Exploration of the GM1 receptor-binding site of heat-labile enterotoxin and cholera toxin by phenyl-ring-containing galactose derivatives

Cholera toxin (CT) and the closely related heat-labile enterotoxin of Esche richia coli (LT) are responsible for numerous cases of diarrhea worldwide, leading to considerable morbidity and mortality. The B subunits of these he terohexameric AB(5) toxins form a pentameric arrangement which is responsib le for binding to the receptor GM1 of the target epithelial cells of the ho st. Blocking these B pentamer-receptor interactions forms an avenue for the rapeutic intervention. Here, the structural characterization of potential r eceptor-blocking compounds are described based on the previously identified inhibitor m-nitrophenyl-alpha -D-galactoside (MNPG). The structure of a CT B-MNPG complex confirms that the binding mode of this inhibitor is identica l in the two homologous toxins CT and LT and is characterized by a glycosyl linkage geometry that leads to displacement of a well ordered water molecu le near the amide group of Gly33 by the O1-substituent of MNPG. This glycos yl geometry is not maintained in the absence of a substituent that can disp lace this water, as shown by a complex of LTB with p-aminophenyl-alpha -D-g alactoside (PAPG). New compounds were synthesized to investigate the feasib ility of maintaining the favorable binding interactions exhibited by MNPG w hile gaining increased affinity through the addition of hydrophobic substit uents complementary to either of two hydrophobic regions of the receptor-bi nding site. The structural characterization of complexes of LTB with two of these compounds, 3-benzylaminocarbonylphenyl-alpha -D-galactoside (BAPG) a nd 2-phenethyl-7-(2,3-dihydrophthalazine-1,4-dione)-alpha -D-galactoside (P EPG), demonstrates a partial success in this goal. Both compounds exhibit a mixture of binding modes, some of which are presumably influenced by the l ocal packing environment at multiple crystallographically independent bindi ng sites. The terminal phenyl ring of BAPG associates either with the pheny l group of Tyr12 or with the hydrophobic patch formed by Lys34 and Ile58. T he latter interaction is also made by the terminal phenyl substituent of PE PG, despite a larger ring system linking the galactose moiety to the termin al phenyl. However, neither BAPG nor PEPG displaces the intended target wat er molecule. Both of the designed compounds exhibit increased affinity rela tive to the galactose and to PAPG notwithstanding the failure to displace a bound water, confirming that additional favorable hydrophobic interactions can be gained by extending the starting inhibitor by a hydrophobic tail. T he insight gained from these structures should allow the design of addition al candidate inhibitors that retain both the glycosyl geometry and water di splacement exhibited by MNPG and the favorable hydrophobic interactions exh ibited by BAPG and PEPG.