Abstract. The Golgi apparatus of plant cells is the site of assembly of glycoproteins, proteoglycans, and complex polysaccharides, but little is known about how the different assembly pathways are organized within the Golgi stacks. To study these questions we have employed immunocytochemical techniques and antibodies raised against the hydroxyproline-rich cell wall glycoprotein, extensin, and two types of complex polysaccharides, an acidic pectic polysaccharide known as rhamnogalacturonan I (RG-I), and the neutral hemicellulose, xyloglucan (XG). Our micrographs demonstrate that individual Golgi stacks can process simultaneously glycoproteins and complex polysacchatides. O-linked arabinosylation of the hydroxyproline residues of extensin occurs in cis-cisternae, and glycosylated molecules pass through all cisternae before they are packaged into secretory vesicles in the monensin-sensitive, trans-Golgi network. In contrast, in root tip cortical parenchyma cells, the anti-RG-I and the anti-XG antibodies are shown to bind to complementary subsets of Golgi cisternae, and several lines of indirect evidence suggest that these complex polysaccharides may also exit from different cisternae. Thus, RG-I type polysaccharides appear to be synthesized in cis-and medial cisternae, and have the potential to leave from a monensin-insensitive, medial cisternal compartment. The labeling pattern for XG suggests that it is assembled in trans-Golgi cisternae and departs from the monensin-sensitive trans-Golgi network. This physical separation of the synthesis/secretion pathways of major categories of complex polysaccharides may prevent the synthesis of mixed polysaccharides, and provides a means for producing secretory vesicles that can be targeted to different cell wall domains.

A tobacco cDNA clone (pCNT1) was characterized that encodes an extensin apoprotein almost entirely composed of the repeats Ser-Pro4(-Lys2), Pro-Tyr2-Pro2-His and Thr-Pro-Val-Tyr-Lys. In healthy plants extensin transcripts are abundant in the roots, less prevalent in the stem and rare in the leaves. In leaves extensin mRNA is induced by wounding, ethylene or virus infection. Tobacco was transformed with pCNT1 cDNA coupled in sense or antisense orientation to the CaMV 35S promoter. Analysis of transgenic plants that over- or underexpressed pCNT1 mRNA demonstrated that the encoded protein constituted the majority of hydroxyproline-rich glycoproteins in roots, stems and leaves. The pCNT1-encoded protein contained at least 50% of total hydroxyproline present in these organs and was abundant in the soluble protein fraction of stems and roots as well as in the cell wall of stem vascular bundles. Analysis of transgenic plants expressing sense or antisense extensin gene constructs showed no correlation between total hydroxyproline concentration or soluble HRGP content and plant development.

No abstract

Antisera raised against the major hydroxyproline-rich glycoprotein (HRCP) in carrot (Daucus carota 1.) taproot, extensin-1, and a minor HRCP, extensin-2, were characterized by western blot analysis, enzyme-linked immunosorbent assay, and periodate oxidation and found to be directed against carbohydrate epitopes shared by both glycoproteins. l h e anti-extensin-1 antibodies (gEl) target periodate-sensitive epitopes and may recognize the terminal a-1,3-arabinoside of extensin-1. l h e anti-extensin-2 antibodies (gE2) recognize periodate-insensitive epitopes, possibly binding the reducing, interna1 &1,2-arabinosides on the carbohydrate side chains. Despite the cross-reactivity of these antibodies, immunolocalization studies of carrot taproot and green bean (Phaseolus vulgaris 1.) leaf tissues reveal a spatial segregation of gEl-and gE2-labeling patterns. l h e gEl antibodies bind only to the celluloserich region of the cell wall (J.P. Staehelin and L.A. Stafstrom [1988] Planta 174: 321-332), whereas gE2 labeling is restricted to the expanded middle lamella at three cell junctions. Periodate oxidation of nonosmicated, thin-sectioned tissue abolishes gEl labeling but leads to labeling of the entire cell wall by gE2, presumably as a result of unmasking cryptic epitopes on extensin-1 in the cellulose layer. Purified extensin-2 protein is more efficient than extensin-1 protein at agglutinating avirulent Pseudomonas strains lacking extracellular polysaccharide. Our results indicate that extensin-2 does not form a heterologous HRCP network with extensin-1 and that, in contrast to extensin-1, which appears to serve a structural role, extensin-2 could participate in passive defense responses against phytopathogenic bacteria.