Developmentally regulated membrane glycoproteins sharing antigenicity with rhamnogalacturonan II are not detected in nodulated boron deficient Pisum sativum


L. Bolaños. Fax: +34 91 497 8344; e-mail:


The peribacteroid membrane (PBM) of symbiosomes from pea root nodules developed in the presence of boron (+B) was labelled by anti-rhamnogalacturonan II (RGII) (anti-rhamnogalacturonan II pectin polysaccharide) antiserum. However, in nodules from plants grown at low boron (−B), anti-RGII pectin polysaccharide did not stain PBMs. Given that RGII pectin binds to borate, and that symbiosomes differentiate aberrantly in −B nodules because of abnormal vesicle traffic, anti-RGII pectin polysaccharide antigens were further analysed. Following electrophoresis and electroblotting, anti-RGII pectin polysaccharide immunostained three bands in +B but not in −B nodule-derived PBMs. A similar banding pattern was observed after the immunostaining of membrane fractions from uninfected roots, indicating that anti-RGII pectin polysaccharide antigens are common to both peribacteroid and plasma membranes. Protease treatment of samples led to disappearance of anti-RGII pectin polysaccharide labelling, indicating that the three immunostained bands correspond to proteins or glycoproteins. The immunochemical study of RGII antigen distribution during nodule development showed that it is strongly present on the PBM of dividing (undifferentiated) symbiosomes but progressively disappeared during symbiosome maturation. In B-deficient nodules, PBMs were never decorated with RGII antigens, and there was an abnormal targeting of vesicles containing pectic polysaccharide (homogalacturanan) to cell membranes. Overall, these results indicate that RGII, boron and certain membrane (glyco)-proteins may interact closely and function cooperatively in membrane processes associated with symbiosome division and general cell growth.


Boron is essential for plant cell wall structure (Matoh 1997) by cross-linking apiose residues in the pectin polysaccharide rhamnogalacturonan II (RGII) (O'Neill et al. 2004). However, this role seems to be insufficient to explain the plethora of rapid biochemical, physiological and anatomical aberrations caused by B deficiency (for a review, see Blevins & Lukaszewski 1998; Brown et al. 2002). Furthermore, recent reports on nutritional effects of B in organisms devoid of cellulose cell walls, including animals and humans (Bennett et al. 1999; Rowe & Eckhert 1999; Nielsen 2002), have led to propose a crucial role of the micronutrient in membrane structure and function (Nielsen & Penland 1999; Brown et al. 2002).

The development of root nodules during legume-rhizobia N2-fixing symbioses (excellently reviewed by Stougaard 2000 or Limpens & Bisseling 2003, among others) is strongly affected by B deficiency (Bolaños et al. 1994, 2001; Bolaños, Brewin & Bonilla 1996; Bonilla et al. 1997) and largely driven by membrane-related functions. Following the pre-infection and infection events, bacteria are released into the cytoplasm of the host cells through an endocytosis-like phenomenon. The rhizobia are now called bacteroids, which are surrounded by a plant-derived membrane (peribacteroid membrane, PBM). They grow, divide and develop into differentiated symbiosomes where biological nitrogen fixation takes place. The PBM harbors a differentiated glycocalyx composed of glycoproteins and glycolipids (Perotto et al. 1991).

Symbiosome maturation implies gradual differentiation of the PBM involving several plant and bacterial glycoconjugates (Kannenberg & Brewin 1994). Some of them have been shown to physically associate in vitro (Bolaños et al. 2004b). Therefore, the characterization of glycoproteins and glycolipids that are developmentally regulated is important for understanding the symbiotic interaction. Moreover, investigating the glycoconjugates of the symbiosome compartment should assist in elucidating a specific role of B in membrane structure and function. Components of the PBM glycocalyx rich in cis-diols could be candidates to react with borate. This reaction is considered the key of B-dependent functions (Bolaños et al. 2004a).


Plant growth and inoculation

Pea (Pisum sativum cv. Lincoln) seeds were surface-sterilized with 70% (v/v) ethanol for 1 min and with 10% (v/v) sodium hypochlorite for 20 min, soaked for 4 h in sterile distilled water and then germinated on wet filter paper at 25 °C. After 4 d, the seedlings were transferred to plastic growth pots and cultivated on B-free perlite with FP medium for legumes (Fahraeus 1957).

For B-free cultures, B was removed from the micronutrient solution. For cultures with the normal content of B, the micronutrient (as H3BO3) was added to a final concentration of 0.1 mg B L−1. All solutions were prepared and stored in polyethylene containers previously tested not to release boron, even under sterilizing conditions. Boron was determined in the solutions and media prior to use, and was found to be below the detection limit (0.02 µg mL−1). B concentration was determined using Azomethine H at pH 5.1 (Wolf 1974) and a Technicon Automatic Analytical System (Tarry Town, New York, NY, USA). Before inoculation, the pea seedlings were grown without added B during 15 d (including germination time) to drain the B stored in the seeds.

Plants were inoculated with 1 mL per seedling of about 108 cells mL−1 of Rhizobium leguminosarum bv. viciae 3841 from an exponential culture in tryptone-yeast (TY extract) medium (Beringer 1974), and were maintained in a growth cabinet at 22 °C day/18 °C night temperatures with a 16 (light)–8 (dark) h photoperiod and an irradiance of 190 µmol m−2 s−1. Relative humidity was kept between 60 and 70%.

Antibodies and antisera

Antiserum anti-RGII pectin polysaccharide (kindly supplied by Dr. T. Matoh) recognizes RGII polysaccharide that is localized at the cell wall–membrane interface (Matoh et al. 1998). Rat monoclonal antibodies MAC 206 and JIM 5 were kindly provided by Dr. N.J. Brewin. MAC 206 recognizes carbohydrate epitopes associated with glycolipids or with a lipid-anchored mini-AGP harbored by the PBM (Perotto et al. 1991). JIM 5 reacts with pectin polygalacturonic acid (Knox et al. 1990).

Fractionation of nodules

Pea nodules were homogenized at 4 °C in Tris-dithiothreitol buffer (50 mM Tris-HCl, pH 7.5, 10 mM DTT) containing 0.5 M sucrose. The soluble fraction of the homogenate was collected by filtration through Miracloth (Calbiochem, Madrid, Spain) and was centrifuged for 30 min at 100 000g to remove the debris. The resulting pellet contained the symbiosome and bacteroid fractions. Bacteroids still enclosed by the PBM were prepared by fractionation through a sucrose gradient, and the fraction corresponding to the PBM was then obtained according to Perotto et al. (1991).

Sodium dodecyl sulphate–polyacrylamide gel electrophoresis (SDS–PAGE) and immunoblotting

Membrane samples for gel separation were extracted by heating in SDS buffer for 10 min at 100 °C. Where indicated, a treatment with Proteinase K was performed (Sindhu, Brewin & Kannenberg 1990). After the centrifugation to remove insoluble debris, the extracts (10 µg protein loaded per lane) were subjected to 12% acrylamide mini-gels (Laemmli 1970). The gels were transferred electrophoretically to membranes of nitrocellulose and were incubated with 5% bovine serum albumin (BSA) in TBS (50 mM Tris-HCl, pH 7.4, 200 mM NaCl) buffer containing an antibody. Immunostaining was visualized using a goat anti-rat IgG (for monoclonal antibodies) or a goat anti-rabbit IgG (for antiserum) peroxidase-conjugated secondary antibody (Bradley et al. 1988).

Periodate oxydation

In order to confirm that epitopes had carbohydrate components, the sensitivity to periodate oxidation (Woodward, Young & Bloodgood 1985) was tested. One microlitre of nodule fractions (1 mg of protein mL−1) was dotted on nitrocellulose sheets. After drying, they were equilibrated with 50 mM sodium acetate buffer, pH 4.5, for 30 min and were treated as previously described (Perotto et al. 1991). The sheets were incubated in the dark for 1 h in the same buffer containing 20 mM sodium metaperiodate, and were washed twice with the same buffer for 30 min and once more with TBS. Finally, the sheets were incubated for 30 min with 50 mM sodium borohydride in TBS and were immunostained as described previously.

Microscopy on nodule sections

Pea nodules developed in the presence or in the absence of B were fixed overnight at 4 °C in 2.5% (v/v) glutaraldehyde in 0.1 M sodium cacodylate buffer, pH 7.2. The nodules were dehydrated in an ethanol series and embedded in LR White resin (London Resin Company, London, UK). Ultra-thin (silver-gold in color) sections of nodules were processed for immunogold staining following transmission electron microscopy according to the methods previously described (Rae et al. 1991). A goat anti-rabbit IgG (to reveal RGII labelling) or an anti-rat IgG (to reveal JIM 5 labelling)-colloidal gold (10 nm gold particles) (Amersham, Little Chalfont, Bucks, UK) was used as a secondary antibody. Additionally, semi-thin (0.5 µm) sections were processed for immunogold staining and silver enhancement as described previously (Perotto et al. 1991). Controls (not shown) without first antibody to identify background staining, and immunostaining with other antibodies, such as MAC 265 (anti-infection thread matrix glycoprotein), MAC 206 (anti-peribacteroid and plasma membrane glycolipid) or MAC 57 (anti-R. leguminosarum 3841 LPS) previously shown to label B-deprived nodule sections (Bolaños et al. 1996) were included to demonstrate that a lack of staining could not be attributed to poor fixation of B-deficient tissues.


Ultra-thin sectioned infected cells of pea nodules (harvested 10 d post-inoculation) developed in the presence (+B) or in the absence (−B) of boron were probed with anti-RGII pectin polysaccharide antiserum raised against the cell wall pectic polysaccharide RGII (Fig. 1). Anti-RGII pectin polysaccharide labelled the symbiosomal PBM in infected cells of +B nodules (Fig. 1a) but not of −B nodules (Fig. 1b). MAC 206 (anti-PBM glycolipid or lipid-anchored mini-AGP), previously shown to label B-deficient tissues (Bolaños et al. 1996), was used as a control for tissue fixation, and the PBMs of both +B (Fig. 1c) and −B symbiosomes (Fig. 1d) were labelled.

Figure 1.

Immunogold labelling with anti-rhamnogalacturonan II (RGII) antiserum (a,b) or with MAC 206 (anti-PBM glycolipid or lipid-anchored mini-AGP) antibody (c,d) of ultra-thin sections of pea nodule infected cells harvested 10 d post-inoculation. (a,c) Nodule from a plant grown in the presence of boron, showing labelling in the PBM region; (b,d) nodule from a plant grown in the absence of boron showing the absence of anti-RGII pectin polysaccharide (b) but the presence of MAC 206 (d) labelling in the PBM. Antibody binding was detected by using anti-rabbit IgG (for anti-RGII pectin polysaccharide labelling) or anti-rat IgG (for MAC 206 labelling) conjugated to 10 nm gold particles. b, bacteroid; pbm, peribacteroid membrane; bars, 0.1 µm.

To confirm that antigens recognized by anti-RGII pectin polysaccharide on the PBM contained sugar epitopes, nodule homogenates or purified PBM fractions were dotted on nitrocellulose sheets and treated with sodium metaperiodate. As expected, anti-RGII pectin polysaccharide bound to +B or −B nodule homogenates (because of the presence of cell wall polysaccharides), and to +B but not −B PBM fractions (Fig. 2a). Figure 2b shows that antibody binding was sensitive to metaperiodate oxidation, confirming that membrane RGII antigens also had carbohydrate components. In order to further analyse the nature of antigens recognized by anti-RGII pectin polysaccharide in the symbiosome, nodule fractions corresponding to the PBM or membranes derived from uninfected tissues were separated by SDS-PAGE, electroblotted and probed with anti-RGII pectin polysaccharide (Fig. 3a). Three bands (highlighted by arrowheads) of about 50, 75 and 150 kD relative molecular weight were immunostained in +B PBM-derived fractions (Fig. 3a, lane 1). A similar immunolabelling pattern appeared in membrane fractions derived from +B pea roots (Fig. 3a, lane 2) or shoots (not shown), indicating that the occurrence of membrane antigens recognized by anti-RGII pectin polysaccharide is not restricted to the PBM. Antigenicity disappeared when membrane extracts were previously treated with protease (Fig. 3a, lane 3). As anti-RGII pectin polysaccharide was raised against sugar epitopes, this indicates that labelled membrane antigens correspond to glycoproteins (thereafter referred to as ‘RGII-glycoproteins’). Clearly, there was no anti-RGII pectin polysaccharide labelling in membrane fractions of −B nodules and roots (Fig. 3a, lane 4). Monoclonal antibody MAC 206, raised against a protease-resistant antigen, as reported by Perotto et al. (1991), was used as a control for loading (Fig. 3b), and both −B derived (Fig. 3b, lane 2) or protease-treated +B PBM samples (Fig. 3b, lane 3) were positively stained.

Figure 2.

Dot immunoassays (1 µL dotted on nitrocellulose sheets) of nodule derived fractions showing anti-rhamnogalacturonan II (RGII) binding (a) and sensitivity of binding to periodate oxidation after incubation of antigens immobilized on nitrocellulose sheets with sodium metaperiodate (b). 1, +B nodule homogenate; 2, −B nodule homogenate; 3, +B nodule-derived peribacteroid membrane (PBM) fraction; 4, −B nodule-derived PBM fraction. The absence of immunostaining in (b) indicates that antigens recognized by anti-RGII pectin polysaccharide have carbohydrate components. Antibody binding was detected by using anti-rabbit IgG conjugated to peroxidase as secondary antibody.

Figure 3.

Immunostaining of membrane fractions from +B or −B pea nodules or roots following sodium dodecyl sulphate–polyacrylamide gel electrophoresis (SDS–PAGE) and electroblotting on nitrocellulose sheets. (a) Anti-rhamnogalacturonan II (RGII) labelling (highlighted by arrowheads). Lane 1, +B peribacteroid membrane (PBM) fractions; lane 2, +B uninfected root plasma membrane fractions; lane 3, +B PBM protease-treated fractions, indicating that stained bands are glycoproteins; lane 4, −B PBM fractions. (b) MAC206 (anti-peribacteroid and plasma membrane glycolipid) staining used as control for loading. Lane 1, +B PBM fractions; lane 2, −B PBM fractions; lane 3, +B PBM protease-treated fractions. Antibody binding was detected by using anti-rabbit IgG (for anti-RGII pectin polysaccharide labelling) or anti-rat IgG (for MAC 206 labelling) conjugated to peroxidase as secondary antibody.

Because B deficiency led to an abnormal targeting of vesicles containing peribacteroid components and to aberrant symbiosome development, we studied whether glycoproteins sharing antigenicity with RGII were developmentally regulated in pea nodules. Immunogold labelling followed by silver enhancement of semi-thin nodule sections 8 days after inoculation with Rhizobium revealed that besides cell walls, all infected cells containing still-young non-N2-fixing symbiosomes were labelled by anti-RGII pectin polysaccharide (Fig. 4a). Figure 4b shows the absence of immunostaining inside the cells of −B 8-day-old nodule sections. In mature 21-day-old +B pea nodules (Fig. 4c), the heaviest anti-RGII pectin polysaccharide labelling corresponded to zones II and III, with the most intense membrane synthesis for enclosing dividing bacteroids. Labelling became weak in infected cells from zone IV where N2 fixation takes place, as revealed by the labelling of serial sections with anti-nitrogenase antiserum (Fig. 4d). Anti-RGII pectin polysaccharide immunostaining in −B 21-day-old nodules was restricted to the cell wall regions (Fig. 4e).

Figure 4.

Immunogold labelling and silver enhancement on longitudinal semi-thin sections of pea nodules harvested at different stages of development. (a) Anti-rhamnogalacturonan II (RGII) staining of +B nodule 8 d post-inoculation, showing the labelling of cell walls and all infected cells [as Fig. 1 shows, labelling is associated with the peribacteroid membrane (PBM) of symbiosomes]. (b) Anti-RGII pectin polysaccharide staining of −B nodule 8 d post-inoculation, showing the labelling in the cell wall regions and the absence of labelling inside the cells. (c) Anti-RGII pectin polysaccharide staining of +B nodule 21 d post-inoculation, showing heavy labelling in cells from zone II and interzone II–III, becoming weak and finally disappearing in zones III and IV. (d) Anti-nitrogenase staining of a serial section (0.5 µm deeper) of +B nodule in (c). (e) Anti-RGII pectin polysaccharide staining of −B nodule 21 d post-inoculation, showing the absence of labelling inside the cells. Antibody binding was detected by using gold-conjugated (10 nm) anti-rabbit IgG as secondary antibody. Ellipse in (c) and (d) highlights cells heavily stained with anti-RGII pectin polysaccharide. I, nodule meristem; II, invasion zone, characterized by an intense bacteroid division; II–III, interzone of bacteroid differentiantion, previous starting nitrogen fixation; III, nitrogen fixation zone; IV, senescence zone; m, nodule meristem; iz, infected zone; c, nodule cortex; vb, vascular bundle. Bar markers: (a–d) 0.5 mm, (e) 0.1 mm.

Because anti-RGII pectin polysaccharide glycoproteins mainly appeared in PBMs of nodule cells that require intense vesicle traffic, and also in plasma membranes, we investigated a possible link between the occurrences of such glycoproteins in membranes and traffic/targeting of vesicle containing cell wall components. Figure 4 shows that anti-RGII pectin polysaccharide apparently stained both +B and −B nodule cell walls. However, following the immunocytochemistry of ultra-thin sections of +B nodule cells (Fig. 5a), anti-RGII pectin polysaccharide labelling was observed throughout the cell wall, being denser in the zone near to the plasma membrane. Vesicles containing anti-RGII pectin polysaccharide labelling (arrowheads) frequently appear targeted to plasma membrane regions. By contrast, anti-RGII pectin polysaccharide staining of −B nodule cells (Fig. 5b) was mostly concentrated near the plasma membrane instead of inside the cell wall, where gold labelling was almost absent (highlighted by double arrowheads). Other wall polysaccharides, as homogalacturonan recognized by JIM 5 monoclonal antibody (Fig. 5c,e), frequently appeared accumulated in the cytosol and were abnormally distributed in the cell wall of −B nodules (Fig. 5d,f), indicating the abnormal trafficking and/or targeting of cell wall polysaccharides containing vesicles under B deficiency.

Figure 5.

Immunolocalization of pectic polysaccharides in +B (a,c,e) or −B (b,d,f) pea nodule sections 21 d post-inoculation. (a,b) Electron micrographs of ultra-thin nodule sections showing the presence of anti-rhamnogalacturonan II (RGII) antigens in vesicles (arrowheads), plasma membrane (pm) or the cell wall (cw) (highlighted by double arrowheads in −B nodule cell wall) regions. (c,d) Immunogold and silver enhancement of semi-thin or immunogolg of ultra-thin (e,f) nodule sections probed with JIM 5 (anti-homogalacturonan) monoclonal antibody showing abnormal wall distribution and accumulation of labelling in the cytosol (cyt) of −B nodule cells. Antibody binding was detected by using anti-rabbit IgG (for anti-RGII pectin polysaccharide labelling) or anti-rat IgG (for JIM 5 labelling) conjugated to 10 nm gold particles. Bar markers: (a,b) 0.2 µm; (c,d) 0.1 mm; (e,f) 0.1 µm. c, nodule cortex; iz, infected zone.


Although the mechanism underlying possible functions of B in membrane-bound processes still has to be elucidated, it is apparent that any B function requires the presence of specific ligand molecules able to form complexes with boric acid/borate (Bolaños et al. 2004a), as apiose cross-linking during cell wall stabilization (O'Neill et al. 2004). Sugar moieties such as mannose, apiose or galactose, as well other hydroxylated ligands such as serine or threonine, may form ester-like complexes with boron (Ralston & Hunt 2000). In membranes, glycoproteins and glycolipids are good candidates for a possible boron function, and in nodules, maturation of the PBM during symbiosome development involves several glycoproteins and glycolipids (Perotto et al. 1991, 1995). Therefore, the role of B on nodule organogenesis and tissue differentiation could be related with the glycocalyx of the PBM.

Anti-RGII pectin polysaccharide is a polyclonal antibody primarily raised against a borate-RG-II complex (Matoh et al. 1998). This antibody was first described to recognize sugar epitopes on RG-II, either in a monomer or in a borate-dimer form, exclusively in cell wall polysaccharides. Immunocytochemistry studies on pea nodules allowed the detection of anti-RGII pectin polysaccharide labelling not only in cell walls (Fig. 5a,b) but also in membranes such as the PBMs (Fig. 1). Sensitivity to metaperiodate oxidation indicated that antigens recognized by anti-RGII pectin polysaccharide in membranes also have sugar components (Fig. 2). Protease treatment and immunoelectroforesis (Fig. 3) confirmed that there are at least three PBM glycoproteins reactive with anti-RGII pectin polysaccharide antigens. These antigens can also be detected in membranes of uninfected root cells (Fig. 3b, lane 2).

During nodule formation and growth, anti-RGII pectin polysaccharide antigens appeared to be developmentally regulated. Whereas other components common to both the PBM and the plasma membrane are present throughout the symbiosome development (Perotto et al. 1991), RGII glycoproteins appeared mainly associated with the PBMs surrounding dividing (undifferentiated) bacteroids that proliferate in young pea nodules (Fig. 4a). In mature nodules, anti-RGII pectin polysaccharide labelling was observed in zone II and interzone II–III that still contain undifferentiated symbiosomes (Fig. 4c). As judged after immunolocalization of nitrogenase in serial sections, RGII immunostaining became weak and almost disappeared in mature symbiosomes that actively fix N2 (zone III) or in those from the senescent zone IV (Fig. 4d). Other antigens, including an interesting glycosyl inositol phopholipid, have also been shown to disappear from the PBM at maturation (Perotto et al. 1995). This suggests that the disappearance of antigens could be a signal that triggers the transition from a plant plasma membrane to the specialized PBM. This assumption is substantiated by the fact that RGII glycoproteins were neither detected in −B pea nodules (Fig. 4b,e), which do not differentiate into functional bacteroids, nor in other −B plant tissues (Fig. 3a, lane 4).

Considering that membrane RGII antigens are common to plasmalemma and immature PBMs, as well as the known consequences of B deficiency on plant cell growth, leads to the hypothesis that anti-RGII pectin polysaccharide reactive membrane-bound glycoproteins are required at a certain stage of the growing process, and that they progressively disappear when either cell enlargement or symbiosome division ceases. Many studies strongly indicate that plant boron requirement (reviewed by Bell, Dell & Huang 2002) is higher for growing (including apical meristems, young leaves or flower tissues) than for mature organs at a maintenance phase. This is in line with a higher B content in nodules than in other legume tissues (Redondo-Nieto et al. 2003). This high B demand for nodule development is due to the high deposition rate of cell wall material during meristematic activity and cell enlargement, but might, at least in part, as well be related to the intense PBM synthesis. Both processes require heavy vesicle traffic and targeting. Several studies link B with vesicle fusion either by playing a membrane-related role or by influencing cytoskeleton function (Goldbach 1997; Yu et al. 2001, 2002). In nodules, the abnormal vesicle targeting under B deficiency accompanied aberrant symbiosome development (Bolaños et al. 2001). RGII is a known B ligand in cell walls; therefore, glycoproteins reactive to anti-RGII pectin polysaccharide are as well likely to interact with borate ions. This interaction could account for a correct vesicle targeting and/or stabilization of a new membrane material during PBM synthesis. Although further analysis is needed, this can be supported by the fact that all dividing symbiosomes in +B (Fig. 4a and zones II and II–III in Fig. 4c), but not in −B (Fig. 4b,e) nodules, were labelled by the RGII antibody.

B deficiency not only affects symbiosome development but also cell growth in nodules (Bolaños et al. 1994; Bonilla et al. 1997). Abnormal pectic polysaccharide anchoring in −B nodule cell walls previously reported (Redondo-Nieto et al. 2003) was attributed to the lack of cross-linking of pectin polysaccharides through borate-diester bonds with apiosyl residues from the RGII. Figure 5 confirmed the abnormal pectic polysaccharide distribution in B-deficient nodules. Considering that RGII glycoproteins are absent in B-deficient membranes (Fig. 3), the residual anti-RGII pectin polysaccharide labelling in −B nodules (Fig. 5b) is likely to show RGII, whereas anti-RGII pectin polysaccharide immunostaining of B-sufficient nodule cells (Fig. 5a) corresponds to both RGII pectin and membrane RGII-glycoproteins. Therefore, it is likely that the highest density of label found in the interface between plasma membrane and apoplast results from the fusion of vesicles (loaded with RGII glycoproteins and/or RGII) during cell growth. Therefore, it is also possible that besides borate-RGII polysaccharide dimers, B-linked glycoproteins and deposited RGII may stabilize walls, at least during cell expansion growth (see also Findeklee & Goldbach 1996). According to our own unpublished observations, the plasma membrane remained usually attached to the cell wall pectic polysaccharides during membrane purification. Apparently, RGII-reactive glycoproteins and (eventually) RGII seem to secure plasma membrane functions and the ordered release of cell wall material when cross-linked by B. The absence of plasmalemma-linked RGII-reactive glycoproteins in B-deficient cells would inhibit the release of homogalacturonan into the apoplast, leading to the observed accumulation of JIM5 antigen in the cytosol (Fig. 5d,f). Moreover, extending this hypothesis, which is in line with earlier observations as summarized by Goldbach (1997), cell wall hydroxyproline/proline-rich glycoproteins are also apparently not secreted under B deficiency and are retained in the cytosol (Bonilla et al. 1997).

Finally, it is important to point out that symbiotic legume nodules undergo a peculiar process of organogenesis and development (Stougaard 2000) with an extensive synthesis of membranes (PBMs) at rates about 30- to 50-fold higher than in other tissues (Robertson & Lyttleton 1984). This membrane progressively differentiates in a highly regulated manner (Perotto et al. 1991) with phases of high B requirements. Thus, legume nodule development provides an excellent model to investigate the role of B in membrane-located processes during differentiation.


This work was supported by Ministerio de Educación y Ciencia BIO2005-08691-C02-01. Miguel Redondo-Nieto is the recipient of a Juan de la Cierva Contract (Ministerio de Educación y Ciencia). María Reguera was granted with a fellowship from Comunidad de Madrid.