Myristoyl-CoA:Protein N-Myristoyltransferase, an Essential Enzyme and Potential Drug Target in Kinetoplastid Parasites*

Co-translational modification of eukaryotic proteins by N-myristoylation aids subcellular targeting and protein-protein interactions. The enzyme that catalyzes this process,N-myristoyltransferase (NMT), has been characterized in the kinetoplastid protozoan parasites, Leishmania andTrypanosoma brucei. In Leishmania major, the single copy NMT gene is constitutively expressed in all parasite stages as a 48.5-kDa protein that localizes to both membrane and cytoplasmic fractions. Leishmania NMT myristoylates the target acylated Leishmania protein, HASPA, when both are co-expressed in Escherichia coli. Gene targeting experiments have shown that NMT activity is essential for viability inLeishmania. In addition, overexpression of NMT causes gross changes in parasite morphology, including the subcellular accumulation of lipids, leading to cell death. This phenotype is more extreme than that observed in Saccharomyces cerevisiae, in which overexpression of NMT activity has no obvious effects on growth kinetics or cell morphology. RNA interference assays in T. brucei have confirmed that NMT is also an essential protein in both life cycle stages of this second kinetoplastid species, suggesting that this enzyme may be an appropriate target for the development of anti-parasitic agents.

Myristoyl-CoA:protein N-myristoyltransferase (NMT 1 ; EC 2.3.1.97) catalyzes the covalent co-translational attachment of the fatty acid, myristate (C14:0) to the amino-terminal glycine residue of a number of eukaryotic cellular and viral proteins of diverse function (1). These include cAMP-dependent serine/ threonine kinases, members of the p60 Src family of tyrosine kinases, retroviral gag polyprotein precursors such as HIV-1pr55, viral capsid components, and the ␣-subunit of many signal-transducing, heteromeric G proteins. Although some myristoylated proteins are cytosolic, many are associated with cellular membranes where myristoylation facilitates membrane attachment. The addition of myristate can also stabilize protein-protein interactions, and many acylated proteins require this modification for full expression of their biological function (2). Genetic studies have shown that the NMT gene is essential for the survival of Saccharomyces cerevisiae (3) and the pathogenic fungi, Candida albicans and Cryptococcus neoformans (4). Given the critical role of N-myristoylation in the cell, NMT has been developed as a target for the development of antifungal chemotherapeutic agents.
The NMT enzyme has an ordered Bi-Bi reaction mechanism (5), binding first to myristoyl-CoA, with the resulting conformational changes generating a peptide-binding site. Subsequent formation of a ternary myristoyl-CoA:NMT-peptide complex leads to catalysis and product release. The myristoyl-CoA binding sites of purified human NMT and the fungal NMTs are highly conserved, but their peptide binding specificities are divergent (6). By exploiting these differences, peptide and peptidomimetic inhibitors with selectivity for fungal NMTs have been designed. For example, an inhibitor based on the N terminus of ADP-ribosylation factor (ARF), the major myristoylated protein detected in [ 3 H]myristate-labeled C. neoformans (causative agent of chronic meningitis in AIDS patients), is competitive for peptide and noncompetitive for myristoyl-CoA binding (7), and a nonpeptide inhibitor based on this sequence is at least ϳ5-fold selective for C. neoformans against human NMT (8). Similarly, peptidomimetic inhibitors based on the N-terminal sequence of C. albicans ARF are 560-fold selective for the fungal NMT as compared with the human enzyme. Gel mobility assays indicate that a single dose of 200 M is sufficient to produce ϳ50% reduction in the myristoylation of C. albicans ARF, which is consistent with fungistatic activity (9). Recent elucidation of the S. cerevisiae and C. albicans NMT crystal structures (10 -12) will facilitate a better understanding of the inhibition mechanisms of these enzymes.
Parasitic kinetoplastid protozoa of the species Leishmania and Trypanosoma are causative agents of some of the most debilitating tropical infections currently affecting world populations (see, on the World Wide Web, www.who.int/tdr/index. html). Metabolic labeling studies have identified at least ten [ 3 H]myristate-labeled proteins in Leishmania major, including the ARF-like protein 1 (53). and the infective stage specific, hydrophilic acylated surface proteins (HASPs) (13,14). HASPB, a candidate vaccine for visceral leishmaniasis (15), requires N-terminal myristoylation and palmitoylation for translocation to the parasite plasma membrane (16). Myristate is also a component of the glycophosphatidylinositol (GPI) lipid anchors that tether the major classes of surface molecules in trypanosomatid parasites. In Leishmania, the myristate-containing GPI anchors of the surface glycoinositolphospholipids are remodeled during synthesis by a reaction involving myristate exchange from a myristoyl-CoA donor (17). In mammalian stages of Trypanosoma brucei, fatty acid remodeling is also important in synthesis of the variant surface glycoprotein GPI anchor, which contains two myristates in its diacylglycerol domain (18). As a consequence, myristate analogues are toxic in T. brucei, probably due to a cumulative effect on GPI anchor, phospholipid, and N-myristoylated protein synthesis (19,20).
Here, we show that myristate analogues (nonspecific inhibitors of NMT) are also toxic to Leishmania and describe the molecular characterization of NMT in both L. major and T. brucei. As with the pathogenic fungi, expression of NMT is essential for viability in these parasites, whereas modulation of enzyme activity inhibits parasite growth, yielding aberrant phenotypes, presumably due to disrupted modification of "downstream" target proteins.

EXPERIMENTAL PROCEDURES
Leishmania Culture and Inhibitor Assays-L. major (MHOM/IL/81/ Friedlin) promastigotes were maintained in vitro at 26°C as described (13). Logarithmically growing parasites at ϳ8 ϫ 10 6 ml Ϫ1 were incubated with different inhibitors (dissolved in 100% ethanol as 100 mM stocks) that had previously been conjugated to fat-free bovine serum albumin as described (16). Ethanol was added to noninhibitor-treated cells to the same final concentration (1%). At the time points indicated, aliquots were removed in triplicate for counting; all motile organisms were scored, including those with abnormal morphology. Each time course was done in triplicate, and the whole experiment was repeated three times.
PCR Amplification, Subcloning, and DNA Analysis-A 677-bp fragment of the NMT gene was amplified from L. major genomic DNA, using Taq DNA polymerase (Promega) at 55°C annealing temperature and the primers, NMT for (5Ј-GAGATYAACTTYCTSTGCGTSCAC-3Ј) and NMT rev (5Ј-RCCRTCRCCIIIRCCRAACTT-3Ј), based on the NMT "signature sequences" (EINFLCVH and KFGXGDG, available on the World Wide Web at www.expasy.org/prosite/) and adjusted to L. major codon bias. The amplified fragment, nmt, was cloned into pGEM-T (Promega) and used to screen a gridded L. major Friedlin genomic cLHYG cosmid library (21). Of the 21 positive clones identified, 12 mapped to the same contiguous sequence on chromosome 33. L. major NMT was subcloned from cosmid L8010 (21) into pBS SKϩ, as 1.6-and 2.0-kb EcoRI/SacI fragments.
For expression in Escherichia coli, the cloned NMT ORF was digested with NdeI/BamHI and subcloned into the expression vector, pET15b (Novagen), to obtain plasmid pNMT. The HASPA ORF was digested with NcoI/XhoI and cloned into the expression vector pET28a (Novagen), generating plasmid pHASPA. The recombinant HASPA expressed from this plasmid has an N-terminal myristoylation motif (MGSSCTK).
Production of Antibodies and Immunoblotting-Expression of N-terminally His-tagged recombinant L. major NMT (from pNMT) was induced by isopropyl-1-thio-␤-D-galactopyranoside in E. coli BL21(DE3)pLysS, and the cells were subsequently lysed in 6 M Gu-HCl, prior to affinity chromatography using Ni 2ϩ -nitrilotriacetic acid-agar-ose (Qiagen). Eluted protein was further purified by SDS-PAGE and used for immunization and generation of rabbit polyclonal antisera (Eurogentec).
Functional Co-expression of Recombinant L. major NMT in E. coli-This assay was based on the co-expression system described in Ref. 2. E. coli BL21(DE3)pLysS cells were cotransformed with pNMT and pHASPA, either singly or in combination. 2-ml cultures were grown to A 600 0.5 in LB media containing ampicillin, kanamycin, and chloramphenicol, and expression of recombinant proteins was induced by the addition of 1 mM isopropyl-1-thio-␤-D-galactopyranoside. Immediately, the cultures were divided into two 1-ml aliquots, with 150 Ci of [9,10-3 H]myristate (Amersham Biosciences) in 10 l of ethanol added to one aliquot and 10 l of ethanol alone added to the second. The cells were then incubated at 37°C for a further 3 h, prior to centrifugation, phosphate-buffered saline washing, lysis, and analysis by SDS-PAGE. Radiolabeled proteins were detected by autoradiography at Ϫ80°C using Hyperfilm MP (Amersham Biosciences).
Parasite Membrane Fractionation-5 ϫ 10 8 midlog phase parasites were lysed and separated into membrane and cytoplasmic fractions as described (23). Proteins in both fractions were analyzed by SDS-PAGE and immunoblotting, as described above.
Targeted Deletion/Complementation of the L. major NMT Gene-To target both NMT alleles, plasmids were generated containing either a hygromycin resistance gene (hyg) or a puromycin resistance gene (pac) flanked with DNA sequences immediately upstream and downstream of the NMT ORF. These constructs, based on the pX63-HYG vector (25), were designated pNMT-HYG and pNMT-PAC, respectively. To generate pNMT-HYG, a 873-bp fragment immediately upstream of the NMT ORF was amplified using primers NKO1F (ATAGCaagcttGCGACGC-GTGGGACTGTGACA) and NKO1R (GTGCATgtcgacCGCAGTATTCG-TGTGCGTGT). Cloning sites are shown in lowercase, and L. major sequence is shown in uppercase. The amplified product was digested with HindIII/SalI and cloned into pX63-HYG, generating plasmid pXN-HYG. Subsequently, a 1110-bp fragment immediately downstream of the NMT ORF was amplified using the primers NKO2F (ATCGTcccg-ggTTTACGCNCCGCCCCACCTTTCCG) and NKO2R (CATGCTagatct-GCACCAGTGCTCGGCGGTCGGC) and cloned into the SmaI/BglII sites of pXN-HYG, generating pNMT-HYG. To construct pNMT-PAC, the pac gene was released from plasmid plmcpb-PAC (26) as previously described (27) and used to replace the hyg gene in pNMT-HYG, following SpeI/BamHI digestion. Linear targeting regions (produced by digesting pNMT-HYG with SapI/MscI and pNMT-PAC with SapI/NcoI) were purified and used sequentially to transfect midlog phase L. major promastigotes as described (28). Transfected parasites were plated in the presence of 32 g ml Ϫ1 hygromycin and 20 g ml Ϫ1 puromycin, clones were amplified, and DNA/protein was extracted for analysis as described (27). The DNA probes HYG (1027 bp) and PAC (600 bp) were generated by digesting pNMT-HYG and pNMT-PUR with SpeI/BamHI. An NMT N-terminal probe (986 bp) was produced by digesting NMT-pTEX with XhoI/SalI. Immunoblotted proteins were detected as described above.
For complementation of transgenic L. major or overexpression in wild type parasites, pNMT was digested with ClaI/SmaI prior to cloning the insertion into the expression vector pTEX (29), generating pTEX NMT NEO. Parasites were electroporated with 20 g of plasmid DNA, prior to plating in the presence of 20 g ml Ϫ1 G418, amplification, and analysis as described above and in Ref. 27.
RNA i in T. brucei-A 543-bp fragment within the NMT ORF was amplified from T. brucei genomic DNA using the primers Tb-NMT-For (5Ј-ACTCCAATAATGGTTCTAGAGAAGAGAGA-3Ј) and Tb-NMT-Rev (5Ј-TCGTCTTTTtctagaCCCACCG-3Ј). T. brucei sequence is shown in uppercase, and cloning sites are underlined. The PCR fragment was digested with XbaI and cloned into the XbaI sites of the RNA interference (RNA i ) vector, p2T7Ti, which supports expression of doublestranded RNA from tetracycline-inducible T7 promoters (30), generating p2T7-NMT.
The procyclic T. brucei strain 29-13 (31), containing the genes expressing T7 RNA polymerase and tetracycline repressor, was maintained in vitro at 27°C in SDM-79 medium supplemented with 10% fetal calf serum, 25 g ml Ϫ1 hygromycin, and 25 g ml Ϫ1 G418. Midlog phase parasites were electroporated with 20 g of NotI-linearized p2T7-NMT DNA, as described (32). The parasites were allowed to recover overnight, and transfectants were selected using 2.5 g ml Ϫ1 phleomycin and grown for 3 weeks until stable cell lines had been established. The expression of double-stranded RNA was induced by the addition of 1 g ml Ϫ1 tetracycline to parasites diluted to 1 ϫ 10 6 ml Ϫ1 in SDM-79 medium. Cell numbers were monitored using a Beckman Coulter counter. Protein samples were prepared at 24-h intervals and analyzed by immunoblotting, as described above.
The bloodstream T. brucei strain 90-13 (31) was maintained in vitro at 37°C with 5% CO 2 in HMI-9 medium supplemented with 10% fetal calf serum, 5 g ml Ϫ1 hygromycin, and 2.5 g ml Ϫ1 G418. Midlog phase parasites were electroporated with 10 g of NotI-linearized p2T7-NMT DNA, as described (33). The parasites were allowed to recover for 6 h, before the addition of 2.5 g ml Ϫ1 phleomycin. The cells were transferred to 24-well culture plates and grown for 10 days until stable cell lines had been established. Expression of double-stranded RNA was induced by the addition of 1 g ml Ϫ1 tetracycline to parasites diluted to 2 ϫ 10 4 ml Ϫ1 in HMI-9 medium. Cell numbers were monitored as above.
Electron Microscopy-Parasites were washed three times in serumfree M199 medium and resuspended at 1 ϫ 10 7 ml Ϫ1 . Parasites were fixed in 4% paraformaldehyde and 2.5% glutaraldehyde in phosphatebuffered saline, pelleted, fixed again in 1% osmium tetroxide in sodium cacodylate buffer, treated with 1% tannic acid, dehydrated through an increasing ethanol series (staining en bloc in 2% uranyl acetate at the 30% ethanol stage), infiltrated, and embedded in TAAB 812 resin (TAAB Laboratories Equipment Ltd.). 60-nm ultrathin sections, cut on a Leica UCT ultramicrotome, were contrasted with uranyl acetate and lead citrate and examined on a Philips CM100 transmission electron microscope. For lipid staining, malachite green was added at 0.5% to the primary fixative solution just prior to fixing on ice for 2 h. Uranyl acetate was used as a light counterstain at the 30% ethanol stage during dehydration, and final contrasting with uranyl acetate and lead citrate, as described above, was omitted.

RESULTS
Effect of Myristoyl-CoA Analogues on the Growth of L. major-Since myristate analogues are nonspecific competitive inhibitors of NMT, we first tested these compounds for their effect on Leishmania growth and division in culture. Specifically, the myristate analogues 2-hydroxymyristate and 4-oxatetradecanoate were incubated with procyclic L. major over a 72-h time course (Fig. 1). Whereas the untreated parasites continued to divide, growth in the inhibitor-treated cultures slowed after 24 h, and viable parasite numbers reached a plateau by 48 h, depending on the inhibitor concentration. At 48 h, there was Ͼ50% reduction in parasite numbers compared with untreated cells, with both inhibitors showing a dose-dependent effect. On microscopic examination, the inhibitor-treated cells showed extensive morphological differences (rounded cells with variable motility and intracellular organelle distention) when compared with the untreated parasites (data not shown).
These results demonstrate that, as with T. brucei, myristate analogues are toxic to Leishmania promastigotes, most likely due to global effects on GPI anchor and phospholipid biosynthesis as well as N-myristoylation of target proteins. It was not possible to correlate the loss of viability observed in the presence of these analogues with reduced N-myristoylation of specific targets, since the single well characterized Leishmania N-myristoylated protein, HASPB, is only expressed later in the growth cycle, following differentiation of procyclic parasites to infective metacyclics (13,14).
Cloning and Characterization of Leishmania N-Myristoyltransferase-The L. major NMT gene was cloned by degenerate PCR amplification of genomic DNA followed by library screening. Contiguous sequence (3480 bp) generated from the subcloned gene fragments contained a 1266-bp ORF, coding for a 48.5-kDa protein of pI 6.1 (accession number AF305956). This deduced protein shares identity with NMT from other species, as shown in the sequence alignment in Fig. 2.
As with all NMTs so far characterized, the L. major and T. brucei enzymes are divergent at their N termini (which are not involved in catalysis) and do not contain N-terminal lysinerich regions (that have been implicated in ribosomal targeting of human and mouse NMTs) (34). Conversely, many of the key fungal NMT residues predicted to be directly involved in the enzyme mechanism or to have critical regulatory roles (6) are conserved in L. major and T. brucei. An exception is Cys 217 of S. cerevisiae NMT (and conserved in other fungal species); mutation of this residue has a selective effect on the S. cerevisiae NMT peptide binding site (35). In L. major and T. brucei NMTs, this residue is replaced by Gly. High resolution structural analysis of the C. albicans enzyme has revealed that two negatively charged residues, Glu 173 and Leu 451 , form the floor of the active site pocket of the enzyme (10). These residues are conserved in both L. major and T. brucei NMTs, but there is also an insertion of ϳ20 amino acids, rich in Lys and Glu, close to the first of these "pocket floor" residues in the parasite enzymes (Fig. 2). An additional insertion of 29 amino acids is found further downstream in T. brucei NMT. These features suggest that the parasite NMTs might form distinctive secondary and tertiary structures that could affect their activities relative to the human and fungal enzymes.
L. major NMT Is a Constitutively Expressed Single Copy Gene-L. major NMT is a single copy gene (as determined by endonuclease digestion, blotting, and hybridization; data not shown). Similarly, the genomes of the New World species, L. amazonensis, and T. brucei also contain single NMT genes.
Leishmania parasites cycle between extracellular promastigotes (within the sandfly vector) and intracellular amastigotes (principally within macrophages of the mammalian host). To investigate NMT expression during the L. major life cycle, equivalent numbers of noninfective (logarithmic or procyclic) and infective (metacyclic) promastigotes and amastigotes (extracted from BALB/c mouse lesions) were lysed and analyzed by SDS-PAGE. After blotting, NMT was immunodetected as a single 48.5-kDa protein, present in all life cycle stages (Fig.  3A). In the same samples, expression of HASPB, a member of the hydrophilic acylated surface protein family, was restricted to infective parasite stages as previously demonstrated (13,14). Expression of the constitutive endoplasmic reticulum marker protein, BiP, was used as a control for equivalent loading. Thus, NMT is constitutively expressed in L. major. Subcellular Localization of L. major NMT-NMT has been localized to the cytoplasm in S. cerevisiae and does not appear to be associated with cellular membranes (36). Conversely, a fraction of Drosophila NMT has been shown to be membranebound (37), and NMTs in mammalian cells (38,39) are all partially associated with membranes, as peripheral membrane proteins. This is also the case for L. major NMT (Fig. 3B). Following separation of parasite extracts into membrane and cytoplasmic fractions, immunoblotting with anti-NMT revealed that Ͼ70% of the L. major enzyme is associated with membranes. As controls for these experiments, GP63 (the major GPI-anchored surface protein of Leishmania species) was detected solely in the membrane fraction, as expected, whereas SHERP, a hydrophilic protein that fractionates with membranes only after in vivo cross-linking (23), was detected only in the cytoplasmic fraction (Fig. 3B).
Functional Analysis of L. major NMT in E. coli-A bacterial co-expression assay system (2) was modified to demonstrate the activity of L. major NMT in vivo. In these experiments, a plasmid expressing L. major NMT (pNMT) and a plasmid expressing the ORF of an NMT substrate were co-transformed into E. coli. Protein expression was induced with isopropyl-1thio-␤-D-galactopyranoside, and N-myristoylation was monitored by radiolabeling. A number of Leishmania substrates have been used in these assays, including ARL-1 (data not shown) and HASPA, a second member of the hydrophilic acylated surface protein family. In the experiment in Fig. 4, E. coli were co-transformed with pHASPA and pNMT prior to metabolic labeling with [ 3 H]myristate. The products were then separated by SDS-PAGE, followed by Coomassie Blue staining and autoradiography.
Following induction of protein expression in the co-transformed bacteria, proteins migrating at 48.5 and 17 kDa were specifically expressed in whole cell lysates (Fig. 4, upper panel,  track 6). These were identified as NMT and HASPA, respectively, by immunoblotting (data not shown). Both proteins were radiolabeled when [ 3 H]myristate was included in the incubation (Fig. 4, lower panel, track 6). The apo-NMT is known to form a high affinity complex with myristoyl-CoA (5) (migrating at 17 kDa) was myristoylated by the recombinant NMT expressed from pNMT. This labeling is resistant to treatment with 1 M hydroxylamine, indicative of an amide linkage, characteristic of N-myristoylation (data not shown). Transformation with pHASPA alone (Fig. 4, lower panel, track 4) generated no detectable radiolabeled product, despite high levels of protein expression (compare tracks 4 and 6). Expression of NMT in the absence of plasmid-encoded substrate resulted in myristoylation of an endogenous 25-kDa E. coli protein (Fig. 4,   lower panel, track 2). Similar results have been observed previously (40,41). Thus, under these assay conditions, recombinant L. major NMT is active in vivo and is required for the N-myristoylation of HASPA.
Leishmania NMT Encodes an Essential Enzyme in Extracellular Parasites-To investigate the importance of NMT expression in the Leishmania life cycle, gene targeting was used to generate transgenic parasites lacking one or both copies of the NMT gene. Linear constructs, containing hygromycin (hyg) or puromycin (pac) resistance genes flanked by sequences from the dihydrofolate reductase locus for constitutive expression (42), were targeted to the NMT locus, using 5Ј and 3Ј intergenic regions for homologous recombination (Fig. 5A). In these experiments, heterozygous lines (containing integrated hyg or pac genes) were readily obtained: of 10 clones analyzed for each integration; six hyg clones and five pac clones resulted from precise chromosomal NMT replacement. However, after a second round of targeting, although it was possible to select clones that were resistant to both drugs, none of these were NMT chromosomal nulls, despite the analysis of Ͼ30 clones by immunoblotting and 20 of these clones by Southern blotting (data not shown). It was only possible to generate true NMT chromosomal nulls when episomal copies of the NMT gene were introduced into heterozygous parasites, prior to targeting the second NMT allele for deletion. Of 17 clones analyzed after complementation in this way, seven had lost both NMT chromosomal copies.
Analysis of a series of typical transgenic clones is shown in Fig. 5B. Integration of the hyg or/and pac gene constructs ( The expression levels of NMT in the four transgenic clones described above were analyzed (Fig. 5C). Using BiP as a control protein, the equivalent NMT expression levels as compared with wild type were 0.4-, 5.9-, and 6.5-fold for transgenic clones 2-4, respectively. Thus, loss of one NMT allele results in ϳ60% decreased protein expression, whereas parasites containing the episome express considerably enhanced levels of NMT, due to the increased gene copy number on the plasmid. Growth of the four clones was also monitored over a 6-day period (Fig. 6D). Transgenic clones 2-4 showed a decrease in growth rate compared with wild type, although total parasite numbers were comparable after 6 days in culture.
Overall, these data resemble those obtained by others when trying to remove sequentially both copies of genes encoding essential proteins in Leishmania (e.g. Refs. [43][44][45]. As a consequence, we conclude that NMT is an essential gene for Leishmania promastigote viability. Overexpression of NMT Is Lethal in Leishmania-To further investigate the effects of increased NMT expression on parasite viability, the pTEX NEO NMT episome was transfected into wild type L. major. The resulting transgenic parasites, selected by their resistance to neomycin, did not differentiate into metacyclics and developed abnormal morphology during growth on plates and in liquid culture. These cells were predominantly rounded, with disrupted subcellular organization and disinte-  2, 4, and 6). The radiolabeled products (the NMTmyristoyl-CoA binary complex and myristoylated HASPA) were detected by autoradiography (lower panel), followed by densitometry to determine relative incorporation levels. The data shown represent one experiment that was repeated three times. Coomassie Blue-stained, SDS-PAGE-separated proteins are shown in the upper panel.
gration of nuclear membranes (Fig. 6, C-E; compare with wild type parasites in Fig. 6, A and B). They contained increased numbers of lysosomes and, as a dominant feature, previously undescribed oval structures, which appeared to be bound by membranes. Staining with malachite green (Fig. 6E) confirmed that these structures are rich in lipids.
Immunoblotting of transfected parasite lysates (prepared from pTEX NEO NMT cells collected at the same time as those processed for electron microscopy) showed a 10.5-fold increase in the expression of NMT protein compared with the wild-type control (Fig. 6F). All clones overexpressing NMT maintained viability for a short time, but cell death occurred after 1-2 weeks in all cases, compared with clones transfected with the pTEX vector alone (Fig. 6G). This is a more extreme phenotype than that seen in Fig. 5D, in which transgenic parasites with a modest 5-6-fold increase in NMT expression showed wild type morphology but had a growth defect. Hence, whereas a 5-fold increase in NMT expression is tolerated by Leishmania, an increase to Ͼ10-fold is lethal. Increasing the drug selection pressure on ϩ/⌬NMT::HYG [pTEX NEO NMT] parasites (transgenic clone 3 in Fig. 5) by 10-fold resulted in increased NMT expression and similar morphological changes to those shown in Fig. 6, C-E.

Expression of the T. brucei NMT Gene Is Essential for Viability in Procyclic and Bloodstream Trypanosomes-Whereas
methods for regulated expression of transgenes are still under development in Leishmania, RNA i has been established as a reproducible technique for gene down-regulation and phenotypic analysis in T. brucei (30,46,47). Thus, to determine the phenotype of trypanosomes deficient in NMT, RNA i constructs were generated to target expression from the T. brucei NMT gene in both vector (procyclic) and mammalian (bloodstream) parasite stages, following induction with tetracycline.
Growth of procyclic trypanosomes of the parental clone, 29-13 (carrying integrated copies of the T7 RNA polymerase and tetracycline repressor genes) and the transgenic clone, p2T7/NMT, were monitored over a time course (Fig. 7A). Both induced and noninduced p2T7/NMT parasites grew at similar rates for the first 3 days, when the cultures were split. From that point, up until 9 days postinduction, growth of the induced p2T7/NMT clone was negligible, whereas the noninduced con-trol replicated as before. Throughout this period, the parental parasites, 29-13, showed a similar (if more rapid) growth pattern to the noninduced control. When cells were split 1 day earlier (at 2 days postinoculation), cell death still occurred at day 4 (data not shown).
Immunoblotting parasite lysates collected at intervals following tetracycline induction demonstrated a decrease in NMT protein expression by 72 h postinduction, correlating with the time at which the cultures were split (Fig. 7B). No further NMT expression could be detected by 96 h, by which time there was a significant reduction in parasite growth and increase in cell death (Fig. 7A). This effect was still apparent at 9 days postinduction, with very low parasite viability and no detectable NMT protein. From 96 h postinduction, there was an accumulation of parasites with aberrant morphology, including rounded or abnormally elongated cells, carrying more than one flagellum and displaying partial plasma membrane detachment ( Fig. 7C and data not shown). This array of phenotypes could be attributed to the effects of NMT depletion on modification of a range of downstream N-myristoylated target proteins. Expression of one of these, the procyclic-specific, cytoskeleton-associated protein CAP5.5 (24), was unperturbed during the time course of this experiment (Fig. 7B).
Similar results were obtained when growth of bloodstream trypanosomes of the parental (90-13) and transgenic (Bp2T7/ NMT) clones were compared (Fig. 7D). The induced parasites grew normally for 72 h and then began to round up, leading to a high rate of cell death by 96 h postinduction. Therefore, we can conclude that, as in L. major, NMT protein is essential for viability of T. brucei in both life cycle stages of the parasite. DISCUSSION The importance of N-myristoylation in the synthesis of functional proteins, often with a role in signaling or other fundamental cellular processes, has focused interest on the enzyme that catalyzes this process, myristoyl-CoA:protein N-myristoyltransferase. The development of peptide and peptidomimetic anti-NMT reagents for potential therapeutic use against fungal diseases has been aided by knowledge of the key structural features of the enzyme. Current published data suggest that both specific and sensitive inhibition of NMT activity can be achieved when compared with host enzyme activity (8,12). Against this background, we have cloned and characterized NMT from several trypanosomatid parasite species, in order to evaluate the functional importance of N-myristoylation in these organisms and to investigate the potential of NMT as a drug target in the diseases that they cause.
The L. major and T. brucei NMTs share high similarity with other characterized NMTs in their primary sequence, including conservation of key amino acids essential for catalysis, including the "pocket floor" residues identified at the active site (10) and the hydrophobic residues (Phe 170 and Leu 171 in S. cerevisiae) critical for myristoyl-CoA binding and product release (11). The noncatalytic N termini of the parasite enzymes show sequence divergence and do not contain the polylysine domains implicated in ribosomal targeting, regions that are also missing from the NMTs of other lower eukaryotic species studied to date. In this context, the recent report of post-rather than co-translational N-myristoylation of the proapoptotic protein, BID, following caspase-8 cleavage, suggests that NMT is localized to more than one distinct intracellular site (48), at least in mammalian species. This observation might correlate with the recent demonstration of a second mammalian NMT cDNA, possibly arising from a second gene, and the expression of different NMT isoforms in mammalian cells (49). Leishmania NMT, encoded by a single copy gene, is unequally distributed between membranes and the cytosol (Fig. 3). NMT is constitutively expressed in Leishmania species and T. brucei, as in the pathogenic fungi. Expression of L. major NMT in E. coli has provided a robust functional assay for enzyme activity in vivo, as originally demonstrated with S. cerevisiae NMT, which has fatty acid and peptide specificities indistinguishable from those of the native enzyme when expressed in bacteria (41). As with S. cerevisiae NMT, two major radiolabeled species are detected in E. coli expressing Leishmania NMT and a suitable substrate: the acyl-enzyme complex and the myristoylated substrate. Both of these are still produced, although less efficiently, when the parasite-specific 20amino acid NMT domain is deleted (data not shown), most likely due to altered higher order structure of the protein. Consistent with this interpretation, construction of a homology model of L. major NMT based on the S. cerevisiae structure has shown that this insertion is on the surface of the protein and does not affect the active site (50). More detailed in vitro analyses with purified soluble enzyme will delineate the importance of this and other parasite-specific regions and residues.
The fungicidal and trypanocidal properties of myristate analogues, nonspecific inhibitors of NMT, have been well documented (7,20,51), although these compounds show distinct organism-specific differences. The two compounds tested here clearly inhibit the growth of L. major promastigotes, presumably due to cumulative effects on various cellular processes. Competitive peptide binding inhibitors will be required to verify this effect and correlate growth inhibition with reduced NMT activity. Using an alternative molecular genetic approach, we have confirmed that NMT expression is essential for the viability of both Leishmania promastigotes and T. brucei procyclic/bloodstream form parasites, whereas increased NMT protein levels also have deleterious effects. In particular, the accumulation of lipid bodies in Leishmania NMT overexpressors, prior to cell death, may be a phenotype specific to cells that express high levels of myristate-containing, GPI-anchored surface molecules. By comparison, 5-6-fold overexpression of NMT activity in S. cerevisiae has no effect on growth kinetics or cell morphology (3). Aberrant N-myristoylation of substrates probably contributes to this unusual parasite phenotype, at the expense of other metabolic fates for the available myristate, which include GPI anchor remodeling (17). However, the recent demonstration of de novo myristate synthesis in T. brucei (52) suggests that fatty acid synthetic pathways may also exist in Leishmania and that myristate concentration may not be a limiting factor contributing to the accumulation of lipid bodies.
Further experiments are under way to assess the importance of NMT expression in intracellular amastigotes of Leishmania and to determine the downstream targets that contribute to the loss of viability observed. It is unlikely that nonmyristoylation of the HASP proteins results in cell death, since transgenic parasites lacking HASP function are as viable as wild type cells (27). By contrast, reduced ARF myristoylation has been shown to cause growth arrest in C. albicans and S. cerevisiae (8,9), suggesting that a member of this protein family may be important for viability in trypanosomatid parasites. Whatever the myristoylated proteins involved, the distinctive features of trypanosomatid NMTs, including their single gene copy number, conserved sequence, and expression in mammalian-infective parasite stages, suggest that these enzymes may be appropriate targets for the development of antiparasitic drugs.