUX/IAAs, and CsSAURs in CsAGA2-RNAi plants (Supplemental Figure S8; Supplemental Table S2). Disruption of rice OsSAUR33 caused reduced germination rates, low seed uniformity and increased sugar-sensitive phenotypes during early seed germination, suggesting that OsSAUR33 possibly affects seed vigor through the sugar pathway (Zhao et al., 2021). Interaction between glucose and auxin signaling could control Arabidopsis root growth and development by mediating the expression of auxin signaling components (Mishra et al., 2009). An exogenous supply of glucose could reactivate target of rapamycin (TOR) activity and promote rapid growth of the Arabidopsis primary root (Yuan et al., 2020). Glucose-TOR energy signaling is important for maintaining a low auxin response region within the elongation zone of the root, which could contribute to promote cell expansion by regulating the expression of the auxin efflux facilitator PIN2 (Yuan et al., 2020). Remarkably, genes in the auxin-related pathway were significantly enriched in the phloem from the cucumber peduncle to the fruit (Sui et al., 2018). Combined with these previous studies, we speculated that low glucose levels caused by silencing of CsAGA2 in cucumber may disrupt auxin synthesis, distribution, and signaling at the site of phloem unloading, which then influences fruit development.

CsAGA2 mediates source–sink communication in cucumber

Communication between the sink (site of import of fixed carbon) and source (site of export of photoassimilates) is a vital factor for plant growth, and this process is regulated by a complex signaling network that includes sugars, hormones, and environmental conditions. There is no doubt that the rate of photosynthesis in the leaf needs to be strictly regulated to reach its optimal level according to the carbohydrate status of the plant. Feedback regulation of photosynthesis by photoassimilates has been widely reported in sucrose-translocating plants (Ainsworth and Bush, 2011). For example, elevated sink demand increases the photosynthetic efficiency in sugarcane (Saccharum spp.; McCormick et al., 2006). Recently, it was reported that the knockout of ClAGA2 in watermelon, a RFOs-translocating plant, results in a lower sugar content in fruit, accompanied by starch accumulation in source leaves (Ren et al., 2021). In this study, folded leaves showed a decline in photosynthetic activity as well as downregulation of RFO biosynthesis-related gene expression and enzyme activity when fruit sink demand for carbohydrate was limited in CsAGA2-RNAi plants. In contrast, increased photoassimilate production and elevated sugar export were detected in source tissues when sink demand was higher in CsAGA2-OE plants. We speculated that this is the result of feedback—negative in RNAi plants or positive in OE lines—related to CsAGA2-mediated changes in sink strength in cucumber.

The number and frequency of PD can affect the rate and mode of phloem loading (van Bel, 1993). In sucrose-translocating plants, mutants defective in carbohydrate partitioning had abnormal vein structure and inhibited sucrose export (Julius et al., 2018; Tran et al., 2019). For example, the maize sucrose export defective1 mutant showed aberrant PD structure, which led to symplasmic interruption and lack of phloem loading capability (Russin et al., 1996). Similarly, the tie-dyed2 mutant of maize also displayed incomplete vascular differentiation and disruption of sucrose movement in the phloem (Slewinski et al., 2012). Most recently, the aberrant vascular cell development and reduced sucrose transport were found in the rice sem1-1/1-2 mutants, which were two allelic mutations in a gene encoding a GLUCAN SYNTHASE-LIKE protein for callose biosynthesis (Wang et al., 2021). In this study, the symplasmic loading pathway in source leaves was influenced as the number of PD changed between MCs and ICs as well as between ICs and SEs. The polymer-trapping model depends on the ability of PD to allow sucrose to enter the ICs but precludes active transport of the larger raffinose and stachyose molecules in the opposite direction (Zhang and Turgeon, 2018). Our results provided a possible explanation that the fruit sink strength could adjust photosynthesis and affect the symplasmic phloem loading efficiency, which then might lead to the subsequent decreases in sugar export from CsAGA2-RNAi or elevations in sugar export from CsAGA2-OE leaves. However, further study is required to determine if and how the source leaves sense sugar signaling or use other pathways to evoke the changes in photosynthesis and phloem loading.

Materials and methods

Plant materials and growth conditions

Monoecious WT plants and transgenic cucumber plants (Cucumis sativus L. “Xintaimici”) were grown for two generations per year (from February to July or from the end of August to December) in a greenhouse under natural sunlight at the China Agricultural University, Beijing. Plants were individually grown in soil. Experiments were conducted on the T1 generation of transgenic plants. Cucumber plants from different geographic groups, including the India group, the Eurasian group, the Xishuangbanna group, and the East-Asian group, were grown in a greenhouse under natural sunlight at the Chinese Academy of Agricultural Sciences, Beijing.

Sequence alignment and molecular phylogenetic analysis

The PCR primers used for cloning CsAGA2 are listed in Supplemental Table S3. Sequence alignment and phylogenetic analysis were performed based on sequences from related α-Gal proteins from other species retrieved using BLAST analysis in Phytozome (http://www.phytozome.net/search.php) or the Arabidopsis Information Resource (http://www.Arabidopsis.org) databases, with the deduced amino acid sequence of cucumber α-Gals. The multiple sequence alignment of cucumber α-Gal and related α-Gal proteins was performed using CLUSTALW within the MEGA version 6 software package. The phylogenetic tree was constructed by the neighbor-joining method with the Poisson model and 1,000 bootstrap replicates using MEGA version 6. The evolutionary distances are in the units of the number of amino acid substitutions per site. Sequence accession numbers are listed in Supplemental Table S4.

Heterologous protein expression in Escherichia coli

The full-length CsAGA2 CDS was cloned into pMAL-C2X vector (to create MBP fusion protein) using homologous recombination, and the recombinant plasmid was then transformed into Escherichia coli BL21 for heterologous protein expression. When the culture reached an optical density₆₀₀ (OD₆₀₀) of 0.6–0.8, 0.2-mM isopropyl-β-d-thiogalactoside (IPTG) was added and shaken at 16°C for 16 h. The culture was centrifuged at 5,000 rpm for 8 min at 4°C, and then the cells were resuspended in 10 mL MBP column buffer [20-mM NaCl, 20-mM Tris–HCl, 1-mM EDTA, 1-mM DTT (pH = 7.4), 1-mM PMSF]. The transformed cells were disrupted via sonication on ice for 15 min. The recombinant protein was purified using 500-µL MBP buffer containing magnetic beads (NEB, Beijing, China), and eluted with 75-µL MBP buffer containing 10-mM maltose (Solarbio, Beijing, China).

Enzymatic characteristic analysis

The substrate specificity of the α-Gals was checked using stachyose or raffinose. The optimum temperature ranged from 0°C to 100°C, the time course ranged from 0 to 90 min, and pH ranged from 2 to 10 in tests using 10-mM raffinose or 10-mM stachyose as the substrate. The K_m and V_max values were confirmed by Eadie–Hofstee after using 250-μM raffinose or 250-μM stachyose as the substrate.

Expression analysis by RT-qPCR

Total RNA was extracted from specified tissues [peduncle, gynophore, fruits from 2 d before anthesis to 9 DAA, and fruit MVB], using the Quick RNA Isolation Kit (HUAYUEYANG, Beijing, China). Sample cDNA was synthesized using the FastKing gDNA dispelling RT SuperMix kit (TIANGEN, Beijing, China). RT-qPCR was performed using SYBR^® Premix Ex Taq (Vazyme Biotech Co., Ltd, Nanjing, China), with an Applied Biosystems 7500 real-time PCR system (Applied Biosystems, Waltham, MA, USA). The cucumber α-TUBULIN (TUB) gene was used as the internal control to normalize the expression data. Three biological replicates using independently prepared RNA/cDNA templates and three technical replicates for each combination of cDNA samples and primer pairs were performed. The gene-specific primers are listed in Supplemental Table S3.

In situ hybridization assay

The experiment was conducted as described (Sui et al., 2017), with some changes. At anthesis, the peduncle and young fruit MVB were excised and fixed with FAA [70% (v/v) ethyl alcohol, 5% (v/v) acetic acid, and 2% (v/v) formalin], embedded in paraffin, sectioned using a microtome, and hybridized with the probes. Digoxigenin-labeled sense and antisense RNA probes were generated by PCR amplification using SP6 and T7 RNA polymerase (Roche, Basel, Switzerland). The primers are listed in Supplemental Table S3.

Immunogold labeling

The experiment was conducted as previously described with minor modification (Sui et al., 2017). Briefly, the fruit MVBs were embedded in resin after double fixing with glutaraldehyde and OsO₄. The samples were cut into a 60-nm thickness and caught with a 200-mesh nickel net. The ultrathin sections were first incubated with rabbit antiserum directed against CsAGA2 (diluted 1:500) of cucumber and then with a secondary antibody (10-nm gold conjugated goat anti-rabbit IgG, diluted 1:200). After staining with uranyl acetate-lead citrate, the sections were examined with an electron microscope (JEM-100S, Japan).

Western blot

For protein extraction, approximately 0.5-g cucumber fruit was ground in liquid nitrogen and homogenized in 150-μL extraction buffer [50-mM Tris–HCl (pH 8.0); 1% (v/v) Triton X-100; 150-mM NaCl; 10-mM MgCl₂; 5- mM EDTA (pH 8.0); 1-mM PMSF; 2% (v/v) β-mercaptoethanol]. After centrifugation at 13,000 g for 30 min at 4°C, a total of 40 μg of protein was separated by sodium dodecyl sulfate polyacrylamide gel electrophoresis (SDS-PAGE) and transferred to membrane as previously described (Nie et al., 2021).

Immunolocalization

Immunolocalization of CsSUS4 was performed as revealed in Sui et al. (2017) and Li et al. (2021). Briefly, sample pretreatment including fixation, dehydration, embedding, and section cutting was done using the same method as for in situ hybridization. The sections were then blocked and incubated with a secondary antibody [goat anti-rabbit IgG–FITC, Jackson, USA, diluted 1:200] after incubation with CsSUS4 (diluted 1:200) primary antibody. Samples were incubated with 0.1% (w/v) aniline blue for 20 min in dark conditions after immunolabeling. Images were obtained under an Olympus fluoview FV1000 confocal laser scanning microscope with visualization at excitation/emission wavelength of 488/510 nm.

Transformation of cucumber

To generate the CsAGA2-RNAi transgenic cucumber plants, 195 nucleotides of the CsAGA2 coding sequence were amplified by PCR using primers containing AscI and SwaI restriction sites, and SpeI and BamHI restriction sites, respectively. The two fragments were inserted into the pFGC1008 vector in the sense and antisense orientations. For CsAGA2-OE plants, the full-length coding sequences of CsAGA2 were amplified with specific primers. The fragments were introduced into the pCAMBIA1300 expression vector after digestion with HindIII and SpeI enzymes. For the generation of CRISPR/Cas9-edited plants of CsAGA2, the specific small guide RNA (sgRNA) target sites were designed by the sgRNA design website (http://crispr.hzau.edu.cn/CRISPR2/). The PCR fragment harboring target site was amplified using four partially overlapping primers and then inserted in the binary CRIPSR/Cas9 vector pKSE402G using BsaI endonuclease and T4 Ligase (New England Biolabs, Ipswich, MA, USA). The resulting CsAGA2-RNAi, CsAGA2-OE, and CsAGA2-CRISPR/Cas9 constructs were then introduced into Agrobacterium tumefaciens strain EHA105 by chemical transformation. Cucumber “Xintaimici” was transformed using a cotyledon transformation approach, as previously described (Hu et al., 2017), with some changes. The primers are listed in Supplemental Table S3.

Determination of sugar and starch content

The samples were ground and extracted with 5 mL 80% (v/v) ethanol for 30 min at 80°C, and the extraction was repeated 3 times. The extracts were pooled and dried at 40°C in an oven, and then decolorized with active carbon. The residues were re-dissolved in 1-mL distilled water and were filtered through the 0.45-μm filter. A high-performance liquid chromatography (HPLC) (Agilent 2100 system; Palo Alto, CA, USA) was used to analyze the sugar content. The retention times and peak heights of standard sugars (Sigma) were used to identify and quantify the eluted sugars. The precipitate was used to extract the starch. The residue was incubated with 10 mL of 30% (v/v) perchloric acid for 15 min and centrifuged at 3,500 g for 10 min. After 3 times centrifugation, the combined supernatants were analyzed with anthrone-H₂SO₄ reagent, and starch content was calculated by the produced glucose content.

Starch staining

After 18 h of dark treatment, the leaves were decolorized in 80% (v/v) ethanol. After staining with 10% (w/v) I-KI solution for 30 min, the samples were washed twice with water for 5 min each time to remove excess stain. The stained leaves were photographed using a stereomicroscope (Leica S8 APO, Germany) and a digital camera (Nikon D7000, Japan).

Enzyme activity assays

The total protein extract was extracted from 0.6 g FW mashed cucumber samples in the buffer containing 50-mM HEPES/NaOH (pH 7.0), 2 mM MgCl₂, 1-mM EDTA, 2% (w/v) polyvinylpyrrolidone (PVP-K30), 2% (w/v) polyethylene glycol (PEG 20,000), 0.1% (v/v) Triton X-100 and 1-mM DDT. Samples were centrifuged at 18,000 g at 4°C for 30 min. The supernatant was then dialyzed at 4°C for 16 h against a buffer containing 25-mM HEPES-NaOH (pH 7.0) and 1-mM DTT. The supernatant was collected for alkaline invertase activity, and the pellet was then resuspended again with 1.5-mL extraction buffer and used for cell wall invertase activity measurement as described in Fan et al. (2019b). The alkaline α-Gal activity reaction mixture contained 10-mM stachyose, 100-mM HEPES buffer (pH 7.4), 80-μL enzyme extract, and lasted 30 min at 37°C, then the next experiment was done as reported by Li et al. (2021). The SUS activity was measured in a mixture containing 80-mM MES (pH 5.5), 5-mM NaF, 100-mM sucrose, 5-mM UDP, and 0.5-mL enzyme extract, and detailed operation was done as described previously by Li et al. (2021). The GolS, RS, and STS activities were measured as follows: Briefly, in addition to 40-μL enzyme extract, the enzyme activity reactions contained 5-mM MnCl₂, 10-mM UDP-Gal, and 40-mM myo-inositol for GolS; or 10-mM galactinol and 40-mM sucrose for RS; or 10-mM galactinol and 40-mM raffinose for STS. The respective reaction mixture was incubated at 30°C for 3 h. The reactions were stopped in boiling water for 10 min. The product was determined using HPLC platform (Agilent 2100 system: Palo Alto, CA, USA).

¹⁴C-labeling experiment

The ¹⁴C-labeling experiment of cucumber was done according to Ma et al. (2019) and Sui et al. (2017). Briefly, when fruits (adjacent to leaf samples) at 9 DAA, the petiole of leaf samples was wrapped to prevent photosynthesis. Then, a sealable plastic bag was attached to those leaves and fed ¹⁴CO₂ generated from NaH¹⁴CO₃ and excess 80% (v/v) lactic acid for 30 min under sunlight. The redundant ¹⁴CO₂ was cleared by 3-M KOH. The leaves were exposed to ambient air for another 24 h after removal of the bag. Leaf, petiole, peduncle, and fruit MVB samples were put in liquid nitrogen, ground, and soluble sugar was extracted. Thin Layer Chromatography was used to separate the sugars on silica plates (Merck, Shanghai, China). The solvent system contained acetic acid, chloroform and water (7:6:1 in vol.). Maximum separation was achieved after running the plates 3 times. The radiolabeled spots were localized with X-ray film (Kodak Biome MR film, Rochester, USA). The spots were removed/cut out and placed in scintillation fluid. The percentage of each scraped radiolabeled sugar was calculated as the proportion of total sugars in the Ecoscint scintillation solution (National Diagnostics, Atlanta, GA, USA).

Transmission electron microscopy

Mature leaves were sampled and fixed in 2.5% (w/v) glutaraldehyde, then washed with 0.1-M phosphate buffer (pH 6.8). The samples were then dehydrated through an acetone gradient (30%, 50%, 70%, 80%, 90%, and 100%, v/v) after fixing with 1% (w/v) osmic acid, washed in 0.1-M phosphate buffer (pH 6.8) again, and embedded in Spurr’s resin. Thin cross-sections (6 μm) were made with a UC6I microtome (Leica) and examined with a JEM-123O scanning transmission electron microscope. High-resolution transmission electron microscope images were obtained to count the number of PD, and frequencies were calculated in per interface between MCs and ICs, as well as between ICs and SEs.

Net photosynthesis measurement

Gas exchange was measured using a LI-6400XT portable photosynthesis system (Li-Cor, Inc., Lincoln, NE, USA) equipped with an infrared gas analyzer (IRGA, 6400-02B). Cucumber plants were first adapted to the light intensity conditions required by the instrument (usually 1 h prior to the measurement), and then the attached leaves were prepared and immediately put into the leaf chamber. CO₂ concentration, relative humidity, air temperature, and air flow rate inside the leaf chamber were maintained automatically at 360 ± 10 μmol mol⁻¹, 50%–60%, 27°C ± 1°C, and 800 μmol s⁻¹, respectively.

RNA-Seq sample preparation and sequencing

The peduncle and fruit MVB at 9 DAA stage, as well as mature leaf samples were collected from the same node of WT and transgenic cucumber plants and immediately placed into liquid nitrogen. Three biological replicates using the specified tissues from independent WT and transgenic plants were collected for each sample. RNA was prepared using RNA Extraction Kit (HUAYUEYANG, Beijing, China). Sequencing was done by the Illumina Novaseq 6000 platform at Biomarker Technologies (BioMarker, Beijing, China). Bioinformatic analysis of RNA-Seq data was performed at BioMarker Cloud platform (www.biocloud.net). The clean reads were mapped to the cucumber reference genome using HISAT2 and StringTie software. Fragments Per Kilobase of transcript per Million fragments mapped (FPKM) was used to show transcript levels. Differently expressed genes were obtained by DEseq2 with fold change ≥1.5 and a false discovery rate (FDR) <0.05. GO enrichment analysis was displayed on the GOseq R package of the BioMarker Cloud platform.

Measurement of IAA content

To measure the levels of endogenous IAA, about 0.5 g of the peduncle and fruit MVB at 9 DAA from the WT and transgenic lines were harvested. Then, the extraction and quantification of auxin were performed using UPLC–MS/MS methods according to Nie et al. (2021).

Statistical analysis

Statistical analysis was conducted by a two-tailed Student’s t test with equal variance using Excel. Data presented are means ± standard error (se) of three independent experimental replicates.

Accession numbers

The accession numbers of all genes used in this article were obtained from the Cucurbit Genomics Database (Cucumber Chinese long genome v2 http://cucurbitgenom ics.org/organism/2) or the NCBI (https://www.ncbi.nlm.nih.gov/) and are listed in Supplemental Table S4.

Supplemental data

The following materials are available in the online version of this article.

Supplemental Figure S1. Sequence analysis of cucumber α-Gal proteins.

Supplemental Figure S2. Phylogenetic tree based on amino acid sequences showing relationships between α-Gals.

Supplemental Figure S3. Specificity analysis of anti-CsAGA2 antibody.

Supplemental Figure S4. Enzymatic analysis of CsAGA2 and CsGAL1 expressed in Escherichia coli.

Supplemental Figure S5. Identification of CsAGA2 mutant line based on CRISPR/Cas9 technology in cucumber.

Supplemental Figure S6. Structural changes in chloroplasts of cucumber leaves from wild type (WT) and CsAGA2 transgenic plants.

Supplemental Figure S7. Starch grain staining in leaves from wild type (WT) and CsAGA2 transgenic plants.

Supplemental Figure S8. CsAGA2 transgenic plants contained different auxin levels.

Supplemental Figure S9. GO terms that were significantly enriched (P < 0.05) in significantly downregulated DEGs (differently expressed genes) in CsAGA2-RNAi leaves compared to WT.

Supplemental Figure S10. Expression analysis of CsSWEET7a in CsAGA2 transgenic lines.

Supplemental Table S1. Downregulated DEGs (differently expressed genes) related to cell wall biosynthesis and regulation in fruit MVB of CsAGA2-RNAi plants compared to WT.

Supplemental Table S2. Downregulated DEGs (differently expressed genes) related to auxin signaling in both peduncle and fruit MVB of CsAGA2-RNAi plants compared to WT.

Supplemental Table S3. List of primers used in this study.

Supplemental Table S4. List of accession numbers used in this study.

Funding

This work was supported by the National Key Research and Development Program of China (2019YFD1000300), the National Natural Science Foundation of China (31972398), the Beijing Innovation Consortium of Agriculture Research System (BAIC01), the China Agriculture Research System of MOF and MARA, and the 111 Project of Ministry of Education of P.R.C. (B17043).

Conflict of interest statement. The authors declare that they have no conflicts of interest.

 

X.S., H.L., and T.L. conceived the project, designed the experiments, and arranged the article framework. H.L., X.L., Y.Z., J.N., X.Y.¹, L.L., N.M., and Y.G. performed most of the experiments. J.Y. and Y.L. provided technical assistance to H.L. H.L., X.S., T.L., and X.Y.² analyzed the data. H.L., T.L., and X.S. wrote the article. T.L. and X.S. agree to serve as the authors responsible for contact and ensure communication.

The author responsible for distribution of materials integral to the findings presented in this article in accordance with the policy described in the Instructions for Authors (https://academic.oup.com/plphys/pages/general-instructions) is Xiaolei Sui (suixiaolei@cau.edu.cn).