Pectic oligosaccharides, produced by microbial enzymes, are well-known oligosaccharins, eliciting defence responses in diseased plants. Regulatory roles in non-diseased plants have also been proposed: e.g. oligosaccharides, generated in vitro by hydrolysis of pectins, promote ripening in tomato (Lycopersicon esculentum L.) and other fruits. Endopolygalacturonase (endo-PG; EC 3.2.1.15) occurs in many fruits and theoretically could generate such oligosaccharides. However, the mere occurrence of an enzyme does not prove that it acts; the proposed products (oligosaccharides) must be sought in vivo. Recent evidence indicates that fruit tissues do naturally produce pectic oligosaccharides, some of which promote ripening when added to unripe fruits. Such studies, complementing earlier work on in-vitro generated oligosaccharides, provide a means of discovering novel oligosaccharins. Unlike fruits, cell cultures of rose (Rosa sp.) produce no pectic oligosaccharides. Their 'absence' is not due to excessively rapid turnover: when [C-14]-oligogalacturonides are added, they undergo only slow hydrolysis. The hydrolysis is by exo-polygalacturonases (exo-PG; EC 3.2.1.67), yielding free galacturonic acid (GalA). If no oligogalacturonides are added, no GalA accumulates in the medium; therefore, exo-PG does not normally operate in healthy rose cells. Exo-PG is presumably 'in reserve', available to trim oligogalacturonides made by phytopathogens. We conclude that the in-vivo action of a wall enzyme is best studied at the level of carbohydrate metabolism in vivo rather than by assay of extracted enzymes in vitro. The major uronic acid-containing oligosaccharide produced by rose cell cultures is alpha-D-mannopyranosyl-(1-->4)-alpha-D-glucuronopyranosyl-(1-->2)-myo-inositol, which may be a novel phytoglycolipid-derived oligosaccharin related to the inositolphosphoglycans that mediate insulin action. (C) 2000 Editions scientifiques et medicales Elsevier SAS.