Search

Paclitaxel resistance mutations in Chinese hamster ovary cells frequently alter a cluster of leucine residues in the H6-H7 loop region of beta-tubulin. To gain further insight into the role of this region in microtubule assembly and drug resistance, site-directed mutagenesis was used to systematically change amino acid L215. The mutated genes were cloned into a tetracycline-regulated expression vector and transfected into wild-type cells. Most of the mutations destabilized microtubule assembly, causing a decreased fraction of tubulin to appear in the microtubule cytoskeleton. In each case, the decreased level of assembly was associated with paclitaxel resistance and increased colcemid sensitivity. In two cases, however, the alteration did not significantly perturb the level of assembled tubulin or confer resistance to paclitaxel. One of these, L215V, produced little or no detectable phenotype, while the other, L215I, conferred increased sensitivity to paclitaxel. The increased drug sensitivity did not extend to epothilone A, a drug that binds to the same site and has a mechanism of action similar to that of paclitaxel, or colcemid, a drug with an opposing mechanism of action and a distinct binding site. Moreover, L215I conferred enhanced paclitaxel sensitivity at very low levels of expression, and sensitivity was not further enhanced in cells with higher levels of expression, implying that paclitaxel acts substoichiometrically. These properties, along with the proximity of L215 to the drug binding site, suggests that the L215I substitution may enhance the binding or effectiveness of paclitaxel. Our studies confirm the importance of the H6-H7 loop of beta-tubulin in microtubule assembly and resistance to antimitotic drugs. They also identify the first mammalian mutation shown to specifically increase sensitivity to paclitaxel.

The multiple drug resistance protein 1 (MRP1 or ABCC1) transports anticancer drugs and normal cell metabolites. Leucotriene C 4 (LTC 4 ) is one of the highest affinity substrates of MRP1. In this study, we have synthesizedand characterized a novel photoreactive azido analogue of LTC 4 (AALTC 4 ). The specificity of AALTC 4 binding to MRP1 was confirmed using an LTC 4 -specific monoclonal antibody. Moreover, binding with radioiodinated [ 1 2 5 I]AALTC 4 (or IAALTC 4 ) to MRP1 was dramatically competed with unmodified LTC 4 and to a lesser degree by glutathione (GSH). Oxidized glutathione (GSSG) slightly increased IAALTC 4 binding to MRP1, while MK571, verapamil, and vincristine inhibited IAALTC 4 binding to MRP1. Using AALTC 4 together with a panel of epitope-specific and LTC 4 -specific monoclonal antibodies, we identified LTC 4 binding sites in MRP1. Western blotting of large tryptic fragments of MRP1 with three well-characterized epitope-specific mAbs (MRPrl, QCRL1, and MRPm6) showed LTC 4 binding in both the N- and C-terminal halves of MRP1. Furthermore, a peptide corresponding to the N-terminal membrane-spanning domain of MRP1 (MSD0) was photoaffinity labeled by AALTC 4 , indicating that MSD0 contains an LTC 4 binding site. Higher resolution mapping of additional LTC 4 binding sites was obtained using eight MRP1 variants with each containing hemaglutanin A (HA) epitopes at different sites (at amino acid 4, 163, 271, 574, 653, 938, 1001, or 1222). MRP1 variants were photoaffinity labeled with IAALTC 4 and digested with trypsin to isolate specific regions of MRP1 that interact with LTC 4 . These results confirmed that sequences in MSD0 interact with IAALTC 4 . Other regions that were photoaffinity labeled by IAALTC 4 include TM 10-11, TM 16-17, and TM 12, shown previously to encode MRP1 drug binding site(s). Together, our results show a high-resolution map of LTC 4 binding domains in MRP1 and provide the first direct evidence for LTC 4 binding within MSD0.

Nonsteroidal anti-inflammatory drugs (NSAIDs) exert their effect by inhibiting the target enzyme cyclooxygenase (prostaglandin H2 synthase); however, little is known about the peptides comprising its NSAID binding site. Hydroxyprostaglandin dehydrogenases also bind NSAIDs, but their NSAID binding sites have not been well characterized. Using existing synthetic strategies, we have incorporated the bromoacetoxy affinity labeling moiety around the perimeter of two potent NSAIDs, indomethacin and mefenamate, a N-phenylanthranilate. The compounds synthesized were 1-(4-(bromoacetamido)benzyl)-5-methoxy-2-methylindole-3-acetic acid (1), 3-(2-(2-bromoacetoxy)ethyl)-1-(4-chlorobenzyl)-5-methoxy-2-methylindole (2), 4-(bromoacetamido)-N-(2,3-dimethylphenyl)anthranilic acid (3), N-(3-(bromoacetamido)phenyl)-anthranilic acid (4), and N-(4-(bromoacetamido)phenyl)anthranilic acid (5). To access whether these compounds have general utility in labeling NSAID binding sites, the compounds were evaluated as affinity labeling agents for 3 alpha-hydroxysteroid dehydrogenase (3 alpha-HSD) from rat liver cytosol. This enzyme displays 9-, 11-, and 15-hydroxyprostaglandin dehydrogenase activity, is inhibited potently by NSAIDs, and is homologous to bovine lung prostaglandin F synthase. Compounds 1-5 were shown to affinity label the NSAID binding site of 3 alpha-HSD. They inactivated 3 alpha-HSD through an E.I complex in a time- and concentration-dependent manner with t1/2 values ranging from seconds to hours. Ligands that compete for the active site of 3 alpha-HSD (NAD+ and indomethacin) afforded protection against inactivation, and the inactivators could demonstrate competitive kinetics against 3 alpha-hydroxysteroid substrates by forming an E.NAD+.I complex.(ABSTRACT TRUNCATED AT 250 WORDS)

One cause of multidrug resistance is the overexpression of P-glycoprotein, a 170 kDa plasma membrane ABC transporter, which functions as an ATP-driven efflux pump with broad specificity for hydrophobic drugs, peptides, and natural products. The protein appears to interact with its substrates within the membrane environment. Previous reports suggested the existence of at least two binding sites, possibly overlapping and displaying positively cooperative interactions, termed the H and R sites for their preference for Hoechst 33342 and rhodamine 123, respectively. In this work, we have used several fluorescence approaches to characterize the molecular interaction of purified P-glycoprotein (Pgp) with the dye LDS-751, which is proposed to bind to the R site. A 50-fold enhancement of LDS-751 fluorescence indicated that the protein binding site was located in a hydrophobic environment, with a polarity lower than that of chloroform. LDS-751 bound with sub-micromolar affinity (Kd ) 0.75 IM) and quenched P-glycoprotein intrinsic Trp fluorescence by 40%, suggesting that Trp emitters are probably located close to the drub-binding regions of the transporter and may interact directly with the dye. Using a FRET approach, we mapped the possible locations of the LDS-751 binding site relative to the NB domain active sites. The R site appeared to be positioned close to the membrane boundary of the cytoplasmic leaflet. The location of both H and R drug binding sites is in agreement with the idea that Pgp may operate as a drug flippase, moving substrates from the inner leaflet to the outer leaflet of the plasma membrane. 1 Abbreviations: ABC, ATP-binding cassette; AMP-PNP, 5'-ade- nylylimidodiphosphate; CHAPS, 3-((3-cholamidopropyl)dimethylam- monio)-1-propanesulfonate; DMSO, dimethyl sulfoxide; H33342, Ho- echst 33342; FRET, fluorescence resonance energy transfer; GuHCl, guanidine hydrochloride; LDS-751, 2-{4-(4-(dimethylamino)phenyl)- 1,3-butadienyl}-3-ethylbenzothiazolium perchlorate; MDR, multidrug resistant or resistance; MIANS, 2-(4-maleimidoanilino)naphthalene-6- sulfonic acid; NATA, N-acetyltryptophanamide; NB, nucleotide bind- ing; NBD-Cl, 7-chloro-4-nitrobenz-2-oxa-1,3-diazole; Pgp, P-glyco- protein; TM, transmembrane; TNP-ATP, 2'(3')-O-(2,4,6-triphenyl)- adenosine 5'-triphosphate. 643

MRP1 is a 190 kDa membrane glycoprotein that confers multidrug resistance (MDR) to tumor cells. Our recent study demonstrated that GSH is required for the labeling of MRP1 9 3 2 - 1 5 3 1 with a photoanalogue of agosterolA (AG-A) and suggested that GSH interacts with the L 0 region of MRP1. In this study, we further characterized the GSH-dependent binding site of azido AG-A on MRP1. Coexpression of the N- and C-terminal halves of MRP1 (residues 1-1222, TM1-16, and 1223-1531, TM17, respectively) in Sf21 insect cells reconstituted a functional drug transporter with a K m for LTC 4 (97 nM) similar to that of intact MRP1. In membrane vesicles from those cells, GSH-dependent photolabeling of the MRP1 fragment (1-1222) required the coexpression of the C-terminal MRP1 fragment (1223-1531). An MRP1 fragment extending from residue 1 to 1295 however could be photolabeled by azido AG-A in a GSH-dependent manner. These data indicate that amino acids 1223-1295 of MRP1 are required for AG-A binding to MRP1 in a GSH-dependent manner. However, cross-linking of the photolabel to MRP1 occurs at a more upstream site. An arginine residue at position 1249 of MRP1 was shown to be important for the GSH-dependent binding of AG-A to MRP1. Mutation of this arginine to alanine (R1249A) resulted in a decreased level of GSH-dependent azido AG-A photolabeling of MRP1. Furthermore, this mutant attenuated MRP1 function by decreasing the level of LTC 4 substrate transport and impairing resistance to the drug vincristine (VCR). In summary, this study demonstrates that a region of MRP1 (amino acids 1223-1295), which includes TM helix 17, is required for azido AG-A binding to MRP1 in a GSH-dependent manner. A GSH-dependent drug binding site may exist in this region. Furthermore, our findings suggest that the charged amino acid Arg 1 2 4 9 proximal to the C-terminus of TM helix 17 is indispensable for MRP1-substrate interaction and the function of MRP1.

The P-glycoprotein multidrug transporter carries out ATP-driven cellular efflux of a wide variety of hydrophobic drugs, natural products, and peptides. Multiple binding sites for substrates appear to exist, most likely within the hydrophobic membrane spanning regions of the protein. Since ATP hydrolysis is coupled to drug transport, the spatial relationship of the drug binding sites relative to the ATPase catalytic sites is of considerable interest. We have used a fluorescence resonance energy transfer (FRET) approach to estimate the distance between a bound substrate and the catalytic sites in purified P-glycoprotein. The fluorescent dye Hoechst 33342 (H33342), a high-affinity P-glycoprotein substrate, bound to the transporter and acted as a FRET donor. H33342 showed greatly enhanced fluorescence emission when bound to P-glycoprotein, together with a substantial blue shift, indicating that the drug binding site is located in a nonpolar environment. Cys428 and Cys1071 within the catalytic sites of P-glycoprotein were covalently labeled with the acceptor fluorophore NBD-Cl (7-chloro-4-nitrobenz-2-oxa-1,3-diazole). H33342 fluorescence was highly quenched when bound to NBD-labeled P-glycoprotein relative to unlabeled protein, indicating that FRET takes place from the bound dye to NBD. The distance separating the bound dye from the NBD acceptor was estimated to be approximately 38 A. Transition-state P-glycoprotein with the complex ADP*orthovanadate*Co2+ stably trapped at one catalytic site bound H33342 with similar affinity, and FRET measurements led to a similar separation distance estimate of 34 A. Since previous FRET studies indicated that a fluorophore bound within the catalytic site was positioned 31-35 A from the interfacial region of the bilayer, the H33342 binding site is likely located 10-14 A below the membrane surface, within the cytoplasmic leaflet of the membrane, in both resting-state and transition-state P-glycoprotein.

Conceptually one may envisage that substrate binding sites on the ABC transporter P-gp cycle between high- and low-affinity conformations in response to signals arising from nucleotide hydrolysis to effect active transport. A radioligand binding assay was used to characterize the interaction of [3H]vinblastine with P-gp and determine how drug binding site parameters are altered during a catalytic cycle of P-gp. In the absence of nucleotide, we show that [3H]vinblastine interacts with a single class of binding site with high affinity (K(d) = 80 +/- 18 nM). In the presence of the nonhydrolyzable ATP analogue AMP-PNP, the drug binding site was in a low-affinity conformation, manifest by a 9-fold increase in K(d) (K(d) = 731 +/- 20 nM). There was no alteration in the binding capacity, reflecting a complete shift in the high-affinity site to a low-affinity form. The posthydrolytic (Mg-ADP-V(i) bound) form of P-gp also exhibited low-affinity substrate binding (K(d) = 446 +/- 57 nM). Restoration of the high-affinity drug binding site conformation (K(d) = 131 +/- 32 nM) did not occur until release of phosphate from the posthydrolysis P-gp-Mg-ADP-P(i). complex. Our results suggest that alteration of the affinity of the vinblastine binding site involves only one nucleotide binding domain per transport cycle. The binding of ATP provides the signal to instigate this change, while release of phosphate post-ATP hydrolysis returns the transporter to its original state to complete the cycle.

P-Glycoprotein is a member of the ABC superfamily of membrane transporters, and functions as an ATP-driven active efflux pump for natural products and chemotherapeutic drugs. Overexpression of P-glycoprotein is a major cause of multidrug resistance in human cancers. Sulfhydryl modification agents are known to inactivate both P-glycoprotein ATPase activity and transport function. In the present study, P-glycoprotein purified from CHRB30 cells was covalently labeled at two conserved Cys residues, one within each of the nucleotide binding domains, using 2-(4-maleimidoanilino)naphthalene-6-sulfonic acid (MIANS). MIANS modification inactivated P-glycoprotein ATPase function, in a concentration-dependent fashion. Increasing concentrations of ATP blocked MIANS labeling with an IC50 of 0.37 mM (similar to the KM for ATP hydrolysis), which suggests that the label is located close to the site of ATP binding within the nucleotide binding domain. A blue shift in the fluorescence spectrum of MIANS bound to P-glycoprotein indicated that the labeled Cys residues are situated in a nonpolar environment. MIANS-labeled P-glycoprotein was still able to bind ATP, as demonstrated by quenching of the fluorescence, with a Kd of 0.46 mM. Addition of a variety of drugs and chemosensitizers to MIANS-labeled P-glycoprotein led to substantial quenching of the probe fluorescence within the nucleotide binding domains. Dissociation constants for drug binding measured by fluorescence quenching were in the range of 0.77 microM for vinblastine to 158 microM for colchicine. Quenching by ATP and drugs was independent and additive, suggesting that each produces a defined change in the protein. The rate of MIANS labeling of Pgp was reduced in the presence of drugs and chemosensitizers, implying that a long-range conformational change arises from drug binding which alters the accessibility of the nucleotide binding domains to MIANS. These results suggest that there is conformational communication between the drug binding site(s) of P-glycoprotein and the ATPase catalytic sites within the nucleotide binding domains.

The biochemical and genetic analyses of P-glycoprotein (P-gp) have indicated that the membrane-associated regions of P-gp play an important role in drug recognition and drug transport. Predicted transmembrane domain 11 (TM11) maps near a major drug binding site revealed by photoaffinity labeling, and mutations in this domain alter the substrate specificity of P-gp. To investigate further the role of TM11 in P-gp function in general, and substrate specificity in particular, each of the 21 residues of TM11 of the P-gp isoform encoded by the mouse mdr3 gene was independently mutated to alanine, or to glycine in the case of endogenous alanines. After transfection and overexpression in Chinese hamster ovary cells, pools of stable transfectants were analyzed for qualitative or quantitative deviations from the profile of resistance to vinblastine, adriamycin, colchicine, and actinomycin D displayed by the wild-type protein. While mutations at eight of the positions had no effect on P-gp function, 13 mutants showed a 2-10-fold reduction of activity against one of the four drugs tested. Although the phenotype of individual mutants was varied, replacements at most mutation-sensitive positions seemed to affect the drug resistance profiles rather than the overall activity of the mutant P-gp. When TM11 was projected in a alpha-helical configuration, the distribution of deleterious and neutral mutations was not random but segregated with a more hydrophobic (mutation-insensitive) face and a more hydrophilic (mutation-sensitive) face of a putative amphipathic helix. The alternate clustering pattern of deleterious vs neutral mutations in TM11 together with the altered drug resistance profile of deleterious mutants suggest that the more hydrophilic face of the TM11 helix may play an important structural or functional role in drug recognition and transport by P-gp. Finally, the conservation of the two residues most sensitive to mutations (Y949 and Y953) in TM11, and in the homologous TM5, of all mammalian P-gps and also in other ABC transporters, suggests that these residues and domains may play an important role in structural as well as mechanistic aspects common to this family of proteins.

The substitution of a single serine to phenylalanine residue within the predicted transmembrane domain 11 of P-glycoproteins (P-gps) encoded by mouse mdr1 (Ser941, 1S;Phe941, 1F) or mdr3 (Ser939, 3S; Phe939, 3F) strongly modulates both the overall activity and substrate specificity of the two P-gps. In cell clones expressing either wild-type (1S, 3S) or mutant P-gps (1F, 3F), we show that the modulating effect of the mutation on the levels of adriamycin (ADM) resistance detected in drug cytotoxicity assays is paralleled by a similar modulation of the intracellular accumulation and extracellular efflux of radiolabeled adriamycin ([14C]ADM) from preloaded cells. Cytofluorescence studies with ADM on live cells produce similar results and demonstrate strong nuclear ADM accumulation only in drug-sensitive LR cells and in the 1F expressing cells, with little if any accumulation in 1S, 3S, or 3F expressing cells. Drug cytotoxicity and drug transport assays carried out in the presence of verapamil or progesterone suggest that the Ser to Phe substitution also reduces the capacity of these two reversal agents to modulate P-gp activity. Labeling experiments with the photoactivatable P-gp ligands iodoarylazidoprazosin and azidopine indicate a strong reduction in binding of these photoactivatable probes to the mutant P-gps (1F, 3F) as compared to their wild-type counterparts (1S,3S). These results indicate that the studied mutations in TM11 reduce drug transport by decreasing initial drug binding to P-gp. This phenotype is opposite to that of a mutation near TM3 in human MDR1 (pst 185), where decreased drug transport is associated with increased drug binding and decreased drug release from P-gp [Safa, A. R., Stern, R. K., Choi, K., Agresti, M., Tamai, I., Metha, N. D., & Roninson, I. B. (1990) Proc. Natl. Acad. Sci. U.S.A. 87, 7225-7229].

Quantitative footprinting methods have been used to measure site-specific binding constants for the anticancer drug actinomycin D on a restriction fragment from pBR 322 DNA. Sites with the highest binding constants, 1–6 x 105M-1, possess the classical binding sequence GC, which is flanked on the 5’ side by either T or C, but not G. The sequence GGC is a weaker binding site, except for one case. A non-classical sequence, CGTC, was found to be a strong actinomycin D site. However, the value of the binding constant for this site and for other classical sites was dependent on the proximity of sites on the 139-mer and/or the nature of flanking sequences. Comparison with earlier studies indicates that the length of DNA may be important, with short oligomers yielding higher binding constants than polymeric DNA. The analysis also revealed that actinomycin D can cause anomalous rate enhancements adjacent to certain of its interaction sites when binding is probed with DNase I and the metalloporphyrin, MnT4MPyP, but not with Fe-MPE. This indicates that regions adjacent to drug binding may have altered minor groove width and/or exist in a partially melted state.

In order to improve the design of HIV-1 protease inhibitors, it is essential to understand how they interact with active site residues, particularly the catalytic Asp25 and Asp125 residues. KNI-272 is a promising, potent HIV-1 protease inhibitor (K(i) approximately 5 pM), currently undergoing phase 1 clinical trials. Because KNI-272 is asymmetric, the complex it forms with the homodimeric HIV-1 protease also lacks symmetry, and the two protease monomers can have distinct NMR spectra. Monomer specific signal assignments were obtained for amino acid residues in the drug binding site as well as for six of the eight Asp residues in the protease/KNI-272 complex. Using these assignments, the ionization states of the Asp carboxyl groups were determined from measurements of (a) the pD dependence of the chemical shifts of the Asp carboxyl carbons and (b) the H/D isotope effect upon the Asp carboxyl carbon chemical shifts. The results of these measurements indicate that the carboxyl of Asp25 is protonated while that of Asp125 is not protonated. These findings provide not only the first experimental evidence regarding the distinct protonation states of Asp25/125 in HIV-1 protease/drug complexes, but also shed light on interactions responsible for inhibitor binding that should form the basis for improved drug designs.