Analysis of methylated and unmethylated polygalacturonic acid structure by polysaccharide analysis using carbohydrate gel electrophoresis

ANALYTICAL BIOCHEMISTRY Analytical Biochemistry 321 (2003) 174–182 www.elsevier.com/locate/yabio

Analysis of methylated and unmethylated polygalacturonic acid structure by polysaccharide analysis using carbohydrate gel electrophoresis Florence Goubet, Brooke Morriswood,1 and Paul Dupree* Department of Biochemistry, University of Cambridge, Building O, Downing Site, Cambridge CB2 1QW, UK Received 14 March 2003

Abstract The electrophoretic migration in polyacrylamide gels of oligogalacturonic acids (OGAs) derivatized by a fluorophore (2-aminoacridone) was studied. We found conditions such that OGAs can be separated up to a degree of polymerization (DP) of 40. The migration was dependent on degree of methylation and DP, because the OGA mobility relies on the charge of the galacturonic acid residues. Since both methylated and unmethylated oligosaccharides can be resolved, polysaccharide analysis using carbohydrate gel electrophoresis (PACE) is a powerful method for studying the fingerprint of pectin hydrolysis. It can be used to characterize endopolygalacturonase (Endo-PG) tolerance of methylation. Furthermore, using an Endo-PG that can distinguish low and highly methylated pectin, PACE can be used to investigate the blockwise or nonblockwise distribution of methylation of polygalacturonic acid. We show that the method can be applied to crude cell wall preparations of Arabidopsis inflorescence stems. Using chemical deesterification before or after Endo-PG digestion, we show that in the Arabidopsis cell wall, the pectins have both nonesterified and highly esterified regions. Ó 2003 Elsevier Inc. All rights reserved. Keywords: Arabidopsis; Hydrolase; Methylation; Oligosaccharide; Pectin; Polysaccharide

Plant cell walls are very complex with regard to the structures and interactions of the different polymers. The cell wall is composed of a microfibrillar cellulose phase and a matrix phase consisting of pectins, hemicelluloses, proteins, and phenolic compounds. The composition varies in different parts of the walls, in different types of cell, in different species, and probably at different stages of the cell cycle. In many cases, polysaccharides are the predominant component. The primary wall is composed of around 90% carbohydrate, and of this around 40% is pectin for all Dicotyledonae and Monocotyledonae with type I primary walls. This percentage is lower for the grass family Monocotyledonae [1].

* Corresponding author. Fax : +44-1223-333345. E-mail address: [email protected] (P. Dupree). 1 Present address: MRC Laboratory of Molecular Biology, Hills Road, Cambridge CB2 2QH, UK.

0003-2697/$ - see front matter Ó 2003 Elsevier Inc. All rights reserved. doi:10.1016/S0003-2697(03)00438-X

The pectins are a group of polysaccharides that are generally charged and rich in galacturonic acid, rhamnose, arabinose, and galactose [2]. The class consists of polygalacturonic acid (PGA),2 rhamnogalacturonan I (RG-I), rhamnogalacturonan II, and apiogalacturonan which has been found so far only in some water plants such as Lemna. PGA is a linear homopolymer of 1,4linked a-D -galacturonic acid which can be methyl esterified at the C-6 carboxyl group [1]. PGA from some plants may also be O-acetylated or substituted by xylosyl residues at C-3. Methylation changes the properties of the pectin since the charged galacturonic acid group is 2 Abbreviations used: AMAC, 2-aminoacridone; ANTS, 8-aminonaphthalene-1,3,6-trisulfonic acid; DM, degree of methylation; DP, degree of polymerization; Endo-PG, endopolygalacturonase; HPLC, high-performance liquid chromatography; NMR, nuclear magnetic resonance; OGA(s), oligogalacturonic acid(s); PACE, polysaccharide analysis using carbohydrate gel electrophoresis; PGA, polygalacturonic acid; PAGEFS, polyacrylamide gel electrophoresis of fluorophore-labeled saccharides; RG-I, rhamnogalacturonan I.

F. Goubet et al. / Analytical Biochemistry 321 (2003) 174–182

esterified. This modification is therefore important for plant growth and development [3] and for industrial properties of pectin [4]. RG-I is a branched pectic polysaccharide and contains a backbone of a-1,4-linked galacturonic acid and a-1,2-linked rhamnose [1,2]. They may be O-acetylated at C-3 or C-2. Many rhamnosyl residues in RG-I are substituted at C-4 and occasionally at C-3, with an oligosaccharide side chain of galactan, arabinan, or arabinogalactan I. Rhamnogalacturonan II has a complex and conserved structure present in all plant cell walls. Various techniques are used to study pectin structure and the enzymes that modify this polysaccharide. Viscosimetry and colorimetry assays are generally used to quantify endopolygalacturonase (Endo-PG) activity, e.g., Gadre et al. [5]. Hydrolases cut polysaccharides at specific sites, and therefore by using them both the structure of polysaccharides and the specificity of the hydrolases can be analyzed. After hydrolysis, the samples can be derivatized and separated by HPLC [6] or capillary electrophoresis [7]. Using this principle, Ishii et al. [8] described RG-I structure, whereas Fu et al. [7] described the specificity of a rhamnogalacturonase. Gas chromatography gives sugar and linkage composition of pectic samples, but studying the degree of methylation (DM) by this technique, although possible, is seldom used [9]. To determine DM, saponification of pectin and methanol quantification is used [9], and this principle is also used to study pectin methyl esterase [10]. Analysis of methylation quantity and distribution in pectin combines enzymatic degradation using Endo-PG with HPLC [10–12] or capillary electrophoresis [13] to separate and quantify the oligogalacturonic acids (OGAs) produced by the enzyme. However, Mort et al. [14] described some difficulties such as aggregation in studying OGAs by HPLC (ionexchange or size-exclusion chromatography). We recently described a new methodology to study plant polysaccharides called polysaccharide analysis using carbohydrate gel electrophoresis (PACE; [15]). It relies on derivatization of reducing ends of sugars with a fluorophore, followed by electrophoresis under optimized conditions in polyacrylamide gels. This method is a variant of the polyacrylamide gel electrophoresis of fluorophore-labelled saccharides (PAGEFS) technique described by Jackson [16]. Charged monosaccharides or oligosaccharides can be studied after derivatization with uncharged fluorophores such as 2-aminoacridone (AMAC; [15,17]) or charged fluorophores such as 8aminonaphthalene-1,3,6-trisulfonic acid (ANTS, [7,18]). The method was optimized and commercialized as the FACE kit to study glycans of glycoproteins [16,19]. This kit has also been used to investigate the structure of hyaluronan and chondroitin using digestion with specific enzymes followed by AMAC derivatization [20]. PAGEFS has now also been used to study glycans of bacteria and algae [21,22]. We optimized PAGEFS to

175

separate the different monosaccharides found in plants, to provide an effective tool for structural analysis and quantitation of charged and uncharged plant polysaccharides, and for studies of plant polysaccharide hydrolase activities [15]. The advantages of PACE over alternatives include the sensitivity and simplicity of the method and the ability to compare many samples in a single gel. Quantification using gel analysis software packages [15] is easier than quantifying the peak areas from HPLC. In our previous work, we showed that charged and neutral monosaccharides and oligosaccharides can be analyzed by PACE [15]. This analysis was optimal for uncharged sugars using ANTS derivatization, and we showed that it was possible to study OGA from DP 1 to DP 3 derivatized with AMAC. In this study, using further standard oligosaccharides, we investigated and optimized the migration of larger-DP OGAs and methylated OGAs, and we optimized PACE to study the structure of methylesterified and unmethylesterified polygalacturonan. We show that it is possible to study the fine structure of pectins by PACE, and we applied the method to define a fingerprint for pectin of the Arabidopsis cell wall.

Materials and methods Materials Citrus PGA, GalU, Endo-PG (EC 3.2.1.15; from Aspergillus japonicus), pectin with 26 and 68% DM, (GalU)3 , and urea were purchased from Sigma (Poole, Dorset, UK). AMAC was purchased from Molecular Probes (Leiden, The Netherlands). Polyacrylamide containing a ratio of acrylamide/N,N 0 -methylenebisacrylamide (29:1) was obtained from Bio-Rad (Hertfordshire, UK). Unmethylated OGAs of DP 5 ((GalU)5 ), 7 ((GalU)7 ), 10 ((GalU)10 ), 11 ((GalU)11 ), and 13 ((GalU)13 ) were given by Paul Knox (University of Leeds, UK). Methylesterified OGAs of DP 4–DM 1 ((GalU)4 Me1 ), DP 5–DM 1 ((GalU)5 Me1 ), DP 5–DM 2 ((GalU)5 Me2 ), and DP 6– DM 2 ((GalU)6 Me2 ) and pectin with 55% of DM were given by Martin Williams (Unilever, Bedford, UK). Plant culture Wild-type Arabidopsis thaliana plants (c.v. Columbia) were sown in soil/sand (3:2, v/v) and grown at 20 °C under fluorescent white light in 16-h-light/8-h-dark cycles. Stems were collected after 7 weeks. Preparation of cell wall materials Plant materials were incubated for 30 min in ethanol/ water (9:1 (v/v)) at 65 °C to inactivate enzymes. They were ground in a Mixer Mill MM200 (Glen Creston,

176

F. Goubet et al. / Analytical Biochemistry 321 (2003) 174–182

Middlesex, UK). The homogenate was centrifuged at 5000 rpm for 15 min. The pellet was briefly washed with ethanol/water (6:4 (v/v); two times), methanol/chloroform (2:3 (v/v); overnight), methanol/chloroform (2:3 (v/ v)), ethanol/water (6:4 (v/v); four times), and ethanol/ water (9:1 (v/v)). The remaining pellet, containing the cell wall, was dried overnight at 80 °C [15]. Enzymatic hydrolysis of cell walls, polysaccharides, or oligosaccharides Cell walls (0.5 mg ml
1 , 200 ll) or polysaccharides (0.5 mg ml
1 ; 100 ll) were treated in suspension with Endo-PG (1 ll; 0.3 U) in buffer in a total volume of 250 ll at room temperature for 1 min to 4 h. For pH 5, the buffer used was 0.1 M ammonium acetate adjusted with glacial acetic acid; for pH 7, the buffer was 10 mM Tris–HCl. Controls without substrates or enzymes were performed under the same conditions to identify any unspecific compounds in the enzymes, polysaccharides/ cell wall materials, or labeling reagents. Boiling for 10 min stopped the reactions. The samples were dried using a centrifugal vacuum evaporator. Any contaminant polysaccharide hydrolase activities were ruled out using PACE with a panel of polysaccharides according to Goubet et al. [15], but pectate or pectin lyase could be present with low activities. Demethylation of oligosaccharides/polysaccharides Polysaccharides and oligosaccharides produced from pectin hydrolysis by Endo-PG were treated by 1 M NaOH (200 mM final concentration). We used 20 or 50 ll of NaOH solution before or after Endo-PG hydrolysis, respectively. The treatment was performed at room temperature for 2 h before a neutralization with HCl (0.5 M) confirmed with pH paper. Then the solutions were dried before either the Endo-PG treatment or the derivatization of samples. Autoclaving of polygalacturonic acid solution PGA (0.5 mg ml
1 , 5–10 ml) in water was autoclaved between one and six times to optimize hydrolysis and produce a large-scale ladder to characterize the resolution limit of OGAs using our method. The autoclave conditions were with a free steam time of 15 min and sterilization of 20 min at 121 °C.

acetic acid/dimethyl sulfoxide (1.5/18.5, v/v) at 50 mM final concentration (made freshly). NaCNBH3 (0.5 M; made freshly and used immediately; toxic) was solubilized in water. To each dry sample, 5 ll of AMAC solution and 5 ll of the appropriate NaCNBH3 solution were added. The reagents were mixed, centrifuged, and incubated at 37 °C overnight. The solution was lyophilized in a centrifugal vacuum evaporator for 2 h at 40 °C. The derivatized sugars were resuspended in 100 ll of 6 M urea, were stored before use at )20 °C, and were stable for around 2 months. Electrophoresis Samples (0.5–4 ll) were separated using a Hoefer SE 660 vertical slab gel electrophoresis apparatus (Amersham, Buckinghamshire, UK) with 24-cm plates, 0.75mm spacer, and well of width 0.25 cm. Standard glass plates were used. Electrophoresis was performed at 10 °C in all cases. Electrophoresis conditions were 14.5, 19.3, 24.2, or 30.9% (w/v) polyacrylamide in the resolving gel containing 0.5, 0.7, 0.8, or 1.1% (w/v) N,N 0 methylenebisacrylamide, respectively. The stacking gel of 10% (w/v) polyacrylamide and 0.4% (w/v) N,N 0 methylenebisacrylamide was the same for all resolving conditions used. Both resolving and stacking gels were prepared in 0.1 M Tris–Cl, pH 8.2. A discontinuous electrophoresis buffer system was used: 0.15 M Tris– glycine, pH 8.5, as the cathode reservoir buffer and 0.1 M Tris–Cl, pH 8.2, as the anode reservoir buffer. The samples were electrophoresed initially at 200 V for 20 min and then at 1000 V for 1.5, 2, and 4 h and at 800 V for 16 h for 15, 20, 25, and 32% polyacrylamide gel, respectively. All gels were performed at least two times with samples made at least two times. The figures in this publication are representative examples of gels obtained. Gel imaging Gels were scanned using a MasterImager CCD (charge-coupled device) camera system (Amersham) with an excitation filter at 400 nm and a detection filter at 530 nm. A 16-bit image of the gel (resolution 100 lm) was obtained and exported in a 8-bit file to PowerPoint.

Results

Derivatization with 2-aminoacridone

Optimization of AMAC derivatization

Derivatization was carried out in the tubes containing dried polysaccharides, oligosaccharides, or monosaccharides [15]. For monosaccharide or oligosaccharide standards, 5 ll of 1 mM sugars were added to a tube and dried before derivatization. AMAC was prepared in

During derivatization of sugars, side reactions of AMAC and NaCNBH3 can occur, causing background signals in the PACE gels [15]. This background varies with commercial supplier and lot of the AMAC. Background in other PAGEFS applications has also been

F. Goubet et al. / Analytical Biochemistry 321 (2003) 174–182

observed by Drummond et al. [23] using a different fluorophore. Therefore, we optimized the derivatization of OGAs with AMAC to improve the clarity of the results. Compared to our previous protocol [15], we decreased acetic acid and NaCNBH3 concentrations by half. Furthermore, less AMAC (50 mM, made freshly, but still in excess of the sugars at 1–5 mM) was used. Use of this protocol significantly increased the signal/background ratio (data not shown). However, when the quantity of reducing sugars was detectable but below the quantification limit of about 5 pmol, the background was still relatively high. We found that in such cases, the background could be reduced by lowering even further the concentrations of AMAC and NaCNBH3 (data not shown). Migration of unmethylated oligogalacturonic acids We previously showed that PACE can be used to separate OGAs of DP 1 to 3 [15]. Analysis of pectin structure requires a range of sizes of OGAs to be analyzable. We studied the mobility of OGAs in variouspercentage polyacrylamide gels, using standards of DP 5 to 13 that had been purified by HPLC from a PGA digest (Fig. 1). The OGA standards contained some contaminant oligosaccharides that were also visible in the PACE gels. Using 15% polyacrylamide in the resolving gel, only GalU was visible, and all other sugars migrated at the front of the gel (data not shown). The protocol previously used was 20% polyacrylamide [15].

177

In these gels, a good separation of the small OGAs (DP 1 to 3) was observed [15], but larger OGAs migrated together at the buffer front (data not shown). In 25% polyacrylamide gels, most oligosaccharides were separated but OGAs of DP 4 and 5 had similar migration (Fig. 1A). Interestingly, between DP 1 and 5, the larger OGAs migrated farther than the smaller ones, but above DP 5 the migration of OGAs reduced as they became larger. Therefore, there was a size-dependent ‘‘turning point’’ in the migration corresponding to DP 5. As a consequence of this turning point, OGAs of DP 7 and 11 migrated similarly to OGAs of DP 3 and 2, respectively. Thus, there is an interesting electrophoresis phenomenon, where a larger DP and thus increased charge of the OGA species results in higher mobility. However, above a certain size, hindrance by the polyacrylamide outweighs the increase in charge, and mobility is decreased. This effect was not due to oligosaccharide dimerization or chelation with divalent cations because the migration was not altered in the presence of 25 mM EDTA in the loading buffer (data not shown). When the polyacrylamide concentration was increased further to 32%, the turning point was decreased to DP 4 (Fig. 1B). The comigrations were also changed; OGAs of DP 2 and 8 now migrated similarly. In addition, all the OGAs moved much more slowly than in 25% polyacrylamide, and consequently the total time required for electrophoresis was greatly increased. Since there were few advantages of the higher-percentage acrylamide condition, we selected 25% to study OGAs. All the OGAs

Fig. 1. Migration of OGAs derivatized with AMAC in polyacrylamide gels. Samples were loaded at the top of the gel and migrated toward the bottom. The acrylamide percentage of the resolving gels was 25% (A) or 32% (B). Lane 1, PGA hydrolyzed by Endo-PG at pH 5.0; lane 2, (GalU)5 ; lane 3, (GalU)7 ; lane 4, (GalU)8 ; lane 5, (GalU)10 ; lane 6, (GalU)13 ; * unspecific band.

178

F. Goubet et al. / Analytical Biochemistry 321 (2003) 174–182

migrated within the gel under this condition. However, this study showed that if it is necessary to separate comigrating OGAs, higher-percentage acrylamide gels can be used. Migration of oligosaccharide ladders derived from EndoPG or autoclave treatment of polygalacturonic acid To determine a size limit of OGA resolution using PACE, mixtures of small and large OGAs were prepared. First, PGA was partially enzymatically hydrolyzed by Endo-PG. Second, PGA was chemically hydrolyzed to varying extents by repeated autoclaving. The OGAs obtained were derivatized by AMAC and analyzed using the optimal 25% acrylamide gel conditions. After partial PGA hydrolysis with Endo-PG at pH 7, a ladder of OGAs was detected (Fig. 2A, lane 2). Comparison with the standards revealed that, above DP 6, each additional GalU residue decreased the mobility of the OGA in a regular fashion. Thus, it is possible to extrapolate the interpretation of OGA mobilities to sizes larger than the largest standard of DP 13. When PGA was incubated longer with Endo-PG, an increase in small OGAs and a corresponding decrease in the large OGAs was observed. However, a DP higher than 20 was

not detected even at short digestion times (data not shown). In contrast, PACE revealed that the OGAs obtained from hydrolysis during autoclaving were of a larger size range (Fig. 2A, lanes 4–6). The proportion of small OGAs increased and the proportion of large OGAs decreased with multiple treatments. OGAs up to DP 40 were separated by PACE and could be observed as specific bands (Fig. 2B). Larger OGAs were also detected but were not always clearly separated and migrated as a smear. Migration of methylated oligosaccharides PGA in plant walls may contain a varying amount of methylated GalU residues, and this modifies its biological and rheological properties. It is therefore of interest to be able to study methylated OGA oligosaccharides. Using the PACE conditions optimized for unmethylated OGA described above, the migration of various methylated OGA standards was studied (Fig. 3). As seen for unmethylated OGA standards, the main band corresponds to the defined OGA, whereas the minor bands are contaminants from the HPLC purification. In general, methylation reduced the migration with respect to the corresponding unmethylated OGA. Methylation

Fig. 2. Migration of OGA ladders produced by partial hydrolysis of PGA either by Endo-PG or by autoclaving. PACE derivatization and gel conditions were as in Fig. 1A. (A) Lane 1, essentially complete hydrolysis of PGA by Endo-PG at pH 5.0; lane 2 partial hydrolysis of PGA by EndoPG at pH 7.0; lane 3, (GalU)13 ; lanes 4, 5, and 6, PGA autoclaved 1, 3, and 6 times, respectively. * unspecific band. (B) Higher magnification of lanes 3, 4, and 5 in A to reveal separation of OGA up to DP 40.

F. Goubet et al. / Analytical Biochemistry 321 (2003) 174–182

179

of OGA modifies its size–charge ratio, since methylation of each GalU decreases the charge by one. For example, (GalU)4 Me migrated less than (GalU)4 and (GalU)5 Me2 migrated less than (GalU)5 Me1 (Fig. 3). However, at the turning point of DP5, (GalU)5 Me1 migrated faster than the unmethylated OGA DP5, suggesting that a subtle conformation effect may be important. The identification of (GalU)6 Me2 was more complex, because two main bands were detectable. These could be contamination by another OGA that coelutes by HPLC. In the future, we aim to investigate this in more detail. Study of Endo-PG specificities using pectins of different DMs

Fig. 3. Migration of methylesterified OGAs is distinct from that of unmethylated OGAs. PACE conditions were as in Fig. 1A. Lane 1, almost complete hydrolysis of PGA by Endo-PG at pH 5.0; lane 2, (GalU)4 Me1 ; lane 3, (GalU)5 ; lane 4, (GalU)5 Me1 ; lane 5, (GalU)5 Me2 ; lane 6, (GalU)6 Me2 .

Since PACE can be used to separate methylated and unmethylated oligosaccharides, we investigated whether it could be used to study the fine structure and enzymatic degradation of methylated pectin. Four pectins of different DMs (0, 26, 55, and 68%) with a random distribution of methylation were treated by Endo-PG, and the fingerprints of the OGAs produced were studied by PACE. Unmethylated PGA hydrolysis produced a simple fingerprint of OGAs DP 1–3 (Fig. 4). In some experiments, a low concentration of DP 4 was detected, suggesting that this is slowly hydrolyzed by the enzyme.

Fig. 4. Differential hydrolysis of four pectins of different DM by Endo-PG. Pectins with DM 0, 26, 55, or 68% were hydrolyzed to completion by Endo-PG at pH 5.0. To demethylate, pectin or oligosaccharides were treated with NAOH before or after digestion with Endo-PG. The released oligosaccharides were studied by PACE as in Fig. 1A. PG, hydrolysis of pectin by Endo-PG; NaOH ! PG, treatment of pectin by NaOH, neutralization and then hydrolysis by Endo-PG; PG ! NaOH, hydrolysis of pectin by Endo-PG and then treatment by NaOH and neutralization. 1, (GalU)5 Me2 ; 2, (GalU)6 Me2 ; 3, (GalU)4 Me1 .

180

F. Goubet et al. / Analytical Biochemistry 321 (2003) 174–182

When methylesterified pectin was hydrolyzed by EndoPG, more bands were detected, some of which corresponded to the standard methylesterified OGAs such as (GalU)4 Me1 , (GalU)5 Me2 , and (GalU)6 Me2 (Fig. 4). There were additional bands that are unidentified methylated oligosaccharides. There was also a smear in the fingerprint of the pectin digests, suggesting the presence of dispersed methylesterified OGAs (i.e., a large number of oligosaccharides with different distributions of methylation and different mobilities). The presence of these in the complete digest indicates that they are resistant to further digestion by the enzyme. The fingerprint of 55% DM pectin was similar to that of 26% DM, but there were a more complex pattern of bands and fewer unmethylesterified OGAs DP 1–3. It is interesting that the highly (68%) methylesterified pectin yielded a lower quantity of both unmethylesterified and methylesterified OGAs. This indicates that this highly methylesterified PGA was not efficiently hydrolyzed by the Endo-PG. To show that these differences were due to esterification, the pectin samples were demethylated by NaOH treatment before Endo-PG digestion. In this case, all the pectin samples yielded an identical pattern and quantity of OGA DP 1–3, with no bands corresponding to the methylated OGA standards. Thus, the differences in digestion were entirely due to the alkalilabile modification. This also indicated that the demethylation was complete with the conditions used. To investigate further whether the smear produced during pectin digestion corresponded to dispersed methylated OGA, the digest was demethylated by NaOH treatment and the demethylated OGAs were studied by PACE (Fig. 4). The demethylation yielded discrete bands that corresponded to an OGA ladder from DP 4 to over 20. Thus, dispersed larger methylated OGAs were present in all the digests of methylesterified pectin samples. The higher-DM pectins yielded more of the larger OGAs than OGAs DP 1 to 3, whereas the reverse was true for the low-DM pectin. This indicates that the digestion of higher-DM pectin produced a higher proportion of methylated oligosaccharides. Consistent with the band intensities prior to demethylation, the quantity of OGAs produced by digestion and then demethylation of the 68% DM pectin was less than those of the other pectins. In further support of the interpretation of the fingerprint of the pectin digests, the quantities of methylesterified oligosaccharides of DP 4– 6 were similar to the quantities of unmethylated OGAs of DP 4–6 after demethylation of the sample (Fig. 4). Study of methylated polygalacturonic acid in Arabidopsis cell walls Since methylated pectin could be distinguished from unmethylated pectin using PACE with Endo-PG, we investigated whether the fine structure of pectin in crude cell

wall samples could be studied using the technique. Cell wall samples from Arabidopsis stems were digested by Endo-PG, and the fingerprint of released were oligosaccharides studied by PACE. The pectin in the cell wall showed a fingerprint similar to that of PGA, yielding OGAs of DP 1–3 (Fig. 5). A very low quantity of OGA DP 4 was also detected. No bands corresponding to methylesterified OGA were detected, but a faint background smear was evident. Complete deesterification by NaOH treatment after Endo-PG digestion barely changed the quantity of DP 1–3 OGAs. However, a few larger OGAs were detected at a low level, suggesting that there was a low quantity of dispersed and highly esterified OGAs released by Endo-PG digest of the Arabidopsis walls. When the deesterification was performed before the Endo-PG hydrolysis, the fingerprint characteristic of PGA digestion was obtained. However, a much higher quantity of OGAs DP 1–3 was produced. This indicates that a major fraction of the cell wall pectin was esterified and thus insensitive to digestion. Since we know that the Endo-PG can cut methylesterified pectin when methylation is randomly distributed (Fig. 4), the resistance to digestion suggests that the enzyme-resistant regions were highly methylated in a blockwise fashion. The low quantity of

Fig. 5. PACE fingerprint of Arabidopsis cell wall treated with Endo-PG at pH 5.0 reveals methylated and unmethylated regions of pectin. Lane 1, hydrolysis of PGA by Endo-PG; lane 2, cell wall from Arabidopsis stems hydrolyzed by Endo-PG; lane 3, as lane 2 but demethylation oligosaccharides was performed with NaOH after Endo-PG hydrolysis; lane 4, demethylation of cell wall with NaOH before Endo-PG hydrolysis. After each NaOH treatment, solutions were neutralized before derivatization.

F. Goubet et al. / Analytical Biochemistry 321 (2003) 174–182

esterified oligosaccharides released by Endo-PG, which became visible after deesterification of the digest, suggests that there is only a low quantity of nonblockwise-esterified pectin in Arabidopsis stems.

Discussion Previously, we developed a new methodology to study plant cell wall polysaccharides [15]. This methodology was optimized for uncharged oligosaccharides, but we were able to investigate mobility of charged oligosaccharides between DP 1 and 3 only. Here, we show that much-larger-DP OGAs and methylated OGAs can be studied using PACE, and we improved the labeling and electrophoresis conditions. Since specific enzymes can be used to release oligosaccharides from methylated pectin, this methodology provides a tool for pectin fine structure analysis and quantitation. We decreased problem background bands from levels previously observed [15] by reducing the concentration of acetic acid, AMAC, and NaCNBH3 in the derivatization buffer. This simplified the interpretation of the digestion fingerprint and improved the quantification of band intensities. This background had been seen particularly when the quantity of sugars in the samples was low. Others have seen similar problems in PAGEFS applications and, to reduce it, have used an alternative reducing agent, triacetoxyborohydride, with an alternative fluorophore, 7-aminonaphtalene-1,3-disulfonic acid [23]. Our conditions allow us to use the uncharged fluorophore AMAC, which is important for separation of oligosaccharides based on native charge. By altering the gel polyacrylamide percentage, we were able to separate OGAs from DP 1 to above 40. Moreover, the methylated OGAs had mobilities differing from those of their unmethylated counterparts. The mobility of the oligosaccharides depends entirely on the native charge because they are derivatized with an uncharged fluorophore. However, the relationship between migration and OGA DP or methylation is complex because both size and charge strongly affect mobility. In 25% polyacrylamide gels, OGAs of increasing DP up to a DP of 5 migrate farther in the gel because their charge increases with DP. Above this turning point, the mobility decreases despite the increase in charge, because the polyacrylamide becomes an increasing obstacle to larger molecules. A related effect has been reported during PAGEFS of hyaluronan oligosaccharides, where the turning point was around DP 6–8 [17]. In contrast, a PAGEFS study of alginate-derivatized oligouronic acids found no similar turning point [22]. However, a charged fluorophore was used, and thus smaller-DP oligosaccharides comigrated near the free fluorophore. A consequence of the turning point is that some OGAs might not be well resolved. However, we showed here that the

181

turning point of OGAs varies with polyacrylamide percentage, and thus different acrylamide percentages can be used to distinguish all OGAs up to DP 40. The oligosaccharide-resolving power of PACE is very high. OGAs of up to DP 40 were separated and larger OGAs were detectable. Using a lower percentage of polyacrylamide or a longer time of migration, we believe that the larger OGAs might become resolvable. Indeed, neutral oligosaccharides of DP 130 have been resolved by PAGEFS using a lower percentage of polyacrylamide gel, and oligosaccharides up to DP 1300 were detected [24]. In contrast, OGAs up to DP 27 have been isolated using HPLC [25]. Using derivatization of OGAs and HPLC, a similar separation is possible [6]. Oligosaccharides from the charged polysaccharide hyaluronan can also be analyzed after derivatization using either PAGEFS or HPLC [21]. By PAGEFS, hyaluronan oligosaccharides are clearly separated up to a DP of 50 [17], whereas by HPLC, the limit of detection was DP 34 [21]. Similarly, using capillary zone electrophoresis, the largest rhamnogalacturonan-derived oligosaccharides separated were DP 30 [7]. Thus, PAGEFS methodologies such as PACE appear to give the best separation of higher-DP oligosaccharides. Methyl-esterified OGAs were separated by PACE, both from unmethylated OGAs and from oligosaccharides of different DM. It has also recently been shown that methylesterified OGAs can be separated by capillary electrophoresis and HPLC [13]. It will be interesting to determine whether OGAs with different positions of the methylester have different mobilities by PACE. This might be expected, as the shape of the OGA might differ. In the future, we will use mass spectrometry to identify methylation of oligosaccharides with different mobilities to resolve this issue. Indeed, Monsarrat et al. [26] showed that mass spectrometry can be used to identify derivatized oligosaccharides. Pectin hydrolysis by Endo-PG can be effectively studied by PACE. There are two important aspects to such studies. First, it is possible to study the specificity of the Endo-PG for methylated and unmethylated regions. Endo-PG from A. japonicus can cut PGA into small OGAs of DP 1–4 (Fig. 4). OGA DP 4 is hydrolyzed only very slowly. It does not cut highly methylated pectin, but can hydrolyze methylated pectin if there are unmethylated regions, yielding a fingerprint of methylated oligosaccharides. With standards, this pattern can be interpreted to yield specificity information. A second aspect of the PACE studies of pectin is that the fine structure can be investigated to quantitate blockwise and nonblockwise methylation. Nonblockwise methyl-esterified pectin is digested by Endo-PG and appears as a series of OGA when deesterified. The blockwise esterified PGA is insensitive to digestion and is digested only after deesterification. Since PACE is quantitative, this can be developed into an assay for the proportion of the pectin under

182

F. Goubet et al. / Analytical Biochemistry 321 (2003) 174–182

each condition: unesterified, nonblockwise esterified, and blockwise esterified. We showed that such an investigation is possible using crude Arabidopsis cell wall material (Fig. 5). Interestingly, the cell wall contains both blockwise and nonblockwise esterification. In combination with a pectin methylesterase or acetylase, it might become possible to use PACE to distinguish methylation and acetylation. Now that we have defined the fingerprint of esterified and nonesterified PGA in the cell wall from Arabidopsis plants, we will quantify these polysaccharides. When the biological variation and quantity is standardized, it will become possible to investigate any altered polysaccharide phenotype of mutant plants [27,28]. Acknowledgments This work was supported by grants from BBSRC. Brooke Morriswood was supported by a Nuffield Foundation summer studentship. We thank Paul Knox (University of Leeds, UK) and Martin Williams (Unilever, Bedford, UK) for providing us with unmethylated and methylated OGAs, respectively. We thank Zhinong Zhang for growing the plant material. We thank our colleagues at the laboratory for their helpful discussions and Peter Jackson for comments on the manuscript. References [1] D. Mohnen, in: B.M. Pinto (Ed.), Comprehensive Natural Products Chemistry, Carbohydrate and their Derivatives Including Tannins, Cellulose and Related Lignins, vol. 3, Elsevier, Oxford, 1999, pp. 497–527. [2] S.C. Fry, The Growing Plant Cell Wall: Chemical and Metabolic Analysis, The Blackburn Press, Caldwell, 2000. [3] J.P. Knox, P.J. Linstead, J. King, C. Cooper, K. Roberts, Planta 181 (1990) 512–521. [4] G.S. Hoondal, R.P. Tiwari, R. Tewari, N. Dahiya, Q.K. Beg, Appl. Microbiol. Biotechnol. 59 (2002) 409–418.

[5] R.V. Gadre, G. Van Driessche, J. Van Beeumen, M.K. Bhat, Enzyme Microbiol. Technol. 32 (2003) 321–330. [6] K. Akita, T. Ishimizu, T. Tsukamoto, T. Ando, S. Hase, Plant Physiol. 130 (2002) 374–379. [7] J. Fu, R. Prade, A. Mort, Carbohydr. Res. 330 (2001) 73– 81. [8] T. Ishii, J. Ichita, H. Matsue, H. Ono, I. Maeda, Carbohydr. Res. 337 (2002) 1023–1032. [9] N.O. Maness, J.D. Ryan, A.J. Mort, Anal. Biochem. 185 (1990) 346–352. [10] P. Massiot, V. Perron, A. Baron, J.-F. Drilleau, Lebensm.-Wiss. u.-Technol. 30 (1997) 697–702. [11] P.J.H. Daas, K. Meyer-Hansen, H.A. Schols, G.A. De Ruiter, A.G.J. Voragen, Carbohydr. Res. 318 (1999) 135–145. [12] P.J.H. Daas, A.G.J. Voragen, H.A. Schols, Carbohydr. Res. 326 (2000) 120–129. [13] M.A.K. Williams, G.M.C. Buffet, T.J. Foster, Anal. Biochem. 301 (2002) 117–122. [14] A.J. Mort, B.M. Moerschbacher, M.L. Pierce, N.O. Maness, Carbohydr. Res. 215 (1991) 219–227. [15] F. Goubet, P. Jackson, M. Deery, P. Dupree, Anal. Biochem. 300 (2002) 53–68. [16] P. Jackson, Anal. Biochem. 216 (1994) 243–252. [17] A. Calabro, M. Benavides, M. Tammi, V.C. Hascall, R.J. Midura, Glycobiology 10 (2000) 273–281. [18] J. Sudor, M.V. Novotny, Anal. Chem. 67 (1995) 4205–4209. [19] P.M. Rudd, R.A. Dwek, Curr. Opin. Biotechnol. 8 (1997) 488– 497. [20] A. Calabro, V.C. Hascall, R.J. Midura, Glycobiology 10 (2000) 283–293. [21] D.J. Mahoney, R.T. Aplin, A. Calabro, V.C. Hascall, A.J. Day, Glycobiology 11 (2001) 1025–1033. [22] T. Shimokawa, S. Yoshida, I. Kusakabe, T. Takeuchi, M. Murata, Carbohydr. Res. 299 (1997) 95–98. [23] K.J. Drummond, E.A. Yates, J.E. Turnbull, Proteomics 1 (2001) 304–310. [24] M.G. OÕShea, M.S. Samuel, C.M. Konik, M.K. Morell, Carbohydr. Res. 307 (1998) 1–12. [25] A.T.J. Hotchkiss, S.L. Lecrinier, K.B. Hicks, Carbohydr. Res. 334 (2001) 135–140. [26] B. Monsarrat, T. Brando, P. Condouret, J. Nigou, G. Puzo, Glycobiology 9 (1999) 335–342. [27] W.-D. Reiter, C. Chapple, C.R. Somerville, Plant J. 12 (1997) 335–345. [28] I. His, A. Driouich, F. Nicol, A. Jauneau, H. H€ ofte, Planta 212 (2001) 348–358.