2016 Volume 85 Issue 1 Pages 23-29
In citrus, fruit bearing affects floral induction and the nutritional condition of the tree. For this reason, bearing too many or too few fruit causes fluctuation in flower number the following spring. This in turn leads to annual alternation between rich and poor crops, known as alternate bearing. To identify the metabolites related to alternate bearing, quantitative metabolomics analysis was conducted with stem tissues of vegetative shoots collected in November. Twelve Satsuma mandarin trees bearing different amounts of fruit were used in this study. Fruit weight per leaf area of these trees was significantly and negatively correlated with the expression of a flowering-related gene, citrus FLOWERING LOCUS T, in the stem in November and the number of flower buds the following spring. In metabolomics analysis, adenosine triphosphate was detected at high concentrations in lightly fruiting trees. Other coenzymes such as uridine triphosphate, nicotinamide adenine dinucleotide phosphate, and ascorbic acid were also more abundant in the off-crop trees. In addition, the off-crop trees accumulated sugar phosphates such as fructose 6-phosphate, glucose 6-phosphate, and ribulose 1,5-diphosphate. Furthermore, heavily fruiting trees accumulated more amino acids. These results indicate that fruit bearing affects the metabolism of coenzymes, sugars, and amino acids in the stem of vegetative shoots.
In the citrus industry, fluctuation of fruit production is one of the main problems because it leads to instability of fruit quality, fruit supply, and price. The change in annual fruit production is caused by suppression of floral induction and nutritional accumulation by fruit bearing. When a tree bears an excessive amount of fruit in one year (on-year), it produces a light bloom the next year followed by low fruit production. Similarly, after a year of light fruit production (off-year), the tree flowers profusely the next year, followed by heavy fruit production. Once trees bear a low or high fruit load, they start to alternate between high and low crops in successive years.
Since alternate bearing has an economic impact on the industry, many studies have been conducted to identify the mechanism behind this phenomenon and to try to prevent it. Starch accumulation is one factor affected by fruiting condition (Sugiyama et al., 2006). Trees bearing an excessive amount of fruit barely accumulate any starch; however, it is unclear whether this deterioration of nutritional condition correlates with flowering and fruit production the following year.
In citrus, flower differentiation is physiologically induced by a low temperature during the fall and winter. It has been indicated that citrus FLOWERING LOCUS T (CiFT) plays an important role in the flowering of citrus and that fruit bearing suppresses floral induction via the suppression of CiFT expression (Nishikawa, 2013; Nishikawa et al., 2007, 2012). In recent years, transcriptomic and proteomic analyses have been conducted regarding alternate bearing in citrus (Muñoz-Fambuena et al., 2013a, b; Shalom et al., 2012, 2014). In transcriptomic analysis, it was demonstrated that the expression of genes involved in abscisic acid (ABA) metabolism and auxin polar transport was altered in buds by fruit removal (Shalom et al., 2014). In proteomic analysis, it was shown that proteins involved in carbohydrate and amino acid metabolism were up-regulated in off-crop buds and that the largest groups of proteins up-regulated in on-crop buds were related to primary metabolism, oxidative stress, and defense responses (Muñoz-Fambuena et al., 2013b). However, comprehensive knowledge of the metabolite status in on- and off-crop trees is indispensable to understand the mechanism causing the alternate bearing.
In 2003, Soga et al. (2003) reported that a method of capillary electrophoresis–mass spectrometry (CE-MS) is available for the simultaneous quantification of metabolites. This method enables quantification of a wide range of compounds such as carbohydrates, organic acids, amino acids, and nucleotides. It was expected that a large number of metabolites would be quantified simultaneously in citrus and that the analysis with CE-MS would give critical information to clarify the mechanism of alternate bearing.
In this study, we conducted metabolomics analysis by CE-MS and identified metabolites closely related to fruit load in citrus. The results of this study showed that several coenzymes and sugar-related metabolites were abundant in the off-crop trees and that many kinds of amino acid accumulated in the on-crop trees. Changes in metabolism in alternate bearing are also discussed.
Twenty-year-old ‘Aoshima’ Satsuma mandarin trees grafted on trifoliate orange [Poncirus trifoliata (L.) Raf.] rootstock were grown in a field at the National Institute of Fruit Tree Science (NIFTS), Kuchinotsu (Minami-shimabara, Nagasaki, Japan). These trees had been harvested at different times in 2009 and had varying flower numbers in the spring of the year of the experiment (2012). They were subjected to customary fertilization and pesticide application. Among these trees, twelve with different amounts of fruit were chosen for this study; some of the trees used in the experiment bore fruit alternately and some did so continuously. From each tree, approximately five vegetative shoots were picked in mid-November 2012, of which time floral induction occurs. The vegetative shoots were divided into stems and leaves. Stem tissue from the five vegetative shoots was mixed and then frozen. All stem samples were homogenized with a Shake Master [Bio Medical Science Inc. (BMS), Tokyo, Japan] and stored at −80°C until use for gene expression, sugar, and metabolomic analyses. In December 2012, all of the fruit were harvested and weighed for each tree. After harvest, the canopy volumes for each tree were measured using a scale. Leaf area density was measured using a plant canopy analyzer (LAI2000; Li-Cor Environmental, Lincoln, NE, USA), and the leaf area was calculated according to the method described by Iwaya et al. (2005). The next spring, about 20 shoots were selected at random from each tree. The numbers of flower buds, nodes, and sprouting nodes were counted for each shoot, and the number of flower buds per sprouting bud and the number of sprouting buds per node were calculated.
For real-time quantitative reverse transcription (qRT)-PCR analysis, total RNA was extracted from stem tissues with an RNeasy Plant Mini Kit (Qiagen, Hilden, Germany) and cleaned by on-column DNase digestion. qRT-PCR reactions were performed with 0.4 μg of purified total RNA and a random hexamer at 37°C for 2 h using a High Capacity cDNA Reverse Transcription Kit (Applied Biosystems, Foster City, CA, USA).
A TaqMan MGB probe (Applied Biosystems) and sets of primers for amplification of total CiFT were designed using Primer Express software (Applied Biosystems) (Nishikawa et al., 2007). A TaqMan Ribosomal RNA Control Reagents VIC Probe (Applied Biosystems) was used as an endogenous control. TaqMan real-time PCR was performed with a TaqMan Universal PCR Master Mix (Applied Biosystems), using an ABI PRISM 7000 Sequence Detection system (Applied Biosystems) according to the manufacturer’s instructions. Each reaction contained 900 nM primers, 250 nM TaqMan MGB Probe, and 2.5 μL of template cDNA. The thermal cycling conditions consisted of: 95°C for 10 min, followed by 40 cycles at 95°C for 15 s and 60°C for 60 s. Standard curves were generated for CiFT using serial DNA dilutions. Relative gene expression levels were analyzed with ABI PRISM 7000 Sequence Detection System Software (Applied Biosystems) using the CT value and normalized to the expression levels of 18S ribosomal RNA. Real-time qRT-PCR was performed in three replicates for each sample, and the mean of the logarithmic value was used in the data analysis.
Sugar extraction and its chromatographic analysis were performed following the method described by Matsumoto and Ikoma (2012) with some modification. To determine the extraction efficiency, 100 μL of a surrogate standard solution (an aqueous solution of norvaline, 1.875 g·L−1) was added to the stem powder (ca. 50 mg) homogenized using a homogenizer (Shake Master; BMS, Tokyo, Japan). About 2 mL of 80% EtOH was added to the homogenate and placed at 80°C for 30 min. Then, the sample was adjusted to 10 mL in a volumetric flask and shaken vigorously. The extract was filtered through 0.2-mm filter paper (EMD Millipore, Bedford, MA, USA), and 100 μL of the aliquot of the filtrate was collected with a gastight syringe and evaporated using a centrifugal evaporator to dryness. The dry residue was redissolved in 100 μL of water, loaded onto an anion exchange (SAX) cartridge (Discovery® DSC-SAX, 1 mL; Sigma Aldrich, St. Louis, MO, USA), and eluted with 250 μL of water. The eluent (100 μL) was mixed with 250 μL of acetonitrile, and 5 μL of the extract was analyzed using liquid chromatography/tandem mass spectrometry (LC-MS-MS) with Agilent 1100 series HPLC (Agilent Technologies, Santa Clara, CA, USA) and an AB SCIEX API2000 triple-stage quadrupole tandem mass spectrometer (AB SCIEX, Foster City, CA, USA) (Matsumoto and Ikoma, 2012).
Fifty milligrams of homogenized stem powder was transferred into 500 μL of methanol containing 50 μL of an external standard solution. After additional homogenization with BMS-M10N21 (BMS) at 1500 rpm, four times for 120 s, 500 μL of chloroform and 200 μL of ultra-pure water were added to the homogenate and mixed well. The samples were then centrifuged at 2300 × g for 5 min at 4°C. The resultant water phases were ultrafiltered with a Millipore Ultrafree-MC PLHCC HMT Centrifugal Filter Device, 5 kDa (Millipore, Billerica, MA, USA). The filtrates were dried and dissolved in 100 μL of ultra-pure water.
The samples were diluted two or 2.5 times and subjected to a CE-MS experiment using the Agilent CE-TOFMS system (Agilent Technologies, Santa Clara, CA, USA) at 4°C. Cation and anion analyses were performed under the conditions described by Soga and Heiger (2000) and Soga et al. (2002, 2003). Cationic metabolites were detected with a fused silica capillary (50 μm i.d. × 80 cm length) with Cation Buffer Solution (Human Metabolome Technologies, Tokyo, Japan). The sample was injected at a pressure of 50 mbar for 10 s. The capillary voltage was set at 4000 V, and the spectrometer was scanned from m/z 50 to 1000.
In the anion analysis, a fused silica capillary (50 μm i.d. × 80 cm length) and an Anion Buffer Solution (Human Metabolome Technologies) were used. The sample was injected at a pressure of 50 mbar for 25 s, and the applied voltage was set at 30 kV. Electrospray ionization mass spectrometry (ESI-MS) was conducted in the negative ion mode, and the capillary voltage was set at 3500 V. The spectrometer was scanned from m/z 50 to 1000.
The areas of the peak detected by CE-MS were used in quantification of the compound. Metabolites were identified according to the m/z and migration time.
In this study, 12 trees bearing different amounts of fruit were used to investigate the relationship between fruit weight per leaf area and metabolites. Our samples showed a strong correlation between fruit weight per leaf area and CiFT gene expression in November, exhibiting a power approximation (Fig. 1). The relationship between fruit load and number of flower buds per sprouting bud the following April showed a strong negative linear correlation (Table 1; Fig. 1). The correlation between the fruit load and the number of sprouting buds per node was lower than with CiFT expression or the number of flower buds (Table 1).
Relationship between fruit weight per leaf area at harvest and CiFT mRNA level (left panel) or number of flower buds per sprouting bud (right panel). From each tree, vegetative shoots were picked in November and stem tissues were used for CiFT mRNA quantification. Real-time qPCR was performed with gene-specific probe and primers. Data of the gene expression are given as the mean (logarithmical value) ± SE (n = 3). The following spring, vegetative and reproductive buds were counted in each tree. Data of the number of flower buds are given as the mean ± SE (n = 20).
Correlation coefficients between fruit load, CiFT expression, and the number of flower buds or sprouting buds.
The contents of fructose, glucose, and sucrose were quantified by LC-MS-MS. They were negatively correlated with the fruit load, with the correlation coefficients being between −0.65 and −0.68 (Table 2).
Correlation coefficients between sugars and fruit load.
In the metabolomics analysis, 154 metabolites were quantified by CE-MS. The correlation coefficients for the linear approximation between fruit weight per leaf area and each metabolite were calculated and compared. Among the 154 metabolites, 57 showed a significant correlation with fruit weight per leaf area (P < 0.05). Of the 57 metabolites, 20 were negatively correlated and 37 were positively correlated with fruit weight per leaf area (Tables 3 and 4).
Metabolites showing significant negative correlation coefficients with fruit weight per leaf area.
Metabolites showing the significant positive correlation coefficients with fruit weight per leaf area.
Figure 2 shows the relationship between fruit weight per leaf area and 5 metabolites showing the highest positive and negative correlation coefficients for the linear approximation with fruit weight per leaf area. Some of the metabolites showed higher correlation coefficients for the power or log approximation than those for the linear approximation (Fig. 2). Among the up-regulated metabolites in the off-crop trees, adenosine triphosphate (ATP) showed the highest negative correlation with fruit load (Table 3; Fig. 2). Other coenzymes such as uridine triphosphate (UTP), nicotinamide adenine dinucleotide phosphate (NADPH, NADP+), adenosine diphosphate (ADP), and ascorbic acid also accumulated in the off-crop trees (Table 3). Other than acting as coenzymes, several sugar phosphates such as fructose 6-phosphate, glucose 6-phosphate, ribulose 1,5-diphosphate, and sucrose 6-phosphate had a tendency to accumulate in the off-crop trees (Table 3). Some other metabolites in sugar/starch metabolism such as gluconolactone, 3-phosphoglyceric, and gluconic acids were also detected at higher levels in the off-crop trees (Table 3). Thus, coenzymes and metabolites from sugar/starch metabolism accumulated in the stem of the off-crop trees.
Correlation between fruit weight per leaf area and metabolites showing the strongest positive (left panels) or negative (right panels) correlation with fruit weight per leaf area. Metabolites were analyzed for stem tissues collected in November. Regression line or curve was applied in each graph.
In the on-crop stems, many amino acids such as histidine, glycine, threonine, tyrosine, serine, alanine, and asparagine accumulated (Table 4). Of the 20 amino acids encoded directly by the genetic code, cysteine, proline, glutamine, and tryptophan showed correlation coefficients between 0.2 and −0.2 with fruit weight per leaf area (data not shown). Among the amino acids accumulated in the on-crop stems, non-proteinogenic amino acids such as 2-aminoisobutyric acid, citrulline, 4-aminobutyric acid (GABA), β-alanine, and ornithine were also included. In addition, some polyamines such as putrescine and spermine accumulated in the on-crop trees.
In our experiments, the accumulation of CiFT mRNA was suppressed in fruiting trees and closely related to the fruit load and number of flower buds the following spring (Table 1); these results corroborated previously reported data (Nishikawa et al., 2012). The suppression of CiFT expression by fruit bearing indicates that CiFT expression is correlated to alternate bearing. Recent studies have attempted to identify the mechanism of alternate bearing through proteomic or transcriptomic analysis (Muñoz-Fambuena et al., 2013a, b; Shalom et al., 2012, 2014). However, further studies are needed to understand fully the alternate bearing in citrus. In this study, the changes in low-molecular-weight metabolites were quantified and the relationship between metabolites and fruit weight per leaf area was investigated.
Generally, plants accumulate energy in the form of starch. It is well known that starch accumulates in off-crop citrus trees. Our research that was conducted with the same trees in 2010 and 2011 showed that the correlation coefficients between starch in the stems of Satsuma mandarin vegetative shoots in November and fruit weight per leaf area at harvest were on average −0.75 (unpublished data). In this study, fructose, glucose, and sucrose were closely correlated with fruit load, although these correlations were weaker than between starch and fruit load. In the present metabolomic analysis, several metabolites showed correlation coefficients of less than −0.75 with fruit load. Among the up-regulated metabolites in off-crop trees, ATP showed the highest negative correlation with the fruit load. This suggests that energy also accumulated in the form of ATP in the off-crop vegetative shoots. ATP is synthesized in photosynthesis, glycolysis, and through respiration in plants. It has been reported that fruit bearing does not affect photosystem II (Syvertsen et al., 2003), implying that ATP accumulation in off-crop trees does not result from the activation of photosystem II. It is thought that ATP accumulation in off-crop trees can be caused by a reduction of its consumption. In plants, ATP is consumed during the assimilation of CO2. Monerri et al. (2011) have reported that fruit load does not affect CO2 fixation. On the other hand, Syvertsen et al. (2003) have reported the reduction of CO2 assimilation in off-crop trees. ATP might accumulate in off-crop trees through the reduction of CO2 fixation.
Generally, ATP is used by many enzymes and affects the rates of their reactions. It is also involved in the amino acid activation in protein synthesis. In addition, it is used in signal transduction pathways by kinases that phosphorylate proteins and lipids in the cell. Recent research has suggested that extracellular ATP functions as a signaling agent (Tanaka et al., 2014). Thus, ATP has several functions and it is thought that its level affects a wide range of metabolic processes. Fruit bearing may affect tree growth and development through regulating the level of ATP.
Besides ATP, several other coenzymes also accumulated in the off-crop trees. Those coenzymes can be bound to the enzymes and, similar to ATP, directly participate in various reactions. It is speculated that various enzyme reactions related to coenzymes that accumulated in off-crop trees proceed smoothly in the stems of non-bearing trees, but not in those with a heavy fruit load.
Sugar phosphates also accumulated in the off-crop trees. They included substrates of the Calvin cycle, glycolysis, and the pentose phosphate pathway. In the Calvin cycle, CO2 is incorporated into sugar phosphates and then the assimilated CO2 is transformed to sucrose. In trees with strong sink tissue such as fruit, sugar phosphates may be swiftly transformed into sucrose, and then sucrose may be promptly transported to fruit. As a result, the amount of sugar phosphates will diminish in the source tissues. By contrast, in the off-crop trees, the presence of a smaller amount of sink tissues may lead to the accumulation of sugar phosphates as well as starch in the stem. It has been reported that the expression of some of the genes and proteins involved in the Calvin cycle is up-regulated in off-crop trees (Muñoz-Fambuena et al., 2013b; Shalom et al., 2012). Therefore, it is suggested that fruit bearing affects not only gene expression or protein levels but also the levels of substrates in the Calvin cycle. Although it is thought that the accumulation of substrates leads to an increase of CO2 fixation in the Calvin cycle, it has been reported that there is no increase of CO2 fixation in off-crop trees (Monerri et al., 2011; Syvertsen et al., 2003). Thus, it seems that the accumulation of substrates and the up-regulation of gene expression or proteins involved in the Calvin cycle do not enhance CO2 fixation in off-crop trees.
Many amino acids accumulated in fruit-bearing trees. They included both proteinogenic and non-proteinogenic amino acids; proteinogenic amino acids are incorporated into proteins and non-proteinogenic ones are those not naturally encoded or found in the genetic code of any organisms. The accumulation of both amino acid types in fruiting trees suggests that fruit bearing affects the amino acid or protein metabolism. Basically, amino acids are synthesized in pathways derived from glycolysis and the citric acid cycle and fruit bearing may affect these pathways. It is also thought that the increase in proteinogenic amino acids may result from the enhancement of protein breakdown or the suppression of protein synthesis. It is known that amino acids are used to synthesize other molecules. For example, methionine is a precursor of ethylene and phenylalanine is a precursor of various phenylpropanoids, which are important in plant metabolism. It has been reported that some of the amino acids have a non-protein role in plants. GABA, which is a non-proteinogenic amino acid, has been suggested to be a signal molecule involved in the responses to stress and in carbon:nitrogen metabolism (Bouché and Fromm, 2004). In watermelon, it has been demonstrated that citrulline is an efficient hydroxyl radical scavenger (Akashi et al., 2001). Thus, some non-protein amino acids are involved in stress responses and some might play a role in the response to fruit bearing in on-crop trees.
Polyamines are involved in a wide range of plant processes, including cell division, morphogenesis, and stress responses (Moriguchi, 2004). In pistachio, it has been reported that free polyamines could have an important physiological function in the development of flower bud abscission, which causes alternate bearing in pistachio trees (Roussos et al., 2004). In olive, it has been shown that conjugated forms of both spermidine and spermine accumulated at significantly higher concentrations in ovaries and leaves during the ‘on’ compared with the ‘off’ year (Pritsa and Voyiatzis, 2005). In this study, several polyamines accumulated in the stem of on-crop trees. Polyamines might thus have a role in the suppression of flowering by fruit bearing in citrus.
In the alternate bearing of citrus, it has been suggested that gibberellin, abscisic acid, and auxin are related to the suppression of flowering by fruit bearing (Koshita et al., 1999; Takagi et al., 1989; Shalom et al., 2014). In the present metabolomic analysis, the conditions did not suit the detection of phytohormones and the levels of those phytohormones in our samples were unclear.
In conclusion, we analyzed 154 low-molecular-weight metabolites in the stem of vegetative shoots in relation to the fruit weight per leaf area of the tree. Fruit load was negatively correlated with several coenzymes and metabolites in sugar/starch metabolism. In addition, fruit weight per leaf area was positively correlated with various kinds of amino acid. Our results suggest that fruit bearing affects metabolic processes involving coenzymes, sugars, and amino acids.