No abstract

Dendritic cell-specific intercellular adhesion molecule-3-grabbing nonintegrin (DC-SIGN) and Langerin are C-type lectins of dendritic cells (DCs) that share a specificity for mannose and are involved in pathogen recognition. HIV is known to use DC-SIGN on DCs to facilitate transinfection of T-cells. Langerin, on the contrary, contributes to virus elimination; therefore, the inhibition of this latter receptor is undesired. Glycomimetic molecules targeting DC-SIGN have been reported as promising agents for the inhibition of viral infections and for the modulation of immune responses mediated by DC-SIGN. We show here for the first time that glycomimetics based on a mannose anchor can be tuned to selectively inhibit DC-SIGN over Langerin. Based on structural and binding studies of a mannobioside mimic previously described by us (2), a focused library of derivatives was designed. The optimized synthesis gave fast and efficient access to a group of bis(amides), decorated with an azide-terminated tether allowing further conjugation. SPR inhibition tests showed improvements over the parent pseudomannobioside by a factor of 3-4. A dimeric, macrocyclic structure (11) was also serendipitously obtained, which afforded a 30-fold gain over the starting compound (2). The same ligands were tested against Langerin and found to exhibit high selectivity towards DC-SIGN. Structural studies using saturation transfer difference NMR spectroscopy (STD-NMR) were performed to analyze the binding mode of one representative library member with DC-SIGN. Despite the overlap of some signals, it was established that the new ligand interacts with the protein in the same fashion as the parent pseudodisaccharide. The two aromatic amide moieties showed relatively high saturation in the STD spectrum, which suggests that the improved potency of the bis(amides) over the parent dimethyl ester can be attributed to lipophilic interactions between the aromatic groups of the ligand and the binding site of DC-SIGN.

DC-SIGN is a dendritic cell-specific C-type lectin receptor that recognizes highly glycosylated ligands expressed on the surface of various pathogens. This receptor plays an important role in the early stages of many viral infections, including HIV, which makes it an interesting therapeutic target. Glycomimetic compounds are good drug candidates for DC-SIGN inhibition due to their high solubility, resistance to glycosidases, and nontoxicity. We studied the structural properties of the interaction of the tetrameric DC-SIGN extracellular domain (ECD), with two glycomimetic antagonists, a pseudomannobioside (1) and a linear pseudomannotrioside (2). Though the inhibitory potency of 2, as measured by SPR competition experiments, was 1 order of magnitude higher than that of 1, crystal structures of the complexes within the DC-SIGN carbohydrate recognition domain showed the same binding mode for both compounds. Moreover, when conjugated to multivalent scaffolds, the inhibitory potencies of these compounds became uniform. Combining isothermal titration microcalorimetry, analytical ultracentrifugation, and dynamic light scattering techniques to study DC-SIGN ECD interaction with these glycomimetics revealed that 2 is able, without any multivalent presentation, to cluster DC-SIGN tetramers leading to an artificially overestimated inhibitory potency. The use of multivalent scaffolds presenting 1 or 2 in HIV trans-infection inhibition assay confirms the loss of potency of 2 upon conjugation and the equal efficacy of chemically simpler compound 1. This study documents a unique case where, among two active compounds chemically derived, the compound with the lower apparent activity is the optimal lead for further drug development.

We explored the use of several breadth‐first and depth‐first algorithms for the computation of Gaussian atomic and molecular surface areas. Our results for whole‐molecule van der Waals surface areas (vdWSAs) were 10 times more accurate in relative error, relative to actual hard‐sphere areas, than those reported by earlier workers. We were also able to extend the method to the computation of solvent‐accessible surface areas (SASAs). This was made possible by an appropriate combination of algorithms, parameters, and preprocessing steps. For united‐atom 3app, a 2366‐atom protein, we obtained an average absolute atomic error of 1.16 Å2 with respect to the hard‐sphere atomic SASA results in 7 s of CPU time on an R10000/194 MHz processor. Speed and accuracy were both optimized for SASA by the use of neighbor‐list reduction (NLR), buried‐atom elimination (BAE), and a depth‐first search of the tree of atomic intersections. Accuracy was further optimized by the application of atom type specific parameters to the raw Gaussian results. ©1999 John Wiley & Sons, Inc. J Comput Chem 20: 688–703, 1999

Up to about half of the atoms in biopolymers are inaccessible to solvents. If such atoms can be rapidly identified, time can be saved in the subsequent computation of atomic surface areas. A quick, approximate method, termed buried atom elimination (BAE), was developed for the detection of such atoms. Following the literature, the method makes use of a Gaussian function to calculate the neighbor density in four tetrahedral directions in 3‐dimensional space, sometimes twice with different orientations. In macromolecules, our method detects between 63 and 81% of the buried atoms but also incorrectly classifies 2–8% of the exposed atoms as buried. These misidentified atoms all have small solvent‐exposed (accessible) surface areas (SASAs): their surfaces sum to a maximum of 0.5% of the molecular SASA, and their maximum atomic SASA is 5.1 Å2. Using our recently reported LCPO method for computing atomic surfaces, which is one of the fastest available, the use of BAE increases the overall speed of computing the atomic SASAs by a factor of up to 1.6 for surfaces only and 1.9 when first and second derivatives are computed. BAE decreases the LCPO average absolute atomic error from about 2.3 Å2 to about 1.7 Å2 (average for larger compounds). BAE was introduced into the MacroModel molecular modeling package and tests show that it increases the efficiency of first‐ and second‐derivative energy minimizations and molecular dynamics simulations without adversely affecting the stability or accuracy of the calculations. BAE parameters were developed for the most important atom types in biopolymers, based on a parameterization set of 18 compounds of different size (33–4346 atoms) and class (organics, proteins, DNA, and various complexes), consisting of a total of 23,186 atoms. ©1999 John Wiley & Sons, Inc. J Comput Chem 20, 586–596, 1999