Planta (2003) 216: 484–495 DOI 10.1007/s00425-002-0869-3
O R I GI N A L A R T IC L E
Jo C. Dumville Æ Stephen C. Fry
Gentiobiose: a novel oligosaccharin in ripening tomato fruit
Received: 27 February 2002 / Accepted: 7 July 2002 / Published online: 10 October 2002 Springer-Verlag 2002
Abstract Two neutral disaccharides, gentiobiose [b-DGlcp-(1fi6)-D-Glc] and nigerose [a-D-Glcp-(1fi3)-DGlc], were detected in tomato (Lycopersicon esculentum Mill.) pericarp and locule. Gentiobiose was present in the locule of green fruit and ripe fruit at 0.88 and 5.8 lmol (kg fresh weight)–1, respectively. When vacuum-infiltrated into green tomato fruit, exogenous gentiobiose (50 or 200 lg per fruit) hastened the initiation of ripening (as judged by colour change) by 1–3 days relative to fruit that were infiltrated with glucose or isomaltose. Nigerose plus gentiobiose was particularly effective, but nigerose alone had no significant effect. The endogenous disaccharides were found to be present in the apoplastic fluid of the fruit, compatible with a proposed intercellular signalling role. The origin and metabolic fate of the disaccharides were investigated. Phenolic esters of these disaccharides were not detectable in tomato fruit and it is therefore unlikely that the free disaccharides were formed from a pool of such esters. An alternative possible biosynthetic origin, via transglycosylation, is discussed. When [14C]gentiobiose was vacuum-infiltrated into unripe or ripe fruit, the disaccharide remained intact for at least 1 h but was largely degraded within 24 h. The results suggest that gentiobiose is a new, naturally occurring oligosaccharin with a rapid turnover rate. Keywords Gentiobiose Æ Lycopersicon Æ Nigerose Æ Oligosaccharin Æ Ripening Æ Turnover Abbreviations BAW, EAW, EPyW: chromatography solvents (see Materials and methods) Æ Glc: glucose Æ Glcp: glucopyranose Æ GPC: gel-permeation chromatogra-
J.C. Dumville Æ S.C. Fry (&) The Edinburgh Cell Wall Group, ICMB, The University of Edinburgh, Daniel Rutherford Building, The King’s Buildings, Edinburgh, EH9 3JH, UK E-mail:
[email protected] Fax: +44-131-6505392
phy Æ PyAW: volatile buffer, pyridine/acetic acid/water (ratio specified in text) Æ XXFG: a nonasaccharide (Glc4.Xyl3.Gal.Fuc) derived from xyloglucan
Introduction Fruit ripening involves changes in texture, flavour, colour and aroma once the growth of the fruit is complete (Brady 1987). Some depolymerisation of hemicelluloses (Sakurai and Nevins 1993) and pectins (Huber and O’Donoghue 1993; Brummell and Labavitch 1997) occurs during ripening. The depolymerisation may be both enzymic and non-enzymic (Fry et al. 2001). Besides probably softening the tissue, such degradation of cell wall polysaccharides can lead to the release of oligosaccharins. The latter are oligosaccharides with biological activity and potential inter-cellular signalling roles (Albersheim and Darvill 1985; Ryan and Farmer 1991; Aldington and Fry 1993; Coˆte´ and Hahn 1994; Dumville and Fry 2000). Exogenous oligosaccharins (synthesised in vitro or isolated from non-plant sources) can influence fruit ripening. For example, certain in-vitro hydrolysis products of plant cell walls can induce ethylene biosynthesis (Baldwin and Biggs 1988). Such products include oligogalacturonides of degree of polymerisation (DP) >8 (Campbell and Labavitch 1991a) and D-galacturonaterich oligosaccharides carrying neutral side-chains (Campbell and Labavitch 1991b; Melotto et al. 1994). In addition, two oligosaccharides from human urine [Man5. GlcNAc and Man3.(Xyl).GlcNAc.(Fuc).GlcNAc] promoted the initiation of ripening in tomato fruit (Priem and Gross 1992). The demonstration of effects of exogenous oligosaccharins does not prove that they play any natural biological role in vivo. Evidence for the occurrence of endogenous oligosaccharins would therefore be particularly valuable in establishing a natural signalling role. Endogenous oligosaccharins are indeed produced in plant cell-suspension cultures, from xyloglucan (McDougall and Fry 1990), xylomannoside-glycoproteins (Priem et al.
485
1990), phytoglycolipids (Smith and Fry 1999), pectins [in some cell cultures (Tani et al. 1992) but not in others (Garcı´ a-Romera and Fry 1997)] and unidentified polymers (Roberts et al. 1997). In tomato fruit pericarp, ten endogenous oligosaccharides have been detected which have a pentasaccharide core (Man3.GlcNAc2) substituted with various combinations of mannose (Man), xylose (Xyl), fucose (Fuc), N-acetylglucosamine (GlcNAc) and galactose (Gal) residues (Priem et al. 1990). These oligosaccharides may have been released in the fruit from N-linked glycoproteins (Priem et al. 1990). The composition of the oligosaccharides appears to alter as ripening continues (Priem et al. 1993), but no biological activity has yet been reported for these naturally occurring fruit oligosaccharides. For plant tissues to maintain a coherent programme of development, any endogenous oligosaccharins with signalling roles should be inactivated once their message has been perceived by the cell. Previous work in this laboratory has documented the turnover of certain oligosaccharins. For example, biologically active oligogalacturonides of DP9 are broken down in vivo, to yield limit-products of DP5, by the action of a-D-galacturonidase (Garcı´ a-Romera and Fry 1995). In spinach cell cultures, the xyloglucan-derived nonasaccharide, XXFG, can be sequestered by incorporation into extracellular polysaccharides by transglycosylation (Baydoun and Fry 1989). Either of these mechanisms would help to control the action of oligosaccharins. In the present work, we have analysed ripening tomato fruit for endogenously occurring oligosaccharides, and tested some of these for possible oligosaccharin activity. We report the occurrence, biological activity and turnover of gentiobiose – a disaccharide not previously considered as an oligosaccharin.
acidified with 4 ll acetic acid. A control sample of the material was treated with equal amounts of NaOH and acetic acid but in reverse order so that no saponification occurred. The aromatic components were then adsorbed to a C18-silica column (bed volume 0.2 ml) in water, and the eluted aliphatic material was analysed by paper chromatography in EPyW2. Gel-permeation chromatography (GPC) Fruit extracts were fractionated on Bio-Gel P-2 (bed volume 740 ml, flow rate 50 ml/h, fraction volume 9.3 ml; or 170 ml, 11 ml/h, 2.1 ml). The void and included volumes were determined with 3 H-hemicellulose (from maize cell cultures; kindly donated by Dr. Ellen Kerr) and [3H]arabinose, respectively. Elution was with either 0.5% chlorobutanol, or pyridine/acetic acid/water (PyAW, 1:1:23, by vol., pH 4.7) containing 0.5% (w/v) chlorobutanol, or 1 M formic acid. Paper chromatography Paper chromatography was performed by the descending method on Whatman No. 1 paper. Solvents used were (all compositions by volume): EAW = ethyl acetate/acetic acid/water, 10:5:6; BAW = butan-1-ol/acetic acid/water, 12:3:5; EPyW1, ethyl acetate/pyridine/water, 8:2:1; EPyW2, ethyl acetate/pyridine/water, 10:4:3. Carbohydrates were stained with silver nitrate or aniline hydrogenphthalate (Fry 2000). Sugars were eluted from preparative paper chromatograms by the method of Eshdat and Mirelman (1972). Thin-layer chromatography TLC was carried out on Merck plastic-backed silica-gel plates in BAW. After development the plate was sprayed with 0.5% (w/v) thymol in ethanol/H2SO4 (19:1, v/v) and heated at 105 C for 5 min to reveal carbohydrates. High-voltage electrophoresis
Tomato seeds (Lycopersicon esculentum Mill., cultivars Ailsa Craig, Moneymaker and Tiny Tim) were from King’s Seeds, Essex, UK. Plants were grown in a greenhouse with supplemented lighting to give 18-h days. Each truss was limited to five fruit to enhance uniformity of fruit growth. Fruit were harvested at selected ripening stages, identified by external colour.
Electrophoresis was performed on Whatman No. 1 paper at pH 3.5 (PyAW, 1:10:189, by vol.; Wright and Northcote 1975) with orange G as an external marker, or in borate buffer at pH 9.4 (1.9% ‘‘borax’’, adjusted to pH 9.4 with NaOH; Narasimham et al. 1980) with picric acid as an external marker. The electrophoretograms were run at 1 kV for 70 min or 1.5 kV for 50 min in tanks with white spirit as an immiscible coolant. Sugars were stained with AgNO3 and sugar-phosphates with molybdate reagent (Fry 2000). On borate electrophoretograms, the alkaline solution used for AgNO3 staining was 80% (v/v) ethanol containing 2% (w/v) NaOH and 4% (w/v) pentaerythritol. Sugars eluted from a borate electrophoretogram were freed of Na+ by cation-exchange chromatography on Dowex-50 (H+) and the eluate was then repeatedly dried from methanol containing 2% acetic acid to remove H3BO3.
Authentic disaccharides
High-pressure liquid chromatography (HPLC)
All the glucose disaccharides were from Sigma Chemical Co., except kojibiose, which was prepared by the method of Kitao et al. (1994) and kindly donated by Dr. T. Ishii (Forestry and Forest Products Research Institute, Ibaraki, Japan).
HPLC was carried out using a Dionex system (Dionex, Camberley, UK) with a CarboPac PA1 column and a pulsed amperometric detector containing a gold electrode. The volume injected was 20 ll. The eluent programme for disaccharides was: 0–42 min, 30 mM NaOH; 42–58 min, 30–600 mM NaOH (linear gradient); 58–63 min, 600 mM NaOH; 63–64 min, 600–30 mM NaOH; 64–75 min, 30 mM NaOH.
Materials and methods Plant material
Hydrolysis of carbohydrates Acid hydrolysis of 20 lg of disaccharide was performed in 2 M trifluoroacetic acid (TFA) at 120 C for 1 h. Alkali hydrolysis of sugar-esters was attempted by incubation of 20 lg of carbohydrate (from late-eluting fractions of Bio-Gel P2) in 20 ll of 0.1 M NaOH at 25 C for 1 h. The sample was then
Quantitative carbohydrate assays Total carbohydrate was estimated by the phenol–H2SO4 method (Dubois et al. 1956). Hexoses were assayed with anthrone (Dische
486 1962) and uronic acids with m-hydroxybiphenyl (Blumenkrantz and Asboe-Hansen 1973). Bulk extraction and fractionation of sugars from tomato fruit Surface-rinsed pericarp or untreated locular material of mature green and red-ripe tomato fruit (cv. Ailsa Craig or Moneymaker) was stored at –80 C. Frozen tissue was immersed completely in 1 vol. of 15% (w/v) formic acid and shaken for 16 h at 4 C. The extract was passed through four layers of muslin, the filtrate was centrifuged in a Sorvall GSA rotor at 10,000 rpm for 15 min, and 35 ml of the supernatant was fractionated on Bio-Gel P-2 (bed vol. 740 ml) in 1 M formic acid. The use of this solvent yielded neutral oligosaccharides uncontaminated by galacturonic acid and oligogalacturonides, which elute after the included volume. All material of Kav 0.2–1.0 was pooled, freeze dried and re-dissolved in water (2 ml). The sample was freed of charged compounds by both anion- and cation-exchange chromatography, freeze dried and then re-fractionated on Bio-Gel P-2 (170 ml gel) in unbuffered 0.5% chlorobutanol; fractions that contained the disaccharides were subjected to preparative paper chromatography in EPyW2. Compounds 1 and 2 were eluted from the paper chromatogram and analysed by paper chromatography, electrophoresis and HPLC. Extraction of apoplastic fluid A ripe tomato fruit was quartered with a razor blade, and the locular and placental material was discarded. The four pieces of pericarp were rinsed with water for 3 min to remove adherent locular material, and then lightly blotted with absorbent paper. Ten to 15 layers of very thin tissue paper (‘‘Rizla’’ cigarette paper; 1·1 cm squares) were then placed on the washed but uncut locular face of each piece of pericarp. These papers were surrounded by a hollow square of Whatman No. 1 filter paper, positioned to absorb any liquid seeping from the cut cells (Fig. 1). The assembly was left in an open dish for 2 h to allow apoplastic sap to soak into the Rizla paper. The bottom layer of Rizla paper (which had been in direct contact with the pericarp surface) was discarded along with the outer square of Whatman No. 1. Sugars were then eluted from the remaining layers of Rizla paper by the method of Eshdat and Mirelman (1972). Material eluted from 200 layers was pooled. Isolation of apoplastic oligosaccharides from tomato fruit The apoplastic sample was fractionated on Bio-Gel P-2 (170 ml). Disaccharide-containing fractions were dried, re-dissolved in water
(1 ml) and paper-chromatographed in EPyW2. The disaccharides (unstained) were eluted with water and concentrated to 150 ll; a portion (40 ll) was re-analysed by paper chromatography in EPyW2 and stained. The remaining 110 ll was freed of charged compounds by anion- and cation-exchange chromatography, dried and re-dissolved in water (150 ll); an aliquot (20 ll) was fractionated by HPLC. To test whether the sample was indeed apoplastic, the material remaining bound to the anion-exchange column after rinsing with water was eluted with PyAW (1:1:23, pH 4.7); this anionic fraction was concentrated and subjected to paper electrophoresis at pH 3.5. The electrophoretogram was stained with molybdate in order to detect any sugar phosphates that could indicate contamination with cytoplasmic material. Preparation of [14C]gentiobiose [14C]Gentiobiose was made by a modification of the transglycosylation method of Cline and Albersheim (1981). Laminariheptaose (0.5 mg; Dextra Laboratories, Reading, UK) was incubated with [U-14C]glucose (32 lg; 11.1 TBq/mol) in the presence of b-glucosidase (0.1 U, from almonds; Sigma Chemical Co.) in a total volume of 50 ll containing 100 mM pyridinium acetate (pH 5.2) for 24 h at 25 C. The formation of [14C]gentiobiose was monitored by silica-gel TLC in BAW (3:1:1) followed by autoradiography. [14C]Gentiobiose was purified by paper chromatography in EAW followed by EPyW2, and then paper electrophoresis in borate at pH 9.4; in each case a radioactive compound that co-migrated with a non-radioactive, external gentiobiose marker was eluted. The identity of the [14C]gentiobiose was confirmed by co-migration (on paper electrophoresis in borate buffer) of a small amount with internal marker gentiobiose, which was stained after scintillationcounting and exhibited exact co-migration. Upon acid hydrolysis and paper chromatography in EPyW1 the purified compound yielded [14C]glucose, as expected. Bioassays Commercial gentiobiose, isomaltose, nigerose and glucose were purified using preparative paper chromatography (EPyW2) and borate electrophoresis. Sodium borate was removed from the samples (see above), which were further purified on Bio-Gel P-2 in unbuffered 0.5% chlorobutanol; the samples were filtered through a 0.45-lm filter, freeze-dried and re-dissolved in water. Each carbohydrate was vacuum-infiltrated into green tomato fruit that had diameters of at least 4 cm (Ailsa Criag) or 2 cm (Tiny Tim). The fruit were surface-sterilised in 10 mM sodium hypochlorite for 30 s, washed and surface-dried. The dose of carbohydrate supplied was 200 lg in 200 ll water per Ailsa Craig fruit and 50 lg in 50 ll water per Tiny Tim fruit. The solution to be infiltrated was pipetted directly on to the stem scar and the fruit was then placed in a chamber, which was evacuated to a pressure of 60 mm Hg for 40–60 s; subsequent gradual release of the vacuum caused the solution to be drawn into the fruit. Each infiltrated fruit was then placed in an individual, tall, clear glass jar, which was covered with Parafilm and incubated at 25 C with constant white fluorescent lighting of 65 lmol photons m–2 s–1. Ripening was assessed at the same time every day, by external colour change. During bioassays the identity of the compound infiltrated into each fruit remained unknown to the assayer until after the experiment had been completed. Turnover of gentiobiose
Fig. 1 Method for sampling apoplastic fluid from tomato (Lycopersicon esculentum) fruit pericarp by capillary action using thin tissue paper
To establish the fate of gentiobiose in vivo we vacuum-infiltrated [14C]gentiobiose into red-ripe Tiny Tim fruit (10 Bq in 50 ll water for each fruit), as described above, and incubated for 1 or 24 h at 25 C. The pericarp and locular tissue were then excised, frozen at –80 C and thawed into 15% formic acid with stirring for 4 h at 4 C. The extract was filtered through muslin and then centrifuged
487 at 2,500 g for 10 min, and the supernatant was fractionated on BioGel P-2.
Results Identification of nigerose and gentiobiose in ripe tomato fruit Formic acid (15%, v/v) was chosen to extract endogenous oligosaccharides because it very effectively inhibits hydrolytic enzymes that could potentially generate or degrade oligosaccharides post mortem. However, the acid itself might be suspected of hydrolysing labile glycosidic linkages. To investigate this, we stored a sample of the xyloglucan-oligosaccharide, [fucosyl-3H]XXFG, in 15% formic acid. After 3 years’ storage at 4 C,
