Capillary zone electrophoresis has been used to monitor the first steps of the dehydrogenative polymerization of coniferyl alcohol, sinapyl aldehyde, or a mixture of both, catalyzed by the horseradish peroxidase (HRP)-H(2)O(2) system. When coniferyl alcohol was the unique HRP substrate, three major dimers were observed (beta-5, beta-beta, and beta-O-4 interunit linkages) and their initial formation velocity as well as their relative abundance varied with pH. The beta-O-4 interunit linkage was thus slightly favored at lower pH values. In contrast, sinapyl aldehyde turned out to be a very poor substrate for HRP except in basic conditions (pH 8). The major dimer observed was the beta,beta'-di-sinapyl aldehyde, a red-brown exhibiting compound which might partly participate in the red coloration usually observed in cinnamyl alcohol dehydrogenase-deficient angiosperms. Finally, when a mixture of coniferyl alcohol and sinapyl aldehyde was used, it looked as if sinapyl aldehyde became a very good substrate for HRP. Indeed, coniferyl alcohol turned out to serve as a redox mediator (i.e. "shuttle oxidant") for the sinapyl aldehyde incorporation in the lignin-like polymer. This means that in particular conditions the specificity of oxidative enzymes might not hinder the incorporation of poor substrates into the growing lignin polymer.