Planta (2003) 217: 1120 DOI 10.

1007/s00425-002-0963-6

O R I GI N A L A R T IC L E

Chun Yao Li David Weiss Eliezer E. Goldschmidt

Effects of carbohydrate starvation on gene expression in citrus root

Received: 24 June 2002 / Accepted: 26 November 2002 / Published online: 28 January 2003 Springer-Verlag 2003

Abstract The roots of alternate-bearing citrus (Murcott, a Citrus reticulata hybrid) trees undergo extreme uctuations of carbohydrate abundance and starvation. Using this system, we investigated the eect of root carbohydrate (total soluble sugar, sucrose and starch) depletion on carbohydrate-related gene expression. A series of genes, including those coding for starch phosphorylase (STPH-L and STPH-H), ADP-glucose pyrophosphorylase, small subunit (Agps), R1, plastidic ADP/ATP transporter (AATP), phosphoglucomutase (PGM-P and PGM-C), sucrose synthase (CitSuS1 and CitSuSA), sucrose transporter (SUT1 and SUT2), hexokinase (HK) and alpha-amylase (a-AMY), have been isolated and their expression analyzed. The genes were found to respond dierentially to carbohydrate depletion. STPH-L, STPH-H, Agps, R1, AATP, PGM-P, PGM-C, CitSuS1 and HK were down-regulated while SUT1 and a-AMY were up-regulated during carbohydrate depletion. Two other genes, CitSuSA and SUT2, did not respond to carbohydrate depletion. Fruit removal, which interrupted the carbohydrate depletion induced by heavy fruiting, reversed these gene expression patterns. Trunk girdling and whole-plant darkening treatments, which brought about root carbohydrate depletion, induced the same changes in gene expression obtained in the alternatebearing system. The possible roles of the up- and down-regulated genes in the metabolism of carbohydratedepleted citrus roots are discussed. Although the specic signals involved have not been determined, the results support the feast/famine hypothesis of carbohydrate regulation proposed by Koch [K.E. Koch (1996) Annu Rev Plant Physiol Plant Mol Biol 47:509540].

Keywords Carbohydrate Citrus Root starvation Starch Sucrose Abbreviations AATP: plastidic ADP/ATP transporter Agps: ADP glucose pyrophosphorylase, small subunit a-AMY: alpha-amylase CHOH: carbohydrate CitSuS1 and CitSuSA: sucrose synthase HK: hexokinase PGM-C: cytosolic phosphoglucomutase PGM-P: plastidial phosphoglucomutase STPH-L: starch phosphorylase (type-L) STPHH: starch phosphorylase (type-H) SUT1 and SUT2: sucrose transporter

Introduction
Carbohydrates (CHOH) play key roles in all aspects of plant life. CHOH starvation initiates various responses in plants and has profound eects on a broad range of metabolic and developmental processes (Yu 1999). Fruit trees, including citrus, face various situations of CHOH depletion. During the annual reproductive cycle, large amounts of CHOH are utilized for owering and fruit set (Bustan and Goldschmidt 1998). Yet, considerable amounts of CHOH, particularly starch, are stored in both canopy and underground organs during the vegetative phase (Goldschmidt and Koch 1996). In alternatebearing fruit trees, the cycle of CHOH levels is turned bi-annually and is best demonstrated in some extreme alternate-bearing citrus cultivars (Monselise and Goldschmidt 1982). During the non-fruiting o year, trees accumulate and store large amounts of CHOH. Following the o year, trees undergo a heavy-fruiting on year. The CHOH reserves, which accumulated during the o year, are mobilized and used for the fruiting process, leading to CHOH depletion in vegetative tree organs and particularly in roots. Both starch and soluble sugar are abundant in roots of o trees and depleted towards the end of the on year (Goldschmidt and Golomb 1982). In extreme cases, CHOH depletion brings about a collapse of trees at the end of the on year (Smith 1976).

C.Y. Li D. Weiss E.E. Goldschmidt (&) The Institute of Plant Sciences and Genetics in Agriculture, Faculty of Agricultural, Food, and Environmental Quality Sciences, The Hebrew University of Jerusalem, 76100, Rehovot, Israel E-mail: goldsmit@agri.huji.ac.il Fax: +972-8-9363731

12

CHOH abundance and depletion result in marked changes in gene expression. In general, CHOH depletion up-regulates genes for photosynthesis, remobilization, and export, while decreasing mRNA levels required for storage and utilization. Abundant sugar levels exert opposite eects through a combination of gene repression and induction (Koch 1996). For example, sugar depletion induces the expression of photosynthetic genes in maize mesophyll protoplasts (Sheen 1990), the chlorophyll a/b-binding protein gene in Arabidopsis (Oswald et al. 2001), malate synthase, and isocitrate lyase genes in cucumber (Graham et al. 1994). The level of elongation factor 2 (eEF2) mRNA decreased in response to CHOH starvation in sugar beet (Vogel et al. 1999) and an increase in ammonium-transporter transcripts is related to CHOH abundance in the root of eld pea and rutabaga (Kubik-Dobosz et al. 2001). CHOH metabolism genes are, evidently among those responding to CHOH depletion or abundance (Koch 1996), including genes that encode proteins for starch phosphorylation, starch synthesis and sucrose metabolism. Roots have always been, and still are the hidden part of the plant. In addition to their acknowledged function in provision of water and mineral nutrients, roots play an important physiological role as storage organs. In fruit trees roots are probably the most important storage organ (Loescher et al. 1990). Roots may play a critical role in the management of CHOH reserves in on and o citrus roots. Roots undergo a dramatic transition from a sink organ which accumulates reserve CHOH during the o year to a source organ which exports its CHOH to the developing fruit during the on year (Goldschmidt and Golomb 1982; Goldschmidt and Koch 1996). Not much is known about root physiology in relation to CHOH depletion. Murcott (a Citrus reticulata hybrid of unknown origin), a strongly alternate-bearing cultivar (Smith 1976), was chosen for the present studies. Using the on and o system of Murcott Sour orange trees has allowed us to identify a series of CHOH metabolic genes which respond dierently to root CHOH depletion. Complementary approaches with trunk girdling and whole-tree darkening, which also cause root CHOH depletion, and fruit removal, which interrupts the CHOH depletion induced by heavy fruiting, support the results obtained with the on/o system. The possible roles of these genes in relation to the build-up, breakdown and mobilization of CHOH reserves in citrus root are discussed.

Plant materials for quantitative CHOH determinations and molecular analyses were sampled during autumn in monthly intervals (September, October, November and December) from on and o trees. Plant material was collected in the morning. Roots 0.5 cm in diameter were selected; samples were immediately cleaned, frozen in liquid nitrogen and stored at )70 C until analyzed.

Fruit removal Field-grown, heavily fruiting Murcott Sour orange trees about 5 years old, were used. Fruit were removed from four trees on 26 October 2000 while the fruit were vigorously growing. The other four trees were kept as a control. Root samples (0.5 cm in diameter) were collected from treated and control trees on 22 January 2001, about 3 months after removal of the fruit.

Girdling experiments Field-grown Murcott Sour orange trees, about 5 years old, were selected for this experiment. Uniform o trees were divided into two groups (four trees each), one for girdling (girdling) and the other kept as a control (non-girdled). Trees were girdled on 25 September 2001 on the trunk 10 cm above the graft union and the girdle was renewed every 3 weeks, in order to prevent rapid healing of the wound. Root samples 0.5 cm in diameter were harvested on the day of girdling (zero-time control) and then on 9 December 2001 for evaluation of the girdling eect.

Darkening experiments Trees (Murcott Sour orange), about 5 years old, grown in containers in the greenhouse were used for this experiment. Three o trees were kept in darkness for 6 weeks before samples (roots 0.5 cm in diameter) were collected. Roots from another three nondarkened trees served as a control.

CHOH determinations Root samples were dried in the oven (80 C) for 48 h and then ground into ne powder. Triplicate dry tissue samples (100 mg each) were extracted four times with 80% (v/v) ethanol at 55 C. All the extracts were collected and brought to a nal volume of 50 ml. Total soluble sugars and sucrose were measured using anthrone, according to the methods of Yemm and Willis (1964) and van Handel (1968), respectively. Starch was measured after glucoamylase hydrolysis of the insoluble faction according to Hassid and Neufeld (1964). Scanning electron microscopy Roots were washed, dried and hand-sectioned with a blade into 3- to 4-mm-long fragments that were then xed with 4% (v/v) glutaraldehyde in 0.1 M sodium phosphate buer (pH 7.2) for 2 h. After rinsing ve times (10 min each) with the same phosphate buer, samples were dehydrated in a graded series of ethanol. The dehydrated samples were critical-point-dried, sputter-coated with gold, observed and photographed under an electron microscope (Jeol-5410 LV; Japan).

Materials and methods

Plant material and experimental design Field-grown Murcott (a Citrus reticulata hybrid of unknown origin) Sour orange trees, about 15 years old, revealing an absolute biennial fruit-bearing habit, were used. Trees were divided into two groups, (i) four trees in the on year (with heavy fruit loading), and (ii) four trees in the o year, the few fruit present were removed at an early stage of fruit development.

Detection of starch with iodine Roots were collected from trees grown in the eld on 15 February 2001. For cross-sections, the roots were hand-sectioned into 4-mmthick fragments; for longitudinal sections the 1-cm-long fragments were cut into two equal pieces. Samples were immediately

13 immersed in Lugol solution (Sigma) containing iodine and potassium iodide. Optimal staining was obtained within 3 min, after which the sections were photographed under an Olympus binocular system (Olympus SZX12). RNA extraction, gene isolation and northern blot analysis Total RNAs were extracted by a standard SDSphenol method. Reverse transcription (RT)PCR and/or library screening strategy were employed for gene isolation. For STPH-L, AATP, HK, SUT1, SUT2, a-amylase, PGM-P and PGM-C isolation, degenerated primers for each gene were designed based on the conserved domains of the known sequences in GenBank. RTPCRs were performed on total RNA isolated from on and o root samples. Amplied fragments were cloned into T-easy vector (Promega) and sequenced. Fragments of STPH-L, AATP, HK, SUT1 and SUT2 were used as probes to screen a citrus cDNA library (Jacob-Wilk et al. 1999) in order to get full gene sequences. For STPH-H and R1 isolation, fragments of Arabidopsis STPH-H and R1 cDNA were used as heterogeneous probes to screen a citrus cDNA library (Jacob-Wilk et al. 1999) under low-stringency conditions (membrane hybridization: 4850 oC, 4SSC, wash: 2SSC + 0.1%SDS). Isolated genes were submitted to GenBank under the following accession numbers: STPH-L: AY098892, STPH-H: AY098895, R1: AY094062, AATP: AY098893, HK: AF196966, SUT1: AY098891, SUT2: AY098894, a-AMY: AY094063, PGM-P: AY112997, PGM-C: AY112996. For RNA blot analysis, total RNA samples (20 lg per lane) were separated by electrophoresis in 1% agarose gels containing formaldehyde and blotted to HybondNX nylon membrane (Amersham Pharmacia Biotech) by capillary transfer. The RNA blots were hybridized with the specic cDNA probes isolated above or taken from the GenBank (Agps: AF184597, CitSuS1: AB022092, CitSuSA: AB022091), labeled using a random priming method (Boehringer Mannheim). Hybridization was carried out at 42 C in a solution with 10% Dextran sulfate (Na salt, MW 500,000), 1%SDS, 2SSC, 50% formamide and 5 Denhardts solution for 16 h followed by a high-stringency wash. The hybridized membranes were exposed and analyzed by Phosphorimager (FUJI BAS 1000). Generally, four membranes loaded with the same RNAs were used for hybridization in each experiment. Each membrane was re-probed several times after stripping and nally re-hybridized with citrus 18S rRNA.

on ratios of 2.2 for total soluble sugar, 3.3 for sucrose and 6.6 for starch, due to depletion of CHOH in on roots and their accumulation in o roots. Iodine staining Starch is a major reserve CHOH in roots of woody plants (Loescher et al. 1990), including citrus (Goldschmidt and Golomb 1982; and Fig. 1c of the present study). To visualize the quantitative dierences and the localization of starch, root tissues were stained with iodine (Fig. 2). Results show that on roots contained very low amounts of starch (Fig. 2a, b) at the end of the productive cycle, as against o roots, which contained a very high amount of starch (Fig. 2c, d). Most of the starch was deposited in the bark and some in xylem ray cells. SEM observations The observations were conducted in mid-winter, the end of the annual reproductive cycle of citrus. Results show

Results
CHOH depletion during the on and o cycle CHOH levels in on and o citrus roots were evaluated in three ways: quantitative analysis, iodine staining and scanning electron microscopic (SEM) observations. Quantitative analysis The levels of CHOH reserves in citrus roots were determined during the later stage of fruit development, from September to December. Figure 1 shows the dynamic changes in total soluble sugar (Fig. 1a), sucrose (Fig. 1b) and starch (Fig. 1c) levels in on and o-roots. Generally speaking, opposite trends are evident; while CHOH levels in on roots were gradually decreasing, those in o roots increased. In September, the dierences in CHOH levels between on and o roots were relatively small. Three months later, however, the dierences in CHOH levels between on and o roots became larger, with o/

Fig. 1 Seasonal changes (September to December) in total soluble sugar (TSS) (a), sucrose (b) and starch (c) levels in on and o roots of Murcott (a Citrus reticulata hybrid) Sour orange citrus trees. Bars indicate means SD of three independent experiments (n=3)

14

that starch granules accumulated mainly in the phloem of o roots (Fig. 3c, d), while the phloem parenchyma cells of on roots (Fig. 3a, b) contained only a few starch granules. In addition, the size of the few starch granules found in on cells (Fig. 3b) was much smaller than those in o cells (Fig. 3d).

Gene expression analyses Genes encoding for CHOH-metabolizing proteins, including genes related to starch metabolism, sucrose cleavage, sugar sensing and sucrose transport, were isolated from citrus for expression analyses. The genes examined can be classied into three groups, genes that are down-regulated under CHOH depletion, genes upregulated in response to CHOH depletion and genes that are insensitive to uctuations in CHOH levels (Fig. 4). Most of the genes tested were down-regulated under CHOH depletion, as indicated in Fig. 4a. STPH-L, STPH-H, Agps, R1, AATP, PGM-P, PGM-C, CitSuS1 and HK are in this group. STPH-L and STPH-H encode two isoforms of starch phosphorylase that are involved in phosphorylytic starch degradation; Agps codes for a small subunit of ADP-glucose pyrophosphorylase that plays a key role in starch synthesis. R1 encodes a dikinase protein that catalyzes starch phosphorylation in potato and Arabidopsis (Lorberth et al. 1998; Yu et al. 2001; Ritte et al. 2002). AATP is a plastidic ATP/ADP transporter gene and its encoded protein is involved in carrying ATP into plastids for starch synthesis. Plastidial phosphoglucomutase and cytosolic phosphoglucomutase, coded by PGM-P and PGM-C, respectively, are involved in starch synthesis. CitSuS1 encodes sucrose synthase (Komatsu et al. 2002), which functions in sucrose cleavage in the cytoplasm. Expression of the above-mentioned genes is very low under depletedCHOH conditions in on roots, in comparison to their counterparts in o roots. The transcripts of several genes, such as STPH-L, Agps, R1 and CitSuS1, show a

Fig. 2 Detection of starch in cross-sections and longitudinal root sections from on (a, b) and o (c, d) citrus trees by iodine staining. B Bark, R xylem ray

Fig. 3 Scanning electron microscopic analyses of crosssections of on (a, b) and o (c, d) citrus roots

15

Fig. 4ac Gene expression in response to CHOH depletion. Total RNAs (20 lg) isolated from on and o citrus tree roots were separated and analyzed by RNA blot. cDNA probes used for hybridization are indicated on the right. Results were from blots of four membranes. The control blot using citrus 18S rRNA was from one of the membranes. a Genes down-regulated in response to CHOH depletion. b Genes up-regulated in response to CHOH depletion. c Genes that did not respond to CHOH uctuations

Fig. 5 Levels of total soluble sugar (TSS) (a), sucrose (b) and starch (c) in citrus roots before and after removal of fruit from on trees; with roots from non-treated on and o trees as controls (the results are means SD of three independent experiments, n=3)

consistent increase in transcription levels from September to December in o roots, correlating with the increases in CHOH levels. The transcript of AATP, which has a relatively high expression in September, decreases towards a very low level in December, correlating with the depletion of CHOH in on roots. Unlike those genes that are down-regulated, two genes were found to be up-regulated in on roots towards CHOH depletion (Fig. 4b): one encodes for a sucrose transporter (SUT1) and the other for alphaamylase (a-AMY). Both genes showed a consistently higher expression in on roots. Not all of the CHOH metabolic genes tested were aected by CHOH depletion. The expression of some genes, such as CitSuSA and SUT2, which encode a second sucrose synthase and a second sucrose transporter, respectively, did not change in response to CHOH depletion (Fig. 4c).

expression patterns induced by heavy fruiting following recovery of the CHOH status. CHOH analysis showed that upon removal of fruit from the on trees, the decline in CHOH stopped and CHOH accumulation resumed (Fig. 5). After about 3 months the levels of CHOH in the roots of the on to o trees were about the same as those of the usual o trees. Correspondingly, the gene expression patterns were also reversed and typical o root expression patterns appeared (Fig. 6). CHOH depletion induced by trunk girdling Trunk girdling of o trees was employed in order to mimic the root CHOH depletion imposed by heavy fruiting. CHOH levels were determined 10 weeks after girdling and results are summarized in Fig. 7, using nongirdled o trees as controls. Girdling caused a marked drop in the levels of soluble sugar, sucrose and starch in citrus roots. The amount was 37.6% for total soluble sugar, 31.6% for sucrose and 16.2% for starch in the roots of girdled trees, as compared with that of control o trees.

Fruit-removal experiments Fruit-removal experiments were undertaken to examine the possibility of reversal of the gene

16

Fig. 6 Gene expression in citrus roots in response to fruit removal. Total RNAs (20 lg) isolated from roots of trees subjected to the fruit removal treatment were separated and analyzed by RNA blot. cDNA probes used for hybridization are indicated on the right. 18S rRNA from citrus was used as an internal control for loading dierences

Fig. 7 Levels of total soluble sugar (TSS) (a), sucrose (b) and starch (c) in roots of girdled and non-girdled control citrus trees (the results are means SD of three independent experiments, n=3)

The expression analyses indicated that genes are responding to girdling-induced CHOH depletion in essentially the same way as found in on roots (Fig. 8). STPH-L, STPH-H, Agps, R1, AATP, PGM-P, PGM-C, CitSuS1 and HK were down-regulated, SUT1 and a-AMY were up-regulated, while the expression of CitSuSA and SUT2 did not change appreciably. Darkening-induced CHOH depletion Prolonged darkening was employed as another means to impose CHOH depletion. As indicated in Fig. 9, 6 weeks of darkening treatment reduced the total soluble sugar from 79.31.8 mg g)1 (means SD of three independent experiments) to 61.81.2 mg g)1, a 20% reduction, sucrose from 320.9 mg g)1 to 19.30.9 mg g)1, a 40% reduction, and starch levels from 112.2 3.3 mg g)1 to 81.91.7 mg g)1, a 27% reduction. The reduction in root CHOH levels in the darkened trees was considerable, although less pronounced than that observed in on roots (Fig. 1) and roots of girdled trees (Fig. 7). Nevertheless, the darkening-induced CHOH depletion (Fig. 10) gave essentially the same patterns of gene expression as those caused by heavy fruiting and girdling.

Discussion
Fluctuations of CHOH abundance and depletion are not uncommon in higher plants, in which storage of CHOH is a major survival strategy. CHOH reserves serve for the initiation of growth following periods of dormancy (Loescher et al. 1990) as well as for processes that require large amounts of carbon and energy, more than current photosynthesis can provide (e.g. owering and fruiting; Bustan and Goldschmidt 1998). Various plant organs may assume a storage function: seed, bulbs, root, trunk and even leaves. In fruit trees, the root is the most important storage organ (Loescher et al. 1990). In fact, the root reverses its physiological role from sink to source according to the changes in carbon demand, as is evident in alternate-bearing trees (Goldschmidt and Golomb 1982). This has been demonstrated up to now through analysis of CHOH levels and carbon budgets. In the present study we went a step further. Our working hypothesis was that the changes in CHOH levels involve changes in gene expression, based on the model of Koch (1996) that suggested dierent patterns of gene expression for CHOH abundance (feast) and depletion (famine) conditions. Thus, we examined the

17

Fig. 8 Gene expression in response to girdling-induced CHOH depletion. RNAs were isolated from roots of girdled and control citrus trees. An equal amount of total RNA (20 lg) was loaded in each lane and analyzed by RNA blot. cDNA probes used for hybridization are indicated on the right. 18S rRNA from citrus was used as an internal control for loading dierences

Fig. 9 Total soluble sugar (TSS) (a), sucrose (b) and starch (c) levels in roots of dark-treated citrus trees and non-darkened controls (the results are means SD of three independent experiments, n=3)

levels of expression of 13 CHOH-metabolism genes under CHOH abundance and depletion conditions. Root CHOH depletion was achieved in the present study with Murcott (a Citrus reticulata hybrid) by dierent means: (i) heavy fruiting, (ii) trunk girdling and (iii) whole-plant darkening. Heavy fruiting induces mobilization of root CHOH to the developing fruit, girdling prevents the transport of CHOH into the root and prolonged darkening brings about an overall decline in CHOH due to lack of photosynthesis. In spite of the dierent mechanisms involved, all three conditions caused a reduction in CHOH levels in the root and produced essentially the same patterns of gene expression (Figs. 4, 8, 10). Interestingly, although the CHOH depletion brought about by darkening (Fig. 9) was less pronounced than that observed with on trees (Fig. 1) and girdling (Fig. 7), similar gene expression patterns were obtained (Fig. 10). Apparently, it is not the absolute level of CHOH that counts but rather the declining trend, shared by all three root CHOH-depletion conditions. The seasonal follow-up in on and

o roots demonstrated the gradual build-up of root starvation (Fig. 1) and the corresponding gene expression analyses revealed a complementary picture (Fig. 4). The fruit-removal experiment showed that CHOH trends (Fig. 5) as well as the gene expression patterns (Fig. 6) are reversible. Starch is the main reserve CHOH in citrus roots. Inhibition of starch synthesis and enhancement of degradation will lead to the depletion of CHOH. In the present study, STPH-L, STPH-H, Agps and R1 were consistently down-regulated under CHOH depletion (Fig. 4a). Most of the proteins encoded by these genes are involved, directly or indirectly, in starch synthesis. Ample evidence indicates that ADP-glucose pyrophosphorylase, encoded by Agps, catalyzes the rate-limiting step for starch biosynthesis in both photosynthetic and non-photosynthetic tissues (Stark et al. 1992; Weber et al. 2000). R1 encodes for an a-glucan, water dikinase in potato tuber and was recently shown to catalyze starch phosphorylation (Ritte et al. 2002). It was found that starch phosphorylation occurs concurrently with de novo biosynthesis of starch in potato tubers (Nielson et al. 1994). Starch phosphorylase, encoded by STPH-L

18

Fig. 10 Gene expression in response to darkening-induced CHOH depletion. Total RNAs (20 lg) isolated from roots of dark-treated and control citrus trees were separated and analyzed by RNA blot. cDNA probes used for hybridization are indicated on the right. 18S rRNA from citrus was used as an internal control for loading dierences

and STPH-H, catalyzes the reversible reaction and can degrade glucose polymer to glucose 1-phosphate or synthesize starch polymers (Stitt and Steup 1985). Although the precise function of starch phosphorylase in starch degradation and synthesis is not yet clear, evidence has shown that increased starch phosphorylase activity is related to rapid starch synthesis in buttercup squash (Irving et al. 1999), and high expression of plastidic starch phosphorylase genes (STPH-L) correlates with starch biosynthesis in potato (Duwenig et al. 1997a), spinach (Duwenig et al. 1997b) and pea (van Berkel et al. 1991). The fact that a plastidic starch phosphorylase was enriched in the amyloplast stroma where other starch-synthetic enzymes are localized supports a role for starch phosphorylase in starch biosynthesis (Yu et al. 2001). Correspondingly, low expression of Agps, STPH-L, STPH-H and R1 might be associated with the lack of starch synthesis in CHOH-depleted on roots (Fig. 4a). AATP, PGM-P, PGM-C, CitSuS1 and HK are also down-regulated under CHOH depletion (Fig. 4a). The plastidic ATP/ADP translocator, encoded by AATP, is

involved in starch synthesis by providing ATP, one of the substrates of ADP glucose pyrophosphorylase, to the plastid (Kampfenkel et al. 1995; Neuhaus et al. 1997). The cytosolic and plastidial phosphoglucomutases, encoded by PGM-C and PGM-P, respectively, which carry out a reversible reaction between glucose 1-phosphate and glucose 6-phosphate, were found to be important in starch accumulation (Harrison et al. 2000; Fernie et al. 2001). Antisense repression of cytosolic phosphoglucomutase in potato also inhibited sucrose synthesis (Fernie et al. 2002). Sucrose synthase, encoded by SuS, which catalyzes the reversible conversion of sucrose and uridine diphosphate (UDP) into fructose and UDP-glucose, was also found to be associated with starch synthesis (Chen and Chourey 1989; Dejardin et al. 1997). However, the regulation of SuS genes is more complicated (Koch 1996). Our study showed that expression of CitSuS1 is inhibited under CHOH depletion, while expression of CitSuSA is not aected. Hexokinase, encoded by HK, is possibly a dual-function enzyme involving both sugar signaling and phosphorylation of hexose (Dai et al. 1999). Suppression of HK expression caused reduction in total soluble sugar and starch in transgenic plants (Dai et al. 1999). Two genes, a-AMY and SUT1, were found to be upregulated by CHOH depletion (Fig. 4b). The a-amylase, encoded by a-AMY, functions in the amylolytic pathway of starch degradation (Steup 1988). It is, presumably, the only enzyme capable of binding and attacking raw starch, and its slow action is responsible for the physiological corrosion of starch granules (Manners 1985). Another gene, SUT1, encodes a sucrose transporter (Riesmeier et al. 1994). This gene is similar (unpublished data) to the one involved in the sucrose loading (Riesmeier et al. 1994), implying its possible role in mobilization of sugar from the root. Although it is yet too early to propose an overall model of CHOH regulation in citrus root, some of our results might be placed within a logical framework. Starch biosynthesis in roots is clearly associated with the o phase and consequently, the genes participating in starch biosynthesis (STPH-L, STPH-H, Agps, R1, AATP, PGM-P, PGM-C and CitSuS1) are strongly expressed during the o phase, when starch accumulation is taking place (Figs. 4a, 8, 10). Starch breakdown is generally less understood and the machinery for starch breakdown must apparently be present at all times. However, the fact that the a-amylase gene (a-AMY) was strongly expressed only during phases of starch decline (Figs. 4b, 8, 10) might indicate that the amylolytic pathway of starch degradation plays a predominant role under declining CHOH conditions. The results of the present study clearly demonstrate the existence of CHOH-related changes in gene expression, thereby conrming our preliminary assumptions and supporting Kochs feast/ famine model (1996).

19 Acknowledgements We thank Prof. Ekkehard Neuhaus (Universita t Kaiserslautern, Germany) for critical reading of the manuscript. This work was supported by a Nidersachsen-Israel Research Grant #0306207 and by funds from the Israeli Citrus Marketing Board to E.E.G. Koch KE (1996) Carbohydrate modulated gene expression in pants. Annu Rev Plant Physiol Plant Mol Biol 47:509 540 Komatsu A, Moriguchi T, Koyama K, Omura M, Akihama T (2002) Analysis of sucrose synthase genes in citrus suggests dierent roles and phylogenetic relationships. J Exp Bot 53:61 71 Kubik-Dobosz G, Bakiewicz M, Gorska A (2001) The importance of root carbohydrate abundance in ammonium uptake. Acta Physiol Plant 23:187192 Loescher WH, McCamant T, Keller JD (1990) Carbohydrate reserves, translocation, and storage in woody plant roots. HortScience 25:274281 Lorberth R, Ritte G, Willmitzer L, Kossmann J (1998) Inhibition of a starch-granule-bound protein leads to modied starch and repression of cold sweetening. Nat Biotechnol 16:4737 Manners DJ (1985) Starch. In: Dey PM, Dixon RA (eds) Biochemistry of storage carbohydrates in green plants. Academic Press, London, pp 149203 Monselise SP, Goldschmidt EE (1982) Alternate bearing in fruit trees. Hort Rev 4:128173 Neuhaus HE, Thom E, Mo hlmann T, Steup M, Kampfenkel K (1997) Characterization of a novel eukaryotic ATP/ADP translocator located in the plastid envelope of Arabidopsis thaliana L. Plant J 11:7382 Nielsen TH, Wischmann B, Enevoldsen K, Moller BL (1994) Starch phosphorylation in potato-tubers proceeds concurrently with de-novo biosynthesis of starch. Plant Physiol 105:111117 Oswald O, Martin T, Dominy PJ, Graham IA (2001) Plastid redox state and sugars: interactive regulators of nuclear-encoded photosynthetic gene expression Proc Natl Acad Sci USA 98:20472052 Riesmeier JW, Willmitzer L, Frommer WB (1994) Evidence for an essential role of the sucrose transporter in phloem loading and assimilate partitioning. EMBO J 13:17 Ritte G, Lloyd JR, Eckermann N, Rottmann A, Kossmann J, Steup M (2002) The starch related R1 protein is an a-glucan, water dikinase. Proc Natl Acad Sci USA USA 99:7166 7171 Sheen J (1990) Metabolic repression of transcription in higherplants. Plant Cell 2:10271038 Smith P (1976) Collapse of Murcott tangerine trees. J Am Soc Hort Sci 101:2325 Stark DM, Timmerman KP, Barry GF, Preiss J, Kishore GM (1992) Regulation of the amount of starch in plant tissues by ADPglucose pyrophosphorylase. Science 258:287 292 Steup M (1988) Starch degradation. The biochemistry of plants 14:255296 Stitt M, Steup M (1985) Starch and sucrose degradation. In: Douce R, Day DA (eds) Encylopedia of plant physiology, NS, vol 18: Higher plant cell respiration. Springer, Berlin Heidelberg New York, pp 347390 Van Berkel J, Conradsstrauch J, Steup M (1991) Glucan-phosphorylase forms in cotyledons of Pisum sativum L. localization, developmental-change, in vitro translation, and processing. Planta 185:432439 Van Handel E (1968) Direct microdetermination of sucrose. Anal Biochem 22:280283 Vogel R, Viereck R, Murmann A, Rausch T (1999) Cloning of a higher plant elongation factor 2 cDNA: Expression of eEF2 and alpha subunit of eEF1B in sugar beet cells during phosphate and carbohydrate starvation. J Plant Physiol 154:192196 Weber H, Rolletschek H, Heim U, Golombek S, Gubatz S, Wobus U (2000) Antisense-inhibition of ADP-glucose pyrophosphorylase in developing seeds of Vicia narbonensis moderately decreases starch but increases protein content and aects seed maturation. Plant J 24:3343

References
Bustan A, Goldschmidt EE (1998) Estimating the cost of owering in a grapefruit tree. Plant Cell Environ 21:217224 Chen YC, Chourey PS (1989) Spatial and temporal expression of the 2 sucrose synthase genes in maize immunohistological evidence. Theor Appl Genet 78:553559 Dai N, Schaer A, Petreikov M, Shahak Y, Giller Y, Ratner K, Levine A, Granot D (1999) Overexpression of Arabidopsis hexokinase in tomato plants inhibits growth, reduces photosynthesis, and induces rapid senescence. Plant Cell 11:12531266 Dejardin A, Rochat C, Wuilleme S, Boutin JP (1997) Contribution of sucrose synthase, ADP-glucose pyrophosphorylase and starch synthase to starch synthesis in developing pea seeds. Plant Cell Environ 20:14211430 Duwenig E, Steup M, Kossmann J (1997a) Induction of genes encoding plastidic phosphorylase from spinach (Spinacia oleracea L.) and potato (Solanum tuberosum L.) by exogenously supplied carbohydrates in excised leaf discs. Planta 203:111120 Duwenig E, Steup M, Willmitzer L, Kossmann J (1997b) Antisense inhibition of cytosolic phosphorylase in potato plants (Solanum tuberosum L.) aects tuber sprouting and ower formation with only little impact on carbohydrate metabolism. Plant J 12:323 333 Fernie AR, Roessner U, Trethewey RN, Willmitzer L (2001) The contribution of plastidial phosphoglucomutase to the control of starch synthesis within the potato tuber. Planta 213:418 426 Fernie AR, Tauberger E, Lytovchenko A, Roessner U, Willmitzer L, Trethewey RN (2002) Antisense repression of cytosolic phosphoglucomutase in potato (Solanum tuberosum) results in severe growth retardation, reduction in tuber number and altered carbon metabolism. Planta 214:510520 Goldschmidt EE, Golomb A (1982) The carbohydrate balance of alternate-bearing citrus trees and the signicance of reserves for owering and fruiting. J Am Soc Hort Sci 107:206208 Goldschmidt EE, Koch KE (1996) Citrus. In: Zamski E, Schaer AA (eds) Photoassimilate distribution in plants and crops. Dekker, New York, pp 797823 Graham IA, Denby KJ, Leaver CJ (1994) Carbon catabolite repression regulates glyoxylate cycle gene-expression in cucumber. Plant Cell 6:761772 Harrison CJ, Mould RM, Leech MJ, Johnson SA, Turner L, Schreck SL, Baird KM, Jack PL, Rawsthorne S, Hedley CL, Wang TL (2000) The rug3 locus of pea encodes plastidial phosphoglucomutase. Plant Physiol 122:1871192 Hassid WZ, Neufeld EF (1964) Quantitative determination of starch in plant tissue. Methods Carbohydr Chem 4:3336 Irving DE, Shingleton GJ, Hurst PL (1999) Starch degradation in buttercup squash (Cucurbita maxima). J Am Soc Hort Sci 124:587590 Jacob-Wilk D, Holland D, Goldschmidt EE, Riov J, Eyal Y (1999) Chlorophyll breakdown by chlorophyllase: isolation and functional expression of the Chlase1 gene from ethylene-treated Citrus fruit and its regulation during development. Plant J 20:653661 Kampfenkel K, Mo hlmann T, Batz O, Van Montagu M, D, Neuhaus HE (1995) Molecular characterization of Inze an Arabidopsis thaliana cDNA-encoding a novel putative adenylate translocator of higher-plants. FEBS Lett 374:351 355

20 Yemm EW, Willis AJ (1964) The estimation of carbohydrates in plant extracts by anthrone. Biochem J 57:508514 Yu SM (1999) Cellular and genetic responses of plants to sugar starvation. Plant Physiol 121:687693 Yu Y, Mu HH, Wasserman BP, Carman GM (2001) Identication of the maize amyloplast stromal 112-kD protein as a plastidic starch phosphorylase. Plant Physiol 125:351359