Leishmania major CorA-like magnesium transporters play a critical role in parasite development and virulence

International Journal for Parasitology 39 (2009) 713–723

Contents lists available at ScienceDirect

International Journal for Parasitology journal homepage: www.elsevier.com/locate/ijpara

Leishmania major CorA-like magnesium transporters play a critical role in parasite development and virulence Ying Zhu *, Antony Davis, Brian J. Smith, Joan Curtis, Emanuela Handman The Walter and Eliza Hall Institute of Medical Research, Melbourne, Vic., Australia

a r t i c l e

i n f o

Article history: Received 12 August 2008 Received in revised form 13 November 2008 Accepted 14 November 2008

Keywords: Leishmania major Ion transporter Magnesium Knockout parasite Virulence Drug target

a b s t r a c t Establishment of infection by Leishmania depends on the transformation of the invading metacyclic promastigotes into the obligatory intracellular amastigotes, and their subsequent survival in the macrophage phagolysosome, which is low in magnesium. We show that two Leishmania major proteins designated MGT1 and MGT2, which play a critical role in these processes, belong to the two-transmembrane domain (2-TM-GxN) cation transporter family and share homology with the major bacterial magnesium transporter CorA. Although both are present in the endoplasmic reticulum throughout the life cycle of the parasite, MGT1 is more highly expressed in the infectious metacyclic parasites, while MGT2 is enriched in the immature procyclic stages. The two proteins, although predicted to be structurally similar, have features that suggest different regulatory or gating mechanisms. The two proteins may also be functionally distinct, since only MGT1 complements an Escherichia coli DCorA mutant. In addition, deletion of one mgt1 allele from L. major led to increased virulence, while deletion of one allele of mgt2 resulted in slower growth and total loss of virulence in vitro and in vivo. This loss of virulence may be due to an impaired transformation of the parasites into amastigotes. Deletion of both mgt1 alleles in the hemizygous MGT2 knockdown parasites reversed the growth defect and partially restored virulence. Our data indicate that the MGTs play a critical role in parasite growth, development and virulence. Ó 2008 Australian Society for Parasitology Inc. Published by Elsevier Ltd. All rights reserved.

1. Introduction Leishmaniasis is a complex disease affecting at least 12 million people globally. Leishmania spp. are transmitted to humans by Phlebotomus sandflies as flagellated metacyclic promastigotes. These are taken up by macrophages where they transform and replicate as intracellular amastigotes in phagolysosomes, which are low in phosphate and magnesium (Eriksson et al., 2003; Kedzierski et al., 2006). The critical role of magnesium and its transporters for microbial survival within endocytic vesicles has been demonstrated for several major intracellular bacterial pathogens, including Salmonella enterica, Mycobacterium tuberculosis and Brucella suis (Moncrief and Maguire, 1998; Maloney and Valvano, 2006; Alix and Blanc-Potard, 2007). Magnesium is the most abundant divalent cation in cells and is essential for life (Smith and Maguire, 1998). It is a cofactor for many enzymes and plays a key role in many biological processes (Maguire and Cowan, 2002). It has been suggested that Leishmania major uses a Mg2+ dependent pathway to proliferate within macrophages and evade their microbicidal activity (Lanza et al., 2004).

* Corresponding author. Present address: 720 Swanston Street, School of Dental Science, The University of Melbourne, Carlton, Vic., Australia. Tel.: +61 3 9341 1576; fax: +61 3 9341 1597. E-mail address: [email protected] (Y. Zhu).

However, little is known about the molecular mechanism(s) of Mg2+ transport in Leishmania. Three classes of magnesium transporters have been identified in Bacteria and Archaea, MgtA/B (p-type ATPases), MgtE and CorA (Knoop et al., 2005). In most prokaryotes, CorA proteins represent the primary constitutive Mg2+ uptake system. They are the most extensively characterized Mg2+ transporters to date and belong to the two-transmembrane domain (2-TM-GxN) type family of transporters (Kehres et al., 1998). They have a large cytoplasmic N-terminal domain and a small C-terminal membrane domain. A universally conserved GxN motif is present at the end of the first of the two transmembrane domains (Niegowski and Eshaghi, 2007). Recently, the three-dimensional structure of the Thermotoga maritima CorA protein has been solved, revealing that it forms a homopentamer and functions as an ion channel (Lunin et al., 2006; Payandeh and Pai, 2006). A subfamily of 2-TM-GxN proteins, the ZntB system of S. enterica mediates Zn2+ efflux, while many other 2-TM-GxN family members have not yet been functionally classified (Papp-Wallace and Maguire, 2007). Several members of the 2-TM-GxN family found in eukaryotes have been shown to be functional homologues of CorA (Knoop et al., 2005). The ALR protein type transporters encode the major plasma membrane Mg2+ uptake system of yeast (Graschopf et al., 2001). The MRS2/LPE10 family in the inner mitochondrial membrane of yeast and mammals are involved in mitochondrial Mg2+

0020-7519/$36.00 Ó 2008 Australian Society for Parasitology Inc. Published by Elsevier Ltd. All rights reserved. doi:10.1016/j.ijpara.2008.11.010

714

Y. Zhu et al. / International Journal for Parasitology 39 (2009) 713–723

uptake (Bui et al., 1999; Zsurka et al., 2001). Homologues of CorA are also found in Arabidopsis thaliana (Gardner, 2003). All these proteins are characterized by the conserved ‘‘GMN” amino acid motif at the end of the first of the two conserved TM domains near the C-terminus (Knoop et al., 2005). However, these eukaryotic CorA-related proteins share little amino acid similarity to bacterial CorA in the N-terminal domain. In our quest for the identification of proteins important for parasite virulence, we have searched the Leishmania genome database for potential magnesium transporters. We identified two 2-TMGxN-type proteins (MGT1 and MGT2), which we believe are the first 2-TM-GxN potential magnesium transporters to be identified and characterized in Leishmania. Our findings demonstrate that MGT1, but not MGT2, functionally substitutes for CorA in Escherichia coli, suggesting that it can transport magnesium. We also show that MGT2 is essential for the development and survival of parasites within macrophage phagolysosomes, and for infection of mice.

2. Materials and methods 2.1. Parasite culture Leishmania major (MHOM/IL/80 FRIEDLIN) promastigotes were cultured at 26 °C in M199 medium containing Hank’s salts (Gibco BRL, USA) and 10% (v/v) heat inactivated foetal bovine serum (FBS) (Trace Biosciences, Australia). Parasites in which MGT genes were deleted by homologous recombination were grown in the same medium with appropriate antibiotics. 2.2. Cloning of the L. major magnesium transporter mgt genes The mgt1 gene (LmjF15.1310) was amplified from genomic DNA of L. major using the forward primer 50 -CGC GGA TCC TCA TCT AAC GCC TCG-30 and the reverse primer 50 -CCG CTG CAG TCA GTG CTT GCG CAG AAA GAG GAC-30 . The full-length product was ligated into pBluescript II SK vector (Stratagene) using the BamHI and PstI restriction sites. The mgt2 gene (LmjF25.1090) was similarly amplified from genomic DNA using the primers 50 -CCG GAA TTC ATG ATG CAC TCC AAG CTT-30 and 50 -CGC GTC GAC TTA GCT GTC CGT CGG GAC-30 and ligated to the pGEM-T (easy) vector (Promega, USA) using the EcoRI and SalI restriction sites. The strategy for the gene deletion experiments is shown in Fig. 5A. For the deletion of the mgt1 gene in L. major the 559-bp 50 -flanking fragment of mgt1 was generated by PCR from L. major genomic DNA with the primers 50 -AAT GCG GCC GCG TTA ACC TTT TCG CCT CGC CGT GC-30 and 50 -AGT ACT AGT AGA TCT CGC GTG TTC TAG GCG TAC30 , introducing NotI and HpaI sites to the 50 end and an SpeI site to the 30 end. The 50 -flanking fragment was then digested with NotI and SpeI, and inserted into NotI/SpeI digested pBluescript II SK vector. The flanking 30 -fragment was generated by PCR using primers 50 -AGT ACT AGT GGA TCC TCC TTG AAA AGT GCG AAG C-30 and 50 CTT GGT ACC GTT ACC CGT CTC ACC AGG AGC ACA G-30 introducing SpeI and BamHI sites to the 50 end and HpaI and KpnI to the 30 end. The resulting 1199 bp fragment was ligated into the SpeI and KpnI sites of the pBluescript vector containing the 50 untranslated region (UTR). For the deletion of mgt2 gene in L. major the 676 bp 50 -flanking fragment of mgt2 was generated by PCR from genomic DNA with primers 50 -AAT GCG GCC GCG TTA ACT CCA CTC CTA CAC TTG CTC-30 and 50 -AGT ACT AGT GAT AGT CGG GGT CTC CGT-30 introducing NotI and HpaI sites to the 50 end and an SpeI site to the 30 end. The fragment was digested with NotI and SpeI, and inserted into NotI/SpeI digested pBluescript II SK vector. The 30 -fragment was generated by PCR using primers 50 -AGT ACT AGT GGA TCC TTC CGT GGC CTG CTC TAC-30 and 50 -CTT GGT ACC GTT

AAC CAG ACG CGG GAC AAG ACA-3, introducing SpeI and BamHI sites to the 50 end and HpaI and KpnI to the 30 end. The resulting 1168 bp fragment was digested by SpeI and KpnI and cloned into the SpeI and KpnI sites of the pBluescript vector containing the 50 UTR. Incorporation of antibiotic resistance genes between the BamHI and SpeI sites eliminated the start codon and a further 1897 and 765 nucleotides of the mgtT1 and mgt2 open reading frames (ORFs), respectively. Constructs were made to house the antibiotic resistance gene for blasticidin and phleomycin. Approximately 8 lg of the knockout cassette DNA was used for each transfection, which was performed as previously described (Cruz et al., 1991). 2.3. Protein expression and generation of antibodies MGT1 was expressed as a fusion protein with the E. coli maltose binding protein (MBP) at its N-terminus, using the pMAL-c2T vector. MGT2 was expressed with a His-tag at its N-terminus using the pProEX-HTB vector (Invitrogen). Proteins were expressed in E. coli BL21 (DE3) by induction for 4 h with 1 mM isopropyl-b-D-thiogalactopyranoside (IPTG) at 37 °C. The MGT1 MBP-fusion protein was purified on amylose resin, and the His-tagged recombinant MGT2 was purified by affinity chromatography using TALON Metal Affinity Resin (BD Biosciences) according to the manufacturers’ recommendations. The purified proteins were used to raise antibodies in rabbits. 2.4. Immunolocalization of MGT proteins Parasites were fixed with 4% paraformaldehyde in PBS pH 7.3 for 15 min at room temperature, washed with PBS and permeabilized by incubation in PBS containing 2% BSA and 0.03% (w/v) saponin. The parasites were then incubated for 1 h at room temperature with FITC-conjugated anti-MGT antibodies. For co-localisation studies, parasites were first incubated with rabbit anti-BIP protein antibodies for 1 h at room temperature, followed by incubation with Alexa Fluor 568-conjugated goat anti-rabbit IgG (Molecular Probes). Then, they were incubated with the directly FITC-conjugated anti-MGT antibodies. 2.5. Infection of macrophages and mice with MGT mutant parasites Bone marrow-derived macrophages from BALB/c mice were grown on coverslips and infected with stationary phase MGT mutant or wild type parasites at a parasite to macrophage ratio of 2:1. After incubation for 5 h at 37 °C, free promastigotes were removed by vigorous washing, and the cells incubated for an additional 24 and 48 h. At each time point, coverslips were washed and stained with Giemsa for microscopic quantitation of infection levels. For in vivo infection studies, either highly susceptible BALB/c mice or immune-deficient hypothymic nude mice were infected with 105 stationary phase wild type or mutant parasites intradermally at the base of the tail and the rate of development of skin lesions was monitored weekly by measuring the lesion diameter as previously described (Elso et al., 2004; Stewart et al., 2005). Experiments were performed in accordance with Institutional Animal Ethics approval AEC-2005-012. 2.6. Disruption of the CorA gene in E. coli and complementation assay The Red recombination system of k phage was used to disrupt the CorA gene in the E. coli MG1655 strain. The gene was replaced by the kanamycin resistance (kan) gene as described by (Datsenko and Wanner, 2000). The kan gene was PCR amplified from plasmid pKD4 using primers carrying 40 bp based on the E. coli MG1655 DNA that flanks CorA. The PCR products were introduced into E. coli MG1655 strain carrying plasmid pKD46, and kanamycin-

Y. Zhu et al. / International Journal for Parasitology 39 (2009) 713–723

715

resistant transformants were selected after induction of the Red genes. The absence of the CorA gene was confirmed by PCR. The kan gene was then eliminated through thermal induction of FLP protein recombinase synthesis by using plasmid pCP20 (Datsenko and Wanner, 2000). Wild type E. coli and the CorA mutant were plated on Luria Bertani (LB)-Agar containing CoCl2 concentrations between 0 and 5 mM. Wild type E. coli did not survive Co2+ concentrations higher than 1 mM, while the CorA mutant withstood Co2+ up to 5 mM. For the complementation experiments, the DCorA E. coli strain was transformed with the pBluescript II SK vector alone, pBluescript II SK-mgt1, pBluescript II SK-mgt2 and pBluescript II SK-LmjF28.1940. The LmjF28.1940 is a gene distantly related to the mgts - it encodes a two-transmembrane domain protein localized to the endoplasmic reticulum (ER) (data not shown). The transformants were selected for ampicillin resistance, and the E. coli colonies were tested for the ability to grow in the presence of 3 mM Co2+.

proteins (Knoop et al., 2005). Analysis of the MGT1 and MGT2 sequences with the FUGUE program (Shi et al., 2001) yielded the Mg2+-transporter structures from T. maritima (PDB accession codes 2HN2, 2IUB and 2BBH) with Z-scores between 43 and 22, indicating extremely high sequence-structure compatibility, and providing confirmation of the structure of these proteins. We have recently identified a distantly related third 2-TM-GxN protein of 678 amino acids with a predicted molecular weight of 72.2 kDa encoded by LmjF28.1940 gene (data not shown). LmjF28.1940 has no sequence similarity to bacterial CorA, but it shows significant similarity to a hypothetical protein of Trypanosoma brucei (Tb09.160.3500) that is predicted to be related to CorA. However, the glycine in the GMN motif, which is condsidered essential for the activity of CorA, has been replaced in LmjF28.1940 and Tb09.160.3500 by an alanine and a threonine, respectively. It is not clear whether the T. brucei protein or the LmjF28.1940 is involved in cation transport.

2.7. Comparative modelling of MGT1/MGT2 and CorA

3.2. Comparative modeling of MGT1/MGT2 and T. maritima CorA shows similarities as well as unique features

The sequences of MGT1 and MGT2 were aligned with the sequence of the T. maritima CorA protein (Lunin et al., 2006) (Protein databank (PDB) identification code 2BBJ) using the FUGUE sequence-structure alignment program (Shi et al., 2001). Homology models of MGT1 and MGT2 were constructed using these sequence alignments (Supplementary Figs. S1 and S2) and the X-ray crystal structure of the CorA protein (2BBJ) as a template with the MODELLER (6v2) program (Fiser and Sali, 2003). A large insertion in the sequence of MGT1 compared with CorA (residues 644 to 682 of MGT1) was omitted from the model construction. Standard quantum-chemical (QM) calculations (Hehre et al., 1986) were performed using the GAUSSIAN 03 program. Geometries were minimized at the HF/6-31G(d) level applying C5 symmetry restraints. 2.8. Image acquisition and processing Immunofluorescence staining was visualised with a Zeiss Axioskop2 fluorescence microscope. Images were assembled with Adobe Photoshop CS 8.0. For Western blots and DNA gels, the images were acquired from films or Kodak paper with an Epson 1680 scanner and processed with Adobe Photoshop. For densitometric scanning, X-ray films were scanned using a BIORAD GS-800 calibrated densitometer and quantitated using BIORAD Quantity One 4.6.1 software.

3. Results 3.1. Identification of candidate magnesium transporter genes in L. major Two Leishmania genes encoding putative 2-TM-GxN transporters were identified by searching the L. major genome (http:// www.genedb.org) using Thermotoga maritima CorA as the query. Leishmania major mgt1 (LmjF15.1310) is a nuclear gene encoding a protein of 832 amino acids with a predicted molecular weight of 92.2 kDa. The C-terminus of MGT1 shares 42% identity with CorA from T. maritima. Leishmania major mgt2 (LmjF25.1090) encodes a protein of 424 amino acids with a predicted molecular weight of 45.8 kDa. The C-terminus of MGT2 shares 26% identity with T. maritima CorA. Consistent with the features of the 2-TM-GxN family of transporters, it is predicted that both MGT1 and MGT2 have highly charged N-terminal domains and two C-terminal hydrophobic regions (Fig. 1A). There is complete conservation of the GMN motif known to be essential for magnesium transport by CorA family

The X-ray structure of CorA from T. maritima was solved almost simultaneously by three research groups (Eshaghi et al., 2006; Lunin et al., 2006; Payandeh and Pai, 2006). All three structures were, unsurprisingly, very similar. Comparative modeling of MGT1 and MGT2 was performed to examine possible structural differences between the two transporters that might shed light on their functional differences. The Mg2+ CorA transporter is composed of a homopentamer, with each monomer consisting of a transmembrane domain (composed principally of two transmembrane helices) and a much larger funnel domain. The ion-conducting pore is formed by one transmembrane helix from each monomer. The solvent-accessible surface of the pore in the models of MGT1 and MGT2 is presented in Fig. 2. Asparagines N795 and N361 of MGT1 and MGT2, respectively, are part of the highly conserved GMN motif of Mg2+ transporters, and form the entrance to the pore. It has been proposed that hydrated Mg2+ may bind this external loop, thereby conferring selectivity for magnesium ions (Lunin et al., 2006). In the model of MGT1, the nitrogen atoms of the N795 amide side-chain point toward the channel axis; an Mg2+ ion would pass within 3.74 Å (atomic separation) of these atoms. This distance is almost identical to the separation of 3.76 Å found in one X-ray structure of CorA (Payandeh and Pai, 2006). In contrast, in the model of MGT2, the oxygen atoms of the amide side-chain of N361 point toward the channel axis; at the narrowest point a Mg2+ ion would pass within 2.69 Å of these atoms. In the X-ray structure of CorA determined by the Lunin group this distance is 2.48 Å, considered to be too small to allow passage of the ion, suggesting a closed state structure. QM calculations allow us to estimate what atomic separations are required to permit a Mg2+ ion to pass a ring of five amide groups. Our calculations indicate that a partially dehydrated Mg2+ ion can accommodate distances as short as 2.27 Å when complexed to the oxygen atoms of five amide groups (Supplementary Fig. S3A). In each of the channels, hydrophobic residues protrude into the channel to form constrictions. These constrictions in CorA are thought to be too small to allow an Mg2+ ion to pass (Maguire, 2006). In our QM calculations of the complex of a Mg2+ ion with five dimethylsulphane molecules (representative of methionine side-chains), the magnesium and sulfur atoms are separated by 3.00 Å (Supplementary Fig. S3B). Similarly, in the complex between Mg2+ and five isobutane molecules (representative of isoleucine side-chains), the separation between Mg2+ and a methyl carbon atom is 3.25 Å (Supplementary Fig. S3C). In the model of MGT1,

716

Y. Zhu et al. / International Journal for Parasitology 39 (2009) 713–723

Fig. 1. Predicted structure of MGT1 and MGT2 proteins. (A) The schematic diagram shows that MGT1 and MGT2 are two-domain proteins with a large N-terminal domain and a small C-terminal domain. The universally conserved GMN motif is present at the end of the first of the two transmembrane domains (stippled) in both proteins. (B) Cartoon representation of the model of MGT1. One monomer of the homopentameric assembly is coloured, from blue at the N-terminus to red at the C-terminus. The transmembrane domain comprises the two C-terminal helices (red and orange), while the funnel domain, external to the membrane, is coloured from pale-orange to blue. Figure produced using the program PyMOL (DeLano Scientific LLC).

the contraction at L775, the greatest narrowing in the model, a Mg2+ ion would make contact with the methyl groups, with a separation of 3.32 Å, very near the optimal separation predicted from the QM calculations. Similarly, in the model of MGT2, the greatest narrowing occurs at M327, where the separation between the Mg2+ ion and the methionine sulfur atom is 3.33 Å, again in close agreement with the QM predictions. Thus, the constrictions in the models of MGT1 and MGT2 are appropriately constructed to allow binding to Mg2+ ions, and therefore these models most likely represent the open form of the ion-conduction pore. D277 of the ‘aspartate ring’ in CorA is replaced with an asparagine in MGT1 (N758). In place of the aspartate ring, a ring of arginines (K754), one turn of a helix N-terminal to N758, forms the apex of the funnel. It is unlikely that such a cationic ring could be involved in substrate dehydration, as speculated for CorA (Niegowski and Eshaghi, 2007), and is more likely to have a regulatory role, perhaps determining the direction of ion transport. In MGT2, the corresponding residue is serine, and does not limit access of ions to the channel opening. A basic sphincter at the membrane-cytosol interface (Lunin et al., 2006) in CorA is not as pronounced in MGT1 (15 basic residues, compared with 30 in CorA), and corresponds with fewer acidic residues in the ‘willow-helices’ in MGT1. In MGT2, the basic sphincter is completely missing.

3.3. MGT1, but not MGT2, functionally substitutes for bacterial CorA protein To investigate whether L. major MGT1 and MGT2 function as cation transporters, we performed functional complementation in E. coli mutants lacking CorA. It has been shown that the E. coli magnesium transporter CorA also transports Co2+, and Co2+ inhibits 28 Mg2+ uptake. The complementation assay was performed in E. coli mutants lacking CorA because they cannot import Co2+, which is toxic to the cells. They thus are protected from the toxic effects of cobalt and can grow at higher concentrations of Co2+ in the medium. Wild type E. coli which are sensitive to Co2+, do not survive at Co2+ concentrations greater than 1 mM, while the DCorA mutant strain can grow in the presence of up to 5 mM Co2+. In order to test whether MGT1 and MGT2 functionally substitute for CorA in the E. coli mutants lacking the gene, we expressed MGT1 and MGT2 using the bacterial multicopy vector pBluescript II SK and incubated the transformants in the presence of 3 mM Co2+. Over-expression of L. major MGT1 restored the Co2+ sensitivity in the DCorA strain. In contrast to the DCorA controls, the MGT1 expressing bacteria could import Co2+ which being toxic to the cells prevented them from growing in 3 mM cobalt (Fig. 3). Albeit indirect, this demonstrates that MGT1 restored the Mg2+ transport ability in the bacteria. Over-expression of L. major MGT2 or the

Y. Zhu et al. / International Journal for Parasitology 39 (2009) 713–723

Fig. 2. Solvent accessible surface of the ion-conduction pore for the models of MGT1 and MGT2 proteins. The positions of the side-chain atoms of the asparagine filter, the constriction sites, and the ‘lysine ring’ in MGT1 are shown. The likely location of the membrane boundaries, together with the position of the basic sphincter and aspartate ring in CorA (corresponding with N758 in MGT1) are indicated. Figure produced using the program DINO (http://www.dino3d.org).

distantly related LmjF28.1940 did not restore the Co2+ sensitivity (Fig. 3). 3.4. MGT1 and MGT2 are expressed in the endoplasmic reticulum throughout the parasite life cycle The possibility of plasma membrane localization has been carefully examined and ruled out based on immunofluorescence of live organisms. Immunofluorescence staining using anti-MGT antibodies showed a reticular pattern. To exclude the possibility that the proteins are localised to the mitochondria, which also have a reticular pattern, double staining with antibodies to MGT and the mitochondrial protein MIX (Uboldi et al., 2006), as well as staining with Mitotracker Red (Molecular Probes) was performed, but did not show any overlap with the MGTs (data not shown). Staining with antibodies to the known ER protein BIP showed that both MGT1 and MGT2 co-localise with BIP (Fig. 4A and B). Antibodies to MGT1 and MGT2 showed that both are expressed in promastigotes as well as amastigotes (Fig. 4). However, Western blot analysis showed that the expression of MGT1 in amastigotes was much reduced compared with stationary phase promastigotes (approximately 8%, data not shown). In contrast, the expression level of MGT2 in amastigotes is approximately 85% of that of stationary phase promastigotes, which was comparable. To investigate how the protein expression level may be regulated during the parasite life cycle, amastigotes obtained from skin lesions of infected mice were allowed to differentiate into promastigotes in culture. The expression of MGT1 and MGT2 was deter-

717

Fig. 3. MGT1 functionally replaces CorA in Escherichia coli. The cobalt resistant DcorA strain, carrying empty vector or Leishmania major MGT1, MGT2 and the gene LmjF28.1940, was streaked on Luria Bertani (LB)-agar containing 3 mM Co2+. All strains, containing vector or L. major genes were able to grow on normal LB medium. The controls and the strain transformed with MGT2 could also grow in the high cobalt concentration of 3 mM, indicating that they could not import cobalt. In contrast, the strain transformed with MGT1 was not able to grow under the toxic high cobalt concentration, demonstrating that MGT1 restored the uptake of the cobalt ions.

mined by immunoblotting of lysates from equal numbers of parasites collected every day over a five-day period. The bulk of the parasite population collected in the logarithmic phase of growth on days 1–3 represented avirulent procyclic promastigotes, while those collected in stationary phase on days four and five represented infective metacyclic promastigotes. MGT1 and MGT2 were expressed during the entire parasite life cycle including the tissue-derived amastigotes, but their amount was differentially regulated (Fig. 5). The level of MGT1 was nearly double in metacyclic promastigotes (approximately 180% of procyclic level) compared with procyclic promastigotes, as measured by densitometry. In contrast, the expression level of MGT2 was reduced in metacyclic promastigotes compared with procyclic promastigotes (approximately 45% of procyclic levels). 3.5. Gene disruption of MGT1 and MGT2 Leishmania spp. are diploid and therefore two rounds of gene disruption are required to generate null mutants. A schematic representation of the mgt knockout constructs is shown in Fig. 6A. We successfully deleted one copy of the mgt1 gene from the Leishmania diploid genome, but we were unable to obtain viable cells in which both copies of the mgt1 gene were inactivated. Similarly, we could not disrupt both copies of the mgt2 gene despite successful introduction of the drug cassette in subsequent rounds of transfection. The presence of the drug cassettes and their correct integration into one allele was demonstrated by PCR with specific primers (Fig. 6A and B) as well as by Southern blotting (data not shown). Heterozygous parasites for the individual genes (Dmgt1/MGT1 and Dmgt2/MGT2) showed significantly reduced expression of MGT1 and MGT2 proteins, respectively, as shown by Western blotting (Fig. 6A and B).

718

Y. Zhu et al. / International Journal for Parasitology 39 (2009) 713–723

Fig. 4. Immunofluorescence microscopy using antibodies to recombinant DNA-derived proteins MGT1 (A), MGT2 (B) and BIP shows the co-localization of the MGTs with BIP in the endoplasmic reticulum (ER). (C) MGT1 and MGT2 are expressed in amastigotes and show a similar pattern of expression as BIP.

To investigate whether MGT1 and MGT2 are functionally related or interacting, we decided to disrupt the mgt1 gene in the mgt2 single-allele knockout (Dmgt2/MGT2) parasites. Surprisingly, after a single round of gene disruption, both PCR (data not shown) and Southern blotting showed the complete absence of the mgt1 gene (Fig. 6C), presumably due to the simultaneous integration of the drug cassette into both alleles. The mutant parasites obtained are thus mgt1 null and mgt2 single-allele knockout (Dmgt1/ Dmgt1-Dmgt2/MGT2). 3.6. MGT2 knockdown parasites show growth defects and loss of virulence MGT mutant parasites were examined to see whether deletion of mgt genes affected parasite growth in culture. As shown in Fig. 7, when wild type and mutant parasites were diluted to 1  105 parasites per ml from stationary phase cultures, Dmgt1/ MGT1 promastigotes grew at a similar rate as the wild type parasites. The growth rate of Dmgt2/MGT2 parasites was much slower than that of the wild type. However, the homozygous deletion of mgt1 from the mgt2 single-allele knockout parasites not only reversed their growth defect, but the Dmgt1/Dmgt1-Dmgt2/MGT2 grew faster than the wild type.

To test whether deletion of the mgts affected parasite virulence, infective stationary phase wild type and mutant parasites that had been cultured for the same period were used to infect bone marrow-derived macrophages in vitro. The initial uptake of the parasites into the cells and subsequent survival over two days of culture were assessed. Interestingly, the mgt1 mutants displayed a gain of initial infectivity compared with the wild type control (Fig. 8A). In this experiment, the infection rate of the wild type control was relatively low, probably due to the long period in culture. Nevertheless, the in vitro infection clearly showed that the initial parasite uptake and subsequent survival of the Dmgt1/MGT1 parasites were higher than those of the control despite the fact that the mutants were cultured for the same time as the wild type parasites. As shown in Fig. 8B, in contrast to the mgt1, the mgt2 single-allele knockout parasites showed a significant loss of virulence. The initial uptake into bone marrow-derived macrophages after 5 h was 25% compared with 44% for wild type parasites. After 24 h only 1% of the cells harboured Dmgt2/MGT2 parasites compared with 33% harbouring wild type organisms (Fig. 8B). In macrophages infected with wild type L. major, at 24 h p.i., all the promastigotes that had been taken up had transformed into amastigotes (Fig. 9). In contrast, at 24 h or even 48 h p.i., many of

Y. Zhu et al. / International Journal for Parasitology 39 (2009) 713–723

719

Fig. 5. MGT1 and MGT2 protein expression is reciprocally regulated during the promastigote developmental cycle of Leishmania major. Amastigotes obtained from skin lesions of infected mice were allowed to differentiate into promastigotes in culture (B), and the expression of MGT1 and MGT2 was determined by immunoblotting on equal numbers of parasites collected on days 1–5. (A) Western blotting using anti-MGT1 and anti-MGT2 antibodies shows that the expression of MGT1 is most highly expressed in stationary phase promastigotes, while MGT2 expression is highest in the early logarithmic phase of growth. Expression of phosphomannomutase (PMM) was used as a control for equal loading.

the Dmgt2/MGT2 parasites remained as promastigotes, with flagella still present (Fig. 9) and actively moving. This suggests that the MGT2 protein is important in the intracellular transition of recently internalized promastigotes into the intracellular amastigote forms. Surprisingly, the complete deletion of mgt1 from the Dmgt2/ MGT2 parasites partially restored the virulence of the mutant parasites in vitro (Fig. 8B). The percentage of infected cells after 5 h of infection with the Dmgt1/Dmgt1-Dmgt2/MGT2 lines was higher than that of cells infected with the wild type parasites, indicating increased invasion. However, the survival of the mutant parasites after 48 h was lower than the wild type (Fig. 8B). Infection rates for wild type parasites were 33% and 28% after 24 and 48 h of infection, respectively. The infection rate for the Dmgt1/Dmgt1-Dmgt2/ MGT2 parasites was only slightly lower after 24 h (23%) but decreased to 12% after 48 h. The in vitro results were mirrored in vivo. In vivo infection using the mouse model of disease showed a similar loss of virulence of the mutant Dmgt2/MGT2 parasites as that detected in vitro (data not shown), but a gain of virulence of the Dmgt1/ MGT1 mirroring the in vitro infection (Fig. 8B). The highly susceptible BALB/c mice were infected intradermally with stationary phase wild type and Dmgt1/MGT1 parasites and the development of skin lesions was monitored weekly. Mice infected with stationary phase Dmgt1/MGT1 parasites developed severe lesions at 5 weeks of infection and had to be killed, while those infected with wild type parasites developed severe lesions only after 10 weeks (Fig. 8C). The Dmgt2/MGT2 and Dmgt1/Dmgt1-Dmgt2/MGT2 lines were used to infect the most susceptible immune-deficient hypothymic nude mice. No lesions developed in any of the mice over a period of 18 weeks (data not shown). However, parasites were

recovered from lymph nodes of mice infected with the Dmgt1/ Dmgt1-Dmgt2/MGT2 line but not from mice infected with the Dmgt2/MGT2 line suggesting that the deletion of MGT1 caused some restoration of virulence in vivo as it did in vitro (data not shown). 4. Discussion It is surprising that despite the fact that magnesium is a critical ion for all life forms, so little is known about magnesium transport in general (Lunin et al., 2006) and in Leishmania spp. in particular. This is largely due to the lack of availability of radioactive magnesium salts and the inability to perform physiological measurements. The remarkable capacity of Leishmania spp. to survive and replicate within the acidified and magnesium-depleted phagolysosomes of macrophages suggests that proteins involved in magnesium transport may play a role in the adaptation to this harsh intracellular environment and may therefore be good targets for anti-parasite drugs. Our search for potential magnesium transporters in the L. major genome database identified two 2-TM-GxN type proteins designated MGT1 and MGT2, which are related to the major bacterial magnesium transporter CorA. CorA is the main transporter of this ion in bacteria and has been shown to play a major role in the virulence of intracellular bacteria. MGT1 and MGT2 are the first 2-TM-GxN type proteins to be studied in Leishmania. Recently, we identified a third member of the 2-TM-GxN family (encoded by Lmj28.1940) which has little similarity to the CorA proteins and which has yet to be characterised. A functional complementation assay showed that over-expression of MGT1 in an E. coli DCorA mutant restored the sensitivity

720

Y. Zhu et al. / International Journal for Parasitology 39 (2009) 713–723

Fig. 6. Generation and characterization of MGT protein mutant parasites. (A) Schematic diagram showing the replacement of the MGTs gene by drug-resistance genes. (B) Characterization of Dmgt1/MGT1 and Dmgt2/MGT2 mutants. Figures (a) and (b) show the replacement of one copy of MGT1 with the blasticidin S-resistance gene (bsr) gene. PCR analysis was carried out on Dmgt1/MGT1 mutant and wild type (WT) parasites using primers to amplify a fragment spanning the bsr gene and the 50 untranslated region (UTR) outside of the construct DNA () to confirm correct integration of the bsr gene into the genomic DNA of the Dmgt1/MGT1 mutant. The Dmgt1/MGT1 mutant showed significantly reduced expression of the MGT1 protein as determined by Western blotting. Figures (c) and (d) show the gene disruption of MGT2. PCR analysis was carried out on Dmgt2/MGT2 mutant and wild type (WT) parasites using primers to amplify a fragment spanning the bsr gene and the 50 UTR outside of the construct DNA () to confirm correct integration of the bsr gene into the genomic DNA. The Dmgt2/MGT2 mutant showed significantly reduced expression of the MGT2 protein as determined by Western blotting. Expression of phosphomannomutase (PMM) was used as a control for equal loading. (C) Generation of Dmgt1/mgt1-Dmgt2/MGT2 parasites. The MGT1 gene has been replaced by the phleomycin-resistance (ble) gene. The Southern blot shows the loss of both copies of the MGT1 gene in Dmgt1/mgt1-Dmgt2/MGT2 parasites.

of the mutant to Co2+, indicating that MGT1 functions as a cation transporter, possibly as a magnesium transporter. Modelling of the MGT1 protein on the three-dimensional structure of the CorA protein from T. maritima which is a proven magnesium transporter showed significant similarity (Figs. 1 and 2). It was therefore surprising that despite the general similarity in the protein organiza-

tion and the presence of the conserved GMN motif (which is essential for transport) in MGT2, MGT1 and CorA, MGT2 did not functionally substitute for the E. coli CorA protein. This may be due to the fact that there is little amino acid similarity between MGT2 and bacterial CorA. Moreover, the lack of complementation in this particular type of experiment which examines influx but

Y. Zhu et al. / International Journal for Parasitology 39 (2009) 713–723

721

Fig. 7. Growth curves of wild type (WT) and MGT protein mutant promastigotes of Leishmania major in liquid culture. No differences in growth were detected between Dmgt1/MGT1 mutants and wild type parasites. The deletion of a single copy of MGT2 (Dmgt2/MGT2) resulted in a retardation of growth. However, when MGT1 was deleted from the Dmgt2/MGT2 line, the resulting Dmgt1/Dmgt1-Dmgt2/MGT2 mutant parasites showed a much higher growth rate than that of the wild type control.

not efflux, does not exclude the possibility that MGT2 does function as a cation transporter. The mechanism of ion transport in bacteria by CorA remains unclear, despite knowledge of its three-dimensional structure. However, several features of this structure have been identified that could mediate ion transport, including the asparagine filter at the pore entrance, the hydrophobic constrictions lining the pore, the aspartate ring at the funnel apex, and the basic sphincter at the cytoplasmic-membrane interface; on all these accounts, the models of MGT1 and MGT2 differ from one another. The asparagine residues at the pore entrance are likely to engage the hydrated Mg2+ ion and participate in partial dehydration of the ion. The different conformations observed here for the residues in MGT1 and MGT2 are likely to reflect alternative states in this process, and as such this is unlikely to represent any functional difference between the two transporters. The ion-conduction pore in MGT1 contains a mixture of hydrophobic and hydrophilic constrictions, whereas the constrictions in the MGT2 pore are predominantly hydrophobic. The hydrophobic constrictions at residues I761, M772 and L775 in MGT1 have homologues in M327, M338 and L341 in MGT2, respectively. Other differences in the pore include (i) T768 in MGT1 is a methionine (M327) in MGT2, (ii) an additional constriction is formed by L349 in MGT2, and (iii) the bulkier N352 forms a tighter constriction in MGT2 than T786 in MGT1. The aspartate ring seen in CorA is replaced by a ring of lysines in MGT1- extending the length of the pore to 67 Å. The model of MGT2 has no such extension to the pore, and is only 50 Å in length. The aspartate ring in CorA is likely to play a regulatory role, and the presence of a similarly charged ring in MGT1 suggests an analogous role. The absence of such a feature in MGT2 and the more hydrophobic nature of the pore in MGT2 implies an alternative regulatory mechanism. The basic sphincter is postulated to regulate gating of the channel (Lunin et al., 2006). The absence of such a feature in MGT2 implies an alternative mechanism of gating. The localisation of MGT1 and MGT2 to the ER is surprising. Our original hypothesis was that L. major MGTs would either be located on the plasma membrane to be involved in the uptake of Mg2+ and/

or other cations from the extracellular environment, as do bacterial CorA proteins and ALR type transporters in yeast. Alternatively, they could have been localised to the mitochondrial membrane like the MRS2/LPE10 family of transporters in yeast and mammals that are involved in mitochondrial Mg2+ uptake (Bui et al., 1999; Zsurka et al., 2001). The lack of 2-TM-GxN type proteins on the plasma membrane of L. major indicates that there must be other molecules and mechanisms involved in Mg2+ uptake through the plasma membrane. We have not found any homologues of the MgtE family in the L. major genome. It is possible that some cell surface p-type ATPases are involved in the import of magnesium into the parasite. In addition to the MGTs, we have recently identified a distantly related third 2-TM-GxN type protein encoded by LmjF28.1940 and surprisingly, our preliminary studies show that this protein is also located in the ER. The importance of the localization of all three L. major 2-TMGxN type proteins to the ER remains a puzzle, but it may suggest that some are involved in influx and others in efflux of ions in this very active organelle (see below). There is evidence for the presence of a calcium pool in the Leishmania ER where it may function in signalling (Gill et al., 1996), but little is known about the magnesium concentration or function in the ER. This is also the case for mammalian cells where the role of the calcium stores in the ER is very well documented, but little is known about the magnesium function in this compartment. Although both MGTs are expressed throughout the parasite life cycle, including the intracellular amastigote, the reciprocal abundance of MGT1 and MGT2 proteins during the promastigote development from the avirulent procyclic to the virulent metacyclic form suggests that the requirement for their function may differ at different stages of the parasite life cycle. The observation that deletion of MGT1 in the MGT2 single-allele knockout parasites not only reversed their growth defect but led to a faster rate of growth, suggests that MGT1 and MGT2 are involved in the control of the cell cycle, but seem to be driving it in opposite directions. They may also have other contradictory effects. Both in vitro and in vivo infection experiments suggest that MGT2 is important for virulence because the loss of even one copy of the gene led to

722

Y. Zhu et al. / International Journal for Parasitology 39 (2009) 713–723

Fig. 9. Macrophages 24 h after infection with wild type (a) or Dmgt2/MGT2 mutant Leishmania major (b). Wild type promastigotes transformed efficiently into amastigotes while many of the Dmgt2/MGT2 mutant parasites remained as promastigotes.

tially different regulation/mechanism of gating raise the tantalizing possibility that MGT1 and MGT2 are involved in opposing aspects of ion transport, for example influx and efflux through the ER, be it of magnesium or other cations. There is evidence that bacterial CorA mediates both influx and efflux of Mg2+ (Smith and Maguire, 1998). The complementation by MGT1 for E. coli CorA strongly suggests that MGT1 is involved in cation influx. It is possible that MGT2 mediates cation efflux, but this could not be examined in the current complementation assay and remains to be determined. The dramatic delay in growth and loss of virulence of Dmgt2/ MGT2 parasites point to the critical role of MGT2 in parasite growth and virulence, and the fact that there is no close mammalian homologue of MGT2 makes it an attractive drug target candidate. Acknowledgements Fig. 8. The effect of MGT gene disruption on virulence of Leishmania major. (A) The Dmgt1/MGT1 parasites lacking a single copy of the MGT1 gene showed higher infection and survival rates in macrophages. (B) The deletion of a single copy of MGT2 resulted in decreased uptake and survival in macrophages. Further deletion of MGT1 from these mutants partially restored virulence. The percentage of infected macrophages was determined by counting 400 cells in duplicate slides. The mean and S.D. are shown. (C) For in vivo studies, BALB/c mice were infected with wild type or Dmgt1/MGT1 parasites intradermally at the base of the tail (six mice per group) and the rate of development of skin lesions was determined by measuring the diameter of the lesion. Mice were killed when lesion size reached a score of four. The mean and S.D. are shown.

complete loss of the ability to survive in macrophages and cause lesions even in the most susceptible of mice. One reason for the loss of virulence may be because MGT2 seems to be involved in the process of promastigote transformation into amastigotes. The inability of the MGT2 knockdown (hemizygous) parasites to complete this process makes the parasites vulnerable to destruction in the phagolysosomes, as is the case for immature procyclic promastigotes. In contrast, the loss of a single copy of MGT1 led to increased virulence. The contradictory effects of MGT1 and MGT2 on parasite growth and virulence as well as their distinct structural features and poten-

We thank James W. Goding for providing the pMALC2T vector, making and providing the FITC-conjugated anti-MGT antibodies and for critical review of the manuscript; Roy Robbins-Brown for the E. coli strains and plasmids and technical advice; Lukasz Kedzierski for critical review of the manuscript. Infrastructure support from the NHMRC IRIISS Grant No. 361646 and the Victorian State Government OIS grant is gratefully acknowledged. This work has been funded by the Australian National Health and Medical Research Council. Appendix A. Supplementary data Supplementary data associated with this article can be found, in the online version, at doi:10.1016/j.ijpara.2008.11.010. References Alix, E., Blanc-Potard, A.B., 2007. MgtC: a key player in intramacrophage survival. Trends Microbiol. 15, 252–256. Bui, D.M., Gregan, J., Jarosch, E., Ragnini, A., Schweyen, R.J., 1999. The bacterial magnesium transporter CorA can functionally substitute for its putative

Y. Zhu et al. / International Journal for Parasitology 39 (2009) 713–723 homologue Mrs2p in the yeast inner mitochondrial membrane. J. Biol. Chem. 274, 20438–20443. Cruz, A., Coburn, C.M., Beverley, S.M., 1991. Double targeted gene replacement for creating null mutants. Proc. Natl. Acad. Sci. USA 88, 7170–7174. Datsenko, K.A., Wanner, B.L., 2000. One-step inactivation of chromosomal genes in Escherichia coli K-12 using PCR products. Proc. Natl. Acad. Sci. USA 97, 6640–6645. Elso, C.M., Roberts, L.J., Smyth, G.K., Thomson, R.J., Baldwin, T.M., Foote, S.J., Handman, E., 2004. Leishmaniasis host response loci (lmr1–3) modify disease severity through a Th1/Th2-independent pathway. Genes and Immun. 5, 93–100. Eriksson, S., Lucchini, S., Thompson, A., Rhen, M., Hinton, J.C., 2003. Unravelling the biology of macrophage infection by gene expression profiling of intracellular Salmonella enterica. Mol. Microbiol. 47, 103–118. Eshaghi, S., Niegowski, D., Kohl, A., Martinez Molina, D., Lesley, S.A., Nordlund, P., 2006. Crystal structure of a divalent metal ion transporter CorA at 2.9 angstrom resolution. Science 313, 354–357. Fiser, A., Sali, A., 2003. Modeller: generation and refinement of homology-based protein structure models. Methods in Enzymol. 374, 461–491. Gardner, R.C., 2003. Genes for magnesium transport. Curr. Opin. Plant Biol. 6, 263– 267. Gill, D.L., Waldron, R.T., Rys-Sikora, K.E., Ufret-Vincenty, C.A., Graber, M.N., Favre, C.J., Alfonso, A., 1996. Calcium pools, calcium entry, and cell growth. Biosci. Rep. 16, 139–157. Graschopf, A., Stadler, J.A., Hoellerer, M.K., Eder, S., Sieghardt, M., Kohlwein, S.D., Schweyen, R.J., 2001. The yeast plasma membrane protein Alr1 controls Mg2+ homeostasis and is subject to Mg2+-dependent control of its synthesis and degradation. J. Biol. Chem. 276, 16216–16222. Hehre, W.J., Radom, L., Schleyer, P.v.R., Pople, J.A., 1986. Ab Initio Molecular Orbital Theory. Wiley, New York. Kedzierski, L., Zhu, Y., Handman, E., 2006. Leishmania vaccines: progress and problems. Parasitology 133 (Suppl.), S87–112. Kehres, D.G., Lawyer, C.H., Maguire, M.E., 1998. The CorA magnesium transporter gene family. Microb. Comp. Genomics 3, 151–169. Knoop, V., Groth-Malonek, M., Gebert, M., Eifler, K., Weyand, K., 2005. Transport of magnesium and other divalent cations: evolution of the 2-TM-GxN proteins in the MIT superfamily. Mol. Genet. Genomics 274, 205–216. Lanza, H., Afonso-Cardoso, S.R., Silva, A.G., Napolitano, D.R., Espindola, F.S., Pena, J.D., Souza, M.A., 2004. Comparative effect of ion calcium and magnesium in the

723

activation and infection of the murine macrophage by Leishmania major. Biol. Res. 37, 385–393. Lunin, V.V., Dobrovetsky, E., Khutoreskaya, G., Zhang, R., Joachimiak, A., Doyle, D.A., Bochkarev, A., Maguire, M.E., Edwards, A.M., Koth, C.M., 2006. Crystal structure of the CorA Mg2+ transporter. Nature 440, 833–837. Maguire, M.E., Cowan, J.A., 2002. Magnesium chemistry and biochemistry. Biometals 15, 203–210. Maguire, M.E., 2006. The structure of CorA: a Mg(2+)-selective channel. Curr. Opin. Struct. Biol. 16, 432–438. Maloney, K.E., Valvano, M.A., 2006. The mgtC gene of Burkholderia cenocepacia is required for growth under magnesium limitation conditions and intracellular survival in macrophages. Infect. Immun. 74, 5477–5486. Moncrief, M.B., Maguire, M.E., 1998. Magnesium and the role of MgtC in growth of Salmonella typhimurium. Infect. Immun. 66, 3802–3809. Niegowski, D., Eshaghi, S., 2007. The CorA family: structure and function revisited. Cell. Mol. Life. Sci. 64, 2564–2574. Papp-Wallace, K.M., Maguire, M.E., 2007. Bacterial homologs of eukaryotic membrane proteins: the 2-TM-GxN family of Mg(2+) transporters. Mol. Membr. Biol. 24, 351–356. Payandeh, J., Pai, E.F., 2006. A structural basis for Mg2+ homeostasis and the CorA translocation cycle. Embo. J. 25, 3762–3773. Shi, J., Blundell, T.L., Mizuguchi, K., 2001. FUGUE: sequence-structure homology recognition using environment-specific substitution tables and structuredependent gap penalties. J. Mol. Biol. 310, 243–257. Smith, R.L., Maguire, M.E., 1998. Microbial magnesium transport: unusual transporters searching for identity. Mol. Microbiol. 28, 217–226. Stewart, J., Curtis, J., Spurck, T.P., Ilg, T., Garami, A., Baldwin, T., Courret, N., McFadden, G.I., Davis, A., Handman, E., 2005. Characterisation of a Leishmania mexicana knockout lacking guanosine diphosphate-mannose pyrophosphorylase. Int. J. Parasitol. 35, 861–873. Uboldi, A.D., Lueder, F.B., Walsh, P., Spurck, T., McFadden, G.I., Curtis, J., Likic, V.A., Perugini, M.A., Barson, M., Lithgow, T., Handman, E., 2006. A mitochondrial protein affects cell morphology, mitochondrial segregation and virulence in Leishmania. Int. J. Parasitol. 36, 1499–1514. Zsurka, G., Gregan, J., Schweyen, R.J., 2001. The human mitochondrial Mrs2 protein functionally substitutes for its yeast homologue, a candidate magnesium transporter. Genomics 72, 158–168.