Dissecting paclitaxel-microtubule association: quantitative assessment of the 2'-OH group.

Paclitaxel (PTX) is a microtubule-stabilizing agent that is widely used in cancer chemotherapy. This structurally complex natural product acts by binding to ß-tubulin in assembled microtubules. The 2'-hydroxyl group in the flexible side chain of PTX is an absolute requirement for activity, but its precise role in the drug-receptor interaction has not been specifically investigated. The contribution of the 2'-OH group to the affinity and tubulin-assembly efficacy of PTX has been evaluated through quantitative analysis of PTX derivatives possessing side chain deletions: 2'-deoxy-PTX, N-debenzoyl-2'-deoxy-PTX, and baccatin III. The affinity of 2'-deoxy-PTX for stabilized microtubules was more than 100-fold lower than that of PTX and only ~3-fold greater than the microtubule affinity of baccatin III. No microtubule binding activity was detected for the analogue N-debenzoyl-2'-deoxy-PTX. The tubulin-assembly efficacy of each ligand was consistent with the microtubule binding affinity, as was the trend in cytotoxicities. Molecular dynamics simulations revealed that the 2'-OH group of PTX can form a persistent hydrogen bond with D26 within the microtubule binding site. The absence of this interaction between 2'-deoxy-PTX and the receptor can account for the difference in binding free energy. Computational analyses also provide a possible explanation for why N-debenzoyl-2'-deoxy-PTX is inactive, in spite of the fact that it is essentially a substituted baccatin III. We propose that the hydrogen bonding interaction between the 2'-OH group and D26 is the most important stabilizing interaction that PTX forms with tubulin in the region of the C-13 side chain. We further hypothesize that the substituents at the 3'-position function to orient the 2'-OH group for a productive hydrogen bonding interaction with the protein.

Characterization of the colchicine binding site on avian tubulin isotype betaVI.

Tubulin, the basic component of microtubules, is present in most eukaryotic cells as multiple gene products, called isotypes. The major tubulin isotypes are highly conserved in terms of structure and drug binding capabilities. Tubulin isotype betaVI, however, is significantly divergent from the other isotypes in sequence, assembly properties, and function. It is the major beta-tubulin isotype of hematopoietic tissue and forms the microtubules of platelet marginal bands. The interaction of the major tubulin isotypes betaI, betaII, betaIII, and betaIotaV with antimicrotubule drugs has been widely studied, but little is known about the drug binding properties of tubulin isotype betaVI. In this investigation, we characterize the activity of various colchicine site ligands with tubulin isolated from Gallus gallus erythrocytes (CeTb), which is approximately 95% betaVI. Colchicine binding is thought to be a universal property of higher eukaryotic tubulin; however, we were unable to detect colchicine binding to CeTb under any experimental conditions. Podophyllotoxin and nocodazole, other colchicine site ligands with divergent structures, were able to inhibit paclitaxel-induced CeTb assembly. Surprisingly, the colchicine isomer allocolchicine also inhibited CeTb assembly and displayed measurable, moderate affinity for CeTb (K(a) = 0.18 x 10(5) M(-1) vs 5.0 x 10(5) M(-1) for bovine brain tubulin). Since allocolchicine and colchicine differ in their C ring structures, the two C ring colchicine analogues were also tested for CeTb binding. Kinetic experiments indicate that thiocolchicine and chlorocolchicine bind to CeTb, but very slowly and with low affinity. Molecular modeling of CeTb identified five divergent amino acid residues within 6 A of the colchicine binding site compared to betaI, betaII, and betaIV; three of these amino acids are also altered in betaIII-tubulin. Interestingly, the altered amino acids are in the vicinity of the A ring region of the colchicine binding site rather than the C ring region. We propose that the amino acid differences in the binding site constrict the A ring binding domain in CeTb, which interferes with the positioning of the trimethoxyphenyl A ring and prevents C ring binding site interactions from efficiently occurring. Allocolchicine is able to accommodate the altered binding mode because of its smaller ring size and more flexible C ring substituents. The sequence of the colchicine binding domain of CeTb isotype betaVI is almost identical to that of its human hematopoietic counterpart. Thus, through analysis of the interactions of ligands with CeTb, it may be possible to discover colchicine site ligands that specifically target tubulin in human hematopoietic cells.

Plant cell-wall cross-links by REDOR NMR spectroscopy.

We present a new method that integrates selective biosynthetic labeling and solid-state NMR detection to identify in situ important protein cross-links in plant cell walls. We have labeled soybean cells by growth in media containing l-[ring-d(4)]tyrosine and l-[ring-4-(13)C]tyrosine, compared whole-cell and cell-wall (13)C CPMAS spectra, and examined intact cell walls using (13)C{(2)H} rotational echo double-resonance (REDOR) solid-state NMR. The proximity of (13)C and (2)H labels shows that 25% of the tyrosines in soybean cell walls are part of isodityrosine cross-links between protein chains. We also used (15)N{(13)C} REDOR of intact cell walls labeled by l-[ε-(15)N,6-(13)C]lysine and depleted in natural-abundance (15)N to establish that the side chains of lysine are not significantly involved in covalent cross-links to proteins or sugars.

(15)N{(31)P} REDOR NMR studies of the binding of phosphonate reaction intermediate analogues to Saccharomyces cerevisiae lumazine synthase.

Lumazine synthase catalyzes the reaction of 5-amino-6-D-ribitylamino-2,4(1H,3H)-pyrimidinedione(1) with (S)-3,4-dihydroxybutanone 4-phosphate (2) to afford 6,7-dimethyl-8-D-ribityllumazine(3), the immediate biosynthetic precursor of riboflavin. The overall reaction implies a series of intermediates that are incompletely understood. The 15N{31P} REDOR NMR spectra of three metabolically stable phosphonate reaction intermediate analogues complexed to Saccharomyces cereVisiae lumazine synthase have been obtained at 7 and 12 T. Distances from the phosphorus atoms of the ligands to the side chain nitrogens of Lys92, His97, Arg136, and His148 have been determined. These distances were used in combination with the X-ray crystal coordinates of one of the intermediate analogues complexed with the enzyme in a series of distance-restrained molecular dynamics simulations. The resulting models indicate mobility of the Lys92 side chain, which could facilitate the exchange of inorganic phosphate eliminated from the substrate in one reaction, with the organic phosphate-containing substrate necessary for the next reaction.

Human factor Xa bound amidine inhibitor conformation by double rotational-echo double resonance nuclear magnetic resonance and molecular dynamics simulations.

Double rotational-echo double resonance (double REDOR) NMR was used to investigate the conformation of a (13)C-, (15)N-, and (19)F-labeled inhibitor (Berlex Biosciences compound no. ZK-806299) bound to human factor Xa. Conformationally dependent carbon-fluorine dipolar couplings were measured by (13)C[(19)F] REDOR. Natural abundance carbon signals in the full-echo spectra were removed by (13)C[(15)N] REDOR. Major and minor binding modes were suggested by the NMR data, but only the former had adequate signal to noise for distance determinations. Molecular dynamics simulations restrained by double-REDOR-determined intramolecular (13)C-(19)F distances revealed two models for the dominant binding mode that are consistent with the NMR data. We conclude that ZK-806299 binds similarly to both FXa. Moreover, it appears to bind to FXa in a fashion previously demonstrated for ZK-807834, a more selective FXa inhibitor.

Characterization of the complex of a trifluoromethyl-substituted shikimate-based bisubstrate inhibitor and 5-enolpyruvylshikimate-3-phosphate synthase by REDOR NMR.

A combination of (15)N[(19)F], (31)P[(15)N], and (31)P[(19)F] rotational-echo double-resonance NMR has been used to characterize the conformation of a bound trifluoromethylketal, shikimate-based bisubstrate inhibitor of 5-enolpyruvylshikimate-3-phosphate synthase. The solid-state NMR experiments were performed on the complex formed in solution and then lyophilized at low temperatures in the presence of stabilizing lyoprotectants. The results of these experiments indicate that none of the side chains of the six arginines that surround the active site forms a compact salt bridge with the phosphate groups of the bound inhibitor.

Rotational-echo double-resonance NMR-restrained model of the ternary complex of 5-enolpyruvylshikimate-3-phosphate synthase.

The 46-kD enzyme 5-enolpyruvylshikimate-3-phosphate (EPSP) synthase catalyzes the condensation of shikimate-3-phosphate (S3P) and phosphoenolpyruvate to form EPSP. The reaction is inhibited by N-(phosphonomethyl)-glycine (Glp), which, in the presence of S3P, binds to EPSP synthase to form a stable ternary complex. We have used solid-state NMR and molecular modeling to characterize the EPSP synthase-S3P-Glp ternary complex. Modeling began with the crystal coordinates of the unliganded protein, published distance restraints, and information from the chemical modification and mutagenesis literature on EPSP synthase. New inter-ligand and ligand-protein distances were obtained. These measurements utilized the native (31)P in S3P and Glp, biosynthetically (13)C-labeled S3P, specifically (13)C and (15)N labeled Glp, and a variety of protein-(15)N labels. Several models were investigated and tested for accuracy using the results of both new and previously published rotational-echo double resonance (REDOR) NMR experiments. The REDOR model is compared with the recently published X-ray crystal structure of the ternary complex, PDB code 1G6S. There is general agreement between the REDOR model and the crystal structure with respect to the global folding of the two domains of EPSP synthase and the relative positioning of S3P and Glp in the binding pocket. However, some of the REDOR data are in disagreement with predictions based on the coordinates of 1G6S, particularly those of the five arginines lining the binding site. We attribute these discrepancies to substantive differences in sample preparation for REDOR and X-ray crystallography. We applied the REDOR restraints to the 1G6S coordinates and created a REDOR-refined xray structure that agrees with the NMR results.

Recognition of an unnatural difluorophenyl nucleotide by uracil DNA glycosylase.

The DNA repair enzyme uracil DNA glycosylase (UDG) utilizes base flipping to recognize and remove unwanted uracil bases from the genome but does not react with its structural congener, thymine, which differs by a single methyl group. Two factors that determine whether an enzyme flips a base from the duplex are its shape and hydrogen bonding properties. To probe the role of these factors in uracil recognition by UDG, we have synthesized a DNA duplex that contains a single difluorophenyl (F) nucleotide analogue that is an excellent isostere of uracil but possesses no hydrogen bond donor or acceptor groups. By using binding affinity measurements, solution (19)F NMR, and solid state (31)P[(19)F] rotational-echo double-resonance (REDOR) NMR measurements, we establish that UDG partially unstacks F from the duplex. However, due to the lack of hydrogen bonding groups that are required to support an open-to-closed conformational transition in UDG, F cannot stably dock in the UDG active site. We propose that F attains a metastable unstacked state that mimics a previously detected intermediate on the uracil-flipping pathway and suggest structural models of the metastable state that are consistent with the REDOR NMR measurements.