Speciation of metal–carbohydrate complexes in fruit and vegetable samples by size-exclusion HPLC-ICP-MS Joanna Szpunar,a Patrice Pellerin,b Alexei Makarov,a Thierry Doco,b Pascale Williamsb and Ryszard £obin�ski*a aCNRS EP132, He�lioparc, 2, av. Pr. Angot, 64 000 Pau, France. E-mail: Ryszard.Lobinski@univpau. fr bInstitut National de la Recherche Agronomique, Institut des Produits de la Vigne, Unite� des Recherches des Polyme`res et des Techniques Physico-Chimiques, 2, pl. Viala, 34 060 Montpellier, France Received 23rd October 1998, Accepted 25th January 1999 Kinetically inert and thermodynamically stable metal complexes with polysaccharides were detected in aqueous leachates and enzymatic digests of apple and carrot samples by size-exclusion chromatography with parallel refractometric and ICP-MS detection.The method developed allowed detection in the water-soluble fraction and the identification of a high molar mass polysaccharide fraction (>50 kDa) containing Pb, Ba, Sr, Ce and B, whereas other metals (Zn, Cu, Mg) eluted as complexes with low molar mass non-carbohydrate compounds.The majority of the metal–carbohydrate complexes were located in the solid water-insoluble fraction of the analysed samples. An extraction procedure with a mixture of pectinolytic enzymes was developed to release these species into the aqueous phase. The metal-binding carbohydrate component was identified as the dimer of rhamnogalacturonan-II, a pectic polysaccharide present in plant cell walls.The unidentified residual metal species contained less than 5% of the metals present. and the quantitative recovery of intact metal species from Introduction solid matrices. Indeed, even for relatively simple organoarsenic Fruits and vegetables are important (and sometimes major) and -selenium compounds, the scarcity of standards has been sources of trace elements in the human diet. In order to be of a considerable shortcoming.Bioligands often exhibit polymorconcern in terms of essentiality and/or toxicity these elements phism and, additionally, mixed metal complexes can be must be bioavailable, i.e. readily absorbable by the gut, and formed. Purification of particular metallocompounds is thus further utilizable in the body. Since the bioavailability depends diYcult and the consequence is again the poor availability of critically on the actual species of an element present, infor- calibration standards. Another critical issue is the sample mation on the total concentration of an element in a foodstuV preparation procedure that should enable the metal species to may be useless with respect to its actual assimilation by be recovered unmodified and to be presented for HPLC in an consumers.1 Precise information regarding the identity, nature aqueous 0.2 mm filtrable solution.To date, most of the studies and concentrations of individual metal compounds present in of the speciation of metal complexes have focused on the a sample (speciation) is therefore required.water-soluble (aqueous buVer) fraction that contained 10–20% Because of the capability of ICP-MS to detect metals and of the elements of concern present in a sample, leaving the metalloids at picogram levels in liquid chromatographic solid residue unexplored. eZuents,2–4 the coupling of HPLC to ICP-MS has been widely FoodstuVs of plant origin contain significant concentrations applied to elemental speciation analysis in foodstuVs of animal of polysaccharides of which the potentially negatively charged and plant origin. The research objectives included the detection oxygen functions can bind cations electrostatically and even and identification of As species in marine edible resources,5 chelate them via polyhydroxy groups.28 In comparison with monitoring of residual As-containing growth additives,6,7 and proteins, however, little is known about the relevance of metal the characterization of Cd– and Zn–metallothionein complexes coordination to carbohydrates that are the most abundant (by in meat and their behaviour during simulated gastrointestinal weight) class of compounds in the biosphere.29 Recently, digestion procedures8–11 and cooking.10 Studies of edible plant attention has been attracted by a structurally complex pectic materials focused on the speciation of Se in nutritional sup- polysaccharide rhamnogalacturonan-II (RG-II).18,20,30–32 This plements12–14 and Se-enriched vegetables,15 B in radish ubiquitous component of primary plant cell walls forms dimers roots,16,17 apple fruit17 and sugar beet,18 metals in tea19 and cross-linked by 152 borate diol esters.30–32 The dimers of Pb in wine.20–22 Speciation of Zn and Cd in vegetables was RG-II borate ester (dRG-II-B) were found to complex in vitro approached by ultra- and dia-filtration techniques23,24 and specific divalent cations31 and the majority of Pb in wine.20–22 SEC25,26 with oV-line detection by AAS23,24,26 or total reflec- The purpose of this paper was to investigate the speciation tion XRF.25 Speciation of trace metals and metalloids in plants of metals in fresh fruits and vegetables by HPLC with both was recently reviewed.27 refractometric and ICP-MS detection. A sample preparation The key challenges to elemental speciation in foodstuVs procedure based on an enzymatic digestion/liquefaction was include the identification of the eluted compounds of which used to gain an insight into metal speciation in the waterinsoluble sample fraction.only the metal component is actually detected by ICP-MS, J. Anal. At. Spectrom., 1999, 14, 639–644 639molecular mass standards (P-5, Mw=5800; P-10, Mw=12 200; Experimental P-20, Mw=23 700; P-50, Mw=48 000, Showa Denko, Tokyo, Instrumentation Japan) using refractometric detection.33,34 P-50 elutes at 8.55 ml, close to the exclusion volume of the column.The Chromatographic separations and flow-injection experiments glucose monomer was injected to obtain the total column were performed using a Series 410 BIO HPLC pump (Perkinvolume of 19.32 ml. The elution volumes for the standards in Elmer, Palo Alto, CA, USA) as the sample delivery system. the fractionation range of the column were 10.05, 11.74 and Injections were performed using a Model 7725 injection valve 14.04 ml for P-20, P-10 and P-5, respectively.It should be with a 50 ml injection loop (Rheodyne, Cotati, CA, USA). noted that pullulans are linear polysaccharides and the use of All the connections were made of PEEK tubing (0.17 mm id). this calibration may lead to some underestimation of the Analyte species were separated on a 13 mm SuperdexTM-75 HR molecular masses of highly ramified polysaccharides. 10/30 SEC column (300×10 mm id) (Pharmacia Biotech, Uppsala, Sweden) with an exclusion limit of 100 kDa and an Speciation analysis of the water-soluble fraction.Apples and eVective separation range between 0.5 and 50 kDa (pollulans). carrots were washed, peeled and sliced. Samples (160 g of apple The presence of polysaccharides was monitored using a and 152 g of carrots) were crushed in aWaring blender for 1 min Model ERC-7512 refractometer (Erma, Tokyo, Japan). after addition of 160 ml (apples) or 152 ml (carrots) of a solution Element-specific detection was realized with an Elan 6000 ICP containing 6 mM of ascorbic acid (to prevent oxidation) and mass spectrometer (Perkin-Elmer SCIEX, Concord, Canada). 0.04% m/v of sodium azide (antibacterial agent).The homogen- The sample introduction system used included a RytonTM ates were centrifuged at 13 000 rpm for 10 min to produce a spray chamber fitted with a cross-flow nebulizer. For total supernatant (180 ml for apples and 117 ml for carrots) and a analyses the samples were fed by means of a Minipuls 3 solid residue (120 g for apples and 187 g for carrots).Aliquots peristaltic pump (Gilson, Villiers-le-Bel, France) that also of the supernatant were analysed by ICP-MS for total metals served for draining the spray chamber. Chromatographic data and by HPLC-ICP-MS for metal species. were processed using the Turbochrom4TM software (Perkin- Elmer). All signal quantifications were performed in the peak Speciation analysis of thr-insoluble fraction. The solid area mode. residues after centrifugation were blended with a solution containing 6 mM of ascorbic acid and 0.04% m/v of sodium Reagents, standards and samples azide (180 and 117 ml, for apple and carrot samples, respect- Analytical-reagent grade reagents purchased from Sigma- ively).Aliquots of 320 and 300 ml, respectively, of the commer- Aldrich (St. Quentin Fallavier, France) were used throughout cial enzyme preparation were added. Samples were incubated unless specified otherwise. Milli-Q water (18 MV) (Millipore, at 30 °C for 24 h.The mixtures were centrifuged as described Bedford, MA, USA) was used throughout. The formate buVer above. Aliquots of the supernatant (301 ml for apples and was prepared by dissolving 30 mM of ammonium formate in 275 ml for carrots) were analysed by ICP-MS and by HPLCwater and adjusting the pH to 5.2 by the addition of formic ICP-MS. The residues (18 g for apples and 30 g for carrots, acid. mainly cellulose) were slurried in an ultrasonic bath and Rapidase Liq+TM (Gist Brocades, Seclin, France) analysed by ICP-MS.and Pectinex Ultra-SPLTM (Novo Nordisk, Copenhagen, Denmark) commercial enzymatic preparations containing pec- Results and discussion tinases, hemicellulases and cellulases were used. These prep- Chromatographic conditions (choice of the column, mobile arations were added to the homogenate samples to give a final phase composition and flow rate) optimized and found success- concentration of 0.1% m/v of each enzyme.ful elsewhere20 for the speciation of biomolecular Pb complexes Two rhamnogalacturonan-II preparations, viz., monomer in wine were adopted in this work. For all peaks shown in the (mRG-II) and dimer (dRG-II ), purified according to Pellerin chromatograms below, the isotopic pattern of the determined et al.,30 were used. The fraction reported earlier30 as RG-II2 element was checked in an independent experiment in order [predominantly (>95%) as monomer] was used as the mRG-II to exclude the possibility of artefacts due to isobaric inter- standard whereas the fraction reported as RG-II3 [predomiferences.Isotopes (one for each element) giving the most nantly (87%) as dimer] was used as the dRG-II standard. intense signals were chosen for the multi-element chromatogra- Apples (Malus domestica; Golden delicious) and carrots phy. The set of eight elements which were monitored included (Daucus carota) were purchased at a local supermarket.Pb as a common toxic element (the signal from Cd was too Procedures small to be monitored), Ba, Sr, and Ce as representative elements known to be complexed by carbohydrates,31 B as an ICP-MS conditions. ICP-MS measurement conditions (nebu- essential nutrient and a component of a diester ligand potenlizer gas flow, rf power and lens voltage) were optimized daily tially binding metals,16–18,20,21,31 and Cu, Zn and Mg (the using a standard built-in software procedure. The isotopes most common essential elements).The fruit and vegetable 138Ba, 139La, 140Ce, 141Pr, 88Sr, 63Cu, 64Zn, 208Pb and 11B were sample homogenates were fractionated by centrifugation into monitored. The dwell time for each isotope was 200 ms and a water-soluble fraction (supernatant) and a water-insoluble the number of replicates allowing for continuous scanning for (solid residue) fraction. The supernatant could be analysed by the duration of the chromatogram was applied. For total size-exclusion HPLC-ICP-MS directly.For the solid fraction, element ICP-MS analyses, a sample was slurried or diluted a procedure to release metal species needed to be developed with HNO3 to reach an acid concentration of 1% m/v in a to allow speciation analysis by HPLC-ICP-MS. 10 ml calibrated flask and placed in an ultrasonic bath for 5 min. Quantification was performed using an external cali- Speciation of metals in fruit and vegetables (water-soluble bration graph.fraction) Molar mass distribution profiles for the diVerent elements. Chromatographic conditions. For size-exclusion HPLC, aliquots of 20–100 ml were injected. The mobile phase was 30 mM Fig. 1 shows a multi-element size-exclusion chromatogram with ICP-MS detection of the water-soluble fraction for apple ammonium formate buVer, pH 5.2, at a flow rate of 0.6 ml min-1. The eluate from the column was fed directly [Fig. 1(a)] and carrot [Fig. 1(b)] samples. The chromatograms (one run for each sample) are each shown in two panels for into the ICP.The column was calibrated with narrow pullulan 640 J. Anal. At. Spectrom., 1999, 14, 639–644column). This fraction represents all the B present in the water-soluble fraction of carrots, but for the apple sample, a small percentage of B is present in the fraction excluded from the column (elution volume 8.0 ml ). The major diVerence between the samples studied is the presence of a strong signal containing Ba, Ce, Pb, Sr and B at an elution volume of 8.0 ml in the chromatogram of the apple water-soluble fraction.This signal corresponds to a compound(s) with an apparent molar mass equal to or above 50 kDa and is absent in the water-soluble fraction of the carrot sample. This compound(s) contains the totality of the Ba, Ce and Sr and the majority (>95%) of Pb present in the sample. For the carrot sample, some Pb co-elutes with the fraction of Zn whereas Ba and Sr elute as a single peak at an elution volume of 11.7 ml corresponding to a compound with a molecular mass of 10–15 kDa. The signals of Ba, Ce, Sr and B for the carrot sample are very low in comparison with those measured for the apple sample, despite similar concentrations of these elements in the initial homogenate.Molar mass distribution profiles for carbohydrates in the water-soluble fraction. Refractometric detection was employed to detect carbohydrates present in the size-exclusion chromatographic eluates of the water-soluble fractions of the apple and carrot samples.The chromatograms are shown in Fig. 2. The presence of polysaccharides excluded from the column (Mw >50 kDa) is detected; the exact elution volume matches, for the apple sample, the elution volume of Ce, Sr, Ba and Pb. No other refractive index signal, except for one from the salts present in the sample (not shown), was observed in any of the analysed samples. It should be noted that the peaks in the chromatograms acquired with refractometric detection are much broader than in those acquired with ICP-MS detection.This suggests that only particular carbohydrates from the polysaccharide fraction bind metals. Identification of metal species in the fruit and vegetable watersoluble fraction. Identification of the metal species detected is hampered by the lack of standards. Nevertheless, in SEC the identity of an element species can be hypothesized on the basis Fig. 1 Size-exclusion HPLC-ICP-MS traces for the supernatants on the centrifugation of (a) an apple sample and (b) a carrot sample. One run for each sample. Two panels are shown for each sample for better clarity of presentation. better clarity of presentation. ForMg, Cu and Zn the molecular size distribution patterns for apple and carrot samples are identical. Magnesium shows a single peak (>99%) at an elution volume exceeding the total volume of the column. The totality of Cu and Zn elute in a single (each element in a diVerent one) fraction.The average molar mass of the Zn-containing fraction is slightly higher than that of the Cu-containing fraction. Boron elutes principally as a low Fig. 2 Size-exclusion HPLC traces with refractometric detection for molar mass fraction interacting strongly with the stationary total soluble polysaccharides obtained for (a) an apple sample, (b) a carrot sample. phase (the elution volume exceeds the total volume of the J.Anal. At. Spectrom., 1999, 14, 639–644 641of its elution volume. Although the precision of the identification complex is released. This was attempted with enzyme preparations that contain high levels of pectinolytic activity. on the basis of the apparent molar mass is relatively poor, it serves well for a preliminary approach. The peak containing the Homogalacturanans are extensively degraded by pectinases whereas dRG-II-B is resistant to such enzymes. majority of the dissolved B is likely to correspond to a mixture of the borate mono- and diesters as identified by Matsunaga and Nagata17 using 10B NMR.Cu and Zn elute in an area of Enzymatic degradation of macromolecular metal–polysaccharide complexes. Enzyme preparations with high levels of pec- molar masses of phytochelatin. The elution volume of the Mg signal suggests that this element may be present as the Mg2+ ion tinolytic activity (hydrolases, lyases, esterases) are widely used in the fruit processing industry to increase yields, and improve which can be further unspecifically retained by the cationexchange sites of the stationary phase.liquefaction, clarification and filtrability of juices.30 A chromatogram [Fig. 4(a)] with the refractometric detection of the super- The similar ionic radius of metals complexed by the excluded (50 kDa) fraction [Fig. 1(a)], which is a polysaccharide natant of an apple homogenate sample subjected to a 4 h treatment at 35 °C with a mixture of such enzymes shows the fraction (cf.Fig. 2), suggests that the species is a complex of these metals (Ba, Ce, Sr, Pb) with a dimer of RG-II as a absence of the high molar mass polysaccharide species seen in Fig. 2(a). Instead, two peakswith elution volumes corresponding ligand, similarly to Pb in wine.20,21 This hypothesis is further corroborated by the co-elution of B which is necessary to form to the RG-II monomer and dimer (cf. Fig. 3) can be seen.Note that only the dimer is able to complex metals.31 This observation a dimer of RG-II,30–32 and hence to form the complex with the metals. The ultimate verification of this hypothesis can be is confirmed by a size-exclusionHPLC-ICP-MS trace [Fig. 4(b)] obtained after enzymatic treatment of the water-soluble apple obtained by comparing the elution volume of this species with that of a dRG-II standard. homogenate fraction. The elution pattern of Cu, Zn and Mg is identical with that Fig. 3(a) shows the elution profiles of the dimer and monomer of RG-II obtained with refractometric detection. The shown in Fig. 1(a) but the excluded peak (50 kDa) containstandard of the dimer used contained an admixture of the monomer. The chromatogram with ICP-MS detection [Fig. 3(b)] shows that only the dimer contains B and Pb. No peak of any of the elements monitored by ICP-MS matches the elution volume of the mRG-II. The chromatographic conditions were the same as in Fig. 1 and 2. The elution volume of the dRG-II standard indicates that the molar mass of the metal species detected in the apple juice exceeds that of dRG-II by a factor of >5. For the carrot sample, however, some Ba and Sr [Fig. 1(b)] elutes apparently as a complex with dRG-II. Taking into account the co-elution of the metals with a particular ionic radius with B and the ubiquitous presence of dRG-II in fruit and vegetables,18,20,30–32 it was considered possible that the metals in the apple water-soluble fraction may indeed be complexed by dRG-II-B but that the complex may be part of an even more complex molecule (high molecular mass polysaccharide). Experiments were therefore designed to disrupt this structure in such a way that the intact dRG-II-B Fig. 4 Size-exclusion HPLC traces for the apple sample water-soluble Fig. 3 Size-exclusion HPLC traces for the RG-II standards. (a) fraction degraded by enzymatic hydrolysis. (a) Refractometric detection; (b) ICP-MS detection (data are shown in two panels for clarity Refractometric detection; (b) ICP-MS detection. 1, RG-II dimer standard; 2, RG-II monomer standard. of presentation). 642 J. Anal. At. Spectrom., 1999, 14, 639–644Fig. 6 Size-exclusion HPLC traces with refractometric detection for an enzymatic digest of the residue of an apple homogenate after centrifugation. methylmercury, alkyllead ) the stability of which could readily be assured during the hydrolysis of the proteinaceous matrix by tetramethylammonium hydroxide35 or proteolytic enzymes.36 For metal complexes, alkaline or acid hydrolysis cannot be accepted since it would aVect the complexation equilibria.Selective degradation of the matrix at the natural pH of the sample studied appeared therefore to be the only valid solution. Fig. 5 Size-exclusion HPLC-ICP-MS traces for an enzymatic digest The polysaccharide matrix of the water-insoluble residue of the residue of an apple homogenate after centrifugation.Data are after centrifugation of fruit and vegetable homogenates is shown in two panels for clarity of presentation. The peak at 12.2 kDa composed of pectic polysaccharides, hemicelluloses and cellu- corresponds to the metal complex with dRG-II. loses. Therefore, the use of enzymatic preparations containing pectinolytic, hemicellulolytic and cellulolytic activities is neces- Table 1 Quantitative distribution of elements in soluble and insoluble sary to solubilize the solid residues. RG-II was reported to be fractions of the fruit and vegetables studied (in mass percent of the initial homogenate) resistant to such treatments.34 This resistance was attributed to the high degree of ramification and diversity of glycosyl Insoluble fraction, linkages of RG-II.Water-soluble liberated by Residue after It was also expected that the complex of dRG-II-B with Element fraction enzymolysis enzymolysis metals, if present in the solid residue, would be resistant to pectinolysis and would be released into the aqueous phase Carrots (Daucus carota)— B 33.0 9.3 56.8 while other polysaccharides would be degraded. Indeed, the Ce 1.0 5.5 93.5 chromatograms in Fig. 5 show that the enzymatic hydrolysis Cu 65.0 31.4 3.6 releases the complex of dRG-II (bridged by B) with Ce, Ba, Pb 3.5 97.0 0.5 Sr and Pb. The enzyme also releases other metals from the Sr 12.2 82.0 5.8 solid fruit and vegetable fraction in the forms in which they Zn 48.6 49.6 1.8 were present in the water-soluble fraction (cf.Fig. 1). The Apples (Malus domestica)— B 19.8 33.0 47.1 chromatogram with refractometric detection (Fig. 6) confirms Ce 73.9 26.1 Below 0.5% that carbohydrates co-elute with the metals and B. Cu 76.6 19.3 4.1 Unlike the water-soluble fraction, the chromatogram of the Pb 42.5 47.5 10.0 enzymatic digests of the carrot sample solid residue is very Sr 28.6 47.4 14.0 similar to that of the apple one. This suggests that the metals Zn 57.2 30.0 12.8 in the solid fraction are bound to a pectic matrix, the metal complexing component of which is dRG-II-B. ing Ba, Ce, B, Pb and Sr has disappeared.Instead, an intense peak can be seen for these elements at an elution volume of Distribution of metal species between the water-soluble and water-insoluble fractions 11.7 ml which matches exactly that of dRG-II-B (cf. Fig. 3). The data lead to the conclusion that Pb, Sr, Ba and Ce in the The quantitative distribution of metals between the water- apple water-soluble fraction are present in the form of a soluble and water-insoluble fractions of the fruit and vegetable complex with dRG-II but this species is, unlike in wine,20–22 samples was investigated.For the water-insoluble fraction, part of high molecular mass polysaccharides, probably due to further distinction was made between the metal fraction that the covalent linkage with homogalacturonans. can be solubilized by enzymatic hydrolysis and the fraction that is retained in the solid residue after pectinolysis.Results, Speciation of metals in fruit and vegetables (water-insoluble expressed as a percentage of the total element concentration fraction) present in the initial sample, are given in Table 1. Because of the slurry sampling technique employed for the analysis of the Approaches to speciation analyses of water-insoluble samples have been scarce because of the diYculties with selective solid residue the error of this analysis can reach 10–15%.Elements in the supernatants could be determined with a destruction of the solid matrix in such a way that the metal complex of interest is preserved intact. To date, the studies precision of 1–3%. The results show that most of the metals are either already have been limited to some organometallic species (organotin, J. Anal. At. Spectrom., 1999, 14, 639–644 6437 S. A. Pergantis, E. M. Heithmar and T. A. Hinners, Analyst, 1997, present in water-soluble forms or can be released as water- 122, 1063. soluble species by enzymatic leaching with pectinolytic 8 L.M. W. Owen, H. M. Crews and R. C. Massey, Chem. Speciat. enzymes. The highest elemental concentrations in the water- Bioavail., 1992, 4, 89. soluble fraction (49–75%) are obtained for Zn and Cu which 9 H. M. Crews, J. R. Dean, L. Ebdon and R. C. Massey, Analyst, are known to be present as complexes with polypeptide ligands 1989, 114, 895. 10 J. R. Dean, S. Munro, L. Ebdon, H. M. Crews and R. C. Massey, (phytochelatins and metallothioneins). A strong signal from J. Anal. At. Spectrom., 1987, 2, 607. Mg in the water-soluble fraction is also observed. The only 11 L. M. Owen, H. M. Crews, R. C. Hutton and A. Walsh, Analyst, elements largely retained in the residue in a form that is 1992, 117, 649. resistant to pectinolytic enzymes are B and Ce. 12 S. M. Bird, H. Ge, P. C. Uden, J. F. Tyson, E. Block and E. The distribution of Pb, Sr, Ba and Ce that are complexed Denoyer, J.Chromatogr., 1997, 789, 349. by dRG-II-B and can be incorporated in larger polysaccharide 13 S. M. Bird, P. C. Uden, J. F. Tyson, E. Block and E. Denoyer, J. Anal. At. Spectrom., 1997, 12, 785. structures between the water-soluble and water-insoluble frac- 14 J. Zheng, W. Go� ssler and W. Kosmus, Trace Elem. Electrol., 1998, tions is diVerent in carrot and apple samples. Whereas in 15, 70. carrots little of these elements is present in the water-soluble 15 H.Ge, X. J. Cai, J. F. Tyson, P. C. Uden, E. R. Denoyer and E. fraction, the majority of Ce and significant proportions of Sr, Block, Anal. Commun., 1996, 33, 279. Pb and Ba in apples are present as water-soluble species. The 16 T. Matsunaga, T. Ishii and H.Watanabe, Anal. Sci., 1996, 12, 673. relative water-soluble fraction is the highest for Ce and the 17 T. Matsunaga and T. Nagata, Anal. Sci., 1995, 11, 889. 18 T. Ishii and T.Matsunaga, Carbohydr. Res., 1996, 284, 1. lowest for Sr. In the carrot samples the majority of Pb, Sr and 19 K. E. Oedegard and W. Lund, J. Anal. At. Spectrom., 1997, 12, Ba can be released (as water-soluble dRG-II complexes) by 403. the enzymatic hydrolysis procedure developed. This is not the 20 J. Szpunar, P. Pellerin, A. Makarov, T. Doco, P. Williams, B. case with Ce, however. The majority of this element in carrots Medina and R. £obin� ski, J. Anal. At. Spectrom., 1998, 13, 749.is retained in the final insoluble residue, either as a complex 21 P. Pellerin and M. A. O’Neill, Analusis, 1998, 26, M32. or more likely as hydrolysed oxide/hydroxide species. 22 P. Pellerin, M. A. O’Neill, C. Pierre, M. T. Cabanis, A. G. Darvill, P. Albersheim and M. Moutounet, J. Int. Sci. Vigne Vin, 1997, 31, 33. Conclusions 23 K. Lange-Hesse, L. Dunemann and G. Schwedt, Fresenius’ J. Anal. Chem., 1991, 339, 240. Size-exclusion HPLC-ICP-MS combined with sample prep- 24 K.Lange-Hesse, L. Dunemann and G. Schwedt, Fresenius’ aration procedures based on hydrolysis using glycosyl hydro- J. Anal. Chem., 1994, 349, 460. lases oVers an attractive tool to study the speciation of trace 25 K. Guenther and A. Von Bohlen, Spectrochim. Acta, Part B, 1991, metals in fruit and vegetables and can be used for monitoring 46, 1413. the fate of metal species during industrial food processes. 26 K. Gu�nther and H. Waldner, Anal. Chim. Acta, 1992, 259, 165. 27 J. Szpunar and R. Lobinski, in Heavy Metal Stress in Plants— Polysaccharides present in samples of plant origin are import- From Molecules to Ecosystem, ed. M. N. V. Prasad and ant ligands able to bind the majority of some metals in stable J. Hagemeyer, Springer, Heidelberg, 1999, ch. 16. water-soluble complexes. The basic ligand seems to be the 28 D. M. Whitfield, S. Stoijkovski and B. Sarkar, Coord. Chem. Rev., dimer of rhamnogalacturonan-II that may be part of larger 1993, 122, 171. polysaccharide structures. The method developed oVers an 29 J. Monreuil, Pure Appl. Chem., 1984, 56, 859. analytical tool to study the stability of the Pb–dRG-II complex 30 P. Pellerin, T. Doco, S. Vidal, P. Williams, J. M. Brillouet and M. A. O’Neill, Carbohydr. Res., 1996, 290, 183. in the gastrointestinal system which is required to evaluate the 31 M. A. O’Neill, D. Warrenfeltz, K. Kates, P. Pellerin, T. Doco, bioavailability of Pb taken in with foodstuVs. A. G. Darvill and P. Albersheim, J. Biol. Chem., 1996, 271, 22923. 32 M. Kobayashi, T. Matoh and J. L. Azuma, Plant Physiol., 1996, References 110, 1017. 33 T. Doco and J. M. Brillouet, Carbohydr. Res., 1993, 243, 333. 1 H. Crews, Anal. Eur., 1995, August, 28. 34 T. Doco, P. Williams, S. Vidal and P. Pellerin, Carbohydr. Res., 2 K. Sutton, R. M. C. Sutton and J. A. Caruso, J. Chromatogr., 1997, 297, 181. 1997, 789, 85. 35 M. Ceulemans, C. Witte, R. Lobinski F. C. Adams, Appl. 3 G. K. Zoorob, J. W. McKiernan and J. A. Caruso, Mikrochim. Organomet. Chem., 1994, 8, 451. Acta, 1998, 128, 145. 36 D. S. Forsyth and J. R. Iyengar, J. Assoc. OV. Anal. Chem., 1989, 4 A. Makarov and J. Szpunar, Analusis, 1998, 26, 44. 72, 997. 5 M.Morita and J. S. Edmonds, Pure Appl. Chem., 1992, 64, 575. 6 J. R. Dean, L. Ebdon, M. E. Foulkes, H. M. Crews and R. C. Massey, J. Anal. At. Spectrom., 1994, 9, 615. Paper 8/08231F 644 J. Anal. At. Spectrom., 1999, 14, 639&nda