Changes in carbohydrate metabolism in Citrus sinensis infected with ‘Candidatus Liberibacter asiaticus’


  • J. Fan,

    1. Key Laboratory of Biorheological Science and Technology, Ministry of Education, College of Bioengineering, Chongqing University, Chongqing 400030, China;
    2. Citrus Research and Education Center, Institute of Food and Agricultural Sciences, University of Florida, Lake Alfred, FL 33850, USA
    Search for more papers by this author
  • C. Chen,

    1. Citrus Research and Education Center, Institute of Food and Agricultural Sciences, University of Florida, Lake Alfred, FL 33850, USA
    Search for more papers by this author
  • R. H. Brlansky,

    1. Citrus Research and Education Center, Institute of Food and Agricultural Sciences, University of Florida, Lake Alfred, FL 33850, USA
    Search for more papers by this author
  • F. G. Gmitter Jr.,

    Corresponding author
    1. Citrus Research and Education Center, Institute of Food and Agricultural Sciences, University of Florida, Lake Alfred, FL 33850, USA
    Search for more papers by this author
  • Z.-G. Li

    1. Key Laboratory of Biorheological Science and Technology, Ministry of Education, College of Bioengineering, Chongqing University, Chongqing 400030, China;
    Search for more papers by this author



Changes in carbohydrate metabolism in sweet orange (Citrus sinensis) infected with ‘Candidatus Liberibacter asiaticus’, a purported cause of citrus Huanglongbing (HLB), were investigated. Starch levels in HLB-infected leaves with and without symptoms increased 3·1- and 7·9-fold, respectively, compared to healthy controls. In symptomless leaves, sucrose and fructose accumulated significantly (< 0·05) in both midribs and lobes, and glucose only in the midribs (greater than fivefold); whereas maltose levels were reduced to 64·6% and 86·8% in the midribs and foliar lobes, respectively, of the values in healthy leaves. In leaves with symptoms, sucrose and glucose remained at high levels compared to healthy leaves, whilst no accumulation of fructose was observed; by contrast, the maltose content decreased to as low as 49·6% of that in healthy leaves. Fourfold induction of cell-wall-bound invertase activity was detected in both types of leaves on diseased plants. Additionally, the expression profiles of starch breakdown genes suggested that the transcription of DPE2 and MEX1 was down-regulated. Together with the reduction of maltose, it is suggested that the impairment of starch breakdown may contribute to the starch accumulation in infected leaves. The imbalance of carbohydrate partitioning and its relation to disease physiology are discussed.


Candidatus Liberibacter’ spp., the presumed causal agent of citrus Huanglongbing (HLB) or citrus greening, includes three species of Gram-negative, phloem-inhabiting α-proteobacteria, namely ‘Ca. Lasiaticus’, ‘Ca. L. africanus’ and ‘Ca. Lamericanus’. ‘Candidatus Liberibacter’ is transmitted by two species of phloem-feeding citrus psyllids, Diaphorina citri and Trioza erytreae, and is now considered the most devastating citrus disease in the world (Bove, 2006; Brlansky & Rogers, 2007). The disease affects most citrus species and no cure is currently available. HLB symptoms include yellow shoots, leaf blotchy mottle, and lopsided fruits with colour inversion and aborted seeds. Furthermore, starch accumulation (Schneider, 1968) and phloem damage (Kim et al., 2009) are also observed in infected citrus plants. Some citrus species and relatives, such as lemon and trifoliate orange, show less severe symptoms and much slower decline than others (Folimonova et al., 2009). Although substantial research efforts have been made to detect the presumed pathogen, little is known about the physiological mechanisms of this disease. Recently, the genome sequence of ‘Ca. L. asiaticus’ has been determined by a metagenomics approach, using DNA extracted from a single ‘Ca. L. asiaticus’-infected psyllid (Duan et al., 2009). Since no toxins, extracellular degrading enzymes or specialized secretion systems were found in the genome, it is more likely ‘Ca. L. asiaticus’ causes host metabolic imbalances by nutrient depletion or interference of transportation, in turn resulting in disease symptoms (Duan et al., 2009). Two reports on host gene expression in response to HLB infection have demonstrated that many genes involved in carbohydrate metabolism are differentially regulated in the infected citrus leaves (Albrecht & Bowman, 2008; Kim et al., 2009). However, an extensive analysis of carbohydrate changes is lacking.

Sucrose is generally the major end product of photosynthetic carbon metabolism and represents, in most plants, the predominant carbohydrate transported in the phloem sieve tubes from mature leaves to sink organs such as young leaves, roots, flowers and fruits (Zimmermann & Ziegler, 1975). Its use relies on hydrolysis into hexoses (glucose and fructose) by specialized enzymes such as invertase (EC There are a range of invertases in higher plants differing in subcellular localization, solubility in buffer, and pH optimum. Cell-wall-bound invertase (acidic), soluble acidic invertase in vacuoles, and soluble neutral/alkaline invertase in the cytosol are the three main isoforms. It is well known that invertase is involved in the breakdown of sucrose and plays a pivotal role in sugar partitioning in plants (Kingston-Smith et al., 1999). Specifically, the induction of cell-wall-bound invertase activity has frequently been shown in response to pathogen infection (Kocal et al., 2008).

The objective of this study was to investigate the effect of HLB on carbohydrate partitioning in sweet orange (Citrus sinensis) by comparing the concentrations of soluble carbohydrates and starch. Furthermore, invertase assays were performed and expression profiles for starch breakdown-related genes determined.

Materials and methods

Plant materials

Ten greenhouse-grown young and healthy sweet orange cv. Madam Vinous plants were used. Five plants were graft-inoculated by grafting with budwood from HLB-affected sweet orange and the other five were grafted with budwood from HLB-free sweet orange. The presence of ‘Ca. L. asiaticus’, the bacterium associated with HLB, was confirmed by quantitative real-time PCR as described by Li et al. (2006). Yellowing and blotchy mottle symptoms began to show 17 weeks after inoculation (WAI). At 30 WAI healthy, mature leaves were collected from healthy plants, and mature leaves with or without HLB symptoms were sampled from infected plants. The leaves were immediately frozen in liquid nitrogen and stored at −80°C until use. Four biological replicates of each leaf type were sampled from four individual plants. Each replicate sample included two to three leaves.

Carbohydrate analysis

The Starch Assay kit SA-20 (Sigma-Aldrich) was used to determine and quantify starch content. Mature leaves from healthy plants, and leaves with and without symptoms from infected plants, were analysed. Whole-leaf samples were ground in liquid nitrogen with a mortar and pestle. According to the manufacturer’s instructions, 100 mg ground tissue was used for extraction and sequential measurement.

The effect of HLB infection on other soluble carbohydrates, i.e. sucrose, glucose, fructose and maltose, were further investigated. Because the HLB-associated bacterium inhabits the phloem sieve tubes, extracts were prepared from both foliar lobes and midribs, the latter representing the phloem-enriched part of the leaves. Soluble sugars were extracted according to Geigenberger & Stitt (1993). Briefly, 50 mg leaf tissue was extracted sequentially at 80°C using 1 mL 80% (v/v) ethanol, 1 mL 50% (v/v) ethanol, and 1 mL H2O for 30 min. The supernatants collected from centrifugation at 18 000 g for 1 min, were pooled and treated with activated charcoal to absorb phenols and chlorophylls that might interfere with enzymes for sugar analysis. Sucrose, glucose and fructose were then quantified using the Sucrose Assay kit and Fructose Assay kit (Sigma-Aldrich), according to the manufacturer’s instructions. Maltose was analysed using maltose epimerase and maltose phosphorylase as described by Shirokane et al. (2000). All enzymes used were purchased from Sigma-Aldrich.

Invertase assay

Three invertases, cell-wall invertase, soluble acidic (vacuolar) invertase and soluble neutral invertase, were investigated in mature leaves using an NADH-linked assay based on Bergmeyer & Bernt (1974). Extracts were obtained by grinding ∼0·2 g leaf tissues in liquid nitrogen. Then, 700 μL extraction buffer [50 mm Tris–HCl (pH 7·5), 150 mm NaCl, 1 mm EDTA, 5 mm DTT, 0·1% Triton X-100] was added to each sample. Each extract was mixed completely, shaken for 30 min at 4°C, and centrifuged at 18 000 g for 10 min at 4°C. The supernatant was transferred to a new tube. Protein content was assayed using Protein Assay kit I from Bio-Rad Laboratories in a NanoDrop Spectrophotometer (NanoDrop Technologies).

For the soluble acidic invertase assay, 20 μL extract, desalted on PD-10 Desalting Columns (GE Healthcare) was added in a reaction buffer comprised of 100 μL sodium acetate buffer (100 mm, pH 4·5) and 40 μL sucrose substrate solution (10%, w/v), and incubated at room temperature for exactly 5 min; then 40 μL Tris Base (300 mm) was added to stop the reaction. Fifty μL of the reaction mixture, which contained glucose and fructose derived from sucrose hydrolysis, was transferred into 1 mL Glucose Assay Reagent (Sigma-Aldrich). After incubation at room temperature for 15 min, NADH was measured at 340 nm (Bergmeyer & Bernt, 1974). Soluble neutral invertase activity was assayed by adjusting the pH of sodium acetate buffer to 7·0. To determine cell-wall invertase activity, cell-wall material was washed three times with diluted (1:40, v/v) extraction buffer and assayed in a similar manner as soluble acidic invertase, except that the first reaction step was performed at 45°C for 15 min.

Statistical analysis

Experiments were performed in a completely randomized design. anova was conducted by sas9.1 (SAS Institute). Mean separation was accomplished using Tukey’s test at the 95% confidence level.

RNA extraction and semiquantitative RT-PCR

Total RNA was extracted from the same samples as for the starch assay, using the RNeasy Plant Mini kit (Qiagen), and treated with DNase I (Qiagen), according to the manufacturer’s instructions.

Expression of starch breakdown-related genes was investigated using semiquantitative RT-PCR. Key genes involved in the pathway of transitory starch degradation in leaves, i.e. GWD1, BAM3, DPE1, DPE2 and MEX1 (referred to in the review of Lloyd et al., 2005), were selected (Table 1). Glyceraldehyde 3-phosphate dehydrogenase (GAPDH) was used as a reference gene in parallel with the target genes. Their sequences were retrieved from the HarvEST database ( The unigene numbers (assembly version C37) and specific primers are listed in Table 1. Using the OneStep RT-PCR kit (Qiagen), RT-PCR was carried out from 200 ng total RNA in a final volume of 20 μL, with the following conditions: reverse transcription at 50°C for 30 min; initial PCR activation step at 95°C for 15 min; 32 (for GAPDH, GWD1 and BAM) or 35 cycles (for DPE1, DPE2 and MEX1) of denaturation at 94°C for 35 s, annealing at 55°C for 35 s and extension at 72°C for 30 s; a final extension step at 72°C for 10 min. PCR products were viewed on a 1·5% agarose gel stained with ethidium bromide.

Table 1.   Protein and gene names, unigene numbers, and primers used for semiquantitative RT-PCR of sweet orange infected with ‘Candidatus Liberibacter asiaticus’
Protein nameGene nameUnigene numberPrimer sequence (5′–3′)a
  1. aF: forward primer; R: reverse primer.

  2. bGenBank number retrieved from NCBI database (

Glyceraldehyde 3-phosphate dehydrogenaseGAPDHCit.122.1F, GGAAGGTCAAGATCGGAATCAA
Glucan-water dikinaseGWD1Cit.3855.1F, AGGCTTACAGACCTGCTCCA
TransglucosidaseDPE2Cit.25124.1F, CAAGGCATTTCCCTCATCAT

Results and discussion

Carbohydrate profiles in HLB-infected mature leaves

As shown in Figure 1, the amount of starch increased significantly in HLB-infected leaves (< 0·05). Specifically, 3·1- and 7·9-fold increases of starch level were observed in leaves with and without symptoms, respectively, compared to levels in healthy leaves (Fig. 1). In addition, HLB infection was associated with a significant increase of sucrose (< 0·05, Fig. 2a) in both midribs and foliar lobes, and a more than fivefold increase of glucose (< 0·05, Fig. 2b), but this was observed only in the midribs. Significant increase of fructose was not detected in leaves with symptoms compared to healthy controls (> 0·05, Fig. 2c), but it was observed in symptomless leaves (< 0·05, Fig. 2c). By contrast, the level of maltose decreased in both leaves with and those without symptoms, especially in leaves with symptoms, where the levels were reduced to 51·9% and 49·6% in the midribs and foliar lobes, respectively, of those in healthy controls (< 0·05, Fig. 2d).

Figure 1.

 Effect of ‘Candidatus Liberibacter asiaticus’ infection on starch concentration in mature leaves of sweet orange. H, healthy control leaf; Asym, infected symptomless leaf; Sym, infected leaf with symptoms. Vertical bars represent SD of four replicates. Different letters on bars indicate significant differences (< 0·05).

Figure 2.

 Effect of ‘Candidatus Liberibacter asiaticus’ infection on the concentrations of soluble carbohydrates (a) sucrose, (b) glucose, (c) fructose and (d) maltose in the midribs and foliar lobes of sweet orange mature leaves. HMid, midribs of healthy control leaves; AsymMid, midribs of infected symptomless leaves; SymMid, midribs of infected leaves with symptoms; HLobe, foliar lobes of healthy control leaves; AsymLobe, foliar lobes of infected symptomless leaves; SymLobe, foliar lobes of infected leaves with symptoms. Vertical bars represent SD of four replicates. Different letters on bars indicate significant differences (< 0·05).

Starch accumulation characterized HLB-infection in sweet orange; this might result from photoassimilate transport blockade prompted by HLB-induced phloem necrosis (Schneider, 1968). Recently, Etxeberria et al. (2009) investigated the extent of starch imbalance throughout HLB-affected Valencia orange trees using light, scanning and transmission electron microscopy. It was shown that starch accumulated extensively in aerial tissues but was depleted in roots. Intriguingly, they also found that bark samples and symptomless leaves contained higher level of starch than controls, but without visible phloem blockage (Etxeberria et al., 2009). In a similar study on the effect of elm yellows on American elm (Ulmus americana), the impairment of carbohydrate partitioning was found even earlier than histopathological symptoms such as phloem necrosis (Braun & Sinclair, 1978). Phloem dysfunction causing starch accumulation might occur prior to visible phloem plugging/necrosis during HLB infection.

Sucrose is the major photoassimilate transported in the phloem sieve tubes from mature leaves to sink organs (Zimmermann & Ziegler, 1975). The results of the present study showed that sucrose accumulated in HLB-infected leaves (Fig. 2a), consistent with the observation of Kim et al. (2009), which implied that photoassimilate translocation is impaired by HLB infection. Sucrose accumulation might contribute to some symptoms of HLB (see below), and so might glucose (Fig. 2b). Sugars such as sucrose and glucose are not only metabolic resources and structural constituents of cells, but they also act as signalling compounds that can alter gene expression in plant growth and development (Smeekens, 2000). Sugar-induced feedback inhibition of photosynthesis was reported in many species (Jang & Sheen, 1994). As reviewed by Smeekens (2000), increased sucrose/glucose levels can lead to the repression of many genes involved in photosynthesis, and cause a decrease of chlorophyll. Moreover, it was demonstrated that photosynthesis/chlorophyll-associated genes, such as those for a photosystem-II 5-kDa protein, photosystem-I subunit O and a chlorophyll A-B binding family protein, were significantly down-regulated by HLB infection (Albrecht & Bowman, 2008). This could contribute to the HLB yellowing leaf mottle symptoms, in addition to the excessive starch buildup causing disintegration of the chloroplast thylakoid system (Schaffer et al., 1986).

Similar to HLB-associated bacteria, phytoplasmas and spiroplasmas are also phloem-restricted pathogens associated with several yellows-type diseases. For example, Spiroplasma citri, the causal agent of citrus ‘stubborn’ disease (Saglio et al., 1973), can infect many other plants such as periwinkle, causing leaf yellows, wilting and stunting. Andre et al. (2005) put forward a hypothesis that S. citri might prefer to use fructose rather than glucose in the phloem sieve tubes, leading to reduced fructose levels in companion cells. As a result, the feedback inhibition of fructose to invertase activity (Walker et al., 1997) is avoided. Increased invertase activity hydrolyzes sucrose and produces more glucose and fructose in infected plants. Fructose is preferentially utilized, but glucose is not used and accumulates, in turn giving rise to the inhibition of photosynthesis and to physiological disorders. It is noteworthy that in the present report glucose accumulated markedly in the midribs but not in the foliar lobes of HLB-infected leaves (Fig. 2b), and no significant fructose accumulation was observed in leaves with symptoms. Similar trends also were found in rough lemon (Citrus jambhiri), Carrizo citrange (Citrus sinensis × Poncirus trifoliata) and Valencia sweet orange infected by HLB (data not shown). It is likely that the pathogenicity of ‘Ca. L. asiaticus’ might fit the model of S. citri suggested by Andre et al. (2005), i.e. preferential use of fructose (see above). Furthermore, since genome analysis of ‘Ca. Lasiaticus’ revealed no toxins, extracellular degrading enzymes or specialized secretion systems (Duan et al., 2009), the lifestyle of ‘Ca. L. asiaticus’ may be more parasitic than pathogenic, leading to impaired partitioning of carbohydrates, starch accumulation and leaf yellowing symptoms in the plant. However, it is also possible that some virulence effectors not yet identified, independent of the type-III secretion system, may exist in the ‘Ca. L. asiaticus’ genome and contribute to its pathogenicity in host plants. One such effector, TENGU, was recently identified in ‘Candidatus Phytoplasma asteris’, which is also a phloem-restricted bacterium lacking a type-III secretion system (Hoshi et al., 2009).

Cell-wall invertase activity

To confirm whether invertase activity was deregulated along with the accumulation of sucrose and glucose in response to HLB infection, invertase assays were carried out. The results showed that soluble acidic and neutral invertase activities were too low or undetectable to be significant, both in healthy control and HLB-infected plants (data not shown). However, cell-wall invertase activity increased fourfold (< 0·05) in the foliar lobes of affected leaves compared to healthy controls. In contrast, only a slight induction was observed on average in the midribs (Fig. 3).

Figure 3.

 Effect of ‘Candidatus Liberibacter asiaticus’ infection on the activity of cell-wall invertase in the midribs and foliar lobes of sweet orange mature leaves. HMid, midribs of healthy control leaves; AsymMid, midribs of infected symptomless leaves; SymMid, midribs of infected leaves with symptoms; HLobe, foliar lobes of healthy control leaves; AsymLobe, foliar lobes of infected symptomless leaves; SymLobe, foliar lobes of infected leaves with symptoms. Vertical bars represent SD of four replicates. Different letters on bars indicate significant differences (< 0·05).

Cell-wall invertase is a highly polymorphic glycoprotein (Faye et al., 1986), normally found at high levels in developing sink organs, such as leaves and roots, where unloading of sucrose from the phloem takes place. In addition to its role in sugar partitioning, it also functions in response to pathogen infection. Accumulation of cell-wall invertase was observed in tomato (Solanum lycopersicum) roots following infection by a fungal wilt pathogen, Fusarium oxysporum f. sp. radicis-lycopersici (Benhamou et al., 1991), which may be part of the plant’s defense response. In contrast, tomato plants in which cell-wall invertase was silenced were found to have delayed symptom development and inhibition of photosynthesis following X. campestris pv. vesicatoria infection (Kocal et al., 2008). von Schaewen et al. (1990) showed that overexpression of a yeast invertase in the cell wall of transgenic tobacco (Nicotiana tabacum) disrupted sucrose export and resulted in a high accumulation of soluble sugars and starch, photosynthesis inhibition, stunted growth and bleached or necrotic leaf areas. As they proposed, sucrose loading to the phloem of the source leaves involves an apoplastic step. In the transgenic plant overexpressing cell-wall invertase, sucrose in the apoplast is hydrolysed into glucose and fructose, which cannot be taken up into the phloem as well as sucrose. These hexoses may accumulate in the apoplast, leading to an increase of osmotic potential and thus to plasmolysis of the cells nearby; or they may be taken up by the mesophyll cells, rephosphorylated by hexokinase and fructokinase and reconverted to sucrose (von Schaewen et al., 1990). Altogether, a hypothesis is suggested that the HLB-associated bacterium induces the cell-wall invertase in mature leaves and preferentially utilizes fructose, leading to the impairment of sucrose loading and accumulation of sucrose and glucose, and thus to photosynthesis inhibition, starch accumulation and sequential typical HLB symptoms.

Expression profiles of starch breakdown genes

Semiquantitative PCR suggested that DPE2 and MEX1 were down-regulated in the leaves of HLB-infected plants, with lower expression levels in leaves with than those without symptoms (Fig. 4). The level of BAM3 expression slightly decreased in some replicates, but did not change in others (Fig. 4). By contrast, the expression of GWD1 and DPE1 seemed not to be affected by HLB infection (Fig. 4).

Figure 4.

 Semiquantitative RT-PCR analysis of genes involved in starch breakdown during ‘Candidatus Liberibacter asiaticus’ infection. Data from two of four replicates are presented as healthy controls (lanes 1 and 2), infected symptomless leaves (lanes 3 and 4) and infected leaves with symptoms (lanes 5 and 6). The GAPDH gene was amplified as the internal control. Gene nomenclature is given in Table 1.

It has been reported that maltose is the predominant form of carbon exported from chloroplasts at night, and maltose increases in leaves when starch breakdown is induced (Lu & Sharkey, 2006). The reduction of maltose in HLB-infected leaves might indicate that starch breakdown is repressed, although enhanced starch biosynthesis cannot be excluded (Albrecht & Bowman, 2008). The expression profiles of DPE2 and MEX1 were positively correlated to maltose level (Fig. 2d) and negatively correlated to starch content (Fig. 1). DPE2 (EC acts as a transglucosidase, using maltose as a donor to transfer one glucose moiety to a polysaccharide, leading to the release of the other glucose moiety. In the studies of Lu & Sharkey (2004), Chia et al. (2004) and Lloyd et al. (2004), the leaves of plants lacking DPE2 were inhibited in starch degradation and accumulated maltose. MEX1, a plastic maltose transporter, is essential in transitory starch breakdown. mex1 mutants accumulate high levels of maltose in their leaves and demonstrate a starch-excess phenotype and growth impairment (Niittyla et al., 2004). However, the results of the present study showed that the decrease of DPE2 and MEX1 expression levels was accompanied by a reduction in maltose levels. It is possible that some other gene(s) upstream of DPE2 and MEX1 may slow down the starch degradation pathway and the production of maltose. In turn, DPE2 and MEX1, and their proteins that use maltose as substrate, might be inhibited to some extent.

In conclusion, the results of this study indicate that ‘Ca. L. asiaticus’ causes the imbalance of carbohydrate partitioning, such as the accumulation of sucrose and glucose, and the reduction of maltose. It is suggested that ‘Ca. L. asiaticus’ pathogenicity might be similar to that of spiroplasmas (preferential use of fructose); consequently, invertase activity was increased, glucose accumulated, and photosynthesis was inhibited. Additionally, the impairment of starch breakdown in infected leaves may be an important factor, although not the only one, leading to starch accumulation.


We thank Mrs M. Wendell and Dr Qibin Yu for their excellent technical support. This work was partially supported by the China Scholarship Council and by a grant from the Florida Citrus Production Research Advisory Council, Florida Department of Citrus (FDACS Contract Number 67), Peter McClure, Chairman.