Research Paper Chloroquine Promotes the Anticancer Effect ...
Chloroquine Transport via the Malaria Parasite’s ...€¦ · Chloroquine Transport via the...
Transcript of Chloroquine Transport via the Malaria Parasite’s ...€¦ · Chloroquine Transport via the...
DOI: 10.1126/science.1175667 , 1680 (2009); 325Science
et al.Rowena E. Martin,Chloroquine Resistance TransporterChloroquine Transport via the Malaria Parasite’s
This copy is for your personal, non-commercial use only.
. clicking herecolleagues, clients, or customers by , you can order high-quality copies for yourIf you wish to distribute this article to others
. herefollowing the guidelines can be obtained byPermission to republish or repurpose articles or portions of articles
(this information is current as of March 8, 2010 ):The following resources related to this article are available online at www.sciencemag.org
http://www.sciencemag.org/cgi/content/full/325/5948/1680version of this article at:
including high-resolution figures, can be found in the onlineUpdated information and services,
http://www.sciencemag.org/cgi/content/full/325/5948/1680/DC1 can be found at: Supporting Online Material
http://www.sciencemag.org/cgi/content/full/325/5948/1680#otherarticles, 7 of which can be accessed for free: cites 21 articlesThis article
http://www.sciencemag.org/cgi/collection/microbioMicrobiology
: subject collectionsThis article appears in the following
registered trademark of AAAS. is aScience2009 by the American Association for the Advancement of Science; all rights reserved. The title
CopyrightAmerican Association for the Advancement of Science, 1200 New York Avenue NW, Washington, DC 20005. (print ISSN 0036-8075; online ISSN 1095-9203) is published weekly, except the last week in December, by theScience
on
Mar
ch 8
, 201
0 w
ww
.sci
ence
mag
.org
Dow
nloa
ded
from
Chloroquine Transport via theMalaria Parasite’s ChloroquineResistance TransporterRowena E. Martin,1* Rosa V. Marchetti,1 Anna I. Cowan,2 Susan M. Howitt,1Stefan Bröer,1 Kiaran Kirk1
The emergence and spread of chloroquine-resistant Plasmodium falciparum malaria parasiteshas been a disaster for world health. Resistance is conferred by mutations in the ChloroquineResistance Transporter (PfCRT), an integral membrane protein localized to the parasite’sinternal digestive vacuole. These mutations result in a marked reduction in the accumulation ofchloroquine (CQ) by the parasite. However, the mechanism by which this occurs is unclear. Weexpressed both wild-type and resistant forms of PfCRT at the surface of Xenopus laevis oocytes. Theresistant form of PfCRT transported CQ, whereas the wild-type protein did not. CQ transport viathe mutant PfCRT was inhibited by CQ analogs and by the resistance-reverser verapamil. Thus,CQ resistance is due to direct transport of the drug via mutant PfCRT.
Malaria, an infectious disease that is prev-alent throughout the tropics, was oncereadily treatable by chloroquine (CQ),
a drug that was cheap, safe, and effective. CQresistance is mediated by the Plasmodiumfalciparum Chloroquine Resistance Transporter(PfCRT) (1, 2) and is associated with a markedreduction in CQ accumulation by the parasite(3, 4). CQ is a diprotic weak base (pKa of 8.1 and10.2, where Ka is the acid dissociation constant),with the relative proportions of the neutral,mono-protonated (CQH+), and di-protonated(CQH2
2+) species varying with pH (table S1).The neutral species enters the parasite and itsinternal compartments via simple diffusion.When the base enters the acidic environment ofthe parasite’s digestive vacuole [pH ~ 5 (5–8)],the equilibrium is shifted toward the CQH2
2+
species, which is unable to diffuse across themembrane and becomes trapped, thereby accu-mulating to high concentrations within this com-partment. CQ is thought to exert its antimalarialeffect here by interfering with the detoxificationof heme, which is released as a byproduct ofhemoglobin digestion.
The key resistance-conferring mutation inPfCRT is the replacement of a lysine (K) witha threonine (T) at position 76 (9). This K76Tmutation occurs in a region of the protein thatis predicted to be involved in substrate recog-nition (10). It is never found in isolation, but isalways accompanied by a number of what arethought to be compensatory mutations in the pro-tein (11). We compared the function of mutantPfCRT from the CQ-resistant (CQR) P. falciparumstrain Dd2 with that of wild-type PfCRT fromthe CQ-sensitive (CQS) strain D10 (fig. S1).
The digestive vacuole is a lysosomal organelle,and the targeting of PfCRT to this compartmentis likely to be mediated by discrete endosomal-lysosomal trafficking motifs. Upon expression ofPfCRT in Xenopus oocytes, such motifs may causethe protein to be targeted to analogous organelles,rendering direct measurements of PfCRT functionimpractical. We therefore identified and removedmultiple putative trafficking motifs from bothtermini of the PfCRT protein sequence (fig. S1).In addition, the PfCRT coding sequence forexpression in Xenopus oocytes was codon-harmonized to facilitate correct folding of theprotein (12, 13). Hemagglutinin (HA)–taggedforms of this modified version of the PfCRTsequence were expressed in Xenopus oocytes, andlocalization of the CQS and CQR forms ofPfCRT (PfCRTCQS and PfCRTCQR, respectively)to the oocyte plasma membrane was confirmed(Fig. 1 and fig. S2, A and B) (14).
The successful expression of (motif-free,codon-harmonized) PfCRT at the oocyte surfaceenabled us to investigate the transport activity ofthe protein. Except where specified otherwise,
uptake of [3H]CQ was measured in an acidicmedium (pH = 6.0), in which the majority of CQwas protonated. Oocytes expressing PfCRTCQR
showed a marked (typically 5-fold, and up to 10-fold) increase in [3H]CQ uptake relative to non-injected controls and to oocytes expressingPfCRTCQS (Fig. 2A). This is consistent withPfCRTCQR, but not PfCRTCQS, mediating thetransport of [3H]CQ. (The membrane potentialand cytosolic pH in PfCRTCQS-expressing oocyteswere the same as those in PfCRTCQR-expressingoocytes; table S2.) In contrast, oocytes injectedwith complementary RNA (cRNA) encodingnative (that is, motif-replete, nonharmonized)PfCRTCQR did not show increased [3H]CQ up-take, nor was the protein present at significantlevels in the plasma membrane (fig. S3, A and B).The sequence modifications made here there-fore facilitated the functional expression ofPfCRT.
T76K and S163R (15) mutations in PfCRTCQR
restore CQ sensitivity to CQR parasites (9, 16).The introduction of either one of these changes toPfCRTCQR, each of which entailed the additionof a positive charge to the putative substrate-binding site of the protein (10, 11), resulted in theloss of CQ transport activity (Fig. 2B). In con-trast, the introduction of K76T to PfCRTCQS didnot result in a significant increase in [3H]CQ up-take (Fig. 2B). (PfCRTCQS K76T did localize tothe oocyte plasmamembrane; fig. S4.) TheK76Tmutation is therefore necessary but not sufficientfor the transport of CQ via PfCRT. This is con-sistent with the other PfCRT mutations acting insynergy with K76T to confer CQ resistance.
CQ transport showed a strong dependence onthe pH of the medium (Fig. 2C). Under alkalineconditions, [3H]CQ was taken up to similarly highlevels in noninjected oocytes and oocytes ex-pressing PfCRTCQR or PfCRTCQS. This is likelyto represent simple diffusion of uncharged CQ(uptake was nonsaturable at pH = 7.4 and 8.4;fig. S5). In contrast, at pH = 5.0 to 6.9, CQtransport in oocytes expressing PfCRTCQR wasmuch higher than that in noninjected oocytes or
1Research School of Biology, The Australian National Uni-versity, Canberra, Australian Capital Territory 0200, Aus-tralia. 2The John Curtin School of Medical Research, TheAustralian National University, Canberra, Australian Capi-tal Territory 0200, Australia.
*To whom correspondence should be addressed. E-mail:[email protected]
Fig. 1. Immunolocalization of PfCRT in theXenopus oocyte. The oocyte plasma mem-brane lies over a band of granules knownas the pigment layer. This, in turn, sur-rounds a cytoplasm crowded with yolk sacsand small endosomal- and lysosomal-typeorganelles (23). Expression of C-terminallyHA-tagged PfCRTCQR or PfCRTCQS results, ineach case, in the appearance of a fluores-cent band external to the pigment layer,indicating that both proteins are expressedin the oocyte plasma membrane. The bandis not present in noninjected oocytes. Simi-lar results were obtained with N-terminallytagged PfCRT (fig. S2A). Pigment
layer Plasmamembrane
Anti-HA Brightfield Merge
PfC
RT
CQ
RP
fCR
TC
QS
No
n-i
nje
cted
25 SEPTEMBER 2009 VOL 325 SCIENCE www.sciencemag.org1680
REPORTS
on
Mar
ch 8
, 201
0 w
ww
.sci
ence
mag
.org
Dow
nloa
ded
from
oocytes expressing PfCRTCQS [P < 0.05, analysisof variance (ANOVA)], with the maximum dif-ference (that is, the pH optimum for PfCRTCQR-mediated CQ transport) observed at pH = 6.0(fig. S6). This pH dependence is consistent withCQ being transported in its mono- or di-protonatedforms. Indeed, depolarization of the membranepotential by the replacement of extracellular Na+
with K+ (at pH = 6.0; table S3) resulted in a 25 T2% reduction in PfCRTCQR-mediated CQ transport(Fig. 2D; P < 0.005, paired t test). The decreasein PfCRTCQR-mediated CQ transport as pH wasreduced from 6.9 to 5.5 (Fig. 2C) may have beendue, at least in part, to the oocyte membrane po-
tential undergoing a depolarization over the samepH range (table S2).
The uptake of [3H]CQ via PfCRTCQR de-creased with increasing concentrations of un-labeled CQ (Fig. 2E), which is consistent with asaturable transport mechanism. In contrast, raisingthe concentration of unlabeled CQ had little ef-fect on [3H]CQ transport in PfCRTCQS-expressingand noninjected oocytes. This is consistent withthe entry of the drug into these oocytes being viasimple diffusion of the neutral species. A least-squares fit of the data to the Michaelis-Mentenequation yielded an apparent Michaelis constantKM(CQ) of 245 T 3 mMand a maximum velocity
Vmax of 67 T 13 pmol hour−1 per oocyte (Fig. 2E,inset) for the transport of CQ via PfCRTCQR.(The presence of a high extracellular concentra-tion of CQ did not affect the membrane poten-tial or cytosolic pH of noninjected PfCRTCQS- orPfCRTCQR-expressing oocytes; table S4.)
Verapamil increases the accumulation of CQby resistant parasites in vitro and thereby increasestheir sensitivity to CQ (4). Verapamil inhibitedthe transport of CQ via PfCRTCQR (Fig. 2F andTable 1; half-maximum inhibitory concentrationIC50 = 30 T 3 mM), as did a range of quinolinesincluding quinine and amodiaquine (table S5). Incontrast, piperaquine and artemisinin (both clin-
A C
[3 H]C
Q u
pta
ke (
% P
fCR
TC
QR c
on
tro
l)
0
20
40
60
80
100
High [K+]Control
E
[Verapamil] (µM)0 50 100 150 200 250
[3 H]C
Q u
pta
ke (
% P
fCR
TC
QR c
on
tro
l)
0
20
40
60
80
100
Time (min)0 50 100 150 200 250
[3 H]C
Q u
pta
ke (
% P
fCR
TC
QR a
t 1
hr)
0
100
200
300
400
0 125 2500
150
300
PfC
RT
CQ
R-m
edia
ted
up
take
(n
orm
alis
ed)
F
[CQ] (µM)0 500 1000 1500
[3 H]C
Q u
pta
ke (
% P
fCR
TC
QR c
on
tro
l)
0
50
100
[CQ] (µM)
0 750 1500
CQ
up
take
(pm
ol /
oo
cyte
/ h
r)
0
25
50
B
pH
5 5.5 6 6.4 6.9 7.4 8.4
[3 H]C
Q u
pta
ke (
% P
fCR
TC
QR p
H 6
.0)
0
50
100
150
200400
600
800[3 H
]CQ
up
take
(%
PfC
RTC
QR c
on
tro
l)
0
20
40
60
80
100
D
ni
PfCRT
CQS
PfCRT
CQR
PfCRT
CQS K76T
PfCRT
CQR S163R
PfCRT
CQR T76K
Fig. 2. Transport properties of PfCRTCQR in Xenopus oocytes. (A) Oocytesexpressing PfCRTCQR (solid circles) showed a marked increase in CQ trans-port relative to noninjected oocytes (solid triangles) and oocytes expressingPfCRTCQS (open circles). Rates of CQ uptake (pmol hour−1 per oocyte; n = 3 TSEM, estimated from uptake at 60 min) were as follows: noninjected, 1.14 T0.16; PfCRTCQS-, 1.14 T 0.19; and PfCRTCQR, 5.54 T 0.44. The PfCRTCQR-mediated uptake of [3H]CQ (obtained by subtracting uptake in oocytesexpressing PfCRTCQR from that in PfCRTCQS-expressing oocytes) was ap-proximately linear with time for at least 4 hours (inset). (B) Introductionof K76T to PfCRTCQS did not increase CQ transport to above that mea-sured in oocytes expressing PfCRTCQS or in noninjected (ni) oocytes (P >0.05, ANOVA). The introduction of T76K or S163R to PfCRTCQR resultedin the loss of PfCRTCQR-associated CQ transport ([3H]CQ uptake in theseoocytes did not differ significantly from that in noninjected oocytes orfrom oocytes expressing PfCRTCQS; P > 0.05, ANOVA). (C) pH depen-dence of [3H]CQ uptake into noninjected oocytes (white bars), oocytesexpressing PfCRTCQS (gray bars) and oocytes expressing PfCRTCQR (black
bars). (D) Effect of depolarization of the oocyte plasma membrane (byreplacement of Na+ with K+ in the extracellular medium) on [3H]CQ up-take into noninjected oocytes (white bars), oocytes expressing PfCRTCQS
(gray bars), and oocytes expressing PfCRTCQR (black bars). PfCRTCQR-expressing oocytes, but not PfCRTCQS-expressing or noninjected oocytes,showed a significant decrease in CQ uptake when depolarized [P < 0.005and P > 0.05 (ANOVA), respectively]. (E) Effect of unlabeled CQ on theuptake of [3H]CQ by noninjected oocytes (solid triangles) and oocytesexpressing either PfCRTCQR (solid circles) or PfCRTCQS (open circles). Theinset shows the [CQ]-dependence of PfCRTCQR-mediated uptake, whichwas calculated by subtracting the uptake measured in oocytes express-ing PfCRTCQS from that in oocytes expressing PfCRTCQR at each CQconcentration. (F) Inhibition by verapamil of the uptake of [3H]CQ byoocytes expressing PfCRTCQR (solid circles) or PfCRTCQS (open circles). Inall panels, uptake is shown as mean T SEM from three to five separateexperiments, within which measurements were made from 10 oocytes pertreatment.
www.sciencemag.org SCIENCE VOL 325 25 SEPTEMBER 2009 1681
REPORTS
on
Mar
ch 8
, 201
0 w
ww
.sci
ence
mag
.org
Dow
nloa
ded
from
ically effective against both CQS and CQR par-asites) had no effect. Amantadine exhibits someantimalarial activity in vitro, particularly againstCQR parasites (16), and also inhibited transportvia PfCRTCQR (table S5).
Several peptides were found to cause a pro-nounced inhibition of CQ transport via PfCRTCQR
(table S5). Most of the peptides that are activeagainst PfCRTCQR have key elements of the CQ-resistance reverser pharmacophore [hydrogen bondacceptor and two hydrophobic aromatic rings (17)](table S6). This pharmacophore can be viewed asdefining the basic elements involved in interactionsbetween PfCRTCQR and substrates or inhibitors.
The concentration dependence of inhibitionof CQ transport was determined for a number ofcompounds (Table 1 and fig. S7). YPWF-NH2
(endomorphin-1; an opioid receptor agonist) wasthe most effective peptide inhibitor of PfCRTCQR-mediated CQ uptake, with an IC50 comparableto that of quinine and verapamil. Measurementsof [3H]YPWF-NH2 uptake in oocytes express-ing different PfCRT constructs revealed thatPfCRTCQR, but not PfCRTCQR-T76K, PfCRTCQR-S163R, or PfCRTCQS, mediates the transport ofthis peptide (fig. S8).
We have demonstrated the transport of CQvia mutant PfCRT, which provides an explana-tion for the phenomenon of CQ resistance, as wellas for the reversal of CQ resistance by reversingagents such as verapamil. The presence of a posi-tive charge (K76 or R163) in the PfCRT substrate-binding site prevents CQH2
2+ (or CQH+) frominteracting with the transporter. The K76T muta-tion removes the positive charge, altering the sub-strate specificity of PfCRT to allow the transportof the protonated drug. In the parasite, the pres-ence of mutant PfCRT on the digestive vacuolewill allow the protonated drug to be transporteddown its electrochemical gradient, out of the vac-uole, and thus away from its site of action (fig. S9).This mechanism is consistent with recent studiesimplicating PfCRTCQR in the transport of [3H]CQin CQR parasites (18–20) and in Dictyosteliumdiscoideum transformants expressing PfCRT atendosomal membranes (21). It is also consistentwith the recent demonstration of a (verapamil-
sensitive) CQ-mediated efflux of H+ from thedigestive vacuole of CQR parasites (22). Theachievement of a robust expression system forPfCRT has the potential to facilitate the rationaldesign of novel CQ-like drugs that bypass theresistance mechanism and/or the design of clin-ically effective resistance-reversing agents.
References and Notes1. D. A. Fidock et al., Mol. Cell 6, 861 (2000).2. A. B. Sidhu, D. Verdier-Pinard, D. A. Fidock, Science 298,
210 (2002).3. C. D. Fitch, Proc. Natl. Acad. Sci. U.S.A. 64, 1181 (1969).4. D. J. Krogstad et al., Science 238, 1283 (1987).5. R. Hayward, K. J. Saliba, K. Kirk, J. Cell Sci. 119, 1016
(2006).6. N. Klonis et al., Biochem. J. 407, 343 (2007).7. Y. Kuhn, P. Rohrbach, M. Lanzer, Cell. Microbiol. 9, 1004
(2007).
8. T. N. Bennett et al., Mol. Biochem. Parasitol. 133, 99(2004).
9. V. Lakshmanan et al., EMBO J. 24, 2294 (2005).10. R. E. Martin, K. Kirk, Mol. Biol. Evol. 21, 1938
(2004).11. P. G. Bray et al., Mol. Microbiol. 56, 323 (2005).12. P. Cortazzo et al., Biochem. Biophys. Res. Commun. 293,
537 (2002).13. A. A. Komar, T. Lesnik, C. Reiss, FEBS Lett. 462, 387
(1999).14. Materials and methods are available as supporting
material on Science Online.15. Single-letter abbreviations for the amino acid residues
are as follows: F, Phe; G, Gly; H, His; K, Lys; L, Leu; P, Pro;Q, Gln; R, Arg; S, Ser; T, Thr; V, Val; W, Trp; and Y, Tyr.
16. D. J. Johnson et al., Mol. Cell 15, 867 (2004).17. D. A. van Schalkwyk, T. J. Egan, Drug Resist. Updat. 9,
211 (2006).18. P. G. Bray et al., Mol. Microbiol. 62, 238 (2006).19. C. P. Sanchez et al., Biochemistry 44, 9862 (2005).20. C. P. Sanchez et al., Mol. Microbiol. 64, 407 (2007).21. B. Naude, J. A. Brzostowski, A. R. Kimmel, T. E. Wellems,
J. Biol. Chem. 280, 25596 (2005).22. A. M. Lehane, K. Kirk, Antimicrob. Agents Chemother. 52,
4374 (2008).23. C. A. Wagner, B. Friedrich, I. Setiawan, F. Lang, S. Bröer,
Cell. Physiol. Biochem. 10, 1 (2000).24. We thank J. Abbey, R. Summers, E. Baker, and R. Slatyer
for technical assistance. This work was supported by theAustralian National Health and Medical Research Council(NHMRC) (grant 471472) and the Australian ResearchCouncil (grant DP0559433). R.E.M. was supported by anNHMRC Australian Biomedical Fellowship (fellowship520320).
Supporting Online Materialwww.sciencemag.org/cgi/content/full/325/5948/1680/DC1Materials and MethodsFigs. S1 to S9Tables S1 to S6References
30 April 2009; accepted 3 August 200910.1126/science.1175667
Global Analysis of Cdk1 SubstratePhosphorylation Sites ProvidesInsights into EvolutionLiam J. Holt,1* Brian B. Tuch,2* Judit Villén,3* Alexander D. Johnson,2Steven P. Gygi,3† David O. Morgan1†
To explore the mechanisms and evolution of cell-cycle control, we analyzed the positionand conservation of large numbers of phosphorylation sites for the cyclin-dependent kinaseCdk1 in the budding yeast Saccharomyces cerevisiae. We combined specific chemical inhibitionof Cdk1 with quantitative mass spectrometry to identify the positions of 547 phosphorylationsites on 308 Cdk1 substrates in vivo. Comparisons of these substrates with orthologs throughoutthe ascomycete lineage revealed that the position of most phosphorylation sites is notconserved in evolution; instead, clusters of sites shift position in rapidly evolving disorderedregions. We propose that the regulation of protein function by phosphorylation oftendepends on simple nonspecific mechanisms that disrupt or enhance protein-protein interactions.The gain or loss of phosphorylation sites in rapidly evolving regions could facilitate theevolution of kinase-signaling circuits.
Cyclin-dependent kinases (Cdks) drivethe major events of the eukaryotic cell-division cycle (1). Comprehensive iden-
tification and analysis of Cdk substrates wouldenhance our understanding of cell-cycle control
and provide insights into the mechanisms andevolution of regulation by phosphorylation. Wetherefore developed methods for comprehensiveidentification of the sites of Cdk1 phosphoryl-ation on large numbers of substrates in vivo. We
Table 1. IC50 values for the inhibition of PfCRTCQR-mediated CQ transport by a number of drugs andpeptides. PfCRTCQR-mediated CQ transport was calculated by subtracting the uptake measured in oocytesexpressing PfCRTCQS from that in oocytes expressing PfCRTCQR. The data are shown in fig. S7 and Fig. 2F.IC50 values were derived by least-squares fit of the equation Y= Ymin + [(Ymax– Ymin)/(1 + ([inhibitor]/IC50)
C],where Y is PfCRTCQR-mediated CQ transport, Ymin and Ymax are theminimum andmaximum values of Y, andC is a constant. All values are mean T SEM from n = 3 or 4 separate experiments, within which mea-surements were made from 10 oocytes per treatment.
Compound IC50 (mM)
Verapamil 30 T 3 (n = 4)Quinine 48 T 6 (n = 3)Amantadine 770 T 63 (n = 3)YPWF-NH2 (endomorphin-1) 64 T 5 (n = 3)WHWLQL (a1-mating factor peptide) 99 T 9 (n = 3)VVYPWTQR (a hemoglobin peptide) 363 T 28 (n = 4)RPPGFSPFR (Bradykinin) 1095 T 71 (n = 3)
25 SEPTEMBER 2009 VOL 325 SCIENCE www.sciencemag.org1682
REPORTS
on
Mar
ch 8
, 201
0 w
ww
.sci
ence
mag
.org
Dow
nloa
ded
from
www.sciencemag.org/cgi/content/full/325/5948/1680/DC1
Supporting Online Material for
Chloroquine Transport via the Malaria Parasite’s Chloroquine Resistance Transporter
Rowena E. Martin,* Rosa V. Marchetti, Anna I. Cowan, Susan M. Howitt, Stefan Bröer, Kiaran Kirk
*To whom correspondence should be addressed. E-mail: [email protected]
Published 25 September 2009, Science 325, 1680 (2009)
DOI: 10.1126/science.1175667
This PDF file includes:
Materials and Methods
Figs. S1 to S9
Tables S1 to S6
References
2
Materials and Methods Cloning of PfCRT. Codon-harmonised versions of the PfCRT sequence encoding ‘motif-
free’ forms of D10 PfCRTCQS and Dd2 PfCRTCQR (synthesised by GenScript; fig. S1) and
the native PfCRTCQR allele were inserted into the oocyte expression vector pGEM-He-Juel
(25). Codon-harmonisation entails changing codons in the target sequence to synonymous
codons with relative abundances in the host organism that approximate that of the original
codon species in situ. Such modifications can significantly increase the proportion of
heterologous proteins that are correctly folded and functional (12, 13). The modified
PfCRTCQS and PfCRTCQR coding sequences contained the synthetic Kozak sequence
GCCGCCACCATG (start codon shown in bold) as well as 5’ (BamHI) and 3’ (EcoRI)
restriction enzyme sites for insertion into pGEM-He-Juel. The coding sequence of native
PfCRTCQR (PfCRTCQR*) was amplified by PCR from P. falciparum K1 cDNA using the
primer pair (forward: 5’CGAGGATCCGCCGCCACCATGAAATTCGCAAGTAAA; reverse:
5’CGAGAATTCTTATTGTGTAATAATTG). These added the Kozak sequence (italics) and 5’
(BamHI) and 3’ (EcoRI) restriction enzyme sites (underlined) to the PfCRTCQR* sequence.
The start/stop codons are shown in bold. The coding sequences of the PfCRTCQR,
PfCRTCQS, and PfCRTCQR* clones were verified by sequencing.
Site-directed mutagenesis of PfCRT. The T76K and S163R mutations were introduced to
the PfCRTCQR coding sequence via site-directed mutagenesis using the primer pairs
(forward: 5’GTTTGTGTCATTGAAaagATTTTTGCGAAGAG; reverse: 5’CTCTTCGCAAAAATcttTT
CAATGACACAAAC) and (forward: 5’GTTCTACAGCTGagaATTCCCATAAAC; reverse: 5’GTTT
ATGGGAATtctCAGCTGTAGAAC), respectively (altered nucleotides shown in lower case).
The K76T mutation was introduced to the PfCRTCQS coding sequence using the primer pair
(forward: 5’GTTTGTGTCATGAACacaATTTTTGCGAAGAG; reverse: 5’CTCTTCGCAAAAATtgtG
TTCATGACACAAAC). The PfCRTCQR T76K, PfCRTCQR S163R, and PfCRTCQS K76T coding
sequences were verified by sequencing.
3
Epitope-tagging of PfCRT. N- or C-terminally HA-tagged forms of PfCRTCQS and
PfCRTCQR were generated by PCR using the primer pairs (forward: 5’CGAGGATCCGCCGCC
ACCATGGTTGCTtacccatatgatgttccagattacgctGTTAAATTCGCCTCTAAGAAAAAC; reverse: 5’CGAG
AATTCTTACTGTGTTGCAGCGCT) and (forward: 5’CGAGGATCCGCCGCCACCATGAAATTC;
reverse: 5’CGAGAATTCTTAAACagcgtaatctggaacatcatatgggtaTGCAACAGCCTGTGTTGCAGCACTG
TCTAC), respectively. The first primer pair was also used to generate N-terminally HA-
tagged PfCRTCQS K76T. PfCRTCQR* was HA-tagged at the N-terminus using the primer
pair (forward: 5’CGAGGATCCGCCGCCACCATGGTTGCTtacccatatgatgttccagattacgc
tGTTAAATTCGCAAGTAAAAAAAAT; reverse: 5’CGAGAATTCTTATTGTGTAATAATTG).
Nucleotides were formatted as follows: HA sequence (lower case), Kozak sequence
(italics), start/stop codons (bold), and restriction enzyme sites (underlined). The coding
sequences of all six clones were verified by sequencing.
Expression of PfCRT in X. laevis oocytes and radioisotope influx measurements. In
vitro transcription, oocyte preparation, and cRNA microinjection (30 ng per oocyte) were
performed as described elsewhere (26). [3H]CQ (0.3 µM; 15 Ci/mmol) and [3H]YPWF-
NH2 (36 nM; 50 Ci/mmol) uptake by oocytes was measured (4-6 days post-injection) as
described previously (26) over 1 hr at 25°C and in medium that, unless specified otherwise,
contained 96 mM NaCl, 2 mM KCl, 1 mM MgCl2, 1.8 mM CaCl2, 10 mM MES, 10 mM
Tris-base (pH 6.0) and either 15 µM unlabelled CQ ([3H]CQ uptake) or 5 µM unlabelled
YPWF-NH2 ([3H]YPWF-NH2 uptake).
Immunofluorescence analysis. Oocytes expressing HA-tagged PfCRT were fixed and
labelled with antibodies three days after injection (50 ng per oocyte) using a protocol
described elsewhere (27). Mouse anti-HA antisera and Alexa Fluor 488 goat anti-mouse
antibody (Molecular Probes) were used at concentrations of 1 in 100 and 1 in 500,
respectively. Oocytes were embedded in an acrylic resin as described previously (25) and
slices viewed using a Leica Microsystems inverted confocal laser microscope.
4
Surface Biotinylation and Western blot analysis. Oocytes expressing HA-tagged PfCRT
were treated with Sulfo-NHS-Biotin (1 mg/ml; Pierce) three days after injection (50 ng per
oocyte). The oocytes were then lysed, the membranes solubilized, and purified biotinylated
protein prepared using streptavidin-coated agarose beads (Pierce), as described elsewhere
(28). Western blot analysis of biotinylated protein preparations was performed as described
previously (29), and membranes were probed with mouse anti-HA antisera (1:1000).
Primary antibody binding was detected using Alexa Fluor 488 goat anti-mouse antibody (1:
10,000; Molecular Probes).
Electrophysiological measurements. The intracellular pH and membrane potential of
oocytes (4-6 days post-injection) was determined using simultaneous recording with two
intracellular electrodes. The tip of the pH-sensitive electrode was coated with
tributylchlorosilane (Fluka). Subsequently the tip was back-filled with H+-ionophore I -
cocktail A (Fluka) followed by a layer of 0.1M Na-citrate pH 6.0 (titrated with HCl), as
described elsewhere (26). The electrode used for measuring membrane potential contained
300 mM KCl. The electrodes were connected via a headstage (0.001X gain for pH-
sensitive electrode (Model H2S, Axon Instruments), 1 X gain for membrane potential
(Model HS-2A 1 X LU)) to an Axoclamp 2A amplifier (Axon Instruments) or to a
Geneclamp 500B amplifier (Axon instruments). The voltage outputs of the two electrodes
were measured relative to an agar reference electrode. The pH electrode sensitivity (58 ± 3
mV/ pH unit) was measured at the beginning and end of each experiment. The intracellular
pH was calculated using pHi = pHo − (VpH−Vm)/S, where pHo is the pH of the
extracellular solution, VpH is the voltage measured by the pH electrode when the oocyte
was impaled, Vm is the membrane potential, and S is the sensitivity of the electrode.
The composition of the Na+-replete and Na+-free solutions used to suspend the oocytes
were identical to those used in the radio-isotope uptake assays (Na+-replete medium: 96
mM NaCl, 2 mM KCl, 1 mM MgCl2, 1.8 mM CaCl2, 10 mM MES, and 10 mM Tris-base
(pH 7.4, 6.9, 6.0, or 5.5); Na+-free medium: 98 mM KCl, 1 mM MgCl2, 1.8 mM CaCl2, 10
mM MES, and 10 mM Tris-base (pH 7.4 or 6.0)). A third solution (pH 7.4 or 6.0)
5
contained concentrations of NaCl (30 mM) and KCl (68 mM) intermediate between those
of the Na+-replete and Na+-free media (but was otherwise identical to these solutions). In
one set of experiments (giving rise to table S4) unlabelled CQ was added to Na+-replete
medium (pH 6.0) to give final concentrations of 15 µM or 500 µM CQ. Oocytes were
incubated in the appropriate solution 10-15 minutes before measurements were made.
Statistics. Statistical comparisons were made with the Student’s t-test for paired samples
or with ANOVA in conjunction with Tukey’s multiple comparisons test.
6
Fig. S1. Putative endosomal/lysosomal trafficking motifs in the PfCRT sequence. The amino acid sequence of PfCRT from the CQ-resistant Plasmodium falciparum strain Dd2 is shown aligned with wild-type PfCRT from the CQ-sensitive strain D10 and homologues from P. chabaudi, P. berghei, P. yoelii, P. vivax, and P. knowlesi. PfCRT is predicted to contain 10 transmembrane domains (TMDs; the positions of TMDs 1 and 10 are indicated) and to be orientated with the N- and C-termini in the parasite cytosol (10). Since the distribution of positively-charged residues (which determines the orientation of proteins in membranes; ‘positive inside rule’ (30, 31)) was not altered, the transporter is likewise expected to orientate with the termini in the cytosol when expressed at the oocyte plasma membrane. Lysosomal/ endosomal targeting motifs (32, 33) are typically found in the cytosolic tails of a membrane protein and an examination of these regions of PfCRT revealed the presence of six putative motifs. Four of these - three “dileucine” motifs (Di-leu) and a “tyrosine-based” motif (Tyr) - are located in the N-terminus and the remainder - an “acidic cluster” and a dileucine motif - are in the C-terminus. All three types of motifs are known to be inactivated by the replacement of key residues with alanine (32-37). We therefore substituted such residues with alanine in all six of the putative motifs (“A” indicates that the residue above was replaced with alanine) in both PfCRTCQS and PfCRTCQR. The positions of the two residues (76 and 163) known to play a primary role in conferring/reversing resistance are indicated. The NCBI accession (gi) number of each protein is given. Residues are shaded as follows: positively charged, blue; negatively charged, red; hydoxyl, orange; amido, gray; proline, green; cysteine, purple; histidine, mid blue; glycine, light blue; tryptophan and tyrosine, olive green; remaining nonpolar, yellow.
7
8
A B
Fig. S2. Surface expression of PfCRT in Xenopus oocytes. (A) Immunofluorescence microscopy assay of oocytes expressing HA-tagged PfCRT. The oocyte plasma membrane lies over a band of granules known as the ‘pigment layer’. This in turn surrounds a cytoplasm crowded with yolk sacs and small endosomal/ lysosomal-type organelles (25). Expression of N-terminally HA-tagged PfCRTCQR or PfCRTCQS resulted, in each case, in the appearance of a fluorescent band external to the pigment layer, indicating that both proteins were expressed in the oocyte plasma membrane. The fluorescent band was not present in non-injected (ni) oocytes. A small fraction of both PfCRTCQR and PfCRTCQS fluorescence was scattered beneath the pigment layer; this is probably attributable to vesicles and/or endosomes ferrying the protein to the oocyte plasma membrane. (NB The presence of PfCRTCQR at intracellular membranes would only have affected CQ accumulation if the lumen of the vesicle/organelle was acidic, and even then, the direction of this effect is predicted to be the opposite of that which was observed, i.e. a decrease, not an increase, in CQ accumulation by oocytes). Addition of the HA-tag did not alter the substrate-specificity of PfCRT; oocytes expressing HA-tagged PfCRTCQR accumulated [3H]CQ significantly above levels in non-injected oocytes, whereas oocytes expressing tagged PfCRTCQS did not (data not shown). Similar results were obtained with C-terminally tagged PfCRT (Fig. 1). (B) Surface biotinylation and Western blot analysis of oocytes expressing HA-tagged PfCRT. A purified preparation of biotinylated proteins was separated by SDS-PAGE and probed with anti-HA antisera. A band (~42 kDa) corresponding to the predicted size of PfCRT (48 kDa; PfCRT is known to migrate slightly faster than might be expected on the basis of its predicted size (1)) was detected in oocytes expressing N-terminally tagged PfCRTCQR (left lane). Biotinylated protein from non-injected oocytes did not bind the HA antibody (right lane). Arrows indicate the positions and sizes of the protein standards.
9
A
B
Fig. S3. Expression of PfCRTCQR* in oocytes. (A) Oocytes injected with a native ‘motif-replete’ PfCRTCQR allele (PfCRTCQR*) did not show a significant increase in CQ transport relative to control oocytes. [3H]CQ uptake was measured in non-injected oocytes (ni), oocytes expressing PfCRTCQS or PfCRTCQR, and oocytes injected with PfCRTCQR*. The
10
PfCRTCQR* sequence was not harmonized for expression in X. laevis and encodes a PfCRTCQR protein replete with all putative trafficking motifs. CQ transport in oocytes expressing PfCRTCQR* was not significantly different from that measured in non-injected oocytes or oocytes expressing PfCRTCQS (P > 0.05). By contrast, oocytes expressing the modified PfCRTCQR sequence accumulated CQ to levels significantly above that seen in non-injected oocytes or oocytes expressing PfCRTCQS or PfCRTCQR* (P < 0.001). P values were calculated using ANOVA in conjunction with Tukey’s multiple comparisons test. Uptake is shown as mean ± s.e.m. from 5 separate experiments, within each of which measurements were made from 10 oocytes per treatment. (B) Immunofluorescence microscopy assay of oocytes expressing N-terminally HA-tagged PfCRTCQR*. Expression of N-terminally HA-tagged PfCRTCQR resulted in the appearance of a fluorescent band external to the pigment layer, indicating that PfCRTCQR was expressed in the oocyte plasma membrane (see also Fig. 1 and fig. S2A). By contrast, N-terminally HA-tagged PfCRTCQR* was not expressed at significant levels in the plasma membrane.
11
Fig. S4. Immunofluorescence microscopy assay of oocytes expressing HA-tagged PfCRTCQS K76T. Expression of N-terminally HA-tagged PfCRTCQR or PfCRTCQS K76T resulted, in each case, in the appearance of a fluorescent band external to the pigment layer, indicating that both proteins were expressed in the oocyte plasma membrane. The fluorescent band was not present in non-injected (ni) oocytes. The finding here that PfCRTCQS K76T did not mediate CQ transport (Fig. 2B) differs from that of Naude et al (21), who reported that this protein transports CQ (albeit at a reduced capacity compared with PfCRTCQR). In their study, Naude and colleagues expressed PfCRTCQR, PfCRTCQS and PfCRTCQS K76T at the membranes of acidic endosomes in Dictyostelium discoideum and found that PfCRTCQR – and to a lesser extent PfCRTCQS K76T – caused a verapamil-sensitive decrease in CQ accumulation in these organelles, whereas PfCRTCQS did not. However, there are key differences between our study and that of Naude et al, aside from the use of different expression systems, that may explain this discrepancy. The D. discoideum genome encodes three homologues of PfCRT, and the product of one of these genes (DdCRTp1) localizes to the acidic endosomes (38). All of the PfCRT constructs Naude et al expressed in D. discoideum were chimeras of DdCRTp1 and PfCRT; residues 1-58 of PfCRT were replaced by residues 1-49 of DdCRTp1 to facilitate expression of the protein at the acidic endosomes. It is possible that replacement of the PfCRT N-terminus with that of DdCRTp1 altered the properties of the transporter, such that introduction of K76T to DdCRTp1-PfCRTCQS resulted in (partial) CQ transport
12
activity. Likewise, it cannot be ruled out that the modifications we made to the termini of PfCRT (mutation of residues in six putative trafficking motifs to alanine) modified the activity of PfCRTCQS K76T. Furthermore, given that both DdCRTp1 and DdCRTp1-PfCRT localize to the same membrane (that of the acidic endosomes) and that members of the CRT family have been postulated to function as dimers (10), it is possible that heterodimers of DdCRTp1 and DdCRTp1-PfCRT form in the D. discoideum expression system. DdCRTp1 contains serine (one of two hydroxy amino acids, the other being threonine) at the position corresponding to the K76T mutation. DdCRTp1 also possesses amino acids associated with CQ resistance at two other key positions – the residues corresponding to 220 and 356 in PfCRT (serine and threonine, respectively). Therefore, a heterodimer formed from DdCRTp1 and DdCRTp1-PfCRTCQS K76T would contain a hydroxy amino acid at position 76 in both monomers, as well as two additional resistance-associated mutations in one monomer, and it is conceivable that such a dimer may transport CQ. By contrast, heterodimers of DdCRTp1 and DdCRTp1-PfCRTCQS and homodimers of DdCRTp1-PfCRTCQS K76T, while containing a hydroxy amino acid at position 76 in one or both of the monomers respectively, would not possess any other resistance-associated mutations. These differences may be sufficient to permit CQ transport via a heterodimer of DdCRTp1 and DdCRTp1-PfCRTCQS K76T, but not via a heterodimer of DdCRTp1 and DdCRTp1-PfCRTCQS or a homodimer of DdCRTp1-PfCRTCQS K76T. Since Xenopus laevis does not encode a member of the CRT family, heterodimers between PfCRTCQS K76T (or any other PfCRT protein) and an endogenous protein are unlikely to form in the oocyte expression system.
13
Fig. S5. Effect of unlabelled CQ on the uptake of [3H]CQ by PfCRTCQR-injected oocytes over the extracellular pH range 5.0 – 8.4. [3H]CQ uptake was measured in the absence or presence of a high extracellular concentration (500 µM) of unlabelled CQ in non-injected oocytes (control, white bars; +500 µM CQ, diagonal-lined bars), oocytes expressing PfCRTCQS (control, grey bars; +500 µM CQ, horizontal-lined bars), and oocytes expressing PfCRTCQR (control, black bars; +500 µM CQ, hatch-lined bars). At both pH 7.4 and 8.4 there was no significant difference in CQ accumulation between non-injected oocytes, oocytes expressing PfCRTCQS, and oocytes expressing PfCRTCQR (P > 0.05). By contrast, under acidic conditions (pH 5.0-6.9) CQ transport in oocytes expressing PfCRTCQR was significantly higher than that in non-injected oocytes and in oocytes expressing PfCRTCQS (P < 0.05). For oocytes of all types suspended at pH 7.4 or 8.4, there was no significant reduction in [3H]CQ accumulation when unlabelled CQ was added to the medium (P > 0.05); i.e. transport was predominantly non-saturable. Likewise, over the pH range 5.0-6.9 the addition of unlabelled CQ had no significant effect on [3H]CQ transport in non-injected oocytes or in oocytes expressing PfCRTCQS (P > 0.05). By contrast, under these acidic conditions (pH 5.0-6.9), the addition of unlabelled CQ caused a significant reduction in the accumulation of [3H]CQ by oocytes expressing PfCRTCQR (P < 0.05); in the presence of unlabelled CQ, there was no significant difference in the accumulation of [3H]CQ between PfCRTCQR-, PfCRTCQS-, and non-injected oocytes (P > 0.05).
14
The pH-dependence of the non-saturable component of CQ uptake seen in oocytes of all types is likely to reflect the pH-dependence of the concentration of uncharged CQ present in the medium. The uncharged CQ species crosses biological membranes via simple-diffusion across the lipid bilayer, and the rate at which [3H]CQ enters the oocyte via this route form will be dependent upon the concentration of the uncharged species. As the pH was increased from 6.0 to 7.4 or 8.4 the concentration of uncharged CQ in the solution increased ~530-fold and >20900-fold, respectively (table S1). Hence, as the pH was shifted from acidic to alkaline the proportion of CQ entering the oocyte via simple diffusion will have increased dramatically, resulting in the high levels of (non-saturable) CQ accumulation shown here. At the more alkaline pH values the rate of transport of protonated CQ via PfCRTCQR was relatively low compared to the rate of entry of [3H]CQ via simple diffusion of the neutral species, and the level of CQ accumulation in oocytes expressing PfCRTCQR was, as a result, indistinguishable from that in control oocytes. It is also worth noting that a preference of PfCRTCQR for CQH2
2+ over CQH+ would result in reduced CQ uptake via the transporter at alkaline pH values; the level of CQH2
2+ decreased by 66.5% when the pH was increased from 6.0 to 8.4. P values were calculated using ANOVA in conjunction with Tukey’s multiple comparisons test. Uptake is shown as mean ± s.e.m. from 3 separate experiments, within each of which measurements were made from 10 oocytes per treatment.
The finding here that [3H]CQ transport was cis-inhibited by unlabelled CQ in oocytes expressing PfCRTCQR (see also Fig. 2E) is in agreement with the report by Sanchez et al (39) that [3H]CQ efflux from CQR parasites is cis-inhibited by unlabelled CQ. Sanchez and colleagues have reported that when parasites are preloaded with increasing concentrations of unlabelled CQ, [3H]CQ uptake in CQR parasites is stimulated (at low preloaded CQ concentrations) and then inhibited (at high preloaded CQ concentrations). Under the same conditions, preloaded CQ inhibits (and does not stimulate) CQ uptake in CQS parasites. Sanchez et al attributed the effect seen in CQR parasites to competitive inhibition of [3H]CQ efflux by unlabelled CQ (i.e. cis-inhibition). That is, the preloaded unlabelled CQ competes with [3H]CQ in the vacuole for efflux via PfCRTCQR, resulting in reduced movement of [3H]CQ out of the vacuole and thus an increase in the accumulation of radiolabel. We have demonstrated the phenomenon of cis-inhibition of PfCRTCQR-mediated [3H]CQ uptake by unlabelled CQ (Fig. 2E and fig. S5), consistent with the Sanchez et al. model. (NB The decrease in [3H]CQ accumulation seen by Sanchez et al, in CQS strains and in CQR parasites, at high concentrations of preloaded CQ, is attributed to the displacement of [3H]CQ from heme binding-sites and the titration of protons in the vacuole. Neither of these phenomena are relevant to the present study.).
15
Fig. S6. CQ uptake in PfCRTCQR-injected oocytes relative to (i.e. divided by) that in either non-injected oocytes (closed symbols) or oocytes expressing PfCRTCQS (open symbols) over the pH range 5.0 – 8.4. The data were derived from those shown in Fig. 2C. Ratios are shown as mean ± s.e.m. from 3 separate experiments, within each of which measurements were made from 10 oocytes per treatment.
16
Fig. S7. Concentration-dependent effects of quinine, amantadine, and the peptides YPWF-NH2 (endomorphin-1), WHWLQL, VVYPWTQR, and RPPGFSPFR on the uptake of [3H]CQ into oocytes expressing PfCRTCQR (closed circles) or PfCRTCQS (open circles). IC50 values derived from these data are shown in Table 1. In all panels uptake is shown as mean ± s.e.m. from 3-4 separate experiments, within each of which measurements were made from 10 oocytes per treatment.
17
A B
Fig. S8. Peptide transport via PfCRTCQR in Xenopus oocytes. (A) Uptake of radiolabelled Endomorphin-1 ([3H]YPWF-NH2) in the presence of either unlabelled Endomorphin-1 (250 µM), verapamil (250 µM), or CQ (750 µM) into non-injected oocytes (white bars) or oocytes expressing, from left-to-right: PfCRTCQS, PfCRTCQR, PfCRTCQR-T76K, PfCRTCQR-S163R. Oocytes expressing PfCRTCQR, but not those expressing any of the other PfCRT constructs, showed a significant increase in [3H]YPWF-NH2 accumulation (P < 0.001 and P > 0.05, respectively). Rates of YPWF-NH2 uptake (pmol hr-1 per oocyte; average ± s.e.m) were as follows: non-injected 0.41 ± 0.07; PfCRTCQS 0.39 ± 0.05; PfCRTCQR 0.65 ± 0.08; PfCRTCQR T76K 0.42 ± 0.05; and PfCRTCQR S163R 0.42 ± 0.05. There was no difference in the rate of YPWF-NH2 uptake between PfCRTCQS-, PfCRTCQR-T76K-, PfCRTCQR-S163R-, and non-injected oocytes (P > 0.05), however the rate of peptide transport in oocytes expressing PfCRTCQR was significantly greater than that measured in all other oocytes (P < 0.001). By contrast, there was no difference in YPWF-NH2 accumulation between the five cell types in the presence of unlabelled peptide, verapamil, or CQ (P > 0.05), indicating that the increased uptake of the peptide into oocytes expressing PfCRTCQR was saturable, and inhibited by both verapamil and CQ, consistent with YPWF-NH2 transport being via PfCRTCQR. (B) pH-dependence of the uptake of radiolabelled Endomorphin-1 into non-injected oocytes (white bars), oocytes expressing PfCRTCQS (grey bars) and oocytes expressing PfCRTCQR (black bars). PfCRTCQR-mediated transport of Endomorphin-1 was observed at pH 6.0, 6.4, and 7.4, but not at pH 5.0 or 9.0 (the rate of YPWF-NH2 uptake in oocytes expressing PfCRTCQR was significantly greater than that measured in PfCRTCQS- and non-oocytes at pH 6.0, 6.4 and 7.4 (P < 0.05), but there was no difference between the three cell types at pH 5.0 and 9.0 (P > 0.05)). At pH 5.0 YPWF-NH2 (a peptide in which the charge on the terminal carboxyl group is masked by an -NH2 group) is present predominantly as a cation; the decrease in PfCRTCQR-mediated transport at pH 5.0 (relative to that at pH 6.0-7.4; P < 0.05) may be due at least in part to depolarisation of the oocyte plasma membrane, as seen with PfCRTCQR-mediated CQ transport (Fig. 2D). P values were calculated using ANOVA in conjunction with Tukey’s multiple comparisons test. Uptake is shown as mean ± s.e.m. from 6 (A) or 5 (B) separate experiments, within which measurements were made from 10 oocytes per treatment.
18
Fig. S9. Model for the molecular mechanism underlying the phenomenon of CQ resistance in P. falciparum malaria. The loss of the positive charge associated with K76 alters the substrate specificity of PfCRT to allow the efflux of CQH2
2+ (and/or CQH+), down its electrochemical gradient, out of the digestive vacuole and away from its site of action. The K76T mutation is necessary but not sufficient for the transport of CQ via PfCRT. The efflux of CQ via PfCRT is blocked by the CQ-resistance reverser verapamil, the antiviral agent amantadine, and by a number of quinoline antimalarials (quinine, quinidine, cinchonine, cinchonidine, amodiaquine, isoquine, mefloquine, and primaquine; table S5). Neither the bisquinoline drug piperaquine (clinically effective against both CQS and CQR parasites) nor the structurally unrelated antimalarial artemisinin inhibit the transport of CQ via PfCRT.
19
Table S1. Structure of di-protonated CQ and the percentages of CQ in the neutral (CQ), mono-protonated (CQH+), and di-protonated (CQH2
2+) forms in solutions of different pH. The percentages were calculated using the Hendersen-Hasselbalch equation, with pKa values of 8.1 for the quinoline nitrogen and 10.2 for the side-chain diethylamine nitrogen.
N+
H
NH
N+
H
Cl
pH % CQ % CQH+ % CQH22+
5.0 0.0000005 0.0799997 99.9199998
5.5 0.0000050 0.2525445 99.7474505
6.0 0.0000497 0.7942781 99.2056722
6.4 0.0003100 1.9714568 98.0282332
6.9 0.0029731 5.9792417 94.0177852
7.4 0.0263210 16.7393496 83.2343294
8.4 1.0392882 66.0955322 32.8651795
20
Table S2. Cytosolic pH and membrane potential of PfCRT-expressing oocytes over the extracellular pH range 7.4-5.5. Measurements were performed on oocytes suspended in Na+-replete medium, as described in the Methods. At each of the extracellular pH values tested there was no significant difference between the cytosolic pH of PfCRTCQS-, PfCRTCQR-, and non-injected oocytes (P > 0.05). The intracellular pH of oocytes suspended at pH 7.4 was not different from that of oocytes suspended at pH 6.9 (P > 0.05). Decreasing the extracellular pH from 7.4 to 6.0 or 5.5, or from 6.9 to 6.0 or 5.5, however, caused a small but significant acidification of the cytosol in all oocytes (P < 0.01 and P < 0.05, respectively). Oocytes expressing PfCRTCQS or PfCRTCQR were depolarized relative to non-injected oocytes at each pH (P < 0.001) but, crucially, the magnitude of depolarization was the same for both PfCRTCQS- and PfCRTCQR-expressing oocytes (P > 0.05) (NB It is not an uncommon observation that oocytes expressing foreign proteins undergo depolarization (40-42). The cause of this phenomenon has not been elucidated, but is likely to be due, at least in part, to high expression levels of foreign protein at the oocyte surface being at the expense of expression of endogenous channels and transporters that play a role in maintaining the normal oocyte membrane potential.) Oocytes suspended at pH 6.0 or pH 5.5 were depolarized relative to oocytes at pH 7.4 (P < 0.01) and pH 6.9 (P < 0.001). In summary, PfCRTCQS- and PfCRTCQR-expressing oocytes shared the same cytosolic pH and membrane potential under conditions in which there was no difference in [3H]CQ accumulation between the two cell types (pH 7.4) as well as under conditions in which PfCRTCQR-expressing oocytes accumulated [3H]CQ to levels significantly higher than those measured in PfCRTCQS-expressing oocytes (pH 6.9-5.5; Fig. 2C). The values are mean ± s.e.m. from 3 separate experiments, within which measurements were made from at least 7 oocytes per treatment. P values were calculated using ANOVA in conjunction with Tukey’s multiple comparisons test. PfCRTCQR-expressing oocytes from the same batch, and on the same day of the electrophysiological measurements, accumulated [3H]CQ significantly above levels in PfCRTCQS-expressing and non-injected oocytes at pH 6.0 (data not shown).
Cytosolic pH Membrane potential (mV)
Extracellular pH Non-injected PfCRTCQS PfCRTCQR Non-injected PfCRTCQS PfCRTCQR
7.4 7.36 ± 0.05 7.35 ± 0.05 7.37 ± 0.05 -33.4 ± 1.0 -22.8 ± 0.4 -21.8 ± 0.4
6.9 7.35 ± 0.05 7.32 ± 0.05 7.32 ± 0.06 -31.7 ± 0.5 -22.7 ± 0.4 -22.1 ± 0.5
6.0 7.21 ± 0.03 7.19 ± 0.03 7.20 ± 0.02 -26.5 ± 0.4 -17.6 ± 0.3 -17.4 ± 0.4
5.5 7.07 ± 0.07 7.09 ± 0.06 7.14 ± 0.02 -24.4 ± 1.1 -17.5 ± 0.5 -17.6 ± 0.2
21
Table S3. Effect of extracellular [Na+] on the cytosolic pH and membrane potential of PfCRT-expressing oocytes at pH 7.4 and 6.0. Measurements were performed on oocytes suspended in medium containing 0, 30, or 96 mM Na+, as described in the Methods. In all oocytes, and at both pH 7.4 and 6.0, decreasing the extracellular concentration of Na+ from 96 to 30 or 0 mM depolarized the cell (P < 0.001) and caused a small but significant alkalisation of the cytosol (P < 0.05). N.B. This very modest change in cytosolic pH did not significantly alter the diffusion of [3H]CQ into the oocyte (c.f. non-injected and PfCRTCQS-expressing oocytes in the presence and absence of 96 mM Na+; Fig. 2D). Under each of the conditions tested there was no significant difference between the cytosolic pH of PfCRTCQS-, PfCRTCQR-, and non-injected oocytes (P > 0.05). Similarly, at each of the conditions tested there was no significant difference between the membrane potentials of PfCRTCQS- and PfCRTCQR-expressing oocytes (P > 0.05). The values are mean ± s.e.m. from 3 separate experiments, within which measurements were made from at least 7 oocytes per treatment. P values were calculated using ANOVA in conjunction with Tukey’s multiple comparisons test. PfCRTCQR-expressing oocytes from the same batch, and on the same day of the electrophysiological measurements, accumulated [3H]CQ significantly above levels in PfCRTCQS-expressing and non-injected oocytes at pH 6.0 (data not shown).
Cytosolic pH Membrane potential (mV)
Extracellular [Na+] (mM) Non-injected PfCRTCQS PfCRTCQR Non-injected PfCRTCQS PfCRTCQR
pH 7.4 96 7.32 ± 0.04 7.32 ± 0.03 7.34 ± 0.04 -33.6 ± 0.9 -22.7 ± 0.2 -21.9 ± 0.2
30 7.48 ± 0.03 7.48 ± 0.03 7.48 ± 0.03 -10.5 ± 0.6 -12.8 ± 0.8 -12.2 ± 0.4
0 7.47 ± 0.03 7.46 ± 0.03 7.48 ± 0.03 -6.6 ± 0.5 -10.5 ± 0.7 -10.4 ± 0.6
pH 6.0 96 7.19 ± 0.02 7.19 ± 0.02 7.20 ± 0.01 -26.7 ± 0.3 -17.2 ± 0.5 -16.8 ± 0.6
30 7.31 ± 0.03 7.33 ± 0.03 7.33 ± 0.04 -12.5 ± 0.9 -12.2 ± 1.0 -12.1 ± 0.9
0 7.38 ± 0.03 7.36 ± 0.02 7.36 ± 0.02 -9.6 ± 0.7 -10.1 ± 1.1 -10.0 ± 1.1
22
Table S4. Cytosolic pH and membrane potential of PfCRT-expressing oocytes in the absence and presence of extracellular CQ at pH 6.0. Measurements were performed on oocytes suspended in Na+-replete medium containing 0, 15, or 500 µM extracellular CQ, as described in the Methods. Under each of the conditions tested there was no significant difference between the cytosolic pH of PfCRTCQS-, PfCRTCQR-, and non-injected oocytes (P > 0.05), nor any significant difference between the membrane potentials of PfCRTCQS- and PfCRTCQR-expressing oocytes (P > 0.05); i.e. CQ had no significant effect on either the cytosolic pH or membrane potential of oocytes, when present at either (i) the low concentration (15 µM) present in the [3H]CQ-uptake assays (e.g. Fig. 2A) or (ii) a high concentration (500 µM) known to inhibit [3H]CQ uptake via PfCRTCQR at pH 6.0 while having no effect on the diffusion of [3H]CQ into PfCRTCQS-expressing or non-injected oocytes (fig. S5). The values are mean ± s.e.m. from 3 separate experiments, within which measurements were made from at least 7 oocytes per treatment. P values were calculated using ANOVA in conjunction with Tukey’s multiple comparisons test. PfCRTCQR-expressing oocytes from the same batch, and on the same day of the electrophysiological measurements, accumulated [3H]CQ significantly above levels in PfCRTCQS-expressing and non-injected oocytes at pH 6.0 (data not shown).
Cytosolic pH Membrane potential (mV)
Extracellular [CQ] (µM) Non-injected PfCRTCQS PfCRTCQR Non-injected PfCRTCQS PfCRTCQR
pH 6.0 0 7.17 ± 0.01 7.17 ± 0.02 7.19 ± 0.02 -26.6 ± 0.4 -16.9 ± 0.8 -16.2 ± 0.8
15 7.14 ± 0.06 7.14 ± 0.06 7.12 ± 0.07 -27.6 ± 0.2 -16.4 ± 0.5 -16.8 ± 0.5
500 7.12 ± 0.07 7.10 ± 0.09 7.11 ± 0.07 -27.1 ± 0.4 -16.4 ± 0.5 -16.7 ± 0.6
23
Table S5. Effects of different antimalarial drugs, amino acids, peptides, and organic cations on the uptake of [3H]CQ into oocytes expressing PfCRTCQR- or PfCRTCQS. The majority of the 43 compounds tested had no effect. However, several peptides were found to cause a pronounced inhibition of CQ transport via PfCRTCQR. While three of the inhibitory peptides correspond to those thought to be generated from the parasite’s digestion of human haemoglobin (43), a further five haemoglobin-derived peptides did not inhibit PfCRTCQR, consistent with the mutant transporter recognizing compounds with specific structural features, rather than recognizing haemoglobin-peptides per se (see table S6). The concentration of the test-compound in the extracellular medium is indicated in parentheses. Stock solutions of the following compounds were made in DMSO (final concentration of DMSO either 0.17% (drugs) or 0.33% (peptides) v/v): quinine, quinidine, cinchonine, cinchonidine, amodiaquine, isoquine, mefloquine, artemisin, and peptides of three or more residues in length. Piperaquine was dissolved in 0.5% lactic acid (w/v) (final concentration of lactic acid 0.003% w/v). Stock solutions of the remaining compounds (CQ, primaquine, amantadine, amino acids, dipeptides, cephalexin, and organic cations) were made in the reaction buffer (see Methods). There was no significant difference in [3H]CQ uptake between solvent controls and solvent-free controls (data not shown). Uptake is shown as mean ± s.e.m. from 3-4 separate experiments, within which measurements were made from 10 oocytes per treatment. Amino acids are indicated by the standard single-letter code. Asterisks indicate significant difference (P < 0.005; paired t-test) between the level of uptake measured in the presence and absence of the test-compound.
24
[3H]CQ uptake
(% PfCRTCQR control)
Test-compound
PfCRTCQS PfCRTCQR A. Drugs (500 µM, except where specified) Control 22.1 ± 1.0 100 Chloroquine 19.1 ± 0.9 33.2 ± 1.9 * Quinine 19.0 ± 1.0 22.8 ± 1.4 * Quinidine 20.4 ± 0.3 27.1 ± 1.4 * Cinchonine 21.2 ± 1.2 35.2 ± 3.7 * Cinchonidine 21.4 ± 0.9 27 ± 4 * Amodiaquine 18.3 ± 1.7 30 ± 3 * Isoquine 22.5 ± 1.0 22.9 ± 0.8 * Mefloquine (50 µM) 20.7 ± 1.1 42 ± 7 * Primaquine (50 µM) 19.8 ± 1.6 63.0 ± 2.5 * Amantadine 20.0 ± 0.2 74.0 ± 0.8 * Piperaquine 20.8 ± 2.2 100 ± 4 Artemisin 22.6 ± 1.3 100 ± 3 B. Amino acids (3 mM), peptides and peptide-related compounds (1 mM) Control 18.9 ± 3.7 100 L 19 ± 4 95 ± 5 F 19 ± 4 86 ± 6 K 20 ± 6 95 ± 11 R 17.7 ± 2.4 97 ± 6 GQ 20 ± 4 95 ± 11 GE 18 ± 4 86 ± 5 G-Sarcosine 19 ± 3 91 ± 7 EVF 18.3 ± 1.5 101 ± 4 GFS 18.1 ± 2.0 105.0 ± 0.3 GGL 19.3 ± 2.0 114 ± 8 IPI 18.5 ± 2.6 102 ± 7 MAS 18.1 ± 2.1 96 ± 5 pyro-EHP-NH2 (Thyrotropin releasing hormone) 19.3 ± 2.0 100 ± 8 ALAL 17.1 ± 1.3 98 ± 5 FGGF 17.5 ± 2.0 95 ± 5 GGYR 19 ± 3 108 ± 3 RFDS 16.7 ± 2.4 104 ± 5 Cephalexin 17.2 ± 1.6 109.0 ± 1.9 YPWF-NH2 (Endomorphin I) 19.5 ± 1.5 20 ± 0.7 * Ac-LLnLA (Calpain inhibitor I) 17.9 ± 1.9 32 ± 6 * Ac-LLMA (Calpain inhibitor II) 18.5 ± 2.6 46.0 ± 3.5 * GLAHL (haemoglobin peptide) 17.9 ± 4.1 99.4 ± 3.0 KEEAE ([Lys23]-thymosin α1 fragment 23-27) 17.4 ± 2.7 106 ± 4 YFPHF (haemoglobin peptide) 17.8 ± 4.1 44.2 ± 2.0 * YGGFL (Leu-Enkephalin) 17.1 ± 1.4 95 ± 5 YIGSR (Laminin peptide) 16.1 ± 2.5 94.4 ± 1.9 YaG(NMe)FG-OH (DAMGO) 17.5 ± 1.5 99 ± 9 RRPYIL (Neurotensin fragment 8-13) 17.7 ± 3.2 22.3 ± 3.6 * WHWLQL (α1-Mating factor fragment 1-6) 18 ± 3 21.9 ± 1.4 * YPFVEPI (β-Casomorphin) 18.1 ± 0.8 98.7 ± 2.6 YGGFMRF (Met-enkephalin-Arg-Phe) 18.4 ± 0.7 30.1 ± 1.1 * DRVYIHPF (Angiotensin II) 17.7 ± 1.1 99 ± 3 aSTTTNYT-NH2 (Peptide T) 19.0 ± 1.4 99.3 ± 2.6 DMPNALSA (haemoglobin peptide) 18.7 ± 5.2 90.2 ± 4.7 LRVDPVNF (haemoglobin peptide) 18.7 ± 4.8 93.3 ± 1.7 LSDLHAHK (haemoglobin peptide) 18.6 ± 4.8 92.4 ± 2.5 LSFPTTKT (haemoglobin peptide) 18.5 ± 4.7 99.3 ± 1.5 VVYPWTQR (haemoglobin peptide) 17.5 ± 3.2 40.9 ± 3.7 * RPPGFSPFR (Bradykinin) 17.0 ± 0.8 57 ± 3 * KLLSHCLLVT (haemoglobin peptide) 18.3 ± 4.9 56.3 ± 3.0 * C. Organic cations (1 mM) Control 18 ± 3 100 Spermine 18 ± 3 94 ± 5 Spermidine 18 ± 4 91 ± 9 Choline 19.5 ± 0.5 98 ± 3
25
Table S6. Most peptides active against PfCRTCQR share a common profile. Eight of the ten peptides identified as inhibitors of PfCRTCQR (see table S5) carry a net positive charge at pH 6.0 and do not contain an acidic residue. Seven of these contain at least two hydrophobic aromatic/heterocyclic rings. The same structural features (a hydrogen bond acceptor and two hydrophobic aromatic rings) are shared by verapamil and the quinoline drugs and represent key elements of the ‘CQ-resistance reverser’ pharmacophore (17). This pharmacophore can be viewed as defining the basic criteria underpinning interactions between PfCRTCQR and substrates or inhibitors. Residues are shaded as follows: negatively charged, red; positively charged, blue; histidine, light blue; proline, green; tryptophan, tyrosine and phenylalanine, olive green. Peptides derived from haemoglobin are shown in bold.
26
Peptide length No inhibition Significant inhibition
Di- GQ GE G-Sarcosine Tri- EVF GFS GGL IPI MAS pyro-EHP-NH2 Tetra- ALAL Ac-LLnLA FGGF Ac-LLMA GGYR YPWF-NH2 RFDS Penta- GLAHL YFPHF KEEAE YGGFL YIGSR YaG(NMe)FG-OH Hexa- WHWLQL RRPYIL Septa- YPFVEPI YGGFMRF Octa- DRVYIHPF VVYPWTQR aSTTTNYT-NH2 DMPNALSA LRVDPVNF LSDLHAHK LSFPTTKT Nona- RPPGFSPFR Deca- KLLSHCLLVT
27
References 25. S. Bröer et al., J. Biol. Chem. 272, 30096 (1997). 26. S. Bröer, Methods Mol. Biol. 227, 245 (2003). 27. A. Weise, H. M. Becker, J. W. Deitmer, J. Gen. Physiol. 130, 203 (2007). 28. S. Chubb, A. L. Kingsland, A. Bröer, S. Bröer, Mol Membr Biol 23, 255 (2006). 29. K. A. Mullin et al., Proc Natl Acad Sci U S A 103, 9572 (2006). 30. W. van Klompenburg, I. Nilsson, G. von Heijne, B. de Kruijff, EMBO J 16, 4261
(1997). 31. G. von Heijne, EMBO J 5, 3021 (1986). 32. F. Nakatsu, H. Ohno, Cell. Struct. Funct. 28, 419 (2003). 33. J. S. Bonifacino, L. M. Traub, Annu. Rev. Biochem. 72, 395 (2003). 34. B. A. Marshall, H. Murata, R. C. Hresko, M. Mueckler, J. Biol. Chem. 268, 26193
(1993). 35. M. Uldry et al., EMBO J. 20, 4467 (2001). 36. N. Mallouk, M. Ildefonse, F. Pages, M. Ragno, N. Bennett, J. Membr. Biol. 185,
129 (2002). 37. R. T. Watson, J. E. Pessin, Exp. Cell Res. 271, 75 (2001). 38. T. Nomura et al., J Infect Dis 183, 1653 (2001). 39. C. P. Sanchez, W. Stein, M. Lanzer, Biochemistry 42, 9383 (2003). 40. C. A. Wagner et al., J Physiol 526 Pt 1, 35 (2000). 41. S. Kowalczuk et al., Biochem J 386, 417 (2005). 42. C. Bohmer et al., Biochem J 389, 745 (2005). 43. K. A. Kolakovich, I. Y. Gluzman, K. L. Duffin, D. E. Goldberg, Mol. Biochem.
Parasitol. 87, 123 (1997).