Background and Aims

Rhamnogalacturonan II (RGII) is a structurally complex pectic sub-domain composed of more than 12 different sugars and 20 different linkages distributed in five side chains along a homogalacturonan backbone. Although RGII has long been described as highly conserved over plant evolution, recent studies have revealed variations in the structure of the polysaccharide. This study examines the fine structure variability of RGII in wine, focusing on the side chains A and B obtained after sequential mild acid hydrolysis. Specifically, this study aims to differentiate intrinsic structural variations in these RGII side chains from structural variations due to acid hydrolysis.

Methods

RGII from wine (Vitis vinifera Merlot) was sequentially hydrolysed with trifluoroacetic acid (TFA) and the hydrolysis products were separated by anion-exchange chromatography (AEC). AEC fractions or total hydrolysates were analysed by MALDI-TOF mass spectrometry.

Key Results

The optimal conditions to recover non-degraded side chain B, side chain A and RGII backbone were 0·1 m TFA at 40 °C for 16 h, 0·48 m TFA at 40 °C for 16 h (or 0·1 m TFA at 60 °C for 8 h) and 0·1 m TFA at 60 °C for 16 h, respectively. Side chain B was particularly prone to acid degradation. Side chain A and the RGII GalA backbone were partly degraded by 0·1 m TFA at 80 °C for 1–4 h. AEC allowed separation of side chain B, methyl-esterified side chain A and non-methyl-esterified side chain A. The structure of side chain A and the GalA backbone were highly variable.

Conclusions

Several modifications to the RGII structure of wine were identified. The observed dearabinosylation and deacetylation were primarily the consequence of acidic treatment, while variation in methyl-esterification, methyl-ether linkages and oxidation reflect natural diversity. The physiological significance of this variability, however, remains to be determined.

## INTRODUCTION

Pectin is a major constituent of primary cell walls of gymnosperms and dicotyledons (O'Neill et al., 2004). It is a multi-component polymer consisting of rhamnogalacturonan I (RGI), homogalacturonan (HG) and HG analogues; these sub-structures are covalently linked to each other (Ishii and Matsunaga, 2001; Ridley et al., 2001; Willats et al., 2001; Mohnen, 2008). The backbone structure of RGI is a repeating dimer of [→2)-α-l-Rhap-(1→4)-α-d-GalpA-(1→]. This heteropolymer is substituted by arabinan, galactan and arabinogalactan side chains in variable amounts (Willats et al., 2001). HG is a linear polymer of (1→4)-linked α-d-GalpA residues, which are usually methyl-esterified at O-6 and sometimes acetyl-esterified at O-2 and/or O-3 (Mohnen, 2008). HG analogues contain a (1→4)-linked α-d-GalpA backbone substituted by sugar monomers, sugar dimers or complex side chains. In xylogalacturonan and apiogalacturonan, the GalpA units are branched at O-3 and O-2 by monomers or dimers of xylose and apiose residues, respectively (Caffall and Mohnen, 2009). Rhamnogalacturonan II (RGII) contains five different side chains, defined as A–E, which decorate a short GalpA backbone. This very complex molecule is composed of more than 12 different sugars and 20 different linkages (O'Neill et al., 2004) constituting a mega-oligosaccharide of 5–10 kDa (Whitcombe et al., 1995; Vidal et al., 2000; Strasser and Amadò, 2001). RGII represents 1–5 % of dicotyledon primary cell walls (Darvill et al., 1978; O'Neill et al., 1990; Matoh et al., 1996). RGII was first characterized in sycamore by Darvill et al. (1978). The extraction method involved treatment with endo-polygalacturonase to release RGII from HG (Ishii and Matsunaga, 2001). Fractionation of the hydrolysate by size-exclusion chromatography (SEC) combined with anion-exchange chromatography (AEC) allowed the purification of RGII. Similar methods using various pectolytic enzymes and chromatographic methods have been used for RGII recovery from rice (Thomas et al., 1989), bamboo (Kaneko et al., 1997), tobacco BY-2 cells (Kobayashi et al., 1997), Cryptomeria japonica (Edashige and Ishii, 1998), ginseng (Shin et al., 1998), Chenopodium album (Fleischer et al., 1999), pumpkin (Ishii et al., 2001), red beet (Strasser and Amadò, 2001), Arabidopsis thaliana (Glushka et al., 2003), lycophytes (Matsunaga et al., 2004), pteridophytes (Matsunaga et al., 2004) and bryophytes (Matsunaga et al., 2004). In wine, enzymatic degradation occurs during fermentation and the remaining non-degraded polysaccharides are mannoproteins from yeasts, arabinogalactan proteins (AGPs), RGI and RGII, the last representing up to 20 % of the recovered wine polysaccharides (Doco and Brillouet, 1993; Pellerin et al., 1996; Vidal et al., 1999, 2003).

The structure of RGII has been determined using different approaches. This complex molecule encompasses specific sugar species, aceric acid (AcefA), methyl fucose (O-Me-Fucp), methyl xylose (O-Me-Xylp), apiose (Apif), 3-deoxy-d-manno-2-octulosonic acid (Dha) and 3-deoxy-d-lyxo-2-heptulosonic acid (Kdo), all of which can be detected and quantified (York et al., 1985; Stevenson et al., 1988; Kobayashi et al., 1996; Edashige and Ishii, 1998; Doco et al., 2001; Glushka et al., 2003; Yapo et al., 2007). Unusual linkages include 3-linked rhamnosyl, 3,4-linked fucosyl, 2-linked glucuronosyl residues and the fully substituted rhamnose, which are RGII-specific and can be detected by glycosyl linkage analysis (Darvill et al., 1978; Thomas et al., 1989; Kobayashi et al., 1996; Pellerin et al., 1996; Edashige and Ishii, 1998; Shimokawa et al., 1999; Strasser and Amadò, 2001; Glushka et al., 2003). The RGII glycosyl sequence was largely determined after chemical fragmentation of the molecule. Lithium dissolved in ethylamine (Stevenson et al., 1988; Edashige and Ishii, 1998), formolysis (Melton et al., 1986) and Smith degradation (Puvanesarajah et al., 1991; Glushka et al., 2003) are different ways that have been used to recover RGII-derived oligosaccharides. However, partial degradation by mild acid hydrolysis – usually with trifluoroacetic acid (TFA) – has been the most widely used method to isolate the RGII components of plant cell walls (Spellman et al., 1983; Stevenson et al., 1988; Thomas et al., 1989; Whitcombe et al., 1995; Shin et al., 1998; Matsunaga et al., 2004; Reuhs et al., 2004; Séveno et al., 2009; Voxeur et al., 2011; Pabst et al., 2013). The recovered oligosaccharides have then been characterized by sugar and linkage analysis, or have been directly studied by mass spectrometry. Such oligosaccharides were also analysed by nuclear magnetic resonance spectroscopy (NMR) to provide information about sugar sequence, anomeric configuration and the location of non-sugar substituents (Melton et al., 1986; Puvanesarajah et al., 1991; Whitcombe et al., 1995; Glushka et al., 2003; Pabst et al., 2013). The proposed distribution of side chains, based on NMR measurements combined with molecular modelling, played a major role in the description of the RGII three-dimensional structure (Vidal et al., 2000; Pérez et al., 2003; Rodríguez-Carvajal et al., 2003).

Around 95 % of RGII molecules from normal plants are dimeric (O'Neill et al., 2001), which was shown to involve boron integration (O'Neill et al., 1996; Fleischer et al., 1999), sequestering about 80 % of the cellular boron in plants (Matoh et al., 1996). Boron was shown to have a key role in plant growth, and RGII structure was shown to be essential for its internalization in the cell wall (O'Neill et al., 2001). Boron deficiency leads to larger cells with swollen walls, and the RGII–dimer–boron complex has therefore been suspected to control cell-wall properties (Matoh et al., 2000).

Interestingly, the structurally highly complex RGII is highly conserved over plant evolution (Matsunaga et al., 2004). However, characterization of oligosaccharides generated by RGII partial hydrolysis highlighted several potential structural alterations (Matsunaga et al., 2004; Pabst et al., 2013). Differentiating intrinsic structural variations from structural variations due to the acid hydrolysis conditions used remains challenging. Here we report on a range of sequential mild acid hydrolyses and their impact on wine RGII side chains, backbone recovery and fine structure. The potential influence of the observed structural variability in muro will be discussed.

## MATERIALS AND METHODS

### RGII extraction

Total colloids were recovered from 21 litres of red wine (Vitis vinifera Merlot) as described by Vidal et al. (2003). The total colloids solution was mixed with one volume of strong cation-exchange resin (IR-120, Amberlite) and left overnight at room temperature. The unbound fraction was mixed with a 0·5 volume of strong cation-exchange gel (Sepharose SP Fast Flow, GE Healthcare, Little Chalfont, UK) and left for 4 h at room temperature. The unbound fraction was finally mixed with one volume of strong anion-exchange resin (IRA-958, Amberlite) and left overnight at room temperature. The unbound fraction, corresponding to non-pigmented total wine polysaccharides, was dialysed against distilled water and freeze-dried. Total wine polysaccharides (7·1 g) were solubilized in 25 mm sodium acetate buffer, pH 4·5 (320 mL), and loaded at 90 mL h−1 in four batches on a DEAE-Sepharose Fast Flow column (5 × 15 cm) equilibrated with the same buffer. An unbound fraction was recovered and bound polysaccharides were eluted by steps of NaCl (50 mm and 250 mm in the starting buffer). The fraction eluted with 250 mm NaCl, corresponding to raw RGII, was dialyzed and freeze-dried. Raw RGII (3 g) was solubilised in 50 mm sodium acetate buffer pH 4·5 (60 mL) and loaded in twelve batches at 90 mL h–1 on a Sephacryl-S400 HR column (2·6 × 100 cm). Appropriate fractions were pooled, dialysed and freeze-dried to yield 1·24 g of purified RGII.

### Sequential mild acid hydrolysis and AEC

RGII (200 mg) was dissolved in 20 mL water and treated with 0·1 m TFA for 16 h at 40 °C. TFA was removed by vacuum rotatory evaporation at 40 °C to dryness. The hydrolysate was solubilized in water and further evaporated to dryness. This last step was repeated three times. The hydrolysate was finally solubilized in water and pH was adjusted to 4·5 with a few drops of 100 mm NaOH. AEC was performed at room temperature on a DEAE-Sepharose Fast Flow column (1·6 × 13·5 cm) equilibrated with degased 25 mm sodium acetate buffer, pH 4·5, at a flow rate of 90 mL h–1. The hydrolysate (20 mL) was loaded onto the column and the gel was washed with 50 mL of 25 mm sodium acetate buffer, pH 4·5. The bound material was eluted with a linear NaCl gradient (0–100 mm NaCl in 25 mm sodium acetate buffer, 300 mL). Sodium acetate buffer containing 500 mm NaCl (50 mL) was then applied. Fractions (3·5 mL) were collected and analysed for their content of GalA and neutral sugars by colorimetry (Thibault, 1979; Tollier and Robin, 1979). Appropriate fractions were combined in pools and concentrated by vacuum rotary evaporation at 40 °C. The pool eluted during the washing step (Pool A) was desalted using a column (1·6 × 100 cm) of Sephadex G-10 run in deionized water at 1 mL min–1. The pool eluted with 500 mm NaCl (Pool B) was dialysed, freeze-dried and further hydrolysed with 0·48 m TFA for 16 h at 40 °C. TFA was removed and pH was adjusted as described above. After fractionation by AEC, pools eluted during the washing step or by the linear NaCl gradient (Pools Ba, Bb and Bc) were desalted using a column of Sephadex G-10 as described above. The pool eluted at 500 mm NaCl (pool Bd) was dialysed and freeze-dried. Pool Bd aliquots (0·25 mg) were further hydrolysed with 1 mL of 0·1 m TFA for 4, 8 or 16 h at 60 °C or for 1, 2 or 4 h at 80 °C (Fig. 1). TFA was removed as described above and hydrolysates were solubilized in 1 mL water.

### Alkaline hydrolysis

Samples (100 μg in 400 μL water) were incubated with 25 mm heptylamine (HA; Sigma-Aldrich, St Louis, MO, USA) in a 1 : 1 ratio (v/v) for 3 d. A tri-methyl-esterified tri-GalpA standard was incubated under the same conditions as a positive control.

### Matrix-assisted laser desorption ionization time-of-flight mass spectrometry

MALDI-TOF-MS spectra were acquired on an Autoflex III TOF/TOF mass spectrometer (Bruker Daltonics, Bremen, Germany), equipped with a Smartbeam Laser (355 nm, 200 Hz) and reflector detection. Freeze-dried samples were solubilized in water (1 mg mL–1). Sample solution (1 mg mL–1 water or HA hydrolysates) and matrix, prepared as previously described (Ropartz et al., 2011), were mixed (1 : 1, v/v) and applied on a polished steel MALDI target plate. Spectra were recorded using FlexControl and processed using FlexAnalysis (Bruker Daltonics, Billerica, MA, USA). Mass spectra were acquired in positive and negative ionization mode.

## RESULTS

### Side chain debranching

RGII is a substituted α-d-GalpA backbone carrying five structurally distinct side chains: (1) the octameric side chain A encompassing three acidic sugars, (2) the hexameric to nonameric side chain B encompassing one acidic sugar, (3) the dimeric side chains C and D encompassing one acidic sugar and (4) the monomeric neutral side chain E. The monomeric side chain E (α-l-Araf), when present, is O-3-linked to the GalpA backbone (Melton et al., 1986; Pérez et al., 2003). The two dimeric side chains C and D are linked to the backbone at O-3 via their ulosonic acid residue (Stevenson et al., 1988). The complex side chains A and B are both attached at O-2 of the backbone via an apiosyl residue (Stevenson et al., 1988). The relative position of the side chains on the GalpA backbone is not fully understood (Yapo, 2011). The three side chains C, B and D are located, in this order, to the first five GalpA residues from the reducing end of the molecule (Vidal et al., 2000). Side chain A is expected to be the last side chain present from the reducing end (Vidal et al., 1999). Because glycosyl linkage analyses detected the presence of (2,3,4)-linked GalpA residues in different species, two side chains are likely to be linked onto the same GalpA unit (Doco and Brillouet, 1993; Pellerin et al., 1996; Doco et al., 1997; Strasser and Amadò, 2001). Side chain E has not yet been precisely located but could be linked to the third GalpA residue from the non-reducing end together with side chain A (Yapo, 2011).

Debranching the various side chains from the GalpA backbone required different hydrolysis conditions. Their ability to be detached probably follows the order C–D > B > A (Table 1). The acid-lability of side chain E is unknown. Hydrolysis conditions have to be chosen carefully to avoid side chain degradation. Dimeric side chains C and D can both be detached with 0·1 m TFA for 6 h at 40 °C (York et al., 1985). When the same hydrolysis conditions (0·1 m TFA at 40 °C) were applied for a longer time (16 or 24 h), side chain B was detached from the GalpA backbone and only trace amounts of degradation products were observed (Table 1). When harsher conditions were used, extensive hydrolysis of chain B was reported (Séveno et al., 2009). In contrast, Pabst et al. (2013) found no significant difference in the data whether the hydrolysis was performed for 16 h at 40 °C or for 1 h at 80 °C. Side chain A was efficiently released from the RGII backbone by increasing the temperature of hydrolysis to 80 °C for 1 h but some degradation products were observed (Table 1). Acid-lability is therefore specific for each branching residue, and sequential hydrolysis is highly recommended for the recovery of RGII side chains (Séveno et al., 2009).

Table 1.

Hydrolysis conditions used for RGII side chain release; 0·1 m TFA was used in all studies

Side chain released Incubation time (h) Temperature (°C) Degraded products* Reference
C and D 0·5 60 – Thomas et al. (1989)
16 40 +/– Whitcombe et al. (1995)
16 40 +/– Matsunaga et al. (2004)
16 40 +/– Séveno et al. (2009)
16 40 Voxeur et al. (2011)
24 40 +/– Shin et al. (1998)
24 40 +/– Reuhs et al. (2004)
24 50 Thomas et al. (1989)
80 ++ Pabst et al. (2013)
80 +++ Séveno et al. (2009)
100 ++++ Séveno et al. (2009)
16 40 – Séveno et al. (2009)
16 40 +/– Voxeur et al. (2011)
24 50 Thomas et al. (1989)
80 +/– Pabst et al. (2013)
80 ++ Séveno et al. (2009)
100 +++ Séveno et al. (2009)
Side chain released Incubation time (h) Temperature (°C) Degraded products* Reference
C and D 0·5 60 – Thomas et al. (1989)
16 40 +/– Whitcombe et al. (1995)
16 40 +/– Matsunaga et al. (2004)
16 40 +/– Séveno et al. (2009)
16 40 Voxeur et al. (2011)
24 40 +/– Shin et al. (1998)
24 40 +/– Reuhs et al. (2004)
24 50 Thomas et al. (1989)
80 ++ Pabst et al. (2013)
80 +++ Séveno et al. (2009)
100 ++++ Séveno et al. (2009)
16 40 – Séveno et al. (2009)
16 40 +/– Voxeur et al. (2011)
24 50 Thomas et al. (1989)
80 +/– Pabst et al. (2013)
80 ++ Séveno et al. (2009)
100 +++ Séveno et al. (2009)

*Refers to the relative amounts of degradation products observed (from −, very low amounts to ++++, very high amounts).

To separate hydrolysis products, various chromatographic methods have been used. SEC has been widely used for purification of side chain B, the detached side chain B (DP 6–9) being well separated from detached dimeric side chains C and D and from RGII remnants consisting of side chain A attached to the GalpA backbone (Spellman et al., 1983; Whitcombe et al., 1995; Shin et al., 1998; Matsunaga et al., 2004; Reuhs et al., 2004). AEC has also been used to separate hydrolysis products based on their charges, which are not equally distributed over the entire molecule, the side chains and backbone (Thomas et al., 1989). Separation on porous graphitized carbon liquid chromatography was recently performed at the analytical scale (Pabst et al., 2013).

In the present study, wine RGII was submitted to sequential acid hydrolysis and hydrolysis products were separated by AEC. After a first hydrolysis step (0·1 m TFA, 16 h, 40 °C), two populations were separated (Fig. 2A). Sugar analysis revealed that Pool A did not contain GalA. Side chains B, C and D lack uronic acids, and were likely to be released under the acid conditions used (Table 1). Pool B, which contained both neutral sugars and GalA, was eluted with an ionic strength compatible with the presence of more than seven charges, and was expected to contain intact and/or partially hydrolysed RGII. Pool B underwent a second hydrolysis step under slightly harsher conditions (0·48 m TFA, 16 h, 40 °C), and AEC of the hydrolysate generated four populations (Fig. 2B). Pool Ba was similar to Pool A with respect to elution volume and global sugar analysis and probably contains side chains B, C and D that were not released during the first acidic treatment (Spellman et al., 1983; Whitcombe et al., 1995; Matsunaga et al., 2004). Pool Bb and Pool Bc both contained GalA and neutral sugars and were eluted with ionic strengths compatible with the presence of two and three charges, respectively. Pool Bd was expected, as with the previously recovered Pool B, to contain RGII remnants that were partially hydrolysed. Detection of neutral sugars in this pool provides evidence that side chains are not totally released after the second hydrolysis step.

Fig. 1.

RGII sequential extraction scheme.

Fig. 1.

RGII sequential extraction scheme.

Fig. 2.

Anion-exchange chromatography elution patterns of RGII hydrolysate (neutral sugars and uronic acid, as indicated in the key). (A) Wine RGII hydrolysed by 0·1 m TFA for 16 h at 40 °C. (B) Pool Bd further hydrolysed by 0·48 m TFA for 16 h at 40 °C.

Fig. 2.

Anion-exchange chromatography elution patterns of RGII hydrolysate (neutral sugars and uronic acid, as indicated in the key). (A) Wine RGII hydrolysed by 0·1 m TFA for 16 h at 40 °C. (B) Pool Bd further hydrolysed by 0·48 m TFA for 16 h at 40 °C.

### Side chain B

Side chain B was first characterized as a heptasaccharide in RGII from sycamore suspension cultured cells (Spellman et al., 1983). However, some variability with respect to side chain B degree of polymerization (DP) was observed, and side chains B of DP 6 to DP 9 have been purified from various plant sources (Table 2; Fig. 3). The length of side chain B depends on variations in glycosyl substitution on the α-l-Arap residue (Fig. 3). The heptasaccharide described by Spellman et al. (1983) is rhamnosylated at O-2 of the α-l-Arap residue (Fig. 3). This structure has been detected in most plants (Table 2). Two different forms of octasaccharides were also found. The first form consists of a disaccharide α-l-Rhap-(2→1)-β-l-Araf substitution at O-2 of the α-l-Arap residue. This type of octasaccharide was detected in red wine, bamboo shoot, suggi and sycamore (Stevenson et al., 1988; Whitcombe et al., 1995; Kaneko et al., 1997; Edashige and Ishii, 1998; Glushka et al., 2003; Pabst et al., 2013). In the second octameric form, α-l-Arap was di-substituted at O-2 and O-3 by α-l-Rhap residues. This structure has been detected in Arabidopsis thaliana and in some lycophytes and pteridophytes (Matsunaga et al., 2004). A nonasaccharide containing both the disaccharide α-l-Rhap-(2→1)-β-l-Araf substitution at O-2 of α-l-Arap and the monomeric α-l-Rhap at O-3 of the α-l-Arap residue was identified in red wine and ginseng (Shin et al., 1998; Glushka et al., 2003; Pabst et al., 2013). A hexasaccharide resulting from the absence of rhamnosylation of the α-l-Arap residue was also detected in sugi, red beet and Arabidopsis thaliana (Edashige and Ishii, 1998; Strasser and Amadò, 2001; Pabst et al., 2013). Different structures are reported to coexist in the same plant and it was recently shown that their proportion varies during plant development and is organ-specific (Pabst et al., 2013).

Table 2.

RGII side chain B structural variability with respect to plant source

Plant source α-l-Rhap linked at O-3 of α-l-Arap α-l-Rhap linked at O-2 of α-l-Arap β-l-Araf linked at O-2 of α-l-Rhap DP References
Sycamore – +/– 7–8 Spellman et al. (1983), Stevenson et al. (1988), Whitcombe et al. (1995)
Red wine +/– +/– +/– 6–9 Glushka et al. (2003), Pabst et al. (2013)
Panax ginseng Shin et al. (1998)
Pectinol AC – – – Stevenson et al. (1988)
Arabidopsis thaliana +/– +/– – 6–8 Glushka et al. (2003), Pabst et al. (2013)
Rice – – Thomas et al. (1989)
Bamboo – +/– 7–8 Kaneko et al. (1997)
Red beet – +/– – 6–7 Strasser and Amadò (2001)
Sugi – +/– +/– 6–8 Edashige and Ishii (1998)
Lycophyte/pteridophyte +/– +/– – 7–8 Matsunaga et al. (2004)
Plant source α-l-Rhap linked at O-3 of α-l-Arap α-l-Rhap linked at O-2 of α-l-Arap β-l-Araf linked at O-2 of α-l-Rhap DP References
Sycamore – +/– 7–8 Spellman et al. (1983), Stevenson et al. (1988), Whitcombe et al. (1995)
Red wine +/– +/– +/– 6–9 Glushka et al. (2003), Pabst et al. (2013)
Panax ginseng Shin et al. (1998)
Pectinol AC – – – Stevenson et al. (1988)
Arabidopsis thaliana +/– +/– – 6–8 Glushka et al. (2003), Pabst et al. (2013)
Rice – – Thomas et al. (1989)
Bamboo – +/– 7–8 Kaneko et al. (1997)
Red beet – +/– – 6–7 Strasser and Amadò (2001)
Sugi – +/– +/– 6–8 Edashige and Ishii (1998)
Lycophyte/pteridophyte +/– +/– – 7–8 Matsunaga et al. (2004)

− absent; +/− partly present; + present.

Fig. 3.

Side chain B structure. Sugars in black are those detected by Spellman et al. (1983). Acetyl groups can esterify the 2-O-Me-α-d-Fucp and/or the α-l-AcefA residue. Methyl groups can etherify terminal α-l-Rhap residues at O-3.

Fig. 3.

Side chain B structure. Sugars in black are those detected by Spellman et al. (1983). Acetyl groups can esterify the 2-O-Me-α-d-Fucp and/or the α-l-AcefA residue. Methyl groups can etherify terminal α-l-Rhap residues at O-3.

Variation in esterification and etherification status was also detected in side chain B. In some lycophytes and pteridophytes, terminal Rhap units O-2- and/or O-3-linked to the α-l-Arap residue were shown to contain 3-O-methyl groups (Matsunaga et al., 2004). Acetyl groups linked to α-l-AcefA and/or 2-O-Me-α-d-Fucp were also identified (Whitcombe et al., 1995). Mono-, di- and no-acetylation generally coexist in all species although mono-acetylation on the 2-O-Me-α-d-Fucp residue appears to be the predominant form (Spellman et al., 1983; Whitcombe et al., 1995). In Arabidopsis thaliana, RGII acetylation was reported not to be organ specific (Pabst et al., 2013).

Pool A and Ba recovered after AEC (Fig. 2) were analysed by mass spectrometry (Fig. 4). No signal was detected for m/z < 1000. Pool A arising from RGII treated with 0·1 m TFA at 40 °C for 16 h exhibited one main signal at m/z 1375 corresponding to a mono-acetylated side chain B of DP 9 (Fig. 4A). All other signals were very weak. Signals at m/z 1333 and 1417 were assigned to non-acetylated and di-acetylated side chain B of DP 9. Signals at m/z 1243 and 1229 were assigned to acetylated side chain B of DP 8 having lost a β-l-Araf residue or α-l-Rhap unit, respectively. The presence of a signal at m/z 1097 indicates that acetylated side chain B of DP 7 had lost the terminal disaccharide α-l-Rhap-(2→1)-β-l-Araf or both the terminal β-l-Araf and α-l-Rhap units. Pool Ba arising from RGII treated with 0·48 m TFA at 40 °C for 16 h exhibited a more complex spectrum (Fig. 4B). The ion corresponding to the di-acetylated side chain B of DP 9 was not detected. The two main signals at m/z 1375 and 1333 were assigned to mono-acetylated and non-acetylated side chain B of DP 9, respectively. Signals at m/z 1243, 1229 and 1097 corresponded to acetylated side chain B lacking a terminal β-l-Araf, terminal α-l-Rhap or the disaccharide (α-l-Rhap-(2→1)-β-l-Araf), respectively. Similar degradation products arising from non-acetylated side chain B were observed at m/z 1201, 1187 and 1055. Both non-acetylated and mono-acetylated side chain B of DP 8 having lost one α-l-Araf residue were also observed and gave strong signals. Signals at m/z 1173 and 1041 were assigned to non-acetylated side chain B fragments having lost a 2-O-Me-α-l-Fucp residue or both a 2-O-Me-α-l-Fucp and a single β-l-Araf unit, respectively. The acetylated form of these two fragments was not detected. It is therefore assumed that the acetyl group is preferably located on the 2-O-Me-α-l-Fucp residue as previously reported (Whitcombe et al., 1995). It is evident that the harsher hydrolysis conditions applied for side chain B generated new fragments, indicating that the β-l-Araf, 2-O-Me-α-l-Fucp and acetyl linkages were particularly acid-labile. In contrast, the rhamnosylation status of side chain B was not greatly affected by hydrolysis conditions; the signal at m/z 1229, corresponding to side chain B lacking one Rhap residue, was weak irrespective of the extraction condition (Fig. 4).

Fig. 4.

MALDI-TOF mass spectra (negative mode) of RGII hydrolysate. (A) Pool A recovered after AEC of 0·1 m TFA for 16 h at 40 °C. (B) Pool Bb recovered after AEC of 0·48 m TFA for 16 h at 40 °C.

Fig. 4.

MALDI-TOF mass spectra (negative mode) of RGII hydrolysate. (A) Pool A recovered after AEC of 0·1 m TFA for 16 h at 40 °C. (B) Pool Bb recovered after AEC of 0·48 m TFA for 16 h at 40 °C.

### Side chain A

Side chain A was first characterized as an octasaccharide (Fig. 5) in sycamore and rice (Stevenson et al., 1988; Thomas et al., 1989). Side chain A is particularly important as it is involved in RGII dimerization via a boron di-ester bond (Shimokawa et al., 1999). Boron is covalently linked to β-d-Apif units at O-2 and O-3 (Ishii and Ono, 1999; O'Neill et al., 1996). This occurs only with Apif residues from side chain A, which has less conformational flexibility than side chain B (Glushka et al., 2003; Rodríguez-Carvajal et al., 2003; O'Neill et al., 2004). Side chain A also contains three uronic acids that constitute a suitable site for chelating cations (Pérez et al., 2003). Two calcium ions, which do not interact with boron, are indeed required to stabilize the RGII complex (O'Neill et al., 1996; Matoh and Kobayashi, 1998; Kobayashi et al., 1999). Possible methylation of side chain A was first hypothesized by Séveno et al. (2009). This was recently confirmed by Pabst et al. (2013), who showed that both methyl-esterification at O-6 of the β-d-GlcpA residue and methylation at O-3 and/or O-4 of the β-d-GalpA residue could occur. Methyl-esterification is likely to affect calcium-mediated RGII complex stabilization.

Fig. 5.

Side chain A structure. The β-d-GlcpA residue can be methyl-esterified (in green). The β-d-GalpA residue can be singly or doubly methyl-etherified (in green). The α-l-Fucp residue can be oxidized in α-l-Galp (green box).

Fig. 5.

Side chain A structure. The β-d-GlcpA residue can be methyl-esterified (in green). The β-d-GalpA residue can be singly or doubly methyl-etherified (in green). The α-l-Fucp residue can be oxidized in α-l-Galp (green box).

Pool Bb and Bc from AEC (Fig. 2B) were analysed by mass spectrometry (Fig. 6A and C). No signal was detected at m/z < 1300. For Pool Bb, a major (M+Na)+ ion corresponding to a mono-methylated intact chain A was detected at m/z 1315 (Fig. 6A). No signal at m/z 1301, corresponding to a non-methylated full chain A, was detected whereas ions at m/z 1329 and 1343, corresponding to di- and tri-methylated full chain A, respectively, were observed. For Pool Bc, a major (M+Na)+ ion corresponding to a non-methylated full chain A was detected at m/z 1301 (Fig. 6C). Weak signals at m/z 1315 and 1329, corresponding to mono- and di-methylated chain A, respectively, were also detected. Pools Bb and Bc were incubated with HA, an alkaline component that can hydrolyse methyl-ester groups. A tri-methyl-esterified tri-GalpA standard was incubated under the same conditions. Only non-methyl-esterified tri-GalpA was detected by mass spectrometry after a 3-day incubation. After HA treatment of Pool Bd, a shift from the mono-methylated (m/z 1315) to the non-methylated (m/z 1301) form of side chain A was observed (Fig. 6B). This, together with the fact that Pool Bd was eluted by AEC for an ionic strength compatible with the presence of two charges, strongly supports the hypothesis that side chain A is mainly singly methyl-esterified in this pool. The fact that ions corresponding to mono- and di-methylated side chain A (m/z 1315 and 1329, respectively) can still be detected after HA treatment suggests that one or two methyl-etherification can exist in addition to methyl-esterification, as recently reported by Pabst et al. (2013). This hypothesis is supported by the disappearance of the signal assigned to tri-methylated side chain A (m/z 1343) (one methyl-ester and two methyl-ether groups) after HA treatment. In contrast, the Pool Bc spectrum remained similar after HA treatment, but lower intensities were observed due to the presence of HA (Fig. 6D). Non-methylated, mono-methylated and di-methylated forms of side chain A were present in similar proportions before and after treatment. In this pool, methyl-etherification only is likely to be present, in agreement with the fact that Pool Bc was eluted by AEC for an ionic strength compatible with the presence of three charges.

Fig. 6.

MALDI-TOF mass spectra (positive mode). (A) Pool Bb recovered from RGII hydrolysate after AEC of 0·48 m TFA for 16 h at 40 °C. (B) Pool Bb incubated with heptylamine. (C) Pool Bc recovered from RGII hydrolysate after AEC of 0·48 m TFA for 16 h at 40 °C. (D) Pool Bc incubated with heptylamine. *(M+Na+); **(M+2Na+); ***(M+3Na+).

Fig. 6.

MALDI-TOF mass spectra (positive mode). (A) Pool Bb recovered from RGII hydrolysate after AEC of 0·48 m TFA for 16 h at 40 °C. (B) Pool Bb incubated with heptylamine. (C) Pool Bc recovered from RGII hydrolysate after AEC of 0·48 m TFA for 16 h at 40 °C. (D) Pool Bc incubated with heptylamine. *(M+Na+); **(M+2Na+); ***(M+3Na+).

Oxidation of the α-l-Fucp or α-l-Galp, adding 16 Da to side chain A, was recently observed by Pabst et al. (2013). Signals 16 Da greater than RGII-derived oligosaccharides were observed in Pool Bb and Bc (Fig. 6A and 6C). Methyl-esterified and/or methyl-etherified forms of the oxidized structure were also evident. In Pool Bb the oxidized structure was only mono- (m/z 1331) and di-methylated (m/z 1345) (Fig. 6A). After incubation with HA, ions at m/z 1317 and m/z 1331 assigned to non-methylated and mono-methylated oxidized side chain A, respectively, were observed (Fig. 6B). In this pool, oxidized side chain A was therefore mono-methyl-esterified and non- or mono-methyl-etherified. In Pool Bc, signals at m/z 1317 and 1331 were assigned to non- and mono-methylated oxidized side chain A, respectively (Fig. 6C). After HA treatment, no significant changes were observed in the mass spectrum (Fig. 6D). In this pool, oxidized side chain A species are either mono-methyl-etherified or non-methylated.

### Backbone and short side chains (C, D, E)

Three short side chains, the dimeric C and D chains and the monomeric chain E, are present on the RGII backbone. Side chains C and D are difficult to extract, due to the acid-lability of ulosonic acid residues, and identifying their (low molecular weight) signals within the matrix peaks (Stevenson et al., 1988; Séveno et al., 2009). Molecule protection through NaBH4 reduction (Stevenson et al., 1988) or derivatization with a fluorescent tag (Séveno et al., 2009) have been used for detection. Side chains C and D are likely to be co-eluted with side chain B in Pool B and Ba as previously observed (Thomas et al., 1989). In the present study these two dimers are expected to have been lost during the desalting steps. Side chain E, consisting of one α-l-Araf residue linked O-3 to a GalpA residue in the backbone, is not always detected. Total or partial removal of this residue can be due to the action of pectolytic enzymes used for RGII extraction and to the TFA hydrolysis conditions used for side chain debranching (Pabst et al., 2013).

Based on the molar composition and molar mass of glycosyl residues, a backbone's length of six to 11 residues has been estimated by Yapo (2011). DP values between 8 and 9 are the most commonly found structures (Melton et al., 1986; Whitcombe et al., 1995; Pellerin et al., 1996). Higher DP values have been reported in sycamore (Whitcombe et al., 1995) and in wine (Pellerin et al., 1996). As RGII is covalently linked (1→4) to α-d-GalpA HG (Ishii and Matsunaga, 2001), the boundary between HG and RGII is difficult to determine. Depending on extraction conditions, the presence of HG at the ends of RGII molecules cannot be precluded (Yapo, 2011). Conversely, oligogalacturonates of low DP ( < 6), which are likely to be due to partial acid degradation of the RGII backbone (Yapo, 2011), have been observed. The GalpA residues can be methyl-esterified, as first reported by Melton et al. (1994).

To gain insight into (1) backbone length, (2) the presence of side chain E and (3) methyl-esterification status, different hydrolysis conditions ranging from 0·1 m TFA for 4 h at 60 °C to 0·1 m TFA for 4 h at 80 °C were applied to Pool Bd (Fig. 2). The different hydrolysates were analysed by mass spectrometry (Supplementary Data Fig. S1). No signal was detected at m/z < 1155. After 0·1 m TFA for 4 h at 60 °C, matrix-related signals only were detected (Fig. S1A). After 0·1 m TFA for 8 h at 60 °C, several signals were observed that were assigned to different forms of side chain A (non-methylated, methylated and oxidized at m/z 1301, 1315 and 1317, respectively) (Fig. S1B) and oligogalacturonates of DP 8 and 9 bearing methyl groups from zero to three and/or an α-l-Araf residue (Fig. S1B, Fig. 7). The presence of di- and tri-methylated oligogalacturonates of DP 7 that exhibit the same molar mass as non-methylated and mono-methylated side chain A, respectively, cannot be precluded. After 0·1 m TFA for 16 h at 60 °C (Fig. S1C), the spectrum obtained was very similar to the data presented in Supplementary Fig. S1B, except for the appearance of a signal at m/z 1155, which can be attributed to a truncated form of side chain A (side chain A lacking 2-O-Me-α-d-Xyl). Interestingly, after hydrolysis at 80 °C (Fig. S1D–F), signals corresponding to oligogalacturonates bearing one β-l-Araf residue could no longer be detected. Furthermore, after the 80 °C treatment the signals corresponding to ‘full length’ side chain A decreased in favour of signals corresponding to the truncated form of the oligosaccharide mentioned above.

Fig. 7.

MALDI-TOF mass spectrum of Pool Bd hydrolysed by 0·1 m TFA for 8 h at 60 °C. Zoom of the m/z 1400–1850 region. The full spectrum is shown on Supplementary Fig. 1B.

Fig. 7.

MALDI-TOF mass spectrum of Pool Bd hydrolysed by 0·1 m TFA for 8 h at 60 °C. Zoom of the m/z 1400–1850 region. The full spectrum is shown on Supplementary Fig. 1B.

## DISCUSSION

In the present study, different mild acid treatments were used to deconstruct wine RGII. AEC allowed good separation of side chain B, and methyl-esterified and non-methyl-esterified side chain A. Hydrolysis conditions were shown to have a significant impact on the structures observed.

After hydrolysis of wine RGII with 0·1 m TFA at 40 °C for 16 h, the main component released was side chain B with a DP of 9 that was mono-acetylated on the 2-O-Me-α-d-Fucp residue. Trace amounts of non- and di-acetylated side chain B and mono-acetylated side chain B of DP 7 or 8 were also detected. The use of harsher hydrolysis conditions induced partial degradation of side chain B. β-l-Araf, 2-O-Me-α-l-Fucp and acetyl groups appeared particularly acid-labile. In a recent study, side chain B of DP 8, having lost its terminal β-l-Araf residue, was reported to be predominant in wine RGII (Pabst et al., 2013). It is likely that in the latter study, the terminal β-l-Araf residue was lost due to the harsh hydrolysis conditions used. The present study provides evidence that the rhamnosylation status of side chain B is not highly affected by hydrolysis conditions. In Arabidopsis thaliana RGII, which lacks a terminal β-l-Araf residue, rhamnosylation was shown to vary according to organ and developmental stage (Pabst et al., 2013). From 2 to 18 days after imbibition, single and double rhamnosylation of side chain B decreases, while a variant of the oligosaccharide lacking any α-l-Rha increases. In adult plants, rhamnosylation of side chain B appears organ-specific; the oligosaccharide decoration lacking α-l-Rha is the predominant structure in stems and siliques, while in leaves chain B generally contains a single α-l-Rha residue (Pabst et al., 2013). The hydrophobic character of side chain B, due to the presence of Rhap, Arap, O-Me and O-Ac groups, has been pointed out by Spellman et al. (1983). Variability in acetylation and/or in the length of side chain B could be a way to modulate the hydrophobic character of RGII and hence its association with other molecules through hydrophobic interactions.

Arabinogalactan proteins, RGII and boron are co-located close to the plasma membrane (Matoh et al., 1998; González-Fontes et al., 2008) and a boron deficiency was shown to prevent the covalent linkage of a hydroxyproline-/proline-rich protein to bean (Phaseolus vulgaris L.) root nodules (Bonilla et al., 1997). More recently, an arabinogalactan protein-extensin (APGE) from a legume nodule was extracted together with RGII using the MAC265 monoclonal antibody that binds to the glycoprotein. Dimeric RGII was shown to bind to the glycoprotein (Reguera et al., 2010).

After hydrolysis of wine RGII with 0·48 m TFA at 40 °C for 16 h, two forms of ‘full-length’ side chain A, one methyl-esterified and one non-methyl-esterified, were recovered and separated by AEC. These two forms of ‘full-length’ side chain A could also be recovered after hydrolysis with 0·1 m TFA at 60 °C for 8 h. The use of harsher extraction conditions rapidly led to the appearance of a truncated form of side chain A having lost the O-Me-α-d-Xylp residue. Analysis by mass spectrometry revealed that side chain A is structurally highly variable. Besides single methyl-esterification, single and double methyl-etherification and substitution of the central α-l-Fucp residue by α-l-Galp were evident, in good agreement with Pabst et al. (2013). In the present study, based on the relative intensities of mass spectrometry signals, the proportion of chain A with substitution of α-l-Fucp by α-l-Galp of 27 % is also in line with Pabst et al. (2013). This naturally occurring substitution provides an explanation for the detection of 3,4-linked-Galp residues in purified wine RGII fractions (Pellerin et al., 1996). Pabst et al. (2013) screened for oxidized side chain A in different plants and showed it constituted up to 45 % of total side chain A. The natural presence of this modified form of side chain A in several plants allowed Pabst et al. (2013) to shed new light on several RGII mutant phenotypes. Indeed, substitution of α-l-Fucp by α-l-Galp had been previously described in the Arabidopsis thaliana mur1 mutant, which lacked the gene encoding GDP-d-Man-4,6-dehydratase required for the conversion of GDP-d-Man to GDP-d-Fuc (Bonin et al., 1997; O'Neill et al., 2001; Reuhs et al., 2004). The mur1 mutant showed a decrease in RGII dimerization, which severely affected plant growth (O'Neill et al., 2001). In this mutant, side chain A was truncated upstream of the central β-l-Rhap, which led Pabst et al. (2013) to claim that the impact of the mutation on RGII dimerization could be due to truncation of side chain A rather than to the α-l-Fucp substitution by α-l-Galp. Silencing of GDP-d-Man-3,5-epimerase in tomato plants also resulted in a lower capacity of RGII to perform in muro cross-linking (Gilbert et al., 2009; Voxeur et al., 2011). The oxidized form of side chain A – encompassing an internal α-l-Galp – represents 25 % of total side chain A in Solanum lycopersicum. It was thereby hypothesized that truncation of side chain A – due to absence of this internal α-l-Galp – rather than a decrease in terminal α-l-Galp content could be responsible for the observed phenotypes (Pabst et al., 2013).

After hydrolysis of wine RGII with 0·1 m TFA at 60 °C for 8 or 16 h, backbone oligogalacturonates of DP 8 and 9 bearing zero or one β-l-Araf residue were released. The use of harsher extraction conditions led to the loss of arabinosylated oligogalacturonates. Methyl-esterification of GalpA residues in the backbone was also evident. Up to two methyl groups were detected on oligogalacturonates of DP 8 and up to three on oligogalacturonates of DP 9. Besides backbone methyl-esterification, the possible single methyl-esterification and single or double methyl-etherification of uronic acids in side chain A, recently shown by Pabst et al. (2013), was confirmed in the present study. Calcium was shown to reinforce RGII dimerization initiated by boron; calcium release promotes hydrolysis of RGII dimers (Fleischer et al., 1999; Kobayashi et al., 1999). Variability in the methylation status of side chain A and backbone could affect RGII dimer stabilization with calcium.

RGII is a very complex structure that is overall well conserved over plant evolution although some structural differences were observed from one land plant to another (Matsunaga et al., 2004). In the present study, several modifications to wine RGII structure were identified. Some of them, such as dearabinosylation and deacetylation, were the consequence of acid treatment. Others, such as methyl-esterification, methyl-etherification and oxidation, reflect a natural diversity. A range of RGII structures exhibiting specific physicochemical properties such as hydrophobicity and charge density were shown to co-exist (this work; Pabst et al., 2013) and to be organ-specific and developmentally regulated in Arabidopsis thaliana (Pabst et al., 2013). The physiological significance of this variability remains, however, to be investigated.

## ACKNOWLEDGEMENTS

This work was supported by the European Union Seventh Framework Programme (FP7 2007–2013) under Grant Agreement no. 263916. This article reflects the author's views only. The European Community is not liable for any use that may be made of the information contained herein.

## LITERATURE CITED

Bonilla
I
Mergold-Villasenor
C
Campos
ME
et al.
The aberrant cell walls of boron-deficient bean root nodules have no covalently bound hydroxyproline-/proline-rich proteins
Plant Physiology

1997
115
1329
1340
Bonin
CP
Potter
I
Vanzin
GF
Reiter
WD
The MUR1 gene of Arabidopsis thaliana encodes an isoform of GDP-d-mannose-4,6-dehydratase, catalyzing the first step in the de novo synthesis of GDP-l-fucose
Proceedings of the National Academy of Sciences of the United States of America

1997
94
2085
2090
Caffall
KH
Mohnen
D
The structure, function, and biosynthesis of plant cell wall pectic polysaccharides
Carbohydrate Research

2009
344
1879
1900
Darvill
AG
McNeil
M
Albersheim
P
Structure of plant cell walls: VIII. A new pectic polysaccharide
Plant Physiology

1978
62
418
422
Doco
T
Brillouet
J-M
Isolation and characterisation of a rhamnogalacturonan II from red wine
Carbohydrate Research

1993
243
333
343
Doco
T
Williams
P
Vidal
S
Pellerin
P
Rhamnogalacturonan II, a dominant polysaccharide in juices produced by enzymic liquefaction of fruits and vegetables
Carbohydrate Research

1997
297
181
186
Doco
T
O'Neill
MA
Pellerin
P
Determination of the neutral and acidic glycosyl-residue compositions of plant polysaccharides by GC-EI-MS analysis of the trimethylsilyl methyl glycoside derivatives
Carbohydrate Polymers

2001
46
249
259
Edashige
Y
Ishii
T
Rhamnogalacturonan II from cell walls of Cryptomeria japonica
Phytochemistry

1998
49
681
690
Fleischer
A
O'Neill
MA
Ehwald
R
The pore size of non-graminaceous plant cell walls is rapidly decreased by borate ester cross-linking of the pectic polysaccharide rhamnogalacturonan II
Plant Physiology

1999
121
829
838
Gilbert
L
Alhagdow
M
Nunes-Nesi
A
et al.
GDP-d-mannose 3,5-epimerase (GME) plays a key role at the intersection of ascorbate and non-cellulosic cell-wall biosynthesis in tomato
The Plant Journal

2009
60
499
508
Glushka
JN
Terrell
M
York
WS
et al.
Primary structure of the 2-O-methyl-α-l-fucose-containing side chain of the pectic polysaccharide, rhamnogalacturonan II
Carbohydrate Research

2003
338
341
352
González-Fontes
A
Navarro-Gochicoa
MT
Camacho-Cristóbal
JJ
Herrera-Rodríguez
MB
Quiles-Pando
C
Rexach
J
Is Ca2+ involved in the signal transduction pathway of boron deficiency? New hypotheses for sensing boron deprivation
Plant Science

2008
217–218
135
139
Ishii
T
Matsunaga
T
Pectic polysaccharide rhamnogalacturonan II is covalently linked to homogalacturonan
Phytochemistry

2001
57
969
974
Ishii
T
Matsunaga
T
Hayashi
N
Formation of rhamnogalacturonan II-borate dimer in pectin determines cell wall thickness of pumpkin tissue
Plant Physiology

2001
126
1698
1705
Ishii
T
Ono
H
NMR spectroscopic analysis of the borate diol esters of methyl apiofuranosides
Carbohydrate Research

1999
321
257
260
Kaneko
S
Ishii
T
Matsunaga
T
A boron-rhamnogalacturonan-II complex from bamboo shoot cell walls
Phytochemistry

1997
44
243
248
Kobayashi
M
Matoh
T
Azuma
J
Two chains of rhamnogalacturonan II are cross-linked by borate-diol ester bonds in higher plant cell walls
Plant Physiology

1996
110
1017
1020
Kobayashi
M
Ohno
K
Matoh
T
Boron nutrition of cultured tobacco BY-2 Cells. II. Characterization of the boron-polysaccharide complex
Plant & Cell Physiology

1997
38
676
683
Kobayashi
M
Nakagawa
H
Asaka
T
Matoh
T
Borate-rhamnogalacturonan II bonding reinforced by Ca2+ retains pectic polysaccharides in higher-plant cell walls
Plant Physiology

1999
119
199
204
Matoh
T
Kobayashi
M
Boron and calcium, essential inorganic constituents of pectic polysaccharides in higher plant cell walls
Journal of Plant Research

1998
111
179
190
Matoh
T
Kawaguchi
S
Kobayashi
M
Ubiquity of a borate-rhamnogalacturonan II complex in the cell walls of higher plants
Plant & Cell Physiology

1996
37
636
640
Matoh
T
Takasaki
M
Takabe
K
Kobayashi
M
Immunocytochemistry of rhamnogalacturonan II in cell walls of higher plants
Plant and Cell Physiology

1998
5
9
Matoh
T
Takasaki
M
Kobayashi
M
Takabe
K
Boron nutrition of cultured tobacco BY-2 cells. III. Characterization of the boron-rhamnogalacturonan II complex in cells acclimated to low levels of boron
Plant & Cell Physiology

2000
41
363
666
Matsunaga
T
Ishii
T
Matsumoto
S
et al.
Occurrence of the primary cell wall polysaccharide rhamnogalacturonan II in pteridophytes, lycophytes, and bryophytes. Implications for the evolution of vascular plants
Plant Physiology

2004
134
339
351
Melton
LD
McNeil
M
Darvill
AG
Albersheim
P
Dell
A
Structural characterization of oligosaccharides isolated from the pectic polysaccharide rhamnogalacturonan II
Carbohydrate Research

1986
146
279
305
Mohnen
D
Pectin structure and biosynthesis
Current Opinion in Plant Biology

2008
11
266
277
O'Neill
M
Albersheim
P
Darvill
AG
Dey
EP
The pectic polysaccharides of primary cell walls
Methods in Plant Biochemistry

1990
London
415
441
O'Neill
MA
Warrenfeltz
D
Kates
K
et al.
Rhamnogalacturonan-II, a pectic polysaccharide in the walls of growing plant cell, forms a dimer that is covalently cross-linked by a borate ester. In vitro conditions for the formation and hydrolysis of the dimer
The Journal of Biological Chemistry

1996
271
22923
22930
O'Neill
MA
Eberhard
S
Albersheim
P
Darvill
AG
Requirement of borate cross-linking of cell wall rhamnogalacturonan II for Arabidopsis growth
Science

2001
294
846
849
O'Neill
MA
Ishii
T
Albersheim
P
Darvill
AG
Rhamnogalacturonan II: structure and function of a borate cross-linked cell wall pectic polysaccharide
Annual Review of Plant Biology

2004
55
109
139
Pabst
M
Fischl
RM
Brecker
L
et al.
Rhamnogalacturonan II structure shows variation in the side chains monosaccharide composition and methylation status within and across different plant species
Plant Journal for Cell and Molecular Biology

2013
76
61
72
Pellerin
P
Doco
T
Vidal
S
Williams
P
Brillouet
J-M
O'Neill
MA
Structural characterization of red wine rhamnogalacturonan II
Carbohydrate Research

1996
290
183
197
Pérez
S
Rodríguez-Carvajal
MA
Doco
T
A complex plant cell wall polysaccharide: rhamnogalacturonan II. A structure in quest of a function
Biochimie

2003
85
109
121
Puvanesarajah
V
Darvill
AG
Albersheim
P
Structural characterization of two oligosaccharide fragments formed by the selective cleavage of rhamnogalacturonan II: evidence for the anomeric configuration and attachment sites of apiose and 3-deoxy-2-heptulosaric acid
Carbohydrate Research

1991
218
211
222
Reguera
M
Abreu
I
Brewin
NJ
Bonilla
I
Bolanos
L
Borate promotes the formation of a complex between legume AGP-extensin and Rhamnogalacturonan II and enhances production of Rhizobium capsular polysaccharide during infection thread development in Pisum sativum symbiotic root nodules
Plant Cell Environment

2010
33
2112
2120
Reuhs
BL
Glenn
J
Stephens
SB
et al.
l-Galactose replaces l-fucose in the pectic polysaccharide rhamnogalacturonan II synthesized by the l-fucose-deficient mur1 Arabidopsis mutant
Planta

2004
219
147
157
Ridley
BL
O'Neill
MA
Mohnen
DA
Pectins: structure, biosynthesis, and oligogalacturonide-related signaling
Phytochemistry

2001
57
929
967
Rodríguez-Carvajal
MA
Hervé du Penhoat
C
Mazeau
K
Doco
T
Pérez
S
The three-dimensional structure of the mega-oligosaccharide rhamnogalacturonan II monomer: a combined molecular modeling and NMR investigation
Carbohydrate Research

2003
338
651
671
Ropartz
D
Bodet
PE
Przybylski
C
et al.
Performance evaluation on a wide set of matrix-assisted laser desorption ionization matrices for the detection of oligosaccharides in a high-throughput mass spectrometric screening of carbohydrate depolymerizing enzymes
Rapid Communications in Mass Spectrometry

2011
25
2059
2070
Séveno
M
Voxeur
A
Rihouey
C
et al.
Structural characterisation of the pectic polysaccharide rhamnogalacturonan II using an acidic fingerprinting methodology
Planta

2009
230
947
957
Shimokawa
T
Ishii
T
Matsunaga
T
Isolation and structural characterization of rhamnogalacturonan II-borate complex from Pinus densiflora
Journal of Wood Science

1999
45
435
439
Shin
K-S
Kiyohara
H
Matsumoto
T
H
Rhamnogalacturonan II dimers cross-linked by borate diesters from the leaves of Panax ginseng C.A. Meyer are responsible for expression of their IL-6 production enhancing activities
Carbohydrate Research

1998
307
97
106
Spellman
MW
McNeil
M
Darvill
AG
Albersheim
P
Dell
A
Characterization of a structurally complex heptasaccharide isolated from the pectic polysaccharide rhamnogalacturonan II
Carbohydrate Research

1983
122
131
153
Stevenson
TT
Darvill
AG
Albersheim
P
Structural features of the plant cell-wall polysaccharide rhamnogalacturonan-II
Carbohydrate Research

1988
182
207
226
Strasser
GR
R
Pectic substances from red beet (Beta vulgaris conditiva). Part I. Structural analysis of rhamnogalacturonan I using enzymic degradation and methylation analysis
Carbohydrate Polymers

2001
44
63
70
Thibault
JF
Automatisation du dosage des substances pectiques par la methode au méta-hydroxydiphenyle
Lebensmittel-Wissenschaft und-Technologie

1979
71
215
223
Thomas
JR
Darvill
AG
Albersheim
P
Isolation and structural characterization of the pectic polysaccharide rhamnogalacturonan II from walls of suspension-cultured rice cells
Carbohydrate Research

1989
185
261
277
Tollier
MT
Robin
JP
Adaptation de la méthode à l'orcinol sulfurique au dosage automatique des glucides neutres totaux: conditions d'application aux extraits d'origine végétale
Annales de Technologie Agricole

1979
28
1
15
Vidal
Sp
Salmon
J-M
Williams
P
Pellerin
P
Penicillium daleae, a soil fungus able to degrade rhamnogalacturonan II, a complex pectic polysaccharide
Enzyme and Microbial Technology

1999
24
283
290
Vidal
S
Doco
T
Williams
P
et al.
Structural characterization of the pectic polysaccharide rhamnogalacturonan II: evidence for the backbone location of the aceric acid-containing oligoglycosyl side chain
Carbohydrate Research

2000
326
277
294
Vidal
Sp
Williams
P
Doco
T
Moutounet
M
Pellerin
P
The polysaccharides of red wine: total fractionation and characterization
Carbohydrate Polymers

2003
54
439
447
Voxeur
A
Gilbert
L
Rihouey
C
et al.
Silencing of the GDP-D-mannose 3,5-epimerase affects the structure and cross-linking of the pectic polysaccharide rhamnogalacturonan II and plant growth in tomato
The Journal of Biological Chemistry

2011
286
8014
8020
Whitcombe
AJ
O'Neill
MA
Steffan
W
Albersheim
P
Darvill
AG
Structural characterization of the pectic polysaccharide, rhamnogalacturonan-II
Carbohydrate Research

1995
271
15
29
Willats
WGT
McCartney
L
Mackie
W
Knox
JP
Pectin: cell biology and prospects for functional analysis
Plant Molecular Biology

2001
47
9
27
Yapo
BM
Pectin rhamnogalacturonan II: On the “Small Stem with Four Branches” in the primary cell walls of plants
International Journal of Carbohydrate Chemistry

2011
2011
1
11
Yapo
BM
Lerouge
P
Thibault
J-F
Ralet
M-C
Pectins from citrus peel cell walls contain homogalacturonans homogenous with respect to molar mass, rhamnogalacturonan I and rhamnogalacturonan II
Carbohydrate Polymers

2007
69
426
435
York
WS
Darvill
AG
McNeil
M
Albersheim
P
3-Deoxy-d-manno-2-octulosonic acid (KDO) is a component of rhamnogalacturonan II, a pectic polysaccharide in the primary cell walls of plants
Carbohydrate Research

1985
138
109
126