In 1933, Boivin et al. extracted an endotoxin from Salmonella typhimurium for the first time, after which a variety of chemical and biological studies on endotoxins have been performed. In 1952, the structural and functional properties of endotoxic lipopolysaccharide (LPS), extracted by a hot phenol and water method devised by Westphal et al., were reported, which led to a number of studies of Gram-negative bacteria in regards to the host defense mechanism. Since 1960, the unique chemical structure and biological activity of Bacteroides species LPS have received a great deal of attention, and there is a long history of such studies. In addition, among oral bacterial strains that have received attention as causative periodontopathic bacteria, many have been classified as Bacteroides species. In particular, a number of researchers have investigated whether LPS of Porphyromonas gingivalis (formerly Bacteroides gingivalis), a black-pigmented oral anaerobic rod, is a virulent factor of the bacterium. The active center of the LPS of these Bacteroides species, the lipid A molecule, is known to be an active participant in endotoxic activation, though its other biological activities are weak, due to its unique chemical structure and action as an antagonist of LPS. On the other hand, many reports have noted that the LPS of those species activate cells in C3H/HeJ mice, which generally do not respond to LPS. We were the first to reveal the chemical structure of P. gingivalis lipid A and, together with other researchers, reported that P. gingivalis LPS and its lipid A have activities toward C3H/HeJ mice. Since that time, because of the popularity of Toll-like receptor (TLR) studies, a great deal of evidence has been reported indicating that P. gingivalis LPS and its lipid A are ligands that act on TLR2. In order to solve such problems as heterogeneity and contamination of the biologically active components of P. gingivalis lipid A, we produced a chemical synthesis counterpart of lipid A and test results indicated it to be a TLR4 agonist. Furthermore, in order to disprove the common belief that P. gingivalis LPS and its lipid A are TLR2 ligands, the TLR2-active component contained in a P. gingivalis LPS fraction was separated and purified, after which we showed its chemical structure to be a lipoprotein consisting of three fatty acid residues, thus answering a longstanding question regarding Bacteroides species LPS. In addition to the field of dentistry, many studies based on the misconception of "TLR2-active LPS/lipid A" still exist in the field of innate immunity. Based on the history of studies of ligands acting on TLR4, Bacteroides species LPS findings were reviewed and are presented here. In particular, we investigated P. gingivalis LPS and its lipid A.