A single amino acid residue can determine the sensitivity of SERCAs to artemisinins
2
A single amino acid residue can determine the sensitivity of SERCAs to artemisinins Anne-Catrin Uhlemann 1 , Angus Cameron 2 , Ursula Eckstein-Ludwig 1 , Jorge Fischbarg 3 , Pavel Iserovich 3 , Felipe A Zuniga 3 , Malcolm East 4 , Anthony Lee 4 , Leo Brady 2 , Richard K Haynes 5 & Sanjeev Krishna 1 Artemisinins are the most important class of antimalarial drugs. They specifically inhibit PfATP6, a SERCA-type ATPase of Plasmodium falciparum. Here we show that a single amino acid in transmembrane segment 3 of SERCAs can determine susceptibility to artemisinin. An L263E replacement of a malarial by a mammalian residue abolishes inhibition by artemisinins. Introducing residues found in other Plasmodium spp. also modulates artemisinin sensitivity, suggesting that artemisinins interact with the thapsigargin-binding cleft of susceptible SERCAs. In the 1970s, Chinese scientists identified artemisinins (qinghaosu) from sweet wormwood (Artemisia annua), thereby giving us our most important class of antimalarial drugs 1 . Use of artemisinin-containing therapies has increased exponentially 2 but the mechanism of action of these sesquiterpene lactone endoperoxides is controversial 3 . Recently, we demonstrated that artemisinins inhibit the sarcoendoplasmic reticulum Ca 2+ ATPase (SERCA) ortholog of P . falciparum (PfATP6) with marked specificity, although the mole- cular determinants of selectivity for Plasmo- dium spp. remained undefined 4 . In contrast, thapsigargin is a sesquiterpene lactone that potently inhibits both mammalian and malarial SERCAs. We hypothesized that arte- misinins interact with a region of PfATP6 that binds thapsigargin, implying that differ- ences in the thapsigargin-binding cleft of malarial and mammalian SERCAs are a determinant of artemisinin-susceptibility. Sequence alignments of the thapsigargin- binding pocket of mammalian SERCAs (for which structures have been solved in Ca 2+ (ref. 5), thapsigargin 6 , ATP-analog 7 and phosphate-bound 8 conformations) and from different Plasmodium spp. show several amino acids (such as Leu263 in PfATP6 Fig. 1a, magenta) that differ among SERCAs. To study Leu263 we mutated PfATP6 and carried out functional assays for artemisinin sensitivity after expression in Xenopus laevis oocytes. Leu263 clearly modulates sensitivity to artemisinins (Table 1). Introduction of a residue in Plasmodium vivax SERCA (PvSERCA, L263A, Fig. 1a) increased sensitivity to artemisinins by three-fold, whereas introduction of a residue in Plasmodium berghei SERCA (PbSERCA, L263S, Fig. 1a) decreased sensitivity by three-fold (P o 0.05 for each comparison with PfATP6). Full-length PvSERCA was ten-fold more sensitive to artemisinin compared with the single- site mutant PfATP6 (L263A, Table 1), suggesting that additional residues in each SERCA also contribute to artemisinin susceptibility. Similarly, full-length PbSERCA was about ten-fold less sensitive to artemisinin compared with the PfATP6 L263S mutant. Published IC 50 values for artemisinins obtained ex vivo for P. berghei 9 , P. vivax 10,11 and P. falciparum 12–14 indicate a correlation with artemisinin-susceptibility of their SERCAs. P. vivax is signifi- cantly more susceptible to artesunate than P. falciparum (geometric means (95% CI) IC 50 D 1.3 nM (1.0–1.8 nM; n ¼ 20) 11 and 4.2 (3.6–4.7 nM; n ¼ 178) 12 ). For P. berghei (mean of two values IC 50 ¼ 28 nM) 9 cultured for 24 h (1 life cycle as for P. vivax and P. falciparum) more limited data are consistent with decreased susceptibility to artemisinins. IC 50 values for parasites are generally lower (by 10- to 200-fold depending on species and first-generation artemisinin derivative) than K i estimates for SERCAs 15 . Many vari- ables including partitioning of artemisinins and chemical stability can explain differences between K i and IC 50 values. When artemisone, which is one of the most potent next-generation semisynthetic P. falciparum 256 256 256 256 248 P. vivax P. yoelii P. berghei SERCA1 P. falciparum 952 921 838 843 740 P. vivax P. yoelii P. berghei SERCA1 P. falciparum 1028 997 914 919 815 P. vivax P. yoelii P. berghei SERCA1 M3 M5 M7 Lys825 Lys252 Lys260 Pro827 Glu255 Leu263 Artemisone Asn1039 Asn1037 Helix M3 Helix M3 Helix M7 Helix M7 TG Helix M5 Helix M5 a b Figure 1 Inhibitor-binding sites. (a) Sequence alignments of M3, M5 and M7 helical regions of SERCAs. Red highlights sequence differences, and magenta indicates residues aligned with Leu263 in PfATP6. Sequences were aligned using ClustalW 20 .(b) Energy-minimized models of the rabbit SERCA1 thapsigargin-binding site (PDB entry 1IWO, left) and the corresponding region of a model of PfATP6 containing artemisone (right). Electrostatic interactions are predicted to exist between Lys252 and Glu255 on helix M3 in the mammalian SERCA1. In PfATP6 Lys260 is predicted to bond to the L6/7 loop above helix M7. Published online 5 June 2005; doi:10.1038/nsmb947 1 Department of Cellular and Molecular Medicine, Infectious Diseases, St. George’s Hospital Medical School, London SW17 0RE, UK. 2 Department of Biochemistry, University of Bristol, University Walk, Bristol BS8 1TD, UK. 3 Departments of Physiology and Ophthalmology, College of Physicians and Surgeons, Columbia University, New York, New York 10032, USA. 4 Department of Biochemistry and Molecular Biology, University of Southampton, Southampton SO16 7PX, UK. 5 Department of Chemistry, The Hong Kong University of Science and Technology, Clear Water Bay, Kowloon, Hong Kong, People’s Republic of China. Correspondence should be addressed to S.K. ([email protected]). 628 VOLUME 12 NUMBER 7 JULY 2005 NATURE STRUCTURAL & MOLECULAR BIOLOGY BRIEF COMMUNICATIONS © 2005 Nature Publishing Group http://www.nature.com/nsmb
Transcript of A single amino acid residue can determine the sensitivity of SERCAs to artemisinins
A single amino acid residue candetermine the sensitivity ofSERCAs to artemisininsAnne-Catrin Uhlemann1, Angus Cameron2,Ursula Eckstein-Ludwig1, Jorge Fischbarg3, Pavel Iserovich3,Felipe A Zuniga3, Malcolm East4, Anthony Lee4, Leo Brady2,Richard K Haynes5 & Sanjeev Krishna1
Artemisinins are the most important class of antimalarial drugs.They specifically inhibit PfATP6, a SERCA-type ATPase ofPlasmodium falciparum. Here we show that a single aminoacid in transmembrane segment 3 of SERCAs can determinesusceptibility to artemisinin. An L263E replacement of amalarial by a mammalian residue abolishes inhibition byartemisinins. Introducing residues found in other Plasmodiumspp. also modulates artemisinin sensitivity, suggesting thatartemisinins interact with the thapsigargin-binding cleft ofsusceptible SERCAs.
In the 1970s, Chinese scientists identified artemisinins (qinghaosu)from sweet wormwood (Artemisia annua), thereby giving us our mostimportant class of antimalarial drugs1. Use of artemisinin-containingtherapies has increased exponentially2 but the mechanism of action ofthese sesquiterpene lactone endoperoxides is controversial3. Recently,we demonstrated that artemisinins inhibit the sarcoendoplasmicreticulum Ca2+ ATPase (SERCA) ortholog of P. falciparum (PfATP6)with marked specificity, although the mole-cular determinants of selectivity for Plasmo-dium spp. remained undefined4. In contrast,thapsigargin is a sesquiterpene lactone thatpotently inhibits both mammalian andmalarial SERCAs. We hypothesized that arte-misinins interact with a region of PfATP6that binds thapsigargin, implying that differ-ences in the thapsigargin-binding cleft ofmalarial and mammalian SERCAs are adeterminant of artemisinin-susceptibility.Sequence alignments of the thapsigargin-binding pocket of mammalian SERCAs(for which structures have been solved inCa2+ (ref. 5), thapsigargin6, ATP-analog7
and phosphate-bound8 conformations) and
from different Plasmodium spp. show several amino acids (such asLeu263 in PfATP6 Fig. 1a, magenta) that differ among SERCAs.
To study Leu263 we mutated PfATP6 and carried out functionalassays for artemisinin sensitivity after expression in Xenopus laevisoocytes. Leu263 clearly modulates sensitivity to artemisinins(Table 1). Introduction of a residue in Plasmodium vivax SERCA(PvSERCA, L263A, Fig. 1a) increased sensitivity to artemisinins bythree-fold, whereas introduction of a residue in Plasmodium bergheiSERCA (PbSERCA, L263S, Fig. 1a) decreased sensitivity by three-fold(P o 0.05 for each comparison with PfATP6). Full-length PvSERCAwas ten-fold more sensitive to artemisinin compared with the single-site mutant PfATP6 (L263A, Table 1), suggesting that additionalresidues in each SERCA also contribute to artemisinin susceptibility.Similarly, full-length PbSERCA was about ten-fold less sensitive toartemisinin compared with the PfATP6 L263S mutant.
Published IC50 values for artemisinins obtained ex vivo forP. berghei9, P. vivax10,11 and P. falciparum12–14 indicate a correlationwith artemisinin-susceptibility of their SERCAs. P. vivax is signifi-cantly more susceptible to artesunate than P. falciparum (geometricmeans (95% CI) IC50 D 1.3 nM (1.0–1.8 nM; n ¼ 20)11 and4.2 (3.6–4.7 nM; n ¼ 178)12). For P. berghei (mean of two valuesIC50 ¼ 28 nM)9 cultured for 24 h (1 life cycle as for P. vivax andP. falciparum) more limited data are consistent with decreasedsusceptibility to artemisinins. IC50 values for parasites are generallylower (by 10- to 200-fold depending on species and first-generationartemisinin derivative) than Ki estimates for SERCAs15. Many vari-ables including partitioning of artemisinins and chemical stability canexplain differences between Ki and IC50 values. When artemisone,which is one of the most potent next-generation semisynthetic
P. falciparum 256256256256248
P. vivaxP. yoeliiP. bergheiSERCA1
P. falciparum 952921838843740
P. vivaxP. yoeliiP. bergheiSERCA1
P. falciparum 1028997914919815
P. vivaxP. yoeliiP. bergheiSERCA1
M3
M5
M7
Lys825
Lys252
Lys260Pro827
Glu255Leu263
Artemisone
Asn1039
Asn1037
Helix M
3
Helix M
3
Helix M
7
Helix M
7
TG
Helix M
5
Helix M
5
a b
Figure 1 Inhibitor-binding sites. (a) Sequence alignments of M3, M5 and M7 helical regions of
SERCAs. Red highlights sequence differences, and magenta indicates residues aligned with Leu263 in
PfATP6. Sequences were aligned using ClustalW20. (b) Energy-minimized models of the rabbit SERCA1
thapsigargin-binding site (PDB entry 1IWO, left) and the corresponding region of a model of PfATP6
containing artemisone (right). Electrostatic interactions are predicted to exist between Lys252 and
Glu255 on helix M3 in the mammalian SERCA1. In PfATP6 Lys260 is predicted to bond to the
L6/7 loop above helix M7.
Published online 5 June 2005; doi:10.1038/nsmb947
1Department of Cellular and Molecular Medicine, Infectious Diseases, St. George’s Hospital Medical School, London SW17 0RE, UK. 2Department of Biochemistry,University of Bristol, University Walk, Bristol BS8 1TD, UK. 3Departments of Physiology and Ophthalmology, College of Physicians and Surgeons, Columbia University,New York, New York 10032, USA. 4Department of Biochemistry and Molecular Biology, University of Southampton, Southampton SO16 7PX, UK. 5Department ofChemistry, The Hong Kong University of Science and Technology, Clear Water Bay, Kowloon, Hong Kong, People’s Republic of China. Correspondence should beaddressed to S.K. ([email protected]).
628 VOLUME 12 NUMBER 7 JULY 2005 NATURE STRUCTURAL & MOLECULAR BIOLOGY
BR I E F COMMUNICAT IONS©
2005
Nat
ure
Pub
lishi
ng G
roup
ht
tp://
ww
w.n
atur
e.co
m/n
smb
artemisinin derivatives under clinical development, was assayedagainst P. falciparum (3D7 strain), the IC50 value was about seven-fold more potent than artesunate (Table 1 and SupplementaryMethods online). Notably, the concentration of artemisone inhibitingPfATP6 in oocytes is highly comparable to that required to killparasites, providing additional evidence that parasite SERCAs are aprimary target for artemisinins. This is also supported by the potencyof artemisone against PvSERCA (o100 pM, Table 1).
Mammalian and X. laevis SERCAs are not susceptible to inhibitionby artemisinins4 and have a glutamic acid residue in the positionthat modulates artemisinin sensitivity in Plasmodium SERCAs(Glu255, Fig. 1a). An L263E mutation in PfATP6 (equivalent toGlu255 in mammalian SERCA) preserves thapsigargin sensitivity(Ki ¼ 213.8 7 79.3 nM), but abolishes artemisinin sensitivity, evenfor 50 mM artemisone. Amino acids in position 263 in PfATP6 arelocated at the apex of the thapsigargin-binding cleft and may thereforeact as ‘gatekeepers’ influencing access or binding of artemisinins tothis site.
The structure of SERCA1 and models of PfATP6 show how such agatekeeper residue might function (Fig. 1b). The thapsigargin cleft isbetween helices M3 and M7, with the L6/7 loop forming the apex ofthis pocket. Interactions between residues in this loop region and helixM7 have the potential to alter access to, and the shape of, this bindingsite. In mammalian SERCA1, the crystal structure reveals minimalelectrostatic interactions between M3 and the L6/7 loop in the vicinityof the thapsigargin-binding site. Energy-minimized models of SERCAsuggest that Glu255 hydrogen-bonds with Lys252 on the same helix(Fig. 1b, left). This salt bridge cannot form in any malarial sequencewhere uncharged residues replace glutamic acid. In models of PfATP6the equivalent Lys260 in M3 may interact with Asn1037 and Asn1039(Fig. 1b, right, modeled in the calcium-free E2 state). The potential forthese interactions correlates with artemisinin sensitivity. Changes inthe conformation of this loop may therefore influence ligand bindingto this site. Mutating Glu255 to leucine in SERCA1 (found in the
equivalent position in PfATP6) could free Lys260 and allow links withL6/7 loop residues. As predicted, this mutation introduces artemisininsensitivity to mammalian SERCA1. As the Ki for inhibition of thisSERCA (Table 1) is less than that for PbSERCA, it is within atherapeutically useful range. The conformation of the L6/7 loop inPfATP6, as regulated by its interactions with the neighboring Lys260region of M3, may be a key determinant of artemisinin binding. AnF256L/S mutation next to the gatekeeper in mammalian SERCA1decreases susceptibility to thapsigargin by four- to five-fold16 andproduces a large change (greater than ten-fold) in artemisinin sensi-tivity in an F264L PfATP6 mutant (Table 1). Some moleculardeterminants of interactions with artemisinin or thapsigargin aretherefore shared between malarial and mammalian SERCAs. Othersteric and charge properties may also determine artemisinin suscept-ibility. For example, a negatively charged aspartic acid (L263D) orpositively charged lysine (L263K, which may interact with Lys260 via awater molecule) did not make PfATP6 susceptible to artemisinins.
Of concern, our studies have shown for the first time thatartemisinin resistance can be introduced into PfATP6. These resultshave important implications for assessment of artemisinin resistancein field isolates17. Mutations that modulate sensitivity of PfATP6 toartemisinins can be monitored in cases where resistance may arise.Functionally silent polymorphisms, as observed between predictedtransmembrane helices M1 and M2 (I89T)18, can be assayed inoocytes (Table 1). The full repertoire of amino acids interactingwith artemisinins can now be explored with guidance from functionaldata. Our structure-function insights also point to model-guidedrefinements in the design of the next generation of artemisinins.Uniquely, we have a lead to monitor for artemisinin resistance beforethis phenotype has become clinically important.
Note: Supplementary information is available on the Nature Structural & MolecularBiology website.
ACKNOWLEDGMENTSThis study was supported by the Wellcome Trust (grant 074395). We thankD. Fidock for invaluable discussions.
COMPETING INTERESTS STATEMENTThe authors declare that they have no competing financial interests.
Received 10 March; accepted 22 April 2005
Published online at http://www.nature.com/nsmb/
1. Klayman, D.L. Science 228, 1049–1055 (1985).2. Arrow, K.J., Panosian, C.B. & Geltband, H. (eds.). Saving Lives, Buying Time:
Economics of Malaria Drugs in an Age of Resistance (National Academic Press,Washington, DC, 2004).
3. Krishna, S., Uhlemann, A.C. & Haynes, R.K. Drug Resist. Updat. 7, 233–244(2004).
4. Eckstein-Ludwig, U. et al. Nature 424, 957–961 (2003).5. Toyoshima, C., Nakasako, M., Nomura, H. & Ogawa, H. Nature 405, 647–655 (2000).6. Toyoshima, C. & Nomura, H. Nature 418, 605–611 (2002).7. Toyoshima, C. & Mizutani, T. Nature 430, 529–535 (2004).8. Toyoshima, C., Nomura, H. & Tsuda, T. Nature 432, 361–368 (2004).9. Janse, C.J., Waters, A.P., Kos, J. & Lugt, C.B. Int. J. Parasitol. 24, 589–594 (1994).10. Russell, B.M. et al. Antimicrob. Agents Chemother. 47, 170–173 (2003).11. Chotivanich, K. et al. Am. J. Trop. Med. Hyg. 70, 395–397 (2004).12. Brockman, A. et al. Trans. R. Soc. Trop. Med. Hyg. 94, 537–544 (2000).13. Ngo, T. et al. Am. J. Trop. Med. Hyg. 68, 350–356 (2003).14. Noedl, H. et al. Am. J. Trop. Med. Hyg. 68, 140–142 (2003).15. Gu, H.M., Warhurst, D.C. & Peters, W. Trans. R. Soc. Trop. Med. Hyg. 78, 265–270
(1984).16. Yu, M. et al. J. Biol. Chem. 273, 3542–3546 (1998).17. Kremsner, P.G. & Krishna, S. Lancet 364, 285–294 (2004).18. Price, R.N. et al. Lancet 364, 438–447 (2004).19. Woodrow, C.J., Penny, J.I. & Krishna, S. J. Biol. Chem. 274, 7272–7277 (1999).20. Thompson, J.D., Higgins, D.G. & Gibson, T.J. Nucleic Acids Res. 22, 4673–4680
(1994).
Table 1 Mutations in SERCAs
Sequence Artemisinin K i (nM) Artemisone K i (nM)
PfATP6
Leu263 (wild type) 169 7 31a / 4.4 7 1.7b 1.7 7 0.6 / 0.6 7 0.2b
L263A (P. vivax) 63 7 12a
L263S (P. berghei) 530 7 84a
L263E (mammalian) 450,000 450,000
L263D 450,000
L263K 450,000
L263Q 552 7 143
F264L 4,150 7 1,850
I89T 122 7 13
SERCA1
E255L (P. falciparum) 314 7 109c
Orthologs
PvSERCA 7.7 7 4.9 0.073 7 0.012
PbSERCA 5,660 7 2,330 153 7 68
Data are mean 7 s.e.m. (at least three experiments) of half-maximal inhibition.Mutation of the numbered amino acid in PfATP6 or SERCA1 was to the residue of theorganism indicated in parentheses. Pv, P. vivax; Pb, P. berghei. ATPase assays werecarried out as described4,19 (Supplementary Methods).aKi; P o 0.05 comparing wild type and mutants. bIC50 values for artesunate or artemisone.cSERCA1 wild-type sequence is not inhibited by artemisinin2.
BR I E F COMMUNICAT IONS
NATURE STRUCTURAL & MOLECULAR BIOLOGY VOLUME 12 NUMBER 7 JULY 2005 629
©20
05 N
atur
e P
ublis
hing
Gro
up
http
://w
ww
.nat
ure.
com
/nsm
b
