Symmetry Requirements for Effective Blocking of Pore-Forming Toxins: Comparative Study with α-, β-, and γ-Cyclodextrin Derivatives

Anthrax toxin inhibition.First, we compared the activities of per-6-amino derivatives carrying unmodified hydroxyl groups at positions 2 and 3 (compounds 1 to 3) with the 2,3-methylated ones (compounds 4 to 6; Fig. 1 shows OH group numberings) (Table 1). None of the methylated derivatives displayed toxin-inhibiting activities, while the unmodified compounds 2 and 3 showed activity at low-micromolar concentrations. These data demonstrate the importance of the H-bonding network for the rigidness of the cyclodextrin core, which is destroyed by methylation of the hydroxyl groups at positions 2 and 3.

We also investigated the activities of various α-, β-, and γ-cyclodextrin derivatives carrying the same modifications. Similarly to results obtained earlier (3), the α-cyclodextrin derivatives showed no (compounds 1, 7, 10, and 13) or low (compound 18) LeTx-inhibiting activity. This finding could be explained by their smaller size or by the features of α-CD's structure, which could provide a less favorable spatial orientation of the substituting groups due to the mismatch of the 6-fold symmetry of α-CD and the 7-fold symmetry of the PA₆₃ pore.

All the β- and γ-cyclodextrin derivatives with free OH groups at positions 2 and 3 displayed anti-LeTx activity in the micromolar range. When we compared structurally similar derivatives of β- and γ-cyclodextrins, it appeared that in many instances the γ-CDs were more active than the β-CD derivatives (compounds 2 and 3, 8 and 9, 14 and 15, and 16 and 17). That could be related to the recently reported observation of PA₆₃ octameric pores having the same 8-fold symmetry as γ-CD (6). It is interesting also that in the case of the structurally related amino and guanidino derivatives, the latter ones were more active (compounds 2 and 14, 3 and 15, 8 and 16, and 9 and 17).

The ability of several key compounds to block PA channels was tested in channel reconstitution experiments. The results showed (Table 1) that there is a general correlation between the cytotoxicity inhibition and channel blocking activity. Similarly to the results of the cytotoxicity experiments, α-CD derivatives (compounds 1, 13, and 18) exhibited a much lower blocking activity than the β- and γ-CD derivatives (compounds 2, 3, 14, 15, 19, and 20).

α-Hemolysin inhibition.We also evaluated the activities of two α- and γ-CDs (compounds 21 and 23) having the same substituents on the macrocycle's primary side as the β-CD derivative (compound 22) which inhibits α-HL (2). The compounds were tested using a rabbit erythrocyte assay. Both α- and γ-CD derivatives did not display any activity against α-HL. Similar results were obtained when we compared the compounds' abilities to block the pores formed by α-HL in planar lipid membranes.

The fact that only derivative 22 with the 7-fold symmetry blocked the heptameric α-HL pore is consistent with the idea that a symmetry match between the blocking molecule and the pore is required for effective inhibition. The difference in the activities of γ-CD derivatives against anthrax and staphylococcal toxins could be explained by the recent observation that in contrast to α-HL, PA presumably can form octameric pores in addition to heptameric ones (6). However, the currently available data are ambiguous. For example, our ion channel measurements showed that both β-CD and γ-CD derivatives independently and completely blocked all PA channels in the planar lipid membranes, bringing the transmembrane current to zero. Earlier we reported two types of PA channel insertions (10) that might be attributed to heptameric and octameric channel formation, but the kinetic characteristics of β-CD binding to these channels were indistinguishable. It should be taken into account that in the channel reconstitution experiments only transmembrane pore blockage is detected. The cell toxicity inhibition could be more complex and also involve inactivation of the prepore. Further studies are required to elucidate the fine mechanisms of the heptameric and octameric pore formations and their relative ratios in the prepores and transmembrane pores as well as the parameters of binding of cyclodextrin derivatives to those forms. To conclude, these data showed that size and conformation as well as the match of the symmetry of a blocking molecule and a pore play important roles in the activity of the inhibitors of pore-forming toxins.