Furthermore, BbGL-IIc induced iNKT cell activation occurs independently of MyD88 and TRIF signaling (49). These results show that BbGL-IIc is a bacterial antigen for the mouse iNKT cell TCR. BbGL-II compounds also stimulate human iNKT cells to release cytokines. Interestingly, BbGL-IIf, which contains linoleic acid (C18:2) in the sn-1 position and oleic acid in the sn-2 position, has been found to PD 332991 be the most potent

antigen for human iNKT cells (49). Data from another study suggest that the different iNKT cell responses to Borrelia glycolipids are due to a difference between human and mouse CD1d molecules (51). These studies show that iNKT cell TCR detects DAGs, another category of glycolipid, in addition to glycosphingolipids.

Moreover, DAG antigen induced iNKT cell activation is dependent on acyl chain length and saturation (49). The TCR of iNKT cells recognizes Sphingomonas GSL and B. burgdorferi DAG as well as αGalCer. Although the structures of these bacterial antigens are similar to that of αGalCer (Fig. check details 5), there are several small structural differences. DAG belongs to a different category of glycolipid than do αGalCer and Sphingomonas GSL. Also, the bacterial antigens are less potent than αGalCer. What determines the antigenic potency of these glycolipids? To address this point, crystal structures of mouse CD1d in complex with Sphingomonas GalAGSL or B. burgdorferi DAG were determined (51, 52). GalAGSL binds to mouse CD1d similarly to αGalCer. Between the α1 and α2 helices, the CD1d molecule has two pockets (A′ and F′) which accommodate Phospholipase D1 the lipid tails of antigens (Fig. 6a, b) (6, 7). The fatty acid and sphinganine tails of GalAGSL extend into the A′ and F′ pockets, respectively (52). However, because of an alternative hydrogen-bonding interaction, the sphinganine tail of GalAGSL, which lacks 4-OH, is more deeply inserted into the F′ pocket (52). The sugar head group of GalAGSL is present in the center of the binding groove at

the CD1d surface where an incoming TCR recognizes antigens (Fig. 6b, c), but it shows a slight lateral shift compared to αGalCer (52). These differences are thought to cause the difference in antigenic potency between Sphingomonas GalAGSL and αGalCer. The binding of B. burgdorferi galactosyl DAG is more flexible than that of Sphingomonas GalAGSL or αGalCer. The sn-1 linked oleic acid and the sn-2 linked palmitic acid of BbGL-IIc are inserted into the A′ and F′ pockets, respectively (51). The glycerol moiety of BbGL-IIc is tilted toward the α1 helix of the CD1d molecule, and the galactose of BbGL-IIc is pointed upward and away from the α2 helix of CD1d. These differences result in the loss of important hydrogen bonding interactions with the amino acids in the α2 helix that are present in the case of αGalCer (51).