Plant Science 154 (2000) 117–126

Xyloglucan mobilisation in cotyledons of developing plantlets ofHymenaea courbaril L. (Leguminosae-Caesalpinoideae)

Marco Aurelio Silva Tine a,b, Angelo Luiz Cortelazzo b, Marcos Silveira Buckeridge a,*a Secao de Fisiologia e Bioquımica de Plantas, Instituto de Botanica, Caixa Postal 4005, CEP 01061-030, Sao Paulo, SP, Brazil

b Departamento de Biologia Celular, UNICAMP, Campinas, SP, Brazil

Received 9 April 1999; received in revised form 16 November 1999; accepted 16 November 1999

Abstract

Many seeds contain storage compounds that are used by the embryo/plantlet as a source of nutrients after germination. In seedsof Hymenaea courbaril, a leguminous tree, the main reserve consists of a structurally unusual xyloglucan stored in thickened wallsof the cotyledon cells. The present work aimed to study H. courbaril xyloglucan metabolism during and after germination in orderto compare its degrading system with the other known xyloglucan containing seeds. Polysaccharide degradation occurred aftergermination between 35 and 55 days after planting. The activities of a-xylosidase, b-glucosidase, b-galactosidase and XET roseduring the period of xyloglucan disassembling but a low level of endo-b-glucanase activity was detected, suggesting that this XEThas high affinity for the oligosaccharides. The pH optimum of b-galactosidase was different from the a-xylosidase, b-glucosidaseand XET optima suggesting that the former may be important in the control of the mobilisation process. A tentative model forxyloglucan disassembling in vivo is proposed, where b-galactosidase allows the free oligosaccharides to bypass a transglycosylationcycle and be disassembled by the other exo-enzymes. Some ecophysiological comparisons among H. courbaril and otherxyloglucan storing seeds are discussed. © 2000 Published by Elsevier Science Ireland Ltd. All rights reserved.

Keywords: Xyloglucan; Hymenaea courbaril ; Seed; Storage; Cell wall; XET

www.elsevier.com/locate/plantsci

1. Introduction

Many legume seeds are known to accumulatecarbohydrates as storage compounds. Usually, thefirst reserves to be broken down during germina-tion are the raffinose family oligosaccharides,whereas cell wall polysaccharides such as galac-tomannan [1], galactan [2] and xyloglucan [3,4] arethought to be reserves for the growing plantlet,being degraded after germination [5]. Amonglegumes, some members of the subfamily Cae-salpinoideae are known to accumulate xyloglucanin the cotyledons [6–9]. These polymers are

present in storage cell walls of species such astamarind (Tamarindus indica) [6], Copaifera langs-dorffii and Hymenaea courbaril [4,8].

Seed storage xyloglucan have a cellulosic b-(1,4)-glucan backbone, branched at some pointswith a-(1,6)-xylopyranosyl (forming with the glu-cose of the backbone a disaccharide assigned X) orb-(1,2)-D-galactopyranosyl-a(1,6)-D-xylopyranosyl(forming a trisaccharide assigned L) [10]. Exceptfor the absence of terminal fucosyl units a-(1,2)-linked to the b-D-galactopyranosyl groups, there isa remarkable similarity between seed storage xy-loglucans and structural xyloglucan from primarywalls of dicotyledons [11]. The basic polymermolecule is composed of heptasaccharide repeat-ing units of XXXG with variation in the substitu-tion with galactose generating oligosaccharidesstructures like XLXG, XXLG and XLLG.

Abbre6iations: b-gal, b-galactosidase; XET, xyloglucan-endo-trans-glycosylase; a-xyl, a-xylosidase; b-glc, b-glucosidase.

* Corresponding author. Tel.: +55-11-5584-6300, ext. 287; fax:+55-11-577-3687.

E-mail address: buckeridge@smtp-gw.ibot.sp.gov.br (M.S. Bucke-ridge)

0168-9452/00/$ - see front matter © 2000 Published by Elsevier Science Ireland Ltd. All rights reserved.

PII: S 0 1 68 -9452 (99 )00245 -9

M.A.S. Tine et al. / Plant Science 154 (2000) 117–126118

Although Leguminosae seems to be the princi-pal family to possess storage xyloglucan in seeds, amember of Tropaeolaceae (Tropaeolum majus) isthe xyloglucan-containing system best studied todate. The work by Reid and collaborators hasshown that the rate of xyloglucan mobilisation inthis species coincided with the increase and laterdecrease in the levels of four hydrolytic enzymes:endo-b-glucanase, b-galactosidase (b-gal), a-xy-losidase (a-xyl) and b-glucosidase (b-glc) [12].Since then, the hydrolases were purified to homo-geneity and some of their properties were studied.Endo-b-glucanase was purified by Edwards et al.[13], and further studies on its mode of action onxyloglucan were performed by Fanutti et al. [14],demonstrating that this enzyme is a xyloglucan-endo-transglycosylase (XET).

As far as we know, the only legume which hadits hydrolases studied is Copaifera langsdorffii.Buckeridge et al. [4] inferred the presence of theseenzymes through the detection of glucose, galac-tose and xylose released by crude enzymic extractscontaining endogenous water soluble xyloglucan.They also detected the exo enzymes (b-gal, a-xyland b-glc) using artificial substrates.

We now present data on the mobilisation of thexyloglucan of H. courbaril, a native tropicallegume known to accumulate up to 40% of theseed dry weight as xyloglucan [8]. Recently a newoligosaccharide was isolated and identified in thisseed [15]. The presence of this oligosaccharidemakes this polysaccharide unique among othersdescribed in the literature, possibly with implica-tions in the metabolism of the cell wall.

2. Material and methods

2.1. Seed origin and germination conditions

Seeds of H. courbaril L. were provided by theSeed Department of the Institute of Botany at SaoPaulo. The quiescent seeds had an average mass of4.7491.18 g and only seeds between 3 and 6.5 gwere used in the experiment to avoid variationscaused by very different seed sizes. The seeds werescarified individually by abrading the seed coatwith sand paper, weighed, soaked in commercialsodium hypochlorite solution for 5 min, washedwith tap water for 10 min, soaked in distilledwater for 12 h and planted in vermiculite. Trays

with the seeds regularly spaced were kept at 25°Cunder a 12 h photoperiod. The regular distributionof the seeds in the trays allowed the calculation ofthe relative dry weight (i.e. the weight of theembryo or the pair of cotyledons divided by theweight of the quiescent seed), because the initialweight of each seed was known. Every 5 days, 20seeds were collected until the abscission of thecotyledons which occurred 75 days after planting.

Of the 20 seeds collected, 15 were used for drymass measurements, two were used for micro-scopic observations and three were subjected tocarbohydrate extraction and analyses and mea-surement of enzyme activities. The seeds used forthe studies of mass allocation were dissected andhad their cotyledons and embryo dried for 24 h at80°C and weighed for the determination of dryweight. The values were expressed in relative dryweight as a result of the difference in the size ofthe seeds used.

2.2. Light microscopy

The seeds collected for microscopic observationswere fixed with p-formaldehyde (4%)/glutaralde-hyde (2.5%) for 24 h at 5°C, dehydrated in a seriesof ethanol (70, 80, 95 and 100%, successively) atroom temperature, included in Paraplast and 8 mmsections were obtained [16]. For staining withiodine, the sections were covered with a solutionof I2/KI (0.5/1%) and photographed immediately.

2.3. Biochemical analyses

All polymeric or oligomeric xyloglucan used inthis work was obtained as described in [15]. Theseeds collected for biochemical analyses were cutinto pieces of approximately 5 mm and the poolwas separated into two groups: one was dried asdescribed before, ground and stored at −20°Cuntil usage for carbohydrate analyses; the otherwas homogenised in 50 mM sodium acetate buffer(pH 5) with NaCl (200 mM), centrifuged(10 000×g, 10 min) and the supernatant was usedas crude extract for determination of enzymeactivities.

Before measurements of the hydrolases activitiescould be made, the optimal pH for each one wasdetermined. For the determination of pH optima,crude extract obtained as described above wasincubated in citrate-phosphate buffer in pHs rang-

M.A.S. Tine et al. / Plant Science 154 (2000) 117–126 119

ing from 2.6 to 7.6 (with increases of 0.2 units).Sodium acetate buffer (pH 5) was used in allassays in the rest of the work. In the viscometricassays, the fall in viscosity, characteristic of endoactivity, was observed only when xyloglucanoligosaccharides were added to the assay medium(see below). Therefore, limit digest xyloglucanoligosaccharides were added to the assay and theenzyme named XET.

The choice of substrates was based on theknowledge gathered after the studies by Reid andco-workers [12–14,17]. For measuring a-xyl activ-ity, the release of pentose (measured according to[18]) from cellulase limit digest oligosaccharides(obtained as in [15]) was used. The enzymaticextract (20 ml) was incubated with 25 ml of sodiumacetate buffer (100 mM pH 5.0) and 20 ml of a 20mg ml−1 oligosaccharide solution for 24 h at 30°Cand the amount of pentose released was measured.b-gal was assayed against pNP-b-D-galactopyra-noside (Sigma) [10 ml of crude extract, 10 ml ofsodium acetate (100 mM pH 5.0) and 10 ml of 50mM substrate, incubated for 30 min at 40°C].b-glc was assayed similar to b-gal, but using pNP-b-D-glucopyranoside as substrate. XET was as-sayed viscometrically using purified H. courbarilxyloglucan as substrate (400 ml of 1.2% xyloglu-

can, 10 ml of a 20 mg ml−1oligosaccharide solu-tion, 50 ml of enzymatic extract and 60 ml of 1 Msodium acetate pH 5, incubated at 30°C and theflow through a 0.2 ml pipette was measured every5 min). The activity of endo-glucanase was mea-sured as described above but without addition ofoligosaccharides. It was calculated as the slope ofthe line obtained from the logarithm of the flowtime versus incubation time.

2.4. Xyloglucan extraction and analysis

For extraction of polysaccharides, 100 mg ofdried seeds were extracted with ethanol (4×750 mlof 80% ethanol, 80°C, 10 min) and then with water(4×1 ml, 80°C, 180 min). The insoluble materialwas extracted with 4 M KOH with 26 mM sodiumborohydride (3×1 ml, 180 min, room tempera-ture). The alkali extracted polymer was neutralisedwith acetic acid and both water and alkali ex-tracted polysaccharides were dialysed against dis-tilled water and freeze dried. The polysaccharideswere hydrolysed with sulphuric acid according toRef. [19] or hydrolysed with cellulase (Megazyme-Australia) according to Ref. [15]. The monosac-charide products of acidic hydrolysis wereanalysed by HPAEC-PAD (Dionex) using iso-cratic elution (23 mM NaOH, 0.8 ml min−1) for20 min with a PA-1 column (CARBOPAK). Theproportions among monosaccharides were cor-rected according to detector sensitivity to eachmonosaccharide, calculated by using equimolarstandards. The xyloglucan oligosaccharides wereanalysed in a gradient of sodium acetate from 75to 116.5 mM in isocratic NaOH (150 mM) for 40min (1 ml min−1) using the same column.

3. Results

3.1. Mass allocation and morphologiccharacteristics

Fig. 1 shows the changes in the dry mass of thecotyledon and embryo. The first drop in thecotyledon dry weight occurred between 10 and 20days and was as a result of the seed coat naturaldetachment. Germination of most of the seedsoccurs 35 days after planting and during the next20 days the embryo dry mass increases while themass of the cotyledons decrease. The primary

Fig. 1. Changes in the relative dry mass of the cotyledons (")and embryo ( ) during germination and seedling establish-ment. The arrow indicates the end of the germination process(radicle protusion). The dry mass was divided by the mass ofquiescent seed, to avoid variation caused by different seedsizes. Each point corresponds to the average of 15 seeds andthe error bars indicate the standard deviation. The barsindicate the average amount of high molecular weight carbo-hydrate extractable with water from the seeds.

M.A.S. Tine et al. / Plant Science 154 (2000) 117–126120

Fig. 2. Sections of parenchymatic cells of the cotyledons ofHymenaea courbaril stained with iodine. In quiescent seeds (a)the storage xyloglucan (W) and protein bodies ( ) can beobserved. During xyloglucan breakdown (b), starch grains( ) appear in the cytoplasm. After xyloglucan mobilisation(c), the storage wall has been completely dissembled and onlya few starch grains can be seen. Bars=10 mm.

leaves expand to their full length by day 45. It isonly after germination that xyloglucan degrada-tion takes place. Mobilisation starts after day 40being complete by day 70, when only residualamounts of xyloglucan are left into severelyshrunken cotyledons. Xyloglucan solubility in-creases from days 25 to 45, prior to degradation.

3.2. Light microscopy

Fig. 2 shows transversal cuts of the storageparenchyma of cotyledons of H. courbaril at dif-ferent stages of xyloglucan mobilisation. In quies-cent seeds, the appearance of the thick storage cellwalls is smooth and protein bodies were observedinto the cytoplasm (Fig. 2a). At the beginning ofxyloglucan disassembling (day 40) the smooth ap-pearance of the cell wall had disappeared. Thestaining with iodine was altered and lamellaecould be observed. At the same time, starch couldbe seen in the cytoplasm and the thickness of thewall increased two fold (Fig. 2b). After nearlycomplete mobilisation (day 60) few xyloglucancould be detected by iodine staining (Fig. 2c) andsome starch remained in the cytoplasm. Wallswere much thinner, accounting for only 10% of theinitial thickness.

3.3. Biochemical analyses

For the optimisation of assay conditions, thepH optima of the hydrolases were determined(Fig. 3). b-glc, a-xyl and XET have maximal activ-ities around pH 4.5, although the latter was activein a broad range of pHs (from 4 to 7). Exceptionwas made for b-gal, which presented a sharp peakof activity at pH 3.2 and less than 30% of itsactivity at pH 5.0, in which the other hydrolasesare most active.

In the XET/endo-b-glucanase assay, the pres-ence of xyloglucan oligosaccharides was requiredfor the detection of activity (Fig. 4). Although amixture of oligosaccharides with different struc-tures and molecular weights was used, an appar-ently Michaelis-Menten kinetics was obtained,which led us to consider the oligosaccharides assubstrates for the transglycosylation reactions.

Using the conditions mentioned above, a timecourse of the activity of each enzyme was obtained(Fig. 5). Although rigidly controlled conditions oflight, temperature and water availability were

M.A.S. Tine et al. / Plant Science 154 (2000) 117–126 121

used, high variability was observed on the enzymeactivities during the time course. Fig. 5 shows thatb-glc and b-gal activities were already present inquiescent seeds (but rise during xyloglucan disas-sembling) whereas a-xyl and XET activities were

not detected at the beginning and rose during andafter germination, respectively. The cellulase activ-ity raised together with transglycosylation activity,indicating that the Hymenaea XET has both activ-ities, like Nasturtium XET [14] or contaminationof the enzyme extract with xyloglucan which, dur-ing the incubation, generates oligosaccharides.

3.4. Carbohydrate analysis

The changes in xyloglucan structure followinggermination and plantlet growth were probed bymonosaccharide and limit digest oligosaccharidesanalyses by HPAE chromatography. Table 1shows monosaccharide analysis of the xyloglucanextracted with hot water and alkali from thecotyledons of quiescent seeds and from 30, 45 and75 days old cotyledons. The proportion of ara-binose in the water-extracted polymer increasedafter mobilisation to a proportion close to the onefound in the alkali extracted polysaccharide. Thisarabinose is probably present in xyloglucanmolecules, because acid hydrolysis of cellulaselimit digest oligosaccharides, which hydrolysednearly 70% of the xyloglucan from H. courbaril,yields arabinose in approximately the same pro-portions (data not shown). Galactose and xylosealso showed changes in proportion. Their propor-tion (amount of monosaccharides for each of thefour glucoses) dropped from 1.2 and 3.2, respec-tively, during imbibition, to 0.9 and 2.9 after 30days and rose back to 1.1–1.2 and 3–3.3, respec-tively, following mobilisation. These data can becompared with the analyses of the limit digestoligosaccharides (Fig. 6), which show that thepeak assigned to oligosaccharide XXLG dramati-cally decreases at the beginning of xyloglucandisassembling, when compared to the other peaks.Together with the change in solubility (see Fig. 1),these data suggest that xyloglucan was somehowprocessed before complete mobilisation to theembryo.

4. Discussion

4.1. Seedling growth and xyloglucan mobilisation

The seeds of the tropical legume H. courbariltake approximately 30 days to germinate. They arelarge seeds and water imbibition takes about one

Fig. 3. Activities of xyloglucan-disassembling enzymes presentin crude extracts of 40 days old cotyledons of Hymenaeacourbaril at different pHs. A-activity of XET (, expressed inarbitrary units) and b-glucosidase (2, assayed upon pNP-b-glucopyranoside). B-activity of b-galactosidase (, assayedupon pNP-b-galactopyranoside) and a-xylosidase (�, assayedupon xyloglucan oligosaccharides).

Fig. 4. XET activity as a function of the mass of xyloglucanoligosaccharides present in the assay. The kinetics shows therequirement of oligosaccharides as a substrate for transgly-cosilation reactions. Other kinetic parameters could not becalculated as a result of the presence of different oligosaccha-rides with distinct molecular weights in the assay. XET activ-ity is expressed as viscometric units (see main text).

M.A.S. Tine et al. / Plant Science 154 (2000) 117–126122

Fig. 5. Enzymatic activities detected in crude extract of Hymenaea courbaril cotyledons following germination and early plantletgrowth. A-a-xylosidase, assayed upon xyloglucan oligosaccharides; B-b-galactosidase, assayed upon pNP-b-D-galactopyranoside;C-b-glucosidase, assayed upon pNP-b-D-glucopyranoside; and D-XET assayed by viscosimetry with () and without ( )xyloglucan oligosaccharides, in arbitrary units.

third of the germination period (10 days), the restconsisting of metabolic activities that culminate inradicle protrusion. Only after germination, themobilisation of xyloglucan from the cotyledonsstarts. This could be concluded from the reductionof the cotyledon average dry mass with an increasein the dry mass of the embryo (Fig. 1), the reduc-tion of the cell wall thickness after this period, theappearance of starch in the cell (indicating a largeinput of carbohydrate into the cell — see Fig. 2)and also by the rise in the activities of the enzymesinvolved in xyloglucan disassembling (Fig. 5). De-spite the presence of other reserves in the cotyle-dons of quiescent seeds (sucrose, raffinose familyoligosaccharides and protein bodies, data notshown), it is only during xyloglucan mobilisationthat the dry mass of the embryo changessignificantly.

The aspect of the cotyledon cell walls afterxyloglucan mobilisation (Fig. 2C) indicates thatsome structural elements of the wall are not mo-bilised after germination. The presence of twopopulations of xyloglucan with different ex-

tractability can be suggested from the data inshown in Table 1. Most of the xyloglucan presentin quiescent seeds can be extracted with hot water.

Table 1Proportion of monosaccharides obtained by acidic hydrolysisof the polysaccharide extracted from the seed powder withwater and KOH (4 M)a

Days after planting

0 30 45 75

1.10.30.2Water 0.2Ara1.2 0.9Gal 1.1 1.2

Glc 4.04.0 4.04.03.03.32.93.2Xyl

KOH Ara 0.7 0.7 1.1 1.4Gal 1.2 0.9 1.0 1.1Glc 4.04.0 4.04.0

3.2Xyl 2.9 3.0 3.3

a The area of the glucose peak was considered 4 and theother areas were calculated proportionally. All the areas werecorrected for the different sensitivity of the detector for eachmonosaccharide, using an equimolar standard.

M.A.S. Tine et al. / Plant Science 154 (2000) 117–126 123

Fig. 6. Changes in the proportions of the limit digest xyloglu-can oligosaccharides before breakdown of the polysaccharide.The known peaks are (a) XXXG; (b) XXLG; (c) XLLG; and(d) XXXXG.

4.2. Biochemical aspects of xyloglucandisassembling

Three facts suggests that a modification in thefine structure of the xyloglucan occurs prior tomobilisation: (1) the change in polymer solubilitythat made the polymer more easily extractableimmediately before disassembling; (2) change inthe appearance the of the cell wall; and (3) changein the polymer structure. The existence of enzymeswith exo-activities, especially b-gal, which is theonly one reported to hydrolyse the polymer [20],could explain the changes in the fine structure ofthe xyloglucan prior to mobilisation. Although afamily of b-galactosidases occurs in seeds, onlysome of its members were purified and character-ised [21,22] and they show differences in specific-ity. The diversity of b-galactosidases and theirmodes of action suggest a high complexity in theprocess of degalactosylation of the xyloglucan.For example, the b-gal from C. langsdorffii [22]was shown to hydrolyse galactose from XLLGand XLXG, but not from XXLG. The characteri-sation of this purified enzyme led to similar resultsfor Hymenaea concerning the pH optimum foractivity, i.e. a peak at pH 3.2.

As a pH optimum around 5 is expected for wallenzymes [23], the difference observed between thepH optima for b-galactosidase (pH 3.2) and theother enzymes (a-xyl, b-glc and XET) suggest thatb-gal activity may limit, under certain circum-stances, xyloglucan disassembling. This can be im-portant in the control of the process, because thedisassembling cannot proceed without the retrievalof the galactose [24]. The acidic growth theory [25]and the model for the control of cell expansionbased on the differences of pH optima of theenzymes involved in the process [26] are two previ-ous examples of regulation of cell wall metabolismbased on pH.

The requirement of xyloglucan oligosaccharidesfor detection of viscosity decrease in the assaysystem for endo-b-glucanase suggests that in H.courbaril, the enzyme responsible for the release ofxyloglucan fragments from the cell wall is a xy-loglucan endo-transglycosylase (XET). However,differently from the enzyme purified from T. ma-jus, this XET has a very low activity againstpurified polymer in the absence of oligosac-charides.

This xyloglucan is poor in arabinose and it de-clines after germination, suggesting a storage role.A smaller part of the xyloglucan content of theseed is extracted only with alkali. This pool ofxyloglucan is rich in arabinose and does not de-cline after germination (data not shown). Theincrease in the proportion of arabinose in thestorage xyloglucan after mobilisation is probably aresult of the decline in storage polymer contentand proportional increase in the contaminationwith the ‘structural’ arabinose containing polymer.In fact, Reis et al. [3] observed an inner and outerwall in cotyledon cells of tamarind. The distinctionwas based on the transmission electron mi-croscopy observations. Although our light mi-croscopy does not permit such an observation, itis possible that those features are also presentin walls of Hymenaea. This separation of xyloglu-can populations would be an important elementfor the characterisation of a ‘pattern’ of mobilisa-tion (see below), because the parenchymatic cellshave to remain functional near to the vascularbundles after their storage xyloglucan has beendegraded.

M.A.S. Tine et al. / Plant Science 154 (2000) 117–126124

4.3. An impro6ed model for storage xyloglucandisassembling

In some aspects, our data appear to differ fromwhat has been observed by Edwards et al. [12] incotyledons of T. majus, in which the activities ofthe four enzymes overlap. On the basis of T. majussystem, Buckeridge and Reid [5] proposed a modelfor xyloglucan degradation in which the polymerwould enter a linear disassembling process whichwould lead from polymer to monomer directly,beginning with the action of one XET (with endo-glucanase activity at low concentrations ofoligosaccharides). The fragments released wouldbe hydrolysed by the sequential action of b-gal,a-xyl and b-glc, producing free galactose, xyloseand glucose.

It is possible that in species of the Leguminosae,the xyloglucan degrading system works differentlyfrom the T. majus one. Thus, some changes in thismodel can be suggested to explain the resultspresented here and by Buckeridge et al. [4], ob-tained with other seeds (from legumes) that storexyloglucan (H. courbaril, C. langsdorffii ). This newtentative model (Fig. 7) proposes that seed xy-loglucan degradation occurs in at least three steps:(1) transglycosylation; (2) degalactosylation; and(3) oligosaccharide disassembling. Such a model is

not linear, but would have two different cycles(transglycosylation and disassembling) whichwould be linked by the action of b-gal. In acondition of low b-gal activity, the fragments gen-erated by the XET would not be disassembled andthe rise in their concentration would lead to anincrease in endo-transglycosylation activity, halt-ing hydrolysis and avoiding an excessive rise in theinput of carbohydrate into the cell. After theincrease in b-gal activity, these fragments of xy-loglucan would be degalactosylated and subse-quently enter the second cycle of degradation,releasing free monosaccharides.

With only a few cell wall storage systems bio-chemically studied so far, it is difficult to suggest ageneral model that describes the disassembling ofthe storage xyloglucan in every seed. As differentspecies are exposed to different environmentalconditions, this is likely to reflect distinct strategiesof adaptation, requiring diverse physiological andbiochemical characteristics. As H. courbaril is atropical tree with seeds much larger than T. majus,the disassembling of xyloglucan at the same timein all the cells might lead to an imbalance of thesource/sink relation, therefore resulting in an over-load of the transport system towards the embryo.Indeed, the accumulation of starch inside cotyle-don cells suggests that the input of carbohydrateinto the cell is greater than the output. If theproposed model is correct, the presence of transg-lycosylation even at low oligosaccharide concen-tration (contrarily to T. majus) might serve tomake the disassembling system of H. courbarilpartially reversible, unless b-gal is active. Thiswould delay degradation and possibly synchroniseit with a relatively slower pace of mobilisation (30days from start to end of degradation). This con-trasts with T. majus which, being a faster grower,degrades xyloglucan within an interval of only 5days [12].

4.4. Ecological meaning of storage xyloglucanmobilisation strategy

Our microscopic observations suggest that thebreakdown of xyloglucan starts in the cells nearthe vascular bundles and spreads towards the restof the cotyledons later on (data not shown).Legume seeds have different patterns of storagemobilisation [27] and it possibly represents anoptimisation of sugar transport. Under the condi-

Fig. 7. Tentative model proposed for the in vivo disassem-bling of the storage xyloglucan in seeds. The polysaccharide isfirst hydrolysed by the XET characterising a period of Trans-glycosylation. Only when certain galactose residues at XLXGand XLLG are hydrolysed (degalactosylation), the remainingoligosacharides of higher fragments containing XXLG andXXXG enter the final cycle of disassembling in which glucoseand xylose are attacked by b-glucosidase and a-xylosidase toproduce free monosaccharides.

M.A.S. Tine et al. / Plant Science 154 (2000) 117–126 125

tions used in this work to detect enzyme activi-ties, these positional differences in activities wereprobably lost and instead of a sharp peak ofactivity, a smooth rise and a plateau were ob-served.

Distinct ecological strategies could also explainsome differences between these two plants. SinceT. majus is herbaceous and germinates in areaswith higher light intensity than H. courbaril, itsplantlets would have to grow faster in order toachieve reproductive success in less than a year.H. courbaril plantlets, on the other hand, growslower under the canopy and may stay undershadowed conditions for much longer periods.This requirement of an internal source of carbo-hydrate for longer periods could explain thelarger seed found in tropical trees, the greateramount of xyloglucan and the longer period forits mobilisation as well as the biochemical differ-ences among the xyloglucan disassembling sys-tems.

References

[1] J.S.G. Reid, Cell wall storage carbohydrates in seeds —biochemistry of the seed ‘gums’ and ‘hemicelluloses’,Adv. Bot. Res. 11 (1985) 125–155.

[2] L.A. Crawshaw, J.S.G. Reid, Changes in cell wallpolysaccharides in relation to seedling development andthe mobilization of reserves in the cotyledons of Lupinusangustifolius Cv, Unicrop. Planta 160 (1984) 449–454.

[3] D. Reis, B. Vian, D. Darzens, J.C. Roland, Sequentialpatterns of intramural digestion of galactoxyloglucan intamarind seedlings, Planta 170 (1987) 60–73.

[4] M.S. Buckeridge, D.C. Rocha, J.S.G. Reid, S.M.C. Di-etrich, Xyloglucan structure and post-germinativemetabolism in seeds of Copaifera langsdorfii from savanaand forest populations, Physiol. Plantarum 86 (1992)145–151.

[5] M.S. Buckeridge, J.S.G. Reid, Major cell wall storagepolysaccharides in legume seeds: structure, catabolismand biological functions, Cienc. Cult. 48 (1996) 153–162.

[6] P. Kooiman, The constitution of Tamarindus-amyloid,Rec. Trav. Chim. Pays-Bas 80 (1961) 849–865.

[7] R.M. Polhill, P.H. Raven, Advances in Legume System-atics. Part 1, Royal Botanic Gardens, Kew, Richmond,1981.

[8] M.S. Buckeridge, S.M.C. Dietrich, Galactomannan fromBrazilian legume seeds, Revta. Bras. Bot. 13 (1990) 109–112.

[9] M.S. Buckeridge, V.R. Panegassi, D.C. Rocha, S.M.C.Dietrich, Seed galactomannan as marker in the taxon-omy and evolution of the Leguminosae, Phytochemistry38 (1995) 871–875.

[10] S.C. Fry, W.S. York, P. Albersheim, T. Hayashi, J-P.

Joseleau, Y. Kato, E. Perez Lorences, G.A. Maclachlan,M. McNeil, J.S.G. Reid, H.U. Seitz, R.R. Selvendran,A.G.J. Voragen, A.E. White, An unambiguous nomen-clature for xyloglucan-derived oligosaccharides, Physiol.Plant 89 (1993) 1–3.

[11] T. Hayashi, Xyloglucans in the primary cell wall, Annu.Rev. Plant Physiol. Plant Mol. Biol. 40 (1989) 139–168.

[12] M. Edwards, I.C.M. Dea, P.V. Bulpin, J.S.G. Reid,Xyloglucan (amyloid) mobilisation in the cotyledons ofTropaeolum majus L. seeds following germination, Planta163 (1985) 133–140.

[13] M. Edwards, I.C.M. Dea, P.V. Bulpin, J.S.G. Reid,Purification and properties of a novel xyloglucan-specificendo-(1,4)-b-glucanase from germinated nasturtium seeds(Tropaeolum majus L.), J. Biol. Chem. 261 (1986) 9489–9494.

[14] C. Fanutti, M.J. Gidley, J.S.G. Reid, Action of a purexyloglucan endo-transglycolase (formerly called xyloglu-can-specific endo-(1,4)-b-glucanase) from the cotyledonsof germinated nasturtium seeds, Plant J. 3 (1993) 691–700.

[15] M.S. Buckeridge, H.J. Crombie, C.J.M. Mendes, J.S.G.Reid, M.J. Gidley, C.C.J. Vieira, A new family ofoligosaccharides from the xyloglucan of Hymenaea cour-baril L. (Leguminosae) cotyledons, Carbohydr. Res. 303(1997) 233–237.

[16] J.A. Kierman, Histological and Histochemical Methods.Theory and Practice, Pergamon, London, 1981, p. 334.

[17] C. Fanutti, M.J. Gidley, J.S.G. Reid, A xyloglucanoligosaccharide specific a-D-xylosidase or exo-oligoxy-loglucan-a-xylohydrolase from germinated nasturtium(Tropaeolum majus L.) seeds. Purification, properties andits interaction with xyloglucan-specific endo-(1,4)-b-glu-canase and other hydrolases during storage xyloglucanmobilisation, Planta 184 (1991) 137–147.

[18] J.H. Roe, E.W. Rice, A photometric method for thedetermination of free pentose in animal tissues, J. Biol.Chem. 173 (1948) 507–512.

[19] J.F. Saeman, J.L. Buhl, E.E. Harris, Quantitative saccha-rification of wood and cellulose, Ind. Eng. Chem. Anal.Ed. 17 (1945) 35–37.

[20] J.S.G. Reid, M. Edwards, I.C.M. Dea, Enzymatic mod-ifications of natural seed gums, in: G.O. Phillips, D.J.Wedlock, P.A.W. Iliams (Eds.), Gums and Stabilisers forthe Food Industry 4, IRL Press, Oxford, 1988, pp.391–398.

[21] M. Edwards, J.L. Bowman, I.C.M. Dea, J.S.G. Reid, Ab-D-galactosidase from nasturtium (Tropaeolum majusL.) cotyledons, J. Biol. Chem. 263 (1988) 4333–4337.

[22] P.H.N. Alcantara, S.M.C. Dietrich, M.S. Buckeridge,Xyloglucan mobilisation and purification of a (XLLG/XLXG) specific b-galactosidase from cotyledons of Co-paifera langsdorffii, Plant Physiol. Biochem. 37 (1999)653–663.

[23] S.C. Fry, Polysaccharide-modifying enzymes in the plantcell wall, Annu. Rev. Plant Physiol. Plant Mol. Biol. 46(1995) 497–520.

[24] H.J. Crombie, S. Chengappa, A. Hellier, J.S.G Reid, Axyloglucan oligosaccharide-active, transglycosilating b-D-

M.A.S. Tine et al. / Plant Science 154 (2000) 117–126126

glucosidase from the cotyledons of nasturtium (Tropae-olum majus L.) seedlings-purification, properties andcharacterization of a cDNA clone, Plant J. 15 (1998)27–38.

[25] F.B. Salisbury, C.W. Ross, Plant Physiology, fourthedition, Wadsworth Publishing Company, California,1992, p. 682.

[26] J. Ricard, G. Noat, Eletrostatic effects and the dynamicsof enzyme reactions at the surface of plant cells, Eur. J.Biochem. 155 (1986) 183–190.

[27] D.L. Smith@, Cotyledons of leguminosae, in: R.M. Pol-hill, P.H. Raven (Eds.), Advances in Legume Systemat-ics. Part 2, Royal Botanic Gardens, Kew, Richmond,1981, pp. 927–940.

.