J.M. East

Artemisinins target the SERCA of Plasmodium falciparum

Artemisinins are extracted from sweet wormwood (Artemisia annua) and are the most potent antimalarials available, rapidly killing all asexual stages of Plasmodium falciparum. Artemisinins are sesquiterpene lactones widely used to treat multidrug-resistant malaria, a disease that annually claims 1 million lives. Despite extensive clinical and laboratory experience their molecular target is not yet identified. Activated artemisinins form adducts with a variety of biological macromolecules, including haem, translationally controlled tumour protein (TCTP) and other higher-molecular-weight proteins. Here we show that artemisinins, but not quinine or chloroquine, inhibit the SERCA orthologue (PfATP6) of Plasmodium falciparum in Xenopus oocytes with similar potency to thapsigargin (another sesquiterpene lactone and highly specific SERCA inhibitor). As predicted, thapsigargin also antagonizes the parasiticidal activity of artemisinin. Desoxyartemisinin lacks an endoperoxide bridge and is ineffective both as an inhibitor of PfATP6 and as an antimalarial. Chelation of iron by desferrioxamine abrogates the antiparasitic activity of artemisinins and correspondingly attenuates inhibition of PfATP6. Imaging of parasites with BODIPY-thapsigargin labels the cytosolic compartment and is competed by artemisinin. Fluorescent artemisinin labels parasites similarly and irreversibly in an Fe2+-dependent manner. These data provide compelling evidence that artemisinins act by inhibiting PfATP6 outside the food vacuole after activation by iron.

Annular and non-annular binding sites on the (Ca2+ + Mg2+)-ATPase.

Quenching of the fluorescence of the (Ca2+ + Mg2+)-ATPase purified from muscle sarcoplasmic reticulum can be used to measure relative binding constants of hydrophobic compounds to the phospholipid-protein interface. We show that the binding constant for cholesterol is considerably less than that for phosphatidylcholine, so that cholesterol is effectively excluded from the phospholipid annulus around the ATPase. However, dibromocholestan-3 beta-ol causes quenching of the fluorescence of the ATPase, and so has access to other, non-annular sites. We suggest that these non-annular sites could be at protein/protein interfaces in ATPase oligomers. Oleic acid can bind at the phospholipid/protein interface, although its binding constant is less than that for a phosphatidylcholine, and it can also bind at the postulated non-annular sites. The effects of these compounds on the activity of the ATPase depend on the structure of the phospholipid present in the systems.

What the structure of a calcium pump tells us about its mechanism

The report of the crystal structure of the Ca(2+)-ATPase of skeletal muscle sarcoplasmic reticulum in its Ca(2+)-bound form [Toyoshima, Nakasako and Ogawa (2000) Nature (London) 405, 647-655] provides an opportunity to interpret much kinetic and mutagenic data on the ATPase in structural terms. There are no large channels leading from the cytoplasmic surface to the pair of high-affinity Ca(2+) binding sites within the transmembrane region. One possible access pathway involves the charged residues in transmembrane alpha-helix M1, with a Ca(2+) ion passing through the first site to reach the second site. The Ca(2+)-ATPase also contains a pair of binding sites for Ca(2+) that are exposed to the lumen. In the four-site model for transport, phosphorylation of the ATPase leads to transfer of the two bound Ca(2+) ions from the cytoplasmic to the lumenal pair of sites. In the alternating four-site model for transport, phosphorylation leads to release of the bound Ca(2+) ions directly from the cytoplasmic pair of sites, linked to closure of the pair of lumenal binding sites. The lumenal pair of sites could involve a cluster of conserved acidic residues in the loop between M1 and M2. Since there is no obvious pathway from the high-affinity sites to the lumenal surface of the membrane, transport of Ca(2+) ions must involve a significant change in the packing of the transmembrane alpha-helices. The link between the phosphorylation domain and the pair of high-affinity Ca(2+) binding sites is probably provided by two small helices, P1 and P2, in the phosphorylation domain, which contact the loop between transmembrane alpha-helices M6 and M7.

Hydrophobic mismatch and the incorporation of peptides into lipid bilayers: a possible mechanism for retention in the Golgi.

Preferential interaction of trans-membrane alpha-helices whose hydrophobic length matches the hydrophobic thickness of the lipid bilayer could be a mechanism of retention in the Golgi apparatus. We have used fluorescence methods to study the interaction of peptides Ac-K2-G-Lm-W-Ln-K2-A-amide (Pm+n) with bilayers of phosphatidylcholines with chain lengths between C14 and C24. The peptide P22 (m = 10, n = 12) incorporates into all bilayers, but P16 (m = 7, n = 9) does not incorporate into bilayers when the fatty acyl chain length is C24 and only partly incorporates into bilayers where the chain length is C22. The strongest binding is seen when the hydrophobic length of the peptide matches the calculated hydrophobic thickness of the bilayer. It is suggested that a too-thin bilayer can match to a too-long peptide both by stretching of the lipid and by tilting of the peptide. However, a too-thick bilayer can only match a too-thin peptide by compression of the lipid, which becomes energetically unfavorable when the difference between the bilayer thickness and the peptide length exceeds about 10 A. The presence of cholesterol in the bilayer leads to a marked reduction in the incorporation of P16 into bilayers where the chain length is C18. Hydrophobic mismatch could explain retention of proteins with short trans-membrane alpha-helical domains in the Golgi, the effect following largely from the low concentration of cholesterol in the Golgi membrane compared to that in the plasma membrane.

Interaction of fatty acids with lipid bilayers

Abstract We present a method by which it is possible to describe the binding of fatty acids to phospholipid bilayers. Binding constants for oleic acid and a number of fatty acids used as spectroscopic probes are deduced from electrophoresis measurements. There is a large shift in p K value for the fatty acids on binding to the phospholipid bilayers, consistent with stronger binding of the uncharged form of the fatty acid. For dansylundecanoic acid, fluorescence titrations are consistent with the binding constants derived from the electrophoresis experiments. For 12-(9-anthroyloxy)stearic acid, fluorescence and electrophoresis data are inconsistent, and we attribute this to quenching of fluorescence at high molar ratios of 12-anthroylstearic acid to phospholipid in the bilayer.

The position of the ATP binding site on the (Ca2+ + Mg2+)-ATPase.

We present a convenient method to calculate the efficiency of fluorescence energy transfer in two-dimensional membrane systems. We apply it to the analysis of energy transfer between phospholipid molecules labelled with fluorescein and rhodamine groups, and of energy transfer in reconstituted membranes containing (Ca2+ + Mg2+)-ATPase purified from sarcoplasmic reticulum, with the ATPase labelled at the ATP binding site with fluorescein as donor, and rhodamine-labelled lipid as acceptor. The ATP binding site is found to be distant from the plane of the lipid/water interface of the membrane. It is suggested that the ATPase is present in the membrane as a dimer, with the two ATP binding sites in the dimer being close to the protein/protein interface. Addition of vanadate causes no change in quenching, suggesting that the ATP binding site does not move significantly with respect to the lipid/water interface in the E1-E2 conformational transition of the ATPase.

Effects of lipid fatty acyl chain structure on the activity of the (Ca2+ + Mg2+)-ATPase

The (Ca2+ + Mg2+)-ATPase purified from rabbit muscle sarcoplasmic reticulum has been reconstituted into a series of phosphatidylcholines in the liquid crystalline phase. For phosphatidylcholines containing monounsaturated fatty acyl chains, optimal activity is observed for a chain length of C18, with longer or shorter chains supporting lower activities. Phospholipids with methyl-branched chain saturated fatty acids support somewhat lower activities than the corresponding phospholipids with mono-unsaturated fatty acids. Mixed chain phospholipids support ATPase activities comparable to those shown by an unmixed chain phospholipid with the same average chain length. However, the response of the ATPase reconstituted with mixed chain phospholipids to the addition of oleyl alcohol is dominated by the longest fatty acyl chain. Based on their ability to displace brominated phospholipids, relative binding constants to the ATPase of a series of phosphatidylcholines have been determined. Binding to the ATPase is virtually unaffected by fatty acyl chain length or the presence of methyl branches.

The potassium channel KcsA and its interaction with the lipid bilayer

The crystal structure of the K+ channel KcsA explains many features of ion channel function. The selectivity filter corresponds to a narrow region about 12 Å long and 3 Å wide, lined by carbonyl groups of the peptide backbone, through which a K+ ion can only move in a dehydrated form. The selectivity filter opens into a central, water-filled cavity leading to a gating site on the intracellular side of the channel. The channel is tetrameric, each monomer containing two transmembrane a helices, M1 and M2. Helix M1 faces the lipid bilayer and helix M2 faces the central channel pore; the M2 helices participate in subunit-subunit interactions. Helices M1 and M2 in each subunit pack as a pair of antiparallel coils with a heptad repeat, but the M2 helices of neighbouring subunits show fewer interactions, crossing at an angle of about –40°. Trp residues at the ends of the transmembrane α helices form clear girdles on the two faces of the membrane, which, together with girdles of charged residues, define a hydrophobic thickness of about 37 Å for the channel. Binding constants for phosphatidylcholines to KcsA vary with fatty acyl chain length, the optimum chain length being C22. A phosphatidylcholine with this chain length gives a bilayer of thickness about 34 Å in the liquid crystalline phase, matching the hydrophobic thickness of the protein. However, a typical biological membrane has a hydrophobic thickness of about 27 Å. Thus either the transmembrane a helices of KcsA are more tilted in the native membrane than they are in the crystal structure, or the channel is under stress in the native membrane. The efficiency of hydrophobic matching between KcsA and the surrounding lipid bilayer is high over the chain length range C10–C24. The channel requires the presence of some anionic lipids for function, and fluorescence quenching studies show the presence of two classes of lipid binding site on KcsA; at one class of site (nonannular sites) anionic phospholipids bind more strongly than phosphatidylcholine, whereas at the other class of site (annular sites) phosphatidylcholines and anionic phospholipids bind with equal affinity.

Evidence for the cytoplasmic location of the N- and C-terminal segments of sarcoplasmic reticulum (Ca2+-Mg2+)-ATPase.

Antibodies were produced against 5 peptides corresponding to segments of the (Ca2+-Mg2+)-ATPase of fast-twitch rabbit skeletal muscle sarcoplasmic reticulum (SR) including the N- and C-terminal regions. With the exception of antibodies directed against the peptide corresponding to residues 567-582 all antibodies bound strongly to the ATPase in intact SR vesicles, indicating that the epitopes were located on the cytoplasmic face of the SR. When the vesicles were disrupted, by solubilisation in SDS, binding of these antibodies was unchanged, further supporting the idea that these epitopes were located on the cytoplasmic face of SR. This is the first demonstration of the location of the N- and C-terminal regions of SR (Ca2+-Mg2+)-ATPase. These observations are discussed in the light of current structural models of the ATPase.

A fast passive Ca2+ efflux mediated by the (Ca2+ + Mg2+)-ATPase in reconstituted vesicles

The (Ca2+ + Mg2+)-ATPase from skeletal muscle sarcoplasmic reticulum was reconstituted into phospholipid bilayers. The permeability of lipid bilayers to Co2+ and glucose was increased slightly by incorporation of the ATPase, and the permeability of mixed bilayers of phosphatidylethanolamine and phosphatidylcholine increased with increasing content of phosphatidylethanolamine both in the presence and absence of the ATPase. The presence of the ATPase, however, resulted in a marked increase in permeability to Ca2+, the permeability decreasing with increasing phosphatidylethanolamine content. Permeability to Ca2+ was found to be dependent on pH and the external concentrations of Mg2+ and Ca2+, was stimulated by adenine nucleotides but was unaffected by inositol trisphosphate. A kinetic model is presented for Ca2+ efflux mediated by the ATPase. It is shown that the kinetic parameters that describe Ca2+ efflux from vesicles of sarcoplasmic reticulum also describe efflux from the vesicles reconstituted from the purified ATPase and phosphatidylcholine. It is shown that the effects of phosphatidylethanolamine on efflux can be simulated in terms of changes in the rates of the transitions linking conformations of the ATPase with inward- and outward-facing Ca2+-binding sites, and that effects of phosphatidylethanolamine on the ATPase activity of the ATPase can also be simulated in terms of effects on the corresponding conformational transitions. We conclude that the ATPase can act as a specific pathway for Ca2+ efflux from sarcoplasmic reticulum.