Side chains of pectic polysaccharides are regulated in relation to cell proliferation and cell differentiation


*For correspondence (fax +44 113 2333144; e-mail


The occurrence and function of the side chains occurring in the rhamnogalacturonan I domain of pectic poly- saccharides have been investigated during carrot cell development using monoclonal antibodies to defined epitopes of (1→4)-β-D-galactan and (1→5)-α-L-arabinan. Immunolocalization studies of carrot root apices indicated that cell walls in the central region of the meristem contained higher levels of (1→5)-α-arabinan than the cell walls of surrounding cells. In contrast (1→4)-β-galactan was absent from the cell walls of the central meristematic cells but appeared abundantly at a certain point during root cap cell differentiation and also appeared in cell walls of differentiating stele and cortical cells. This developmental pattern of epitope occurrence was also reflected in a suspension-cultured carrot cell line that can be induced to switch from proliferation to elongation by altered culture conditions. (1→4)-β-galactan occurred at a low level in cell walls of proliferating cells but accumulated rapidly in cell walls following induction, before any visible cell elongation, while (1→5)-α-arabinan was present in cell walls of proliferating cells but was absent from cell walls of elongated cells. Immunochemical assays of the cultured cells confirmed the early appearance of (1→4)-β-galactan during the switch from cell proliferation to cell elongation. Anion-exchange chromatography confirmed that (1→4)-β-galactan was attached to acidic pectic domains and also indicated that it was separate from a distinct homogalacturonan-rich component. These results indicate that the neutral components of pectic polysaccharides may have important roles in plant cell development.


Cell walls are a distinctive and specialized feature of plant cells and are involved in many plant-specific mechanisms for growth and cell development. Recent work in a variety of systems has indicated that cell walls are not merely a structural component of plants but may also retain and confer information important for the direction of cell fate ( Berger et al. 1994 ; McCabe et al. 1997 ; Pennell 1998; Reinhardt et al. 1998 ).

The pectic polysaccharides are a class of structurally complex polysaccharides that occur in the matrix of the primary cell walls of all land plants. One aspect of their complexity is the occurrence of several distinct structural domains that may or may not be covalently linked together within the cell wall matrix. These domains include homogalacturonan (HG), rhamnogalacturonans I and II (RGI and RGII) and xylogalacturonan ( Albersheim et al. 1996 ; Lerouge et al. 1993 ; Mohnen 1999; O’Neill et al. 1990 ; O’Neill et al. 1996 ; Schols et al. 1995 ). RGI has a backbone of alternating rhamnose and galacturonic acid residues with between 20 and 80% of the rhamnose residues substituted with side chains that are predominantly (1→4)-β-galactan, (1→5)-α-arabinan and/or arabinogalactans ( Albersheim et al. 1996 ; Guillon & Thibault 1989; Lerouge et al. 1993 ; Ros et al. 1996 ). The nature of the links between pectic domains and their links with other cell wall molecules, and the arrangement and precise structures of their side chains are far from clear. The reasons for such complexity of the pectic network are unknown ( Jarvis 1984) but are likely to relate to its role in maintaining a dynamic operating environment in the primary cell wall matrix: pectic polysacharides are implicated in regulation of cell wall ionic status, cell wall porosity, cell-to-cell adhesion and cell expansion. Oligosaccharide fragments, of HG in particular, are thought to have roles in signalling, development and defence ( Aldington & Fry 1993; Darvill et al. 1992 ).

Due to the structural complexity and multi-functionality of pectic polysaccharides, it is difficult to relate defined structures to precise functions. Composition and linkage analyses have provided the most valuable information concerning pectic polysaccharide structure, but such an approach does not provide information on possible modulations associated with cell development or on pectin structure in relation to the architecture of individual cell walls. In view of this, antibodies to defined pectic antigens and epitopes are important probes for the study of the function, organization and architecture of plant cell walls in a developmental context ( Freshour et al. 1996 ; Knox 1997; McCabe et al. 1997 ; Willats et al. 1999 ; Williams et al. 1996 :). This report describes the use of two precisely defined monoclonal antibodies to (1→4)-β-galactan and (1→5)-α-arabinan to demonstrate that the side chains of pectic polysaccharides are modulated in relation to cell proliferation and cell differentiation.


Specificity of anti-(1→4)-β-galactan and anti-(1→5)-α-arabinan in carrot root and cell systems

Two monoclonal antibodies have been generated, using neoglycoproteins, to defined epitopes occurring in the side chains of the RGI component of pectic polysaccharides. LM5 is a monoclonal antibody that recognizes four residues in (1→4)-β- d-galactan and LM6 is a monoclonal antibody that recognizes five residues in (1→5)-α- l-arabinan ( Jones et al. 1997 ; Willats et al. 1998 ). The specific recognition of the appropriate neoglycoproteins by these antibodies in an immuno-dot-assay is shown in Fig. 1(a). It has previously been shown by immuno-dot-assays that LM5 binds to a galactan-rich pectin from lupin ( Jones et al. 1997 ) and LM6 binds to an arabinan-rich pectin from sugar beet ( Willats et al. 1998 ), and that both antibodies bind to citrus pectic polysaccharides ( Jones et al. 1997 ; Willats et al. 1998 ). However, LM5 and LM6 do not recognize a sample of HG that is recognized by the anti-pectin monoclonal antibody JIM5 ( Knox et al. 1990 ) as shown in Fig. 1(a).

Figure 1.

Specificities of anti-(1→4)-β-galactan (LM5) and anti-(1→5)-α-arabinan (LM6).

(a) Immuno-dot-assay of monoclonal antibody (LM5, LM6, JIM5) recognition of (1→4)-β-galacto-tetrasaccharide–BSA and (1→5)-α-arabino-heptasaccharide–BSA neoglycoproteins and HG.

(b) Immuno-blots of carrot root apices (Roots), carrot suspension-cultured cells in a proliferative state (+2,4-D) and 70 h following removal of 2,4-D (−2,4-D (70 h)), (1→4)-β-galacto-tetrasaccharide–BSA ((1→4)-β-Gal.) and (1→5)-α-arabino-heptasaccharide–BSA ((1→5)-α-Ara.). Blots were probed with LM5, LM6, LM2 (anti-arabinogalactan-protein), LM1 (anti-hydroxyproline-rich glycoprotein) and JIM5 (anti-HG), or stained with the total protein stain Protogold. Loading levels were 15 μg protein per lane for plant material samples and 5 μg protein per lane for the neoglycoproteins. The tops of the stacking and resolving gels are indicated by S and R, respectively.

This report describes the occurrence of the (1→4)-β- d-galactan and (1→5)-α- l-arabinan epitopes in carrot cell walls. To show that the epitopes of LM5 and LM6 are not also present in glycoproteins or proteoglycans in the carrot system, the antibodies were used to probe immunoblots. Figure 1(b) shows immunoblots of preparations of material from carrot root apices, carrot suspension-cultured cells and the neoglycoprotein immunogens. The approximately equivalent loading levels of the plant material were confirmed by staining with the general protein stain Protogold. LM5 and LM6 did not bind to any components in the root or cell culture material that entered the gel, but did bind to the appropriate neoglycoproteins, although the major material recognized by LM6 did not migrate fully into the gel, possibly due to alteration of the neoglycoprotein during processing for SDS–PAGE. Equivalent blots were also probed with the anti-arabinogalactan-protein monoclonal antibody LM2 ( Smallwood et al. 1996 ) and the anti-hydroxyproline-rich glycoprotein monoclonal antibody LM1 ( Smallwood et al. 1995 ), both of which bound material that entered the gels. The anti-HG monoclonal antibody JIM5 did not bind to any material on the immunoblots. To confirm that pectic polysaccharides were extracted during preparation for SDS–PAGE, and that pectic epitopes were not destroyed during this processing, aliquots of all the preparations used for SDS–PAGE (but without bromophenol blue marker dye) were applied directly onto nitrocellulose, and, after drying, probed with all the antibodies described above, all of which bound to the applied material (data not shown). These observations demonstrate that LM5 and LM6 do not cross-react with carrot glycoproteins or proteoglycans, and in the light of the epitopes recognized and the observation that the antibodies bind to pectic polysaccharides in other systems, we are confident that LM5 and LM6 are specific to pectic polysaccharides in the carrot systems used in this study.

(1→4)-β-galactan and (1→5)-α-arabinan are developmentally regulated at the carrot root apex

Indirect immunofluorescent labelling of medial longitudinal cryosections of root apices of carrot seedlings indicated that the LM5 (1→4)-β-galactan and LM6 (1→5)-α-arabinan epitopes were developmentally regulated as shown in Fig. 2. (1→4)-β-galactan was absent from the walls of cells in the centre of the meristem and also from root cap cells adjacent and proximal to the central meristematic region and the developing epidermis. (1→4)-β-galactan occurred abundantly in the cell walls of the distal root cap cells and occurred in tissues proximal to the centre of the meristem, appearing first in the emerging vascular cylinder and cortex as shown in Fig. 2(a), and occurred ultimately in all cell walls of the developing root body prior to elongation (data not shown). In contrast, (1→5)-α-arabinan was most abundant in the central cells of the meristem as shown in Fig. 2(b). (1→5)-α-arabinan occurred to a lesser extent in the cell walls of the files of emerging cortical cells, at a low level in cell walls of the vascular cylinder and was absent from epidermal cells and the root cap ( Fig. 2b). (1→4)-β-galactan and (1→5)-α-arabinan were therefore both developmentally regulated and their patterns of occurrence were to some extent mutually exclusive. This is in distinct contrast to the distribution of HG which occurs in all cell walls throughout the carrot root apex ( Knox et al. 1990 ).

Figure 2.

Indirect immunofluorescent labelling of (1→4)-β-galactan and (1→5)-α-arabinan at the carrot root apex.

(a) Longitudinal section of a carrot root apex probed with anti-(1→4)-β-galactan LM5 showing strong labelling of cells in distal root cap (rc) and in files of emerging cortical (c) and stele (s) cells. The centre of the meristem is indicated by an asterisk.

(b) Longitudinal section of carrot root apex probed with anti(1→5)-α-arabinan LM6 showing strong labelling in the meristem (*) and weaker labelling in files of emerging cortical cells (c) and stele cells (s). Scale bars = 100 μm.

The occurrence of (1→4)-β-galactan and (1→5)-α-arabinan in a suspension-cultured carrot cell line is modulated in response to a developmental switch

The labelling of sections of carrot root apices provided information on the location of (1→4)-β-galactan and (1→5)-α-arabinan, but not on their relationships to developmental changes nor the structural relationships between the antigens bearing the epitopes. Since it is not possible to excise discrete cell populations from root apices in sufficiently large amounts for biochemical analysis, we studied the occurrence of these epitopes in a suspension-cultured carrot cell line which can be manipulated to undergo changes that reflect some aspects of cell development in planta. In the Oxford/L2 line of suspension-cultured carrot cells, proliferation as small clusters of non-vacuolated cells can be maintained by the presence of 2,4-D, but when sub-cultured at low cell density in medium without 2,4-D, a high proportion of the cells elongate dramatically within 3 days ( Lloyd et al. 1980 ; McCann et al. 1993 ; Willats & Knox 1996). This provides a readily manipulated system with which to study aspects of a switch from cell proliferation to cell elongation.

The unadhered surfaces of cells in intact cell clusters or of single intact suspension-cultured cells may be probed with monoclonal antibodies using an immersion immunofluorescence approach. Cell populations immunolabelled when in a proliferative state (+2,4-D) and at 70 h after being induced to elongate by the removal of 2,4-D are shown in Fig. 3. (1→4)-β-galactan was absent or occurred at low levels on the surface of cells/cell clusters maintained in a proliferative state as shown in Fig. 3(a), but was highly abundant at the surface of 70 h post-induction elongated cells as shown in Fig. 3(f). (1→4)-β-galactan was detectable at the outer wall surface of cells at all stages of elongation, including early stages, at which the length ratio of the longest to shorts axis was approximately 2 : 1 (data not shown), but in the small number of cells that remained non-elongated even 70 h post-induction, the epitope was absent or present at low levels ( Fig. 3f). The occurrence of (1→5)-α-arabinan was more variable. In a subset of proliferative cells, (1→5)-α-arabinan was detectable at the cell wall surface as shown in Fig. 3(b). Moreover, a subset of the cells that had remained non-elongated at 70 h post-induction were also labelled as shown in Fig. 3(g), but the elongated cells that formed the majority of the post-induction cell population were not labelled. The HG epitope recognized by JIM5 was abundant at the surface of all cells, both in an proliferative state as shown in Fig. 3(c), and at 70 h post-induction as shown in Fig. 3(h). Control cells that were probed with only secondary FITC-conjugated antibody were not labelled, as shown in Fig. 3(d,i).

Figure 3.

Immunofluorescent labelling of carrot suspension-cultured carrot cells.

Whole, fixed suspension-cultured cells in a proliferative state (+2,4-D) (a–d) and 70 h following removal of 2,4-D (f–i) were labelled with anti-(1→4)-β-galactan LM5 (a,f), anti-(1→5)-α-arabinan LM6 (b,g) and anti-HG JIM5 (c,h). Control cells (d,i) were labelled with secondary antibody only. Scale bar = 100 μm. Resin-embedded carrot suspension-cultured cells in a proliferative state (e) and 70 h following the removal of 2,4-D (j) were sectioned and labelled with LM5. Scale bar = 100 μm.

On resin-embedded sections of proliferative cells, (1→4)-β-galactan was absent from, or occurred at very low levels in, the cell walls but had a punctate distribution in the interior of cells as shown in Fig. 3(e), possibly resulting from labelling of the Golgi apparatus. On sections of cells 70 h post-induction, the (1→4)-β-galactan was abundant in the cell walls, and the interiors of cells were unlabelled as shown in Fig. 3(j). These results indicate that the differences in labelling observed in the immersion labelling experiments were not due merely to differences in permeability to the antibody probes between proliferative and elongating cells but reflected the abundance of the epitopes throughout the cell walls of the suspension-cultured cells. This was explored further by immunogold electron microscopy of 70 h post-induction cells as shown in Fig. 4. At the adhered interface of elongated and non-elongated cells, (1→4)-β-galactan was present in cell walls of elongating cells but absent from the walls of non-elongated cells as shown in Fig. 4(a,b). At this magnification, the inner region of the cell wall of elongated cells appeared as a less dense, more dispersed fibrous structure when compared with the cell wall of a non-elongated cell ( Fig. 4b). In both cases, the outer region of the cell wall had a highly dispersed fibrous structure ( Fig. 4a,b).

Figure 4.

Immunogold electron microscopy of carrot suspension-cultured cells probed with anti-(1→4)-β-galactan LM5 70 h after induction to elongate following the removal of 2,4-D.

Where the cell walls of elongated and non-elongated cells were adhered, immunogold particles were present in the cell wall of elongated cells (e) but absent from cell walls of non-elongated cells (ne). The cell wall of elongated cells appeared more diffuse than that of non-elongating cells. Immunolabelling of de-esterified sections by anti-HG JIM5 (data not shown) indicated that the unstained region between the cell wall and the plasma membrane is artifactual space. Scale bar = 200 nm.

(1→4)-β-galactan appears rapidly after developmental switching of suspension-cultured cells and is attached to an acidic pectic component

To establish a time course for the appearance of the galactan epitope in association with the switch from cell proliferation to cell elongation in the carrot cell culture system, LM5- (anti-(1→4)-β-galactan) and JIM5- (anti-HG) reactive material was quantified using ELISAs. Extracts of homogenates of the suspension-cultured cells isolated at various time points subsequent to removal of 2,4-D, standardized to equivalent levels of protein, were used as immobilized antigens. A representative binding profile of LM5 and JIM5 is shown in Fig. 5(a), which indicates that while the abundance of both epitopes increased from the beginning of the induction period, the abundance of the JIM5 epitope increased in an approximately linear fashion over the first 50 h after the removal of 2,4-D, while the abundance of the LM5 epitope increased exponentially over this period.

Figure 5.

Immunochemistry of the appearance of (1→4)-β-galactan in cells induced to switch from proliferation to elongation.

(a) ELISA of the appearance of pectic epitopes in carrot cell suspension cultured cells following the removal of 2,4-D from the culture. Cells were removed from culture at the time points indicated, homogenized, and supernatants coated onto microtitre plates prior to antibody capture assays with anti-HG JIM5 and anti-(1→4)-β-galactan LM5. In all cases, standard errors of the mean were less than 0.05 absorbance units. (b) Immunoprofiling of pectic polysaccharides from carrot suspension-cultured cells. Samples of lupin galactan (LG), HG and supernatant from a homogenate of suspension-cultured carrot cells (70 h after the removal of 2,4-D) were applied to nitrocellulose and probed with anti-HG JIM5 and anti-(1→4)-β-galactan LM5. (c) Immuno-dot/immunoprofiling analyses of the appearance of pectic epitopes in carrot cell suspension cultured cells following the removal of 2,4-D from the culture. Cells were removed from culture at the indicated time points, homogenized, and supernatants applied to nitrocellulose. Blots were probed with anti-HG JIM5, anti-(1→4)-β-galactan LM5, anti-(1→5)-α-arabinan LM6 and anti-arabinogalactan-protein LM2.

We also analysed the pectic polysaccharides in these cultures using the recently developed technique of immuno-profiling in which immuno-reactive components occurring in preparations of pectic polysaccharides can be resolved on the basis of their differing mobilities on nitrocellulose membranes ( Willats & Knox 1999). Smaller and/or less branched pectic components are relatively mobile and migrate further from the point of application than larger and/or more branched components before attaching to nitrocellulose during drying. The differing mobilities of pectic components when resolved in this way are shown in Fig. 5(b), in which HG resolves as a single large diameter ring when probed with JIM5, while lupin galactan resolves as a small diameter dot when probed with LM5. Supernatant from a homogenate of carrot suspension-cultured cells (70 h after the removal of 2,4-D) resolved into two discrete components when probed with JIM5, indicating that the HG epitope recognized by JIM5 was present on at least two distinct pectic components. The same sample was also probed with LM5 which bound to a central dot of the same diameter as the inner ring bound by JIM5. Carrot suspension-culture time course samples equivalent to those analysed by ELISAs were immunoprofiled as shown in Fig. 5(c). The relative abundance of the JIM5 epitope in both pectic components increased over the time course. However, while the JIM5-reactive, more mobile component that formed an outer ring was present in proliferating cells and all samples taken following the removal of 2,4-D, the less mobile components (seen as the LM5-reactive inner ring) were present only in samples taken from 1 h post-induction onwards. The increase in the LM5 reactivity of the samples reflected the increase in the less mobile inner ring component recognized by JIM5. By contrast, the abundance of (1→5)-α-arabinan did not significantly alter during the time course ( Fig. 5c). Prior to blotting, all samples were adjusted with respect to protein concentration, and the even labelling of the anti-arabinogalactan protein monoclonal antibody LM2 ( Smallwood et al. 1996 ) also indicated equivalent loading of all samples as shown in Fig. 5(c). The ELISAs and immunoprofiling procedure both confirmed that the increase in (1→4)-β-galactan was a very early event, appearing within 6 h of manipulation of cell culture conditions, and well before appearance of elongated cells within the culture which occurs from around 24 h post-induction onwards.

In order to confirm that the up-regulated (1→4)-β-galactan in the carrot cell culture occurred as a side chain to an acidic pectic backbone and not as a free (1→4)-β-galactan, CDTA-soluble material from cells 70 h following the removal of 2,4-D was applied to an anion-exchange chromatography column and eluted with a gradient of NaCl. Fractions were collected and probed using an immunodot assay. Representative results are shown in Fig. 6. Material bound by LM5, LM6 and JIM5 co-eluted over several fractions (starting at an NaCl concentration of 0.64 m), although an outer ring of only JIM5-reactive material eluted in the first of these fractions. This indicated that (1→4)-β-galactan and (1→5)-α-arabinan were attached to acidic components, likely to be RGI and HG.

Figure 6.

Anion-exchange chromatography of pectic polysaccharides isolated from carrot cell suspension cultures.

Carrot cells were removed from culture 70 h after the removal of 2,4-D from the medium, homogenized and supernatants applied to an anion-exchange column. Pectic polysaccharides were eluted with an NaCl gradient and fractions applied to nitrocellulose prior to probing with anti-HG JIM5, anti-(1→4)-β-galactan LM5 and anti-(1→5)-α-arabinan LM6.


The observations reported here present a new perspective on pectic polysaccharides, and particularly RGI, in relation to cell development. Previously, several anti-HG antibodies have been widely used and have indicated that changes in HG methyl esterification are associated with aspects of cell adhesion, cell expansion and cell development (see Knox 1997 for a review). The neutral components of pectin are known to be rich in galactose and arabinose, to occur as side chains of RGI and have been thought of as branched or ‘hairy’ regions, probably occurring in covalent association with all pectin. Carbohydrate and metabolic analyses have indicated that the turnover of pectic galactose residues is related to extension growth ( Labavitch 1981) and also that the galactose and arabinose content of pectic polysaccharides varied in relation to cell cluster size in carrot ( Kikuchi et al. 1996 ). However, few details of the occurrence or function of pectic side chains have been reported. The developmentally regulated occurrence of RGI epitopes in the root of Arabidopsis thaliana seedlings has been demonstrated, although the precise structures of the epitopes recognized were not determined ( Freshour et al. 1996 ). The LM5 (1→4)-β-galactan and the LM6 (1→-5)-α- l-arabinan epitopes are the first defined epitopes occurring in the side chains of pectic polysacharides to be immunolocalized in developing systems, and the observations reported here indicate dynamic expression patterns for the RGI pectic domains that are closely correlated with developmental events. A regulated appearance of the (1→4)-β-galactan epitope in root cap cells at the flax root apex has already been reported ( Vicréet al. 1998 ).

One of the most important observations reported here is that (1→4)-β-galactan and (1→5)-α-arabinan can have distinct and to a large extent separate locations in relation to cell development at the carrot root apex and in suspension-cultured cells. They may therefore have distinct functions. From this work, it is now also clear that there is no single structural form to RGI and these observations indicate that pectin biochemistry within a developing organ is likely to be considerably more complex than previously appreciated. Aspects of this biochemistry have been explored in the suspension-cultured carrot cell system where the dynamics of pectic galactan and pectic arabinan occurrence can be dissected temporally. The removal of 2,4-D switches cells from proliferation to elongation (which is perhaps better viewed in this context as an aspect of differentiation). Monitoring the appearance of the (1→4)-β-galactan epitope subsequent to 2,4-D removal indicated that it appeared well before visible elongation and that it occurred in association with HG and/or RGI domains, but distinct from a separate HG-rich polysaccharide that also increases in relative abundance during this developmental switch. Although (1→4)-β-galactan appeared well before obvious morphological changes to the cells within the cultures, it is not clear whether this relates most closely to the cessation of proliferation or the subsequent elongation of the cells. This connection of (1→4)-β-galactan and (1→5)-α-arabinan with the processes of cell division and cell enlargement is to some extent reflected by the occurrence of these pectic structures at the carrot root apex.

The function(s) of the pectic side chains containing these epitopes are far from clear. Such side chains are known to be much more flexible than the HG components that form the backbones of pectic polysaccharides ( Foster et al. 1996 ; Ha et al. 1996 ) and they may therefore be involved in modulating HG alignment and contributing to the regulation of the porosity of the cell wall matrix. The presence of side chains attached to RGI domains may influence the access of proteins with wall-modifying properties to sites of action within the cell wall matrix. Such proteins may act to modify pectin structure, e.g. polygalacturonases ( Hadfield & Bennett 1998) or may be proteins such as expansins and xyloglucan endotransglycosylase capable of altering the cellulose–hemicellulose network leading to cell enlargement ( Cosgrove 1997). A study of the structure of the pectic polysaccharides carrying these side chains and the enzymes that are responsible for the synthesis and any modifications of the neutral side chains of RGI will be of considerable interest.

Experimental procedures


Carrot seeds (Daucus carota L. cv. Early Nantes) were germinated on moist filter paper in the dark at 22°C for 5–10 days. A carrot (Daucus carota L.) suspension-cultured cell line known as L2 or ‘Oxford’ ( Knox et al. 1991 ; Lloyd et al. 1980 ; McCann et al. 1993 ) was cultured in 4.43 g l−1 Murashige and Skoog basal medium supplemented with 25 g l−1 sucrose and 0.5 mg l−1 2,4-dichlorophenoxyacetic acid (2,4-D) and sub-cultured by 10-fold dilution into fresh medium every 7 days. In order to induce elongation in the cell cultures, cells were washed once in medium without 2,4-D, then diluted 50-fold into medium without 2,4-D.

The preparation of the (1→4)-β-galacto-tetrasaccharide–BSA and (1→5)-α-arabino-heptasaccharide–BSA neoglycoproteins and the generation of monoclonal antibodies LM5 and LM6 are described elsewhere ( Jones et al. 1997 ; Willats et al. 1998 ). The other monoclonal antibodies used for comparison purposes were anti-HG, JIM5 ( Knox et al. 1990 ; VandenBosch et al. 1989 ), anti-arabinogalactan-protein, LM2 ( Smallwood et al. 1996 ) and antihydroxyproline-rich glycoprotein, LM1 ( Smallwood et al. 1995 ). All were rat antibodies and were used unpurified from hybridoma supernatants.

Immunolabelling procedures and microscopy

Carrot root apices were excised from seedlings 5–10 days after germination and formaldehyde-fixed for cryosectioning, immunofluorescent labelling and microscopy as described previously ( Smallwood et al. 1994 ).

For immunofluorescent immersion labelling of intact suspension-cultured cells, cells were fixed in 4% paraformaldehyde in 50 m m PIPES (1,4-piperazine-diethanesulphonic acid), 5 m m MgSO4, and 5 m m EGTA (ethylene glycol bis (β-aminoethylether)-N,N,N′,N′-tetraacetic acid) pH 6.9 overnight at 4°C. Cells were washed once in phosphate-buffered saline (PBS, 0.14 m NaCl, 2.7 m m KCl, 7.8 m m Na2HPO4·12 H2O, 1.5 m m KH2PO4, pH 7.2) prior to blocking for 1 h with PBS containing 5% fat free milk powder (MPBS). After a further wash with PBS, cells were incubated in primary antibodies (hybridoma supernatants of LM5, LM6 and JIM5, diluted 1/10 in MPBS) for 1.5 h then washed three times with PBS prior to incubation in secondary antibody (anti-rat conjugated to fluorescein isothiocyanate (anti-rat/FITC), Sigma, UK, diluted 1/100 in MPBS) for 1.5 h. Cells were washed three times in PBS before being mounted in Citifluor antifade agent (Agar Scientific) and examined on an Olympus BH-2 microscope with epifluorescence illumination. Images were recorded on 400 ASA colour slide film (Kodak).

In certain cases, cells were dehydrated in a graded ethanol series and embedded in LR White resin as described below. Sections (0.5–1 μm thick) were applied to multi-well slides coated with Vectabond (Vector Laboratories, UK) and blocked with 5% MPBS for 30 min prior to application of 25 μl per well of primary antibody for 1.5 h. Following washing with PBS, 25 μl per well of secondary antibody (anti-rat/FITC diluted 1/100 in MPBS) was applied for 1.5 h. Slides were washed with PBS and mounted and examined as described above.

For immunogold electron microscopy, supension-cultured cells were pelleted by centrifugation (100 g) and resuspended in fixative (1% glutaraldehyde, 2% paraformaldehyde in 0.1 m sodium cacodylate buffer pH 7.0) for 1 h at room temperature on a specimen rotator. After three 5 min washes in 0.1 m sodium cacodylate buffer pH 7.0, cells were resuspended in buffer containing molten 3% low melting point agar at 40°C and vortexed. Agar blocks (0.5 cm3) were processed using a progressive low temperature method ( VandenBosch 1991), consisting of an ethanol dehydration series with 20 min at each stage (10% and 20% EtOH steps were undertaken at room temperature, 30% at 4°C and all subsequent steps at −20°C) and embedding in LR White resin (London Resin Co., Reading, Berkshire, UK) containing 0.5% benzoin methy ether with regular changes of resin over 2 days. Resin was polymerized in gelatin capsules (Agar Scientific Ltd, Stansted, Essex, UK) using UV light for 12 h at −20°C and 12 h at room temperature. Ultra-thin sections (approximately 80 nm) were cut using a Reichert Ultracut 4 ultramicrotome and a Diatome diamond knife. Sections were collected on 200-mesh nickel grids coated with Formvar (Agar Scientific Ltd). Sections were blocked by flotation on droplets of 1% (w/v) bovine serum albumin (BSA) in phosphate-buffered saline containing 2% Tween 20 (PBST) and washed with PBST. Primary antibodies (LM5, LM6 or JIM5) were diluted fivefold in PBST containing 1% BSA and used at room temperature for 2 h incubations (control sections were incubated with PBST). After washing with PBST, sections were incubated for 1 h at room temperature in anti-rat IgG (whole molecule) coupled to 10 nm gold (Sigma Chemical Company, Poole, Dorset, UK) diluted 1:20 in PBST containing 1% BSA. Sections were washed with PBST followed by water before post-staining with lead citrate and 2% (w/v) uranyl acetate. Immunolabelled sections were viewed using a Jeol 1200 EX11 electron microscope operating at 80 V. Electron micrographs were recorded on Scientia EM film 23D56 (Agar Scientific Ltd).

Immuno-dot-assays and enzyme-linked immunosorbent assays

At selected time points during subculture, suspension-cultured carrot cells were removed and washed once in Tris-buffered saline (TBS, 0.14 m NaCl, 2.7 m m KCl, 24.8 m m Tris base, pH 7.2). For each time point, a packed cell volume of approximately 100 μl of cells was homogenized (in a glass hand homogenizer with Teflon plunger powered by an electric screwdriver at 200 rev min−1) in 100 μl of buffer containing 50 m m Tris–HCl pH 7.2, 50 m m trans-1,2-diaminocyclohexane-N,N,N′,N′-tetraacetic acid (CDTA) and 25 m m dithiothreitol, and centrifuged at 7000 g for 10 min on ice. The protein concentration of the resulting supernatants was determined and all samples were adjusted to a protein concentration of 1 mg ml−1.

For immuno-dot-assays, 1 μl or 5 μl aliquots of adjusted supernatants were applied to nitrocellulose membrane (Scheicher & Schuell, Dassel, Germany) and left to dry for 1 h. Samples of pea HG (obtained from Professor J.F. Thibault, INRA, Nantes, France) and lupin galactan (Megazyme, Bray, Ireland) were dissolved in de-ionized water to a concentration of 1 mg ml−1 and applied to nitrocellulose as 5 μl aliquots. Nitrocellulose membranes were blocked with MPBS prior to incubation in primary antibody (hybridoma supernatants diluted 1/10 in MPBS) for 1.5 h. After washing extensively under running tap water followed by 10 min in PBS containing 0.1% (v/v) Tween 20 (PBST), membranes were incubated in secondary antibody (anti-rat horseradish peroxidase conjugate, Sigma, Poole, UK) diluted 1/1000 in MPBS for 1.5 h. Membranes were washed as described above prior to development in substrate solution (25 ml de-ionized water, 5 ml methanol containing 10 mg ml−1 4-chloro-1-naphthol, 30 μl 6% (v/v) H2O2). In all cases experiments were performed at least three times with essentially the same results.

For enzyme-linked immunosorbent assays (ELISAs), samples were adjusted to a protein concentration of 10 μg ml−1 by the addition of sodium carbonate coating buffer (pH 9.6) before application to 96-well microtitre plates (Maxisorb, Nunc, Denmark), 75 μl per well, and incubated overnight at 4°C. Plates were blocked with 3% BSA in PBS for 2 h prior to the addition of 100 μl per well of primary antibody (serially diluted hybridoma supernatants) for 1.5 h. Plates were washed 15 times in tap water prior to addition of secondary antibody (anti-rat horseradish peroxidase conjugate, diluted 1/1000 in 3% BSA in PBS) for 1.5 h. Plates were washed 15 times with tap water and developed by the addition of substrate (18 ml of de-ionized water, 2 ml of 1 m sodium acetate buffer, pH 6, 200 μl of tetramethylbenzidene and 20 μl of 6% H2O2). Colour development was stopped by the addition of 35 μl per well of 2 N sulphuric acid and absorbances were determined at 450 nm. At least five replicates were used for each quantification and standard errors of the mean calculated.

SDS–PAGE and immunoblotting

Root apices of 10-day-old carrot seedlings and carrot suspension cultures in a proliferative state (+2,4-D) or 70 h after the removal of 2,4-D were homogenized to a fine powder in liquid nitrogen, immediately suspended in SDS–PAGE sample buffer and boiled for 5 min. Samples were vortexed for 2 min, centrifuged for 5 min at 10 000 g and aliquots of the supernatants were removed for protein determination ( Bradford 1976). (1→4)-β-galactan and (1→5)-α-arabinan neoglycoproteins were also boiled in sample buffer. All samples were separated on 1.5 mm 8% polyacrylamide gels essentially according to Laemmli (1970). Loading levels per lane were 15 μg for plant material samples and 5 μg for the neoglycoproteins. After separation, material was transferred to nitrocellulose essentially according to Towbin et al. (1979) . Nitrocellulose membranes were washed once in PBS prior to being blocked in MPBS for 1 h or stained with the total protein stain Protogold (British BioCell International, UK). Selected membranes were incubated in primary antibodies (LM5, LM6, LM1, LM2 and JIM5) diluted 1/10 in MPBS for 1.5 h. After washing for two periods of 10 min in PBS, blots were incubated in secondary antibody (anti-rat horseradish peroxidase conjugate, Sigma, Poole, UK) diluted 1/1000 in MPBS for 1.5 h. Membranes were washed as described above prior to development in substrate solution previously described for immuno-dot-binding assays.

Anion-exchange chromatography of CDTA-soluble material

Suspension-cultured carrot cells, 70 h post-induction (−2,4-D) were homogenized as described for immunochemical assays. The homogenate was centrifuged at 10 000 g for 10 min and 250 μl of supernatant (with a protein concentration of approximately 2.5 mg ml−1) was loaded onto a 5 ml anion-exchange column (Econo-Pac, Q cartridge, Bio-Rad) that had been equilibrated with 0.1 m Tris–HCl pH 7.2. The column was eluted with a gradient of 0–1.0 m NaCl over 30 min and 1 ml fractions collected. Aliquots of each fraction (2 μl) were applied to nitrocellulose and probed by the method previously described for immuno-dot-assays.


We are grateful for financial support from the EU Framework IV Biotechnology programme and the UK Biotechnology and Biological Sciences Research Council.