ABCMdb

PMID: 14596601

Carrier I, Julien M, Gros P
Analysis of catalytic carboxylate mutants E552Q and E1197Q suggests asymmetric ATP hydrolysis by the two nucleotide-binding domains of P-glycoprotein.
Biochemistry. 2003 Nov 11;42(44):12875-85., 2003-11-11 [PubMed]

Sentences

No. Mutations Sentence Comment

48 ABCB3 p.Glu552Gln In an effort to gain insight into the mechanism of ATP hydrolysis by P-gp, including testing the role of the two NBDs in catalysis, we have previously mutated the glutamate residues homologous to E179 of HisP in the mouse Mdr3 enzyme (E552Q and E1197Q in NBD1 and NBD2, respectively) (43). Login to comment 51 ABCB3 p.Glu552Gln In this study, we have attempted to characterize the molecular basis of the defect in the E552Q and E1197Q mutants. Login to comment 53 ABCB3 p.Glu552Gln The wild-type mouse Mdr3 (WT) and the E552Q, E1197Q, and D551N mutants were created by site-directed mutagenesis and modified by in-frame addition of a six-histidine tag (His6) at the C-terminus of the protein and were expressed in the yeast Pichia pastoris after cloning in the expression plasmid pHIL-D2 (Invitrogen, license 145457), as previously described (43). Login to comment 108 ABCB3 p.Glu552Gln RESULTS During a previous search for catalytic carboxylate residues in the NBDs of P-glycoprotein two mutants of the mouse Mdr3 isoform at homologous positions in NBD1 and NBD2 (E552Q and E1197Q) showed a similar loss-of-function phenotype, which featured inability to convey multidrug resistance and abrogation of steady-state ATPase activity (43). Login to comment 110 ABCB3 p.Glu552Gln For this, wild-type (WT) Mdr3 and the E552Q and E1197Q mutants were expressed in the yeast P. pastoris as recombinant proteins bearing an in-frame polyhistidine tail (His6) at the carboxyl terminus. Login to comment 116 ABCB3 p.Glu552Gln As before, the purified E552Q and E1197Q mutants show a very low basal ATPase activity (0.13-0.18 µmol min-1 mg-1 ) that is not stimulated by drugs (0.13-0.20 µmol min-1 mg-1 ) and which is comparable to the activity seen in the ATPase inactive mutant D551N (49, 51) that is considered to be background. Login to comment 117 ABCB3 p.Glu552Gln To determine whether the loss of ATPase activity seen in E552Q and E1197Q was caused by an effect of the mutations on affinity for nucleotides, the nucleotide-binding properties of the WT and Mdr3 variants were compared by photoaffinity labeling. Login to comment 123 ABCB3 p.Glu552Gln Together, these results suggest that the E552Q, E1197Q, and D551N mutations do not have a major effect on nucleotide binding to Mdr3 and are therefore unlikely to cause major nonspecific structural changes in the NBDs. Login to comment 124 ABCA4 p.Lys429Arg This is in agreement with previous studies of catalytic residue mutants of the Walker A and B signature motifs (K429R, K1072R, D551N, and D1196N) which severely affect the catalytic activity of mouse Mdr3 but have little effect on the nucleotide-binding affinity of the protein (49). Login to comment 133 ABCB3 p.Glu552Gln Purified and activated wild-type (WT) and mutant Mdr3 variants (E552Q, E1197Q, D551N) were UV-irradiated on ice in the presence of 3 mM MgCl2 and 5, 10, 20, 40, and 80 µM 8-azido-[R-32P]ATP. Login to comment 137 ABCB3 p.Glu552Gln Despite the observed lack of ATPase activity of E552Q and E1197Q measured by Pi release, 8-azidonucleotide trapping that is stimulated both by drug and by Vi is readily detectable in these mutants. Login to comment 140 ABCB3 p.Glu552Gln Although studies of WT Mdr3 [Figure 1 and (43)] and of the inactive mutant D551N [Figure 1 and (43)] suggest that under the hydrolysis conditions used (37 °C; see Experimental Procedures) little if any of the photolabeling is due to 8-azido-[R-32 P]ATP binding, additional experiments were undertaken to verify that the labeling seen in E552Q and E1197Q ((Vi) was due to trapping of hydrolyzed nucleotide, as opposed to simple binding of the label to Mdr3. Login to comment 141 ABCB3 p.Glu552Gln As formation of the Mg-8-azido-ADP‚Vi complex is optimum at 37 °C and requires Mg2+ ions (28, 53), the temperature dependence and EDTA sensitivity of E552Q and E1197Q photolabeling by 8-azido-[R-32 P]ATP were investigated. Login to comment 142 ABCB3 p.Glu552Gln Results in Figure 2A,B show that labeling of WT, E552Q, and E1197Q by 8-azido-[R-32 P]ATP under all conditions tested ((Vi, (drugs) was either completely eliminated or largely reduced (>90%) when the incubation and washing steps of the labeling reaction were carried out at 4 °C (Figure 2B) as opposed to 37 °C (Figure 2A). Login to comment 145 ABCB3 p.Glu552Gln These results are consistent with 8-azido-[R-32 P]ATP hydrolysis in the WT and E552Q and E1197Q mutants with concomitant trapping of 8-azido-[R-32 P]ADP. Login to comment 152 ABCB3 p.Glu552Gln To obtain further evidence that the E552Q and E1197Q mutants can form the {MgADP‚Vi} transition-state complex, similar nucleotide trapping experiments were carried out in the presence of two other transition-state analogues: aluminum fluoride (AlF4 - ) and beryllium fluoride (BeFx). Login to comment 157 ABCB3 p.Glu552Gln Results in Figure 3A indicate that AlF4 - can also induce nucleotide trapping in the WT and E552Q and E1197Q mutants, with characteristics similar to those observed when Vi is used to induce trapping (43). Login to comment 160 ABCB3 p.Glu552Gln These results show that the different transition states revealed by distinct Pi analogues can all be formed in the E552Q and E1197Q mutants in a comparable manner to WT Mdr3. Login to comment 161 ABCB3 p.Glu552Gln Finally, results in Figure 3C show that BeFx- and AlF4 - -induced nucleotide trapping in the WT and E552Q and E1197Q mutants is EDTA sensitive. Login to comment 162 ABCB3 p.Glu552Gln Together, these results expand observations with Vi [Figure 3 (43)], showing that in the mutants E552Q and E1197Q activation and cleavage of the bond between the β- and γ-phosphates of ATP seem to take place. Login to comment 164 ABCB3 p.Glu552Gln As seen in Figure 4, 8-azido- [R-32 P]ADP can be detected following incubation of the WT and E552Q and E1197Q mutants with 8-azido-[R-32 P]ATP and vanadate, while it is absent in the catalytically inactive D551N mutant. Login to comment 165 ABCB3 p.Glu552Gln Furthermore, formation of ADP does not occur in any of the enzymes when the trapping reaction is carried out at 4 °C. Thus, it appears that the E552Q and E1197Q mutants are able to hydrolyze at least one molecule of ATP. Login to comment 166 ABCB3 p.Glu552Gln In these experiments, we also detected the presence of ATP in both WT and the E552Q and E1197Q mutants. Login to comment 182 ABCB3 p.Glu552Gln ABCB3 p.Glu552Gln The nature of the molecular defect in the E552Q and E1197Q mutants was further investigated by comparing the Vi dependence of nucleotide trapping of the WT and the E552Q and E1197Q mutants in a dose-response experiment (0, 0.05 µM e Vi e 100 µM) (Figure 5). Login to comment 184 ABCB3 p.Glu552Gln As expected from the results in our previous publication (43), both E552Q and E1197Q could trap 8-azido-[R-32 P]nucleotide at all Vi concentrations tested. Login to comment 186 ABCB3 p.Glu552Gln E552Q showed low levels of trapping in absence of Vi, but labeling increased in a dose-dependent fashion (similar to the WT enzyme) with very intense labeling seen at 100 µM Vi (>50-fold stimulation). Login to comment 188 ABCB3 p.Glu552Gln ABCB3 p.Glu552Gln The distinct Vi dose-response behaviors of the two homologous mutants E552Q and E1197Q suggest that NBD1 and NBD2 are not catalytically symmetrical, with NBD2 (intact in E552Q) showing a more robust Vi- dependent trapping than its NBD1 counterpart (intact in E1197Q). Login to comment 189 ABCB3 p.Glu552Gln To further investigate a possible functional asymmetry between NBD1 and NBD2, suggested by the Vi dose-response study (Figure 5), we attempted to semiquantitatively assess in which of the NBD(s) of E552Q and E1197Q the nucleotide was trapped, in both the absence and presence of Vi. Login to comment 194 ABCB3 p.Glu552Gln For E1197Q, results in panels F and H clearly show that all of the label is in the MD7-reactive, C-terminal half of the protein, with little, if any, overlap with the MD13 reactive species detected in panel D. Conversely, and although the overall photolabeling signal is much weaker than that seen for E1197Q, the label incorporated in E552Q colocalizes with the N-terminal and MD13-reactive half (panels D and H) and does not overlap with the faster migrating MD7 positive fragment (panel F). Login to comment 204 ABCB3 p.Glu552Gln In the presence of Vi, nucleotide trapping was seen in both NBD1 and NBD2 of the E552Q and E1197Q mutants (albeit at different ratios, depending on the position of the mutation). Login to comment 206 ABCB3 p.Glu552Gln These results indicate that, in WT Mdr3 and the E552Q and E1197Q mutants, both NBDs can hydrolyze at least one FIGURE 6: Trypsin digestion of Mdr3 NB site mutants photolabeled with Mg-8-azido-[R-32P]ATP in the absence of vanadate. Login to comment 221 ABCB3 p.Glu552Gln Results expressed in the present study suggest that the E552Q and E1197Q mutants are not completely inactive (as opposed to the Walker B mutant D551N) and that these mutants can indeed cleave ATP to ADP and Pi, undergoing partial reactions toward a full cycle of catalysis but never fully turning over. Login to comment 227 ABCB3 p.Glu552Gln Finally, TLC analysis of the nucleotides bound to the enzymes following Vi trapping of 8-azido-ATP shows formation of 8-azido-ADP by the WT and E552Q and E1197Q mutants (Figure 4). Login to comment 228 ABCB3 p.Glu552Gln Together, these results indicate that E552Q and E1197Q can indeed cleave ATP to ADP and Pi and that 8-azido-ADP is the nucleotide trapped in the photolabeled enzymes. Login to comment 229 ABCB1 p.Glu1201Gln ABCB1 p.Glu556Gln ABCB1 p.Glu1201Ala ABCB1 p.Glu556Ala Moreover, in their recent work, Sauna and colleagues also demonstrated using Rand γ-labeled 8-azido-[R-32 P]ATP that mutants at the equivalent positions of the human MDR1 protein (E556Q and E556A, E1201Q and E1201A, and the double mutants) are indeed capable of ATP hydrolysis and single-site catalysis (59). Login to comment 230 ABCB3 p.Glu552Gln Therefore, in the E552Q and E1197Q mutants, steps downstream from the formation of the transition state, such as the release of MgADP and/or Pi, or others, must be impaired. Login to comment 234 ABCB1 p.Glu556Ala ABCC4 p.Glu1202Ala Indeed, they suggest that the single E556A/Q and E1202A/Q mutants are defective in the second ATP hydrolysis event of the catalytic cycle, which should reset the protein after hydrolysis at the first site. Login to comment 243 ABCB3 p.Glu552Gln The study of E552Q and E1197Q reported here clearly argues in favor of two NBDs that are not functionally equivalent in full-length P-gp. Login to comment 245 ABCB3 p.Glu552Gln ABCB3 p.Glu552Gln In this experiment it can be observed that E552Q and E1197Q trap nucleotide in the absence of Vi and that the response of each mutant to Vi is completely different, with E552Q showing a strongly dose-dependent increase in labeling (similar to WT), while Vi has little effect on nucleotide trapping in E1197Q. Login to comment 247 ABCB3 p.Glu552Gln Since nucleotide trapping in the mutants occurs after cleavage of ATP (with ADP trapped; see above), the differential Vi dose-response observed in E552Q and E1197Q is most easily explained by differential sensitivity of the nonmutant site to inhibition by Vi. Login to comment 252 ABCB3 p.Glu552Gln In summary, the present work supports the finding that both NBDs are essential for function with complete cooperativity between the two sites and suggests that the E552 and E1197 residues of mouse Mdr3 are probably not the catalytic residues, as the E552Q and E1197Q mutants can hydrolyze ATP in both nucleotide-binding domains. Login to comment