In a previous report from this laboratory (N. J. Laible and G. R. Germaine, Infect. Immun. 48:720-728, 1985), evidence was presented to suggest that the bactericidal actions of both reduced (i.e., muramidase-inactive) human placental lysozyme and the synthetic cationic homopolymer poly-D-lysine involved the activation of a bacterial endogenous activity that was inhibitable by N,N',N"-triacetylchitotriose (chitotriose). In the present investigation however, we found that the bactericidal and bacteriolytic action of poly-D-lysine could be prevented only by some commercially available chitotriose preparations and not by others. Analysis by physical and chemical methods failed to distinguish protective chitotriose (CTa) and nonprotective chitotriose (CTi) preparations. CTi and CTa preparations displayed equal capacities to competitively inhibit binding of [3H]chitotriose by immobilized lysozyme and were indistinguishable in their abilities to block the lytic activity of lysozyme against Micrococcus lysodeikticus cells. Elemental analysis revealed significantly higher levels of phosphorus, calcium, iron, sodium, manganese, and copper in CTa. Removal of metals from CTa by chelate chromatography completely abolished the poly-D-lysine-protective capacity. Of the metals detected, only ferric iron (5 to 10 microM) mimicked the protective action of CTa. A Fe(III) concentration of 50 microM was required to inhibit lysozyme (5 micrograms/ml). Both Fe(III) and CTa (but not CTi) quantitatively blocked the labeling of poly-D-lysine by fluorescamine, suggesting that the primary amino groups of the lysine residues participate in iron binding. Thus, it appears that the poly-D-lysine-protective capacity of certain chitotriose preparations was due not to the chitotriose itself but to contaminating metal ions which interact directly with the polycationic agent. In contrast, Fe(III) cannot account for inhibition of either the bactericidal or bacteriolytic activity of lysozyme by chitotriose.