Hevea leaves at four progressive developmental stages (Fang et al., 2016), i.e., bronze (I), color-change (II), pale green (III), and bright green (mature; IV) were sampled at 10 a.m. from untapped 4-year-old tissue-cultured plants of the cultivar Reyan7-33-97 that were cultivated in the experimental plantation of the Rubber Research Institute, Chinese Academy of Tropical Agricultural Sciences (Danzhou, Hainan, China). Three individual trees with similar girth and growth vigor were chosen, each of which was used as a separate biological repeat.

Leaf soluble proteins were extracted according to the method of Walker and Huber (1989) with slight modifications. Fresh leaves (0.2 g) were ground using a mortar and pestle in 2 mL ice cold extraction buffer containing 50 mM HEPES-NaOH (pH 7.5), 1 mM EDTA, 5 mM Mg²⁺, 10 mM β-mercaptoethanol, 0.05% (v/v) Triton X-100 and 1% PVP40 (w/v). The homogenate was centrifuged at 20,000 g for 10 min. The supernatant was desalted and made free of soluble sugars on a Sephadex G-25 column equilibrated with extraction buffer minus the Triton X-100 and PVP40 prior to four enzyme assays, viz. VIN, NIN, Sus, and SPS.

To extract CWI, the centrifugation pellet was suspended in 100 μL of 1 M NaCl and 400 μL of extraction buffer, and kept at 4°C overnight. The suspension was centrifuged at 20,000 g for 10 min with the supernatant used for the CWI assay.

Protein concentrations of enzyme extract were determined according to the method of Bradford (1976).

Invertase activities were determined as described previously with modifications (Tupý, 1969; Miron and Schaffer, 1991). Assays were tested in a reaction mixture of 100 μL containing 30 μL enzyme extract and 100 mM sucrose in the reaction buffer (50 mM HEPES, 5 mM Mg²⁺ and 1 mM EDTA). The pHs of the reaction buffers were adjusted to 7.5, 4.5, and 3.5, respectively, for the three invertases: NIN, VIN and CWI. The mixture was chilled on ice immediately after incubation at 37°C for 1 h and then adjusted to 1mL by adding 900 μL of 3, 5-dinitrosalicylic acid reagent (DNS), followed by 5 min of incubation in a boiling water bath. The blank control contained 30 μL of boiled enzyme extract. Contents of reducing sugars produced in the reaction mixture were measured using dinitrosalicylic acid reagent in conformity with Miller (1959). Enzyme activity was expressed in units per mg of protein, where one unit (U) was defined as 1 μmol of sucrose cleaved per hour under the experiment conditions.

The activity of SPS was determined by measuring the sucrose (plus sucrose-6-P) formed in the assay mixture (100 μL) that contained 30 μL enzyme extract, 5 mM UDPG, and 40 mM fructose-6-P in the reaction buffer (50 mM HEPES, 5 mM Mg²⁺ and 1 mM EDTA, pH 6.5) (Huber, 1981). After incubation at 37°C for 1 h, 100 μL of 1 M NaOH was added to terminate the reaction and destroy the unreacted hexose simultaneously by placing the tubes in boiling water for 10 min. The blank differed from the assay mixture by lacking UDPG, and it was kept on ice prior to the addition of NaOH. After chilling on ice, 250 μL of 0.1% (w/v) resorcinol in 95% ethanol and 750 μL of 30% HCl were added to each tube, followed by incubation at 80°C for 10 min. The tubes were then centrifuged at 20,000 g for 5 min, and the OD520 was measured by a Spectramax 250 microplate reader (Molecular Devices, Sunnyvale, CA, United States). Enzyme activity was expressed in units per mg of protein, where one unit (U) was defined as 1 μmol of sucrose produced per hour.

Sus catalyzes a reversible reaction that both hydrolyzes and synthesizes sucrose. Sus activity in the synthesis direction was assayed as for SPS except that the pH of reaction buffer was changed to 8.5, and fructose-6-P substituted with fructose. Sus activity in the cleavage reaction was assayed as for invertase except that the pH of reaction buffer was changed to 6.5, and 5 mM UDP was supplemented. The blank control contained no UDP. Enzyme activities were expressed in units per mg of protein, where one unit (U) was defined as 1 μmol of sucrose produced or cleaved per hour.

In line with standard practices in enzymology, blanks are expected to contain substrates as well. However, leaving out the substrates UDPG (SPS, SSS) or UDP (SSC) from the blanks did not alter the OD₅₂₀ readouts (Supplementary Table S1). Accordingly, HPLC-ELSD analysis of the UDPG and UDP substrates showed complete absence of sucrose and fructose (Supplementary Figure S1).

Leaves were desiccated at 105°C for 30 min and then at 80°C to constant weight. The dried leaves were ground into powder using a mortar and pestle in room temperature, and kept in a dryer prior to use. Soluble sugars were extracted according to Ma et al. (2014) with modifications. Approximately 100 mg was sonicated in 800 μL double-distilled water (ddH₂O) for 30 min and then incubated at 80°C for 30 min with intermittent shaking for every 5 min. The mixture was centrifuged at 4°C at 20,000 g for 10 min. The supernatant was kept at -20°C, whereas the pellet was subjected to a second round of extraction with 200 μL of ddH₂O. The supernatant pooled from the two rounds of extraction was dried under vacuum at 60°C, then dissolved in 80% ethanol (v/v) and kept at -20°C overnight. The suspension was centrifuged at 4°C in 20,000 g for 10 min. The supernatant was dried again, dissolved in 50% acetonitrile (v/v), and stored at -20°C before use.

To prepare phloem exudate, 1-year-old Hevea branches bearing mature leaves were excised with a sharp knife, and the cuts were cleaned with sterile water, and then subjected to the extraction of phloem exudate following the protocol of King and Zeevaart (1974). The exudates obtained were lyophilized, and dissolved in 50% acetonitrile (2 mL per branch), and stored at -20°C before use.

The high-performance liquid chromatography (HPLC)-evaporative light scattering detection (ELSD) method was used to determine soluble sugars in Hevea leaves and phloem exudate. HPLC analysis was performed on a Waters e2695 separations module (Waters, Milford, MA, United States) equipped with an Alltech 3300 ESLD detector (Alltech, Deerfield, IL, United States). Separation was achieved on a XBridge^TM Amide column [4.6 mm × 250 mm i.d., 3.5 μm particle size (Waters, Milford, MA, United States)]. All samples and standards were filtered through 0.45 μm Millipore filters before loading samples of 10 μL onto the machine. The HPLC-ELSD conditions were optimized following Ma et al. (2014) with a solvent ratio of 85 acetonitrile:15 water (v/v), a flow rate of 1 mL/min, the column and drift tube temperatures set at 45 and 82°C, respectively, and the nebulizer gas flow rate set at 2 L/min. Peaks were quantified using calibration standards of HPLC grade sugars, viz. glucose, fructose, sucrose, raffinose, and stachyose (Sigma–Aldrich, Shanghai, China), and sorbitol and quebrachitol (Shanghai Yuan ye Bio-Technology Co., Ltd., Shanghai, China).

Starch extraction and content assay in Hevea leaves were performed following the procedures of the Total Starch Assay Kit (Megazyme, Bray, Ireland).

Total RNA of Hevea leaves was extracted using the Plant Total RNA Extraction Kit (BioTeke Corporation, Beijing, China) following the manufacturer’s instructions. RNA integrity was checked by agarose gel electrophoresis while RNA concentration and purity were examined using a NanoDrop 2000 spectrophotometer (Thermo Fisher Scientific, Wilmington, DE, United States). RNA samples were prepared DNA-free, and reverse transcribed to cDNA using the PrimeScript^TM First Strand cDNA Synthesis Kit (TaKaRa, Dalian, China) following the manufacturer’s protocol.

The RNA-Seq data of Hevea leaves at four developmental stages (I to IV) in our previous study (Fang et al., 2016) were divided into two sets: developing (stage I to III) and mature (stage IV). The two data sets were exploited to determine the expression in FPKM (Fragments Per Kilobase of transcript per Million fragments mapped) of sucrose-metabolizing genes in developing and mature leaves by using the RSEM software (Li and Dewey, 2011).

Specific PCR primers for the main sucrose-metabolizing genes (Supplementary Table S2) were designed by Primer Premier 5, and their specificity was confirmed by agarose electrophoresis and sequencing of the PCR products. Quantitative real-time PCR (qPCR) was carried out using the SYBR Premix Ex Taq II (2×; Tli RNaseH Plus; TaKaRa, Dalian, China) on a CFX96 Touch^TM Real-Time PCR detection system (Bio-Rad, Hercules, CA, United States). Quantification cycle values, PCR efficiency, correlation coefficients and the relative fold change of expression were determined as previously described (Long et al., 2015, 2016). The Hevea YLS8 gene was selected as an internal control based on the results of our previous work (Li et al., 2011).

All statistical analyses were conducted using SPSS Statistics 22. Duncan’s test was used to compare sugar concentrations, enzyme activity, and gene expression in Hevea leaves across the four developmental stages. Differences were accepted as significant at P < 0.05. Correlation coefficients were determined between activity of sucrose-metabolizing enzymes and soluble sugar contents or gene expression in the course of Hevea leaf development. Correlation was accepted as significant at P < 0.05 or 0.01.

In the process of Hevea leaf development, leaf colors changed from bronze (stage I) to color change (stage II), pale-green (stage III), and then to bright green (stage IV), identical to our previous report (Fang et al., 2016). Leaf size (cm²) increased rapidly from stages I (9.6 ± 0.6) to III (83.0 ± 10.6), and remained unchanged thereafter (Table 1). On the contrary, specific leaf weight (mg fresh weight per cm²) decreased progressively from stages I (12.5 ± 1.0) to III (9.8 ± 1.7), but resumed to a high value of 12.9 ± 1.5 in mature leaves (stage IV) (Table 1).

To determine soluble sugar composition in Hevea leaves and branch phloem exudates, the HPLC-ELSD system was used. Six sugars, namely glucose, fructose, sorbitol, sucrose, raffinose, and stachyose that are reported to be abundant or translocated in most plant species (Zamski and Schaffer, 1996) served as reference standards, together with quebrachitol which is the predominant non-rubber carbon compound in Hevea latex (Bealing, 1981). Raffinose and stachyose, easily separated from the five other sugars by the HPLC-ELSD method (Supplementary Figure S2A), were not detected in Hevea leaves and phloem exudate. Glucose, fructose, sucrose, sorbitol, and quebrachitol were well separated in 28 min by the optimized HPLC-ELSD method (Supplementary Figure S2B). Fructose, glucose, and sucrose turned out to be the main soluble sugars in Hevea leaves although their relative abundance varied with leaf development (Supplementary Figures S3A–D). Interestingly, Hevea leaves contained also small amounts of quebrachitol, but no sorbitol. Hevea phloem exudate from branches exhibited a profile of sugar composition similar to that of the leaf (Supplementary Figure S3E), except that sucrose was in far greater abundance than the other sugars, consistent with a previous assumption of sucrose being the major transported carbohydrate in Hevea (Tupy, 1989).

Changes in leaf contents of glucose, fructose, sucrose, and starch during leaf development are depicted in Figure 1. Starch accumulated in the course of leaf development, with a 2.5-fold increase from stage I to stage IV. At all leaf stages, however, starch was much lower than the other three sugars. In mature leaves (stage IV), starch comprised only 5.7% of total sugars, whereas sucrose accounted for 63.5%. Glucose and fructose had comparable concentrations except for stage III where the content of fructose was almost double that of glucose. Concentrations of fructose and glucose increased progressively with leaf development, and their combined content was more than 1.5-fold that of sucrose, but then decreased abruptly at maturation to less than 0.5-fold that of sucrose at the same stage (Figure 1). The content of sucrose remained almost unchanged from stage I to II, and then increased noticeably at stage III, before peaking at maturation (stage IV).

All sucrose-cleaving enzymes, i.e., VIN, NIN, CWI, and Sus in the cleavage direction (SSC) showed comparable activities in developing leaves (stages I to III) that were significantly higher than in mature leaves (stage IV) (Figure 2). Higher activity of sucrose-cleaving enzymes corresponded well to a higher hexose content in developing leaves (Figure 1). Examined in detail, the four sucrose-cleaving enzymes exhibited obvious variation in activity. VIN activity increased almost two-fold between stages I and II, and then decreased progressively to stage IV, reaching a level of only one fourth of its highest at stage II (Figure 2A). NIN activity remained constantly high throughout the three developing stages (I to III), but decreased significantly by 40% at maturation (stage IV) (Figure 2B). CWI and SSC activities shared a similar pattern of variation, both of which increased progressively with leaf development until stage III, and then declined abruptly at stage IV by 50% and 80%, respectively (Figures 2C,D). In mature leaves, NIN and CWI activities were significantly higher than those of VIN and SSC. Interestingly, sucrose-synthesizing enzymes, i.e., SPS and Sus in the synthesis direction (SSS) behaved similarly to sucrose-cleaving enzymes by displaying higher activity in developing leaves than in mature ones (Figures 2E,F). Correlations between activities of sucrose-metabolizing enzymes and the soluble sugar content were investigated in Hevea leaves during their development (Table 2). CWI activity correlated positively with fructose content at a significant level, whereas SSC and SSS activities showed significant positive correlation with glucose content. Interestingly, NIN activity exhibited a significant negative correlation with sucrose content, suggesting that the extent of its cleavage by NIN influenced its presence in the leaf.

To investigate the transcriptional regulation of sucrose-metabolizing enzymes and their roles in Hevea leaf development, the principal encoding genes expressed in developing and mature leaves were first identified among respective gene families of sucrose-metabolizing enzymes (VIN, NIN, CWI, SPS and Sus) using the RNA-Seq data (Figure 3). The main members thus identified from different gene families, viz. HbVIN1 to 3, HbNIN1, 2, 6 and 8, HbCWI1 to 3, HbSus2 to 5, and HbSPS2 and 3 were explored for their expression dynamics with leaf development by qPCR analysis (Figure 4).

Expressions of all three HbVIN genes in Hevea leaves decreased significantly between stages III and IV (Figure 4A), corresponding well to an abrupt depression in VIN activity at stage IV (Figure 2A). However, the three HbVIN genes exhibited variation during the three developing leaf stages (I to III), with a constant rise for HbVIN2 but different fluctuations for HbVIN1 and 3. Except for HbNIN8 that accumulated transcripts throughout the leaf development, the other three HbNIN genes analyzed, namely HbNIN1, 2 and 6 showed consistently higher expressions in developing leaves (stages I to III) than in mature leaves (stage IV) (Figure 4B), consistent with the higher NIN activity observed in developing leaves (Figure 2B). All three HbCWI genes displayed high expressions at stage III, but expression decreased significantly at stage IV in HbCWI1 and 3 (Figure 4C); this trend was consistent with the changes in CWI activity (Figure 2C). The four HbSus genes displayed different patterns of expression during the four developmental leaf stages (Figure 4D). Expressions of both HbSus2 and 3 increased constantly from stages I to III. At stage IV, HbSus3 decreased abruptly to a barely detectable level whereas HbSus2 increased further. Expression of HbSus4 was the highest at stage I, and maintained high levels through stages II and III, but decreased significantly to a very low level at stage IV. HbSus5 expression was much higher at late stages (III and IV) than the early stages (I and II). Both HbSPS2 and 3 exhibited high gene activity in mature leaves at stage IV (Figure 4E), despite the lowest SPS enzyme activity observed at this stage (Figure 2F). Correlation analyses revealed that among the 16 sucrose-metabolizing related genes explored, expressions of only three genes, viz. HbNIN2, HbCWI1 and HbSus3, correlated significantly with the activities of their respective enzymes (Table 3).