Reaction chemistry ABC-style

For a time, membrane proteins proved intractable to X-ray crystallography. There were difficulties of overexpression and purification, multiple choices of solubilizing detergent that could only be decided by trial and error, unwillingness to crystallize, and often, poor diffraction from resultant crystals. However, the tide now seems to be turning, and in PNAS, Oldham and Chen (1) present an X-ray crystallography study of the entire WT maltose uptake protein from Escherichia coli, in which four different structures are described at commendable 2.2- to 2.4-Å resolution. Maltose uptake protein is a well-characterized model (2) for the large ABC transporter family, members of which occur in all living organisms. ABC transporters, whether exporters or importers, display a common architecture consisting of two transmembrane domains (TMDs) involved directly in transport and two nucleotide binding domains (NBDs) that cooperatively hydrolyze ATP (actually MgATP) to power transport (3). Internationally, numerous researchers study ABC transporters because of their wide involvement in human and plant disease and consequent importance in medicine and agriculture. Maltose uptake protein has subunit structure MalGFK₂, with GF forming the TMDs and K₂ forming the head to tail NBD dimer, with its two ATP binding sites. Bacterial ABC importers use a binding protein that captures the requisite transport substrate from the periplasm and delivers it to the transporter. In the work by Oldham and Chen (1), maltose binding protein is bound to the MalGFK₂ complex in its open state, and the just-delivered maltose molecule is clearly seen in the transport pathway at midmembrane.

ABC Reaction Possibilities

The work by Oldham and Chen (1) focuses on defining a specific facet of mechanism certain to be common to all ABC transporters, namely, the chemical reaction mechanism of ATP hydrolysis. ABC transporters are defined by several highly conserved consensus amino acid sequences within the NBDs. They are the Walker A and Walker B, the LSGGQ or signature sequence, the D loop, the Q loop, and the switch region or H loop. We can confidently state that all ABC transporters use the same ATPase pathway, because earlier X-ray structures of isolated soluble NBDs and more recently, entire transporters, from varied eukaryotic and prokaryotic sources, showed that the consensus sequences interact closely with bound ATP. Three hypothetical reaction mechanisms have been considered. The first and most popular mechanism is general base catalysis promoted by a conserved glutamate (Glu) occurring at the end of the Walker B sequence. Suggested by the first X-ray structure of an isolated NBD (4), this idea gained traction from subsequent mutagenesis of the Glu residue (5, 6). In this associative mechanism, attack by polarized water on the γ-phosphorus of ATP forms a pentacovalent

Oldham and Chen present an X-ray crystallography study of the entire WT maltose uptake protein from Escherichia coli.

phosphorus transition state intermediate with trigonal bipyramidal geometry. Subsequent cleavage of the ADP-O to P bond generates ADP and Pi (figure 2a of ref. 1). The Glu carboxyl directly H-bonds to the attacking water, polarizes it, and accepts a proton in the process. The second mechanism, suggested in ref. 7, is substrate-assisted catalysis, in which the attacking water is polarized by the γ-phosphate of ATP and the conserved histidine (His) of the switch or H-loop sequence acts to stabilize the transition state. A third mechanistic possibility is that the conserved serine (Ser) in the LSGGQ sequence plays a direct catalytic role. In myosins and kinesins catalytic sites, no potential catalytic base is found in suitable proximity to ATP. This finding led to proposals that a conserved active site Ser directly bonds and stabilizes the attacking water (8, 9).

Answers from Structures

Metallofluorides and orthovanadate (Vi) have been used extensively in biochemical and structural studies of ATPase enzymes. They form complexes with ADP that bind tenaciously at catalytic sites, thus producing very stable and inactive enzyme states. Vi and aluminum tetrafluoride form complexes with ADP (trigonal bipyramidal in ADP-Vi and octahedral in ADP-AlF₄⁻) in which the ADP-O to V or Al apical bond is longer by about 0.4 Å than the corresponding ADP-O to P bond in ATP. Therefore, these complexes mimic the ATPase transition state. Beryllium fluoride forms a complex with ADP that, with tetrahedral geometry, mimics the ATP-bound ground state. The work by Oldham and Chen (1) presents structures of maltose uptake protein with each of these three complexes bound in catalytic sites and a fourth structure in which the nonhydrolysable analog of ATP, called AMPPNP, is bound. Together, the four structures give clear insight into the role of the consensus sequences and the mechanism of ATP hydrolysis. The ADP-AlF₄⁻ structure shows an appropriately positioned attacking water bonded to Glu159, the putative general base. The ADP-Vi structure shows the same Glu directly bonded to a Vi apical oxygen in the expected position of the attacking water. Notably, this Glu residue does not interact directly with substrate in either ground state (AMPNP or ADP-BeF₃) structure. Neither the His of the H loop nor the Ser of the LSGGQ sequence is directly bonded to the attacking water, although both residues interact with substrate in the ground and transition states. Therefore, we have a clear answer. ATP hydrolysis in the ABC transporter family uses general base catalysis, with the Glu residue at the end of the Walker B sequence binding to and accepting a proton from the attacking water. Substrate-assisted catalysis (7) and Ser-assisted catalysis (8, 9) are eliminated. Thus, ABC transporters resemble recA (10) and F₁F₀-ATPase (11) in mechanism, and they differ from myosins and kinesins (and also, G proteins).

The consensus sequences are seen to immobilize and orient ATP in the ground and transition states, forming a network of H-bond and electrostatic interaction with the substrate and with each other, which results in optimal placement for catalysis. In all four structures, both sites seem similar. It may be noted, however, that some ABC transporters, notably but not exclusively in the ABCC family, have one ATP site that contains atypical consensus sequences (12, 13), extending, in some cases, to lack of the catalytic Glu. Such sites clearly cannot carry out catalysis through the normal pathway. Isolated NBDs, which usually have low, nonphysiological ATPase rates, probably fall into this category as well, because they lack structural coupling to TMDs through the Q loop, which is shown in the work by Oldham and Chen (1) to be involved in binding both ground and transition states through the Mg of MgATP and a γ-phosphate oxygen.

A second major outcome of this study is that all four structures are generally similar to each other, indicating that no major conformational rearrangement accompanies the ground state to transition state conversion. Because it is established that ABC transporters undergo reorientation of the transport pathway from extracellular-facing with ATP-bound NBDs to inward-facing with ADP-bound or empty NBDs (2, 3), we may conclude that product release, specifically, Pi-release, is geared to a conformational change of the transport pathway, which was previously suggested (14, 15). Of relevance, in the F₁F₀-ATPase rotary mechanism, Pi-release generates 40° of rotation during each 120° step (16).

A discrepancy that Oldham and Chen (1) note is that ADP-Vi and ADP-metallofluorides are seen bound in both catalytic sites (i.e., 2 mol/mol transporter), whereas previous biochemical experiments detected only one ADP-Vi (mol/mol). The discrepancy is germane to the important question, which is still unsettled, of whether one or two molecules of ATP are hydrolyzed per transport cycle. They discuss two explanations. First, maybe occupation by Vi in the crystal structure is only 0.5 at each catalytic site, which would be consistent with an alternating catalytic sites mechanism (14). Second, partial dissociation of ADP-Vi during separation of protein from free ligand before analysis of bound ADP-Vi in biochemical experiments might yield underestimated stoichiometry. The crystals were obtained in high concentrations of ligand, perhaps sufficient to keep both sites filled. In the multidrug resistance-conferring P-glycoprotein, another well-researched ABC transporter, it was found that protein with ADP-Vi trapped at just one site is able to bind nucleotides at the other site (17), and the two sites have different affinities. In a catalytically arrested mutant (18) or using a slowly hydrolyzed ATP analog (17), it was seen that, after initial binding of ATP at both catalytic sites, one site became occluded and seemingly committed to hydrolysis, implying that only one site enters the transition state at any one time. However, it also implies an asymmetry not evident in the work by Oldham and Chen (1). This question, therefore, remains one for future experiments.

For the Future

Finally, lest we miss the forest for the trees, let me emphasize that this work has generated a large amount of atomic level structural data, essential knowledge for biophysicists and biochemists eager to understand mechanism in detail. I hope within this decade to watch real-time videos of single molecules of ABC transporters hydrolyzing ATP and transporting substrate across the membrane. Perhaps transport substrate pathway specificity can be modified to generate organisms that can take up pollutants or use waste products to make new fuels. Maybe multidrug resistance in pathogens or human cancer can be overcome by designing new small molecules that bind and disrupt the NBD dimer interface or transport pathway. Whatever are your favorite dreams, this work surely provides information foundational for such future accomplishments. I urge you to turn the pages or click the mouse, go to the snappy title, and read the real paper.