The cell-wall pectic domain rhamnogalacturonan-II (RG-II) is cross-linked via borate diester bridges, which influence the expansion, thickness and porosity of the wall. Previously, little was known about the mechanism or subcellular site of this cross-linking. Using polyacrylamide gel electrophoresis (PAGE) to separate monomeric from dimeric (boron-bridged) RG-II, we confirmed that Pb2+ promotes H3BO3-dependent dimerisation in vitro. H3BO3 concentrations as high as 50 mm did not prevent cross-linking. For in-vivo experiments, we successfully cultured ‘Paul's Scarlet’ rose (Rosa sp.) cells in boron-free medium: their wall-bound pectin contained monomeric RG-II domains but no detectable dimers. Thus pectins containing RG-II domains can be held in the wall other than via boron bridges. Re-addition of H3BO3 to 3.3 μm triggered a gradual appearance of RG-II dimer over 24 h but without detectable loss of existing monomers, suggesting that only newly synthesised RG-II was amenable to boron bridging. In agreement with this, Rosa cultures whose polysaccharide biosynthetic machinery had been compromised (by carbon starvation, respiratory inhibitors, anaerobiosis, freezing or boiling) lost the ability to generate RG-II dimers. We conclude that RG-II normally becomes boron-bridged during synthesis or secretion but not post-secretion. Supporting this conclusion, exogenous [3H]RG-II was neither dimerised in the medium nor cross-linked to existing wall-associated RG-II domains when added to Rosa cultures. In conclusion, in cultured Rosa cells RG-II domains have a brief window of opportunity for boron-bridging intraprotoplasmically or during secretion, but secretion into the apoplast is a point of no return beyond which additional boron-bridging does not readily occur.

Summary Dimerization of rhamnogalacturonan‐II (RG‐II) via boron cross‐links contributes to the assembly and biophysical properties of the cell wall. Pure RG‐II is efficiently dimerized by boric acid (B(OH)3) in vitro only if nonbiological agents for example Pb2+ are added. By contrast, newly synthesized RG‐II domains dimerize very rapidly in vivo. We investigated biological agents that might enable this.We tested for three such agents: novel enzymes, borate‐transferring ligands and cationic ‘chaperones’ that facilitate the close approach of two polyanionic RG‐II molecules. Dimerization was monitored electrophoretically.Parsley shoot cell‐wall enzymes did not affect RG‐II dimerization in vitro. Borate‐binding ligands (apiose, dehydroascorbic acid, alditols) and small organic cations (including polyamines) also lacked consistent effects. Polylysine bound permanently to RG‐II, precluding electrophoretic analysis. However, another polycation, polyhistidine, strongly promoted RG‐II dimerization by B(OH)3 without irreversible polyhistidine–RG‐II complexation. Likewise, partially purified spinach extensins (histidine/lysine‐rich cationic glycoproteins), strongly promoted RG‐II dimerization by B(OH)3 in vitro.Thus certain polycations, including polyhistidine and wall glycoproteins, can chaperone RG‐II, manoeuvring this polyanionic polysaccharide domain such that boron‐bridging is favoured. These chaperones dissociate from RG‐II after facilitating its dimerization, indicating that they act catalytically rather than stoichiometrically. We propose a natural role for extensin–RG‐II interaction in steering cell‐wall assembly.

Rhamnogalacturonan-II (RG-II), a domain of plant cell wall pectins, is able to cross-link with other RG-II domains through borate diester bridges. Although it is known to affect mechanical properties of the cell wall, the biochemical requirements and lifecycle of this cross-linking remain unclear. We developed a PAGE methodology to allow separation of monomeric and dimeric RG-II and used this to study the dynamics of cross-linking in vitro and in vivo. Rosa cells grown in medium with no added boron contained no RG-II dimers, although these re-appeared after addition of boron to the medium. However, other Rosa cultures which were unable to synthesize new polysaccharides did not show dimer formation. We conclude that RG-II normally becomes cross-linked intraprotoplasmically or during secretion, but not post-secretion.