Cloning of Trypanosoma brucei and Leishmania major Genes Encoding the GlcNAc-Phosphatidylinositol De-N-acetylase of Glycosylphosphatidylinositol Biosynthesis That Is Essential to the African Sleeping Sickness Parasite*

The second step of glycosylphosphatidylinositol anchor biosynthesis in all eukaryotes is the conversion of D-GlcNAcα1–6-d-myo-inositol-1-HPO4-sn-1,2-diacylglycerol (GlcNAc-PI) tod-GlcNα1–6-d-myo-inositol-1-HPO4-sn-1,2-diacylglycerol by GlcNAc-PI de-N-acetylase. The genes encoding this activity are PIG-L and GPI12 in mammals and yeast, respectively. Fragments of putative GlcNAc-PI de-N-acetylase genes from Trypanosoma bruceiand Leishmania major were identified in the respective genome project data bases. The full-length genes TbGPI12and LmGPI12 were subsequently cloned, sequenced, and shown to complement a PIG-L-deficient Chinese hamster ovary cell line and restore surface expression of GPI-anchored proteins. A tetracycline-inducible bloodstream form T. brucei TbGPI12conditional null mutant cell line was created and analyzed under nonpermissive conditions. TbGPI12 mRNA levels were reduced to undetectable levels within 8 h of tetracycline removal, and the cells died after 3–4 days. This demonstrates thatTbGPI12 is an essential gene for the tsetse-transmitted parasite that causes Nagana in cattle and African sleeping sickness in humans. It also validates GlcNAc-PI de-N-acetylase as a potential drug target against these diseases. Washed parasite membranes were prepared from the conditional null mutant parasites after 48 h without tetracycline. These membranes were shown to be greatly reduced in GlcNAc-PI de-N-acetylase activity, but they retained their ability to make GlcNAc-PI and to processd-GlcNα1–6-d-myo-inositol-1-HPO4-sn-1,2-diacylglycerol to later glycosylphosphatidylinositol intermediates. These results suggest that the stabilities of other glycosylphosphatidylinositol pathway enzymes are not dependent on GlcNAc-PI de-N-acetylase levels.

A significant proportion of eukaryotic cell-surface glycoproteins are attached to the plasma membrane by covalent linkage to a glycosylphosphatidylinositol (GPI) 1 membrane anchor. The structure and biosynthesis of GPI membrane anchors and related molecules have been reviewed recently (1)(2)(3)(4). The basic GPI core structure attached to protein comprises NH 2 CH 2 CH 2 PO 4 H-6Man␣1-2Man␣1-6Man␣1-4GlcN␣1-6-D-myoinositol-1-HPO 4 -lipid, where the lipid can be diacylglycerol, alkylacylglycerol, or ceramide. This minimal GPI structure may be embellished with additional ethanolamine phosphate groups and/or carbohydrate side-chains in a species-and tissuespecific manner (5).
Protozoa tend to express significantly higher densities of cell-surface GPI-anchored proteins than do higher eukaryotes (1,4,6). For example, Trypanosoma brucei, the causative agent of African sleeping sickness, expresses a dense cell-surface coat consisting of ϳ5 ϫ 10 6 dimers of a GPI-anchored variant surface glycoprotein that protects the parasite from the alternative complement pathway of the host and, through antigenic variation, from specific immune responses (7). The related kinetoplastid parasite Leishmania sp. expresses lower copy numbers of GPI-anchored glycoproteins, such as the promastigote surface protease (Psp or gp63), gp42, and GPI-anchored proteophosphoglycans, but high copy numbers of the GPI-related structures lipophosphoglycan and the glycoinositolphospholipids (4,6,8,9). It has been suggested by several groups that inhibitors able to arrest the formation of GPI-anchored proteins and/or GPIrelated molecules on the plasma membrane of parasitic protozoa might prove useful in the development of antiparasitic agents. This notion has been validated for T. brucei, where disruption of the TbGPI10 gene encoding the third mannosyltransferase of GPI anchor biosynthesis has been shown to be lethal for the bloodstream form of the parasite (10,11). The situation is less clear in Leishmania sp., where, for example, L. mexicana is infective without lipophosphoglycans, glycoinositolphospholipids, and GPI-anchored glycoproteins, whereas L. major is significantly attenuated in the absence of lipophosphoglycan (12)(13)(14).
The sequence of events underlying GPI biosynthesis has been studied in T. brucei (15)(16)(17)(18)(19)(20), Trypanosoma cruzi (21), Toxoplasma gondii (22), Plasmodium falciparum (23), Leishmania sp. (24 -26), Saccharomyces cerevisiae (27,28), and mammalian cells (29 -31), and references therein. In all cases, GPI biosynthesis involves the addition of GlcNAc to phosphatidylinositol (PI) to give GlcNAc-PI, which is then de-N-acetylated by N-acetyl-D-glucosaminylphosphatidylinositol deacetylase (EC 3.1.1.69), referred to here as GlcNAc-PI de-Nacetylase, to form GlcN-PI (32)(33)(34)(35). De-N-acetylation is a prerequisite for the mannosylation of GlcN-PI to form later GPI intermediates (34,36). The GlcNAc-PI de-N-acetylases from protozoan and mammalian sources are similar with regard to their specificities for the acyl (R) group removed from GlcNR-PI substrates (36), but differ with regard to their specificity for the myo-inositol residue. Thus, the trypanosomal enzyme can de-N-acetylate GlcNAc-PI containing either D-or L-myo-inositol and ␣or ␤-D-GlcNAc, whereas the human (HeLa) enzyme strictly requires ␣-D-GlcNAc (1-6)D-myo-inositol (37,38). These differences, and the ability of the trypanosomal enzyme to tolerate a C8 O-alkyl substituent on C2 of the D-myo-inositol residue, were recently exploited in the design and synthesis of two parasite-specific GlcNAc-PI de-N-acetylase suicide substrate inhibitors (38). The gene encoding the rat de-N-acetylase (PIG-L) was the first to be cloned (34) and a yeast homologue (GPI12) has been shown to complement PIG-L-deficient mammalian cells and vice versa (35). Here, we describe the molecular cloning of the T. brucei and L. major homologues TbGPI12 and LmGPI12, demonstrate functional complementation in a PIG-L-deficient mammalian cell line, and describe the creation of a T. brucei TbGPI12 conditional null mutant. We further demonstrate that membranes from the conditional null mutant are deficient in GlcNAc-PI de-N-acetylase activity under non-permissive conditions and that TbGPI12 is an essential gene in bloodstream form T. brucei.

EXPERIMENTAL PROCEDURES
Cloning of T. brucei and L. major GPI12-The 401-bp end-sequence of an Institute for Genomic Research genome survey sequence clone (AQ644232), returned from the tBLASTn search with yeast GPI12p (accession number P23797), was used to design a reverse PCR primer (5Ј-cgcGGATCCtcatgcgacccccaattccttcacttc-3Ј (capital letters indicate a BamHI site)) that was used with Pfu polymerase, blood-stream form T. brucei cDNA, and a forward primer based on the 5Ј mini-exon (5Ј-ggcccgctattattagaacagtttctgta-3Ј) to amplify an ϳ0.8-kb fragment containing the entire TbGPI12 ORF. Amplification conditions were 95°C for 45 s, 60°C for 1 min, and 72°C for 3 min for 30 cycles. The PCR product was purified from an agarose gel (QIAEX II kit) and ligated into a pUC18 cloning vector using a SureClone ligation kit (Amersham Biosciences). Twelve representative clones were used for DNA sequencing, revealing a 759-bp ORF. The same PCR product (the TbGPI12 probe) was fluorescein-labeled by random priming (Gene Images kit; Amersham Biosciences) or labeled with 32 P (Prime-It RmT random primer labeling kit; Stratagene) for use in Southern blotting and for probing a BAC library filter, respectively (see below).
A tBlastn search with the sequence LVIAHPDDEAMFFAP, a sequence strictly conserved in rat and human PIG-L and substantially conserved in yeast GPI12, identified an L. major expressed sequence tag sequence (AA728250) with 85% similarity. The corresponding cDNA clone was kindly provided by Prof. J. M. Blackwell (Cambridge University) and fully sequenced. The clone contained the full-length LmGPI12 gene.
Functional Complementation-For expression in mammalian cells, the TbGPI12 gene was PCR-amplified using Pfu in two segments. The 5Ј-end of the ORF was amplified using forward primer 5Ј-gagAAGCT-TCATATGcatggtgctttggcgtttggg-3Ј and reverse primer 5Јcatggcggaaaa-gctGgtgaacaatgag-3Ј and the 3Ј-end of the ORF was amplified using forward primer 5Ј-ctcattgttcacCagcttttccgccatg-3Ј and reverse primer 5Ј-cgGGATCCtcaCAGGTCCTCCTCCGAGATTAGCTTCTGTTCGTTA-ATTAAtgcgacccccaattcctt-3Ј. The two PCR products were used together in a further Pfu PCR reaction to yield a product containing a silent mutation that removed a HindIII site (capital italic letters indicate the mutation), a myc epitope tag fused to the C terminus of TbGPI12 (underlined letters), and 5Ј-HindIII and 3Ј-BamHI restriction sites (capital letters). The purified construct was digested with HindIII and BamHI and ligated into the respective cloning sites of the pcDNA3.1/Hygro (ϩ) (Invitrogen) mammalian expression vector.
The LmGPI12 gene was also PCR-amplified using Pfu in two segments. The 5Ј-end of the ORF was amplified using forward primer 5Ј-cccAAGCTTgggatgcacagtatcacagtt-3Ј and reverse primer 5Ј-gcaggtg-gaggatGcctggaggcatgttc-3Ј and the 3Ј-end of the ORF was amplified using forward primer 5Ј-gaacatgcctccaggCatcctccacctgc-3Јand reverse primer 5Ј-cgcGGATCCgcgctagagctcttcgatctc-3Ј. The two PCR products were used together in a further Pfu PCR reaction to yield a product containing a silent mutation that removed a BamHI site (capital italic letters indicate the mutation) and 5Ј-HindIII and 3Ј-BamHI restriction sites (capital letters). The purified construct was digested with HindIII and BamHI and ligated into the respective cloning sites of pcDNA3.1/Hygro (ϩ).
Southern Blots-T. brucei genomic DNA (5 g/lane) was digested with various restriction enzymes and the products were resolved on a 0.7% agarose gel. After transfer to nitrocellulose and UV-cross-linking, the blot was hybridized with fluorescein-labeled TbGPI12 probe (16 h, 60°C) and washed twice with 1ϫ SSC, 0.1% SDS for 15 min and twice with 0.5ϫ SSC, 0.1% SDS for 15 min. Blots were developed with horseradish peroxidase-conjugated anti-fluorescein antibody according to the manufacturers instructions (gene images CDP-Star kit; Amersham Biosciences).
Generation of a Bloodstream Form T. brucei TbGPI12 Conditional Null Mutant-A T. brucei strain 427 BAC library filter (CHORI RPCI-102), representing 46-fold genome coverage, was probed with a 32 Plabeled TbGPI12 probe under the same conditions described for the Southern blot. Fifty positive clones were identified and the corresponding TbGPI12-containing BAC plasmids were purified from 3-ml cultures of four clones, using the "DNA isolation from BAC & PAC clones" protocol recommended by CHORI BACPAC Resources (www.chori.org/ bacpac). The presence of the TbGPI12 gene was confirmed by PCR using 5Ј-gagAAGCTTCATATGcatggtgctttggcgtttggg-3 and 5Ј-cgcGGATCCtcatgcgacccccaattccttcacttc-3Ј forward and reverse primers. One clone was selected and the purified BAC plasmid DNA (1 g) was used as template for DNA sequencing using primers from within the gene, 5Ј-catcgcttcatcgtccgggtgtgc-3Ј and 5Ј-attccgccgacctcattgttcaca-3Ј. Consequently, 466 bp of 5Ј-UTR and 648 bp of 3Ј-UTR sequence were obtained. Based on these data, 426 bp of 5Ј-UTR immediately upstream of the start codon were PCR-amplified using Pfu and genomic DNA template with the forward primer 5Ј-ataagaatGCGGCCGCcctccccccgcgcctacggatg-3Ј and reverse primer 5Ј-gtttaaacttacggaccgtcaagcttgtgtatgagcgactccctcaac-3Ј. Likewise, 478 bp immediately downstream of the stop codon were PCR-amplified using with the forward primer 5Ј-gacggtccgtaagtttaaacggatccatcgaagaaatttagcccccgc-3Ј and reverse primer 5Ј-ataagtaaGCGGCCGCcgactccggcatcttgtaaattg-3Ј. The two PCR products were used together in a further PCR reaction to yield a product containing the 5Ј-UTR linked to the 3Ј-UTR by a short HindIII, PmeI, and BamHI cloning site (underlined letters) and NotI restriction sites at each end (capital letters). Subsequently, the PCR product was cloned into the NotI site of pGEM-5Zf(ϩ) vector (Promega) and the hygromycin phosphotransferase (HYG) and puromycin acetyltransferase (PAC) drug resistance genes were introduced into the targeting vector via the HindIII/BamHI cloning site. The previously described HindIII-silenced, C-terminally myc-tagged TbGPI12 construct was ligated into the HindIII/BamHI cloning site of the pLew100 tetracycline-inducible ex-pression vector (39). Plasmids were prepared (Qiagen Maxi-Prep), digested with NotI, precipitated with ethanol, redissolved in sterile water, and used for electroporation of bloodstream form T. brucei strain 427 (variant 221), which are stably transfected to express T7 RNA polymerase and tetracycline repressor protein under continuous G418 selection (39). Cell culture, transformation and drug selection conditions were as described previously (39 -42). Tet-system approved fetal calf serum (Clontech) was used in experiments on the effects of tetracycline-removal.
Northern Blots-Total RNA was prepared (Qiagen RNeasy Protect Midi kit) from 2 ϫ 10 8 cells. Samples of RNA (5 g) were run on formaldehyde agarose gel and transferred to Hybond-N nylon membrane (Amersham Biosciences) for hybridization with [␣-32 P]dCTP labeled TbGPI12 probe (Stratagene Prime-It RmT random primer labeling kit). As a loading control, a ␤-tubulin probe was used on the same blot.
Cell-free System Experiments-Bloodstream form T. brucei membranes (cell-free system) were prepared (15,19,43) from wild-type cells and TbGPI12 conditional null mutant cells grown continuously in 1 g/ml tetracycline and grown tetracycline-free for 48 h. Trypanosome membranes were washed twice and resuspended at 5 ϫ 10 8 cell-equivalents/ml in 2ϫ incorporation buffer supplemented with 10 mM N-ethylmaleimide or 2 mM dithiothreitol (43,44). The lysates were briefly sonicated and aliquots of 10 7 cell equivalents were added to an equal volume of GDP- HPTLC and Enzyme Digests-Samples were digested with jack bean ␣-mannosidase and Bacillus thuringiensis phosphatidylinositol-specific phospholipase C (both from Glyko) as described in (46). Samples and glycolipid standards were applied to 10 cm aluminum-backed silica gel-60 HPTLC plates (Merck) and developed with chloroform/methanol/1 M ammonium acetate/13 M ammonia/water (180:140:9:9:23, v/v). Radiolabeled components were detected by fluorography at Ϫ70°C after spraying with En 3 Hance (PerkinElmer) using Kodak XAR-5 film and an intensifying screen.

Cloning of T. brucei and L. major GlcNAc-PI De-N-acetylase
Genes (TbGPI12 and LmGPI12)-A partial gene sequence was found in The Institute for Genomic Research T. brucei data base with a tBLASTn search, using the Saccharomyces cerevisiae GPI12 protein sequence as the query. The putative TbGPI12 gene fragment contained the 3Ј-end of the gene (377 bp), including a stop codon, followed by 24 bp of putative 3Ј-UTR. A cDNA clone was obtained by PCR using bloodstream form T. brucei cDNA as the template, a forward primer based on the 5Ј-spliced leader (a 35-bp sequence trans-spliced onto all T. brucei mRNA), and a reverse primer based on the 3Ј-end sequence found in the data base. The 0.8-kb product contained a 759-bp ORF (GenBank TM accession number AY157267). A genomic clone was also obtained by PCR using genomic DNA as the template, a forward primer to the 5Ј-end of the gene based on the cDNA clone, and a reverse primer based on the 3Ј-UTR sequence found in the data base. Both the cDNA and genomic DNA ORF sequences were identical.
A tBlastn search with the sequence LVIAHPDDEAMFFAP, a conserved sequence in rat and human PIG-L that is largely conserved in yeast GPI12 (34), identified an L. major expressed sequence tag sequence. The corresponding cDNA clone was fully sequenced and found to contain the full-length LmGPI12 gene (GenBank™ accession number AY157268).
The predicted amino acid sequences of the two parasite putative GlcNAc-PI de-N-acetylases, aligned with related sequences, are shown in Fig. 1. All of the sequences predict proteins with the majority of their sequence in the cytoplasm and anchored to the endoplasmic reticulum via a single Nterminal transmembrane domain, an arrangement that has been demonstrated experimentally for rat PIG-L (34). The PIG-L/GPI12 sequences have similarity to the pfam02585 family of prokaryote sequences that include Rv1170, a GlcNAc␣1-1-Dmyo-inositol de-N-acetylase involved in mycothiol synthesis in Mycobacterium tuberculosis (47).
Complementation of a PIG-L-deficient CHO-K1 Reporter Cell Line-The LmGPI12 and TbGPI12 genes, the latter fused to a C-terminal myc-tag, were cloned into the pcDNA3.1 mammalian expression vector. The recombinant plasmids were used to transiently transfect a previously established CHO-K1 reporter cell line (34) deficient in PIG-L. This reporter cell line is stably transfected to express two GPI-anchored proteins: CD59 and DAF. These proteins fail to receive a GPI anchor because of the PIG-L deficiency and cannot be detected on the cell surface by immunofluorescence microscopy ( Fig. 2A) unless the cells are complemented by transfection with a gene (e.g. rat PIG-L) encoding a functional GlcNAc-PI de-N-acetylase (Fig. 2B). Transient transfection with both parasite putative GlcNAc-PI de-N-acetylase genes produced similar results (Fig. 2, C and D), demonstrating that TbGPI12 and LmGPI12 encode functional GlcNAc-PI de-N-acetylases.
TbGPI12 Is an Essential Gene in Bloodstream Form T. brucei-Southern blot analysis revealed that the TbGPI12 gene was present as a single copy per haploid genome (Fig. 3). Replacement of both alleles of the TbGPI12 gene from the diploid genome of bloodstream from T. brucei parasites was attempted. Gene replacement by homologous recombination of a single TbGPI12 allele with the drug-resistance gene PAC or HYG, and selection for transformants with the appropriate antibiotic, was successful. However, attempts to replace the second TbGPI12 allele in a ⌬TbGPI12::PAC clone with HYG failed, suggesting that TbGPI12 may be essential. To test this, we created a conditional, tetracycline-inducible null mutant. The "wild-type" trypanosome cell line used in this study is a transgenic parasite that constitutively expresses T7 RNA polymerase and the tetracycline repressor (TETR) protein under G418-selection (39). Thus, a tetracycline-inducible (Ti) ectopic myc-tagged TbGPI12 gene was introduced into the trypanosome ribosomal DNA locus (using the pLew100 expression vector) downstream of a trypanosome (procyclin) promoter and two tetracycline operator sequences (Fig. 4A). Several TbGPI12-myc Ti ⌬TbGPI12::PAC clones were isolated, and one of these was induced with tetracycline and used for a second round of homologous recombination to replace the second endogenous TbGPI12 allele with HYG. Five TbGPI12-myc Ti ⌬TbGPI12::PAC/⌬TbGPI12::HYG conditional null mutant clones were obtained. A Southern blot of one of these clones, and its TbGPI12-myc Ti ⌬TbGPI12::PAC, ⌬TbGPI12::PAC, and wild-type parent cell-lines, confirmed the loss of both chromosomal TbGPI12 alleles and the introduction of an ectopic gene copy (Fig. 4B).
The TbGPI12 conditional null mutant cells grew continuously in culture in the presence of 1 g/ml tetracycline. Cells were counted daily and cultures were split when densities approached 2 ϫ 10 6 cells/ml (Fig. 5A). However, cells washed three times and cultured in medium without tetracycline grew for about 3-4 days and then died (Fig. 5B). The few surviving cells (below the limits of detection by light microscopy) failed to divide unless tetracycline was reintroduced at day 6, whereupon they resumed sustained growth (Fig. 5C), or until around day 14, when some cultures spontaneously started to grow once more at normal rates (Fig. 5B). The spontaneous recovery of conditional null mutant cultures in the absence of tetracycline has been reported previously (40,42,48) and is caused by the loss of tetracycline control through, for example, deletion of the TETR gene (42).
Northern blot analysis showed that TbGPI12 mRNA levels were undetectable within 8 h of tetracycline removal (Fig. 6). However, the cells continued to divide for 2-3 days, suggesting that it takes this length of time for de-N-acetylase loss (because of dilution by cell division and protein turnover) to reach unsustainable levels. Taken together, these data demonstrate that expression of the TbGPI12 gene is essential for parasite growth.
Biochemical Phenotype of the TbGPI12 Conditional Null Mutant-The TbGPI12 conditional null mutant was grown with and without tetracycline for 48 h, harvested, and made into cell-free system (washed membrane preparations) (15) to analyze aspects of GPI biosynthesis. Incubation of the cell-free systems prepared from tetracycline-induced (ϩTet) cells with UDP-[ 3 H]GlcNAc, followed by glycolipid extraction and analysis by HPTLC and fluorography, revealed the formation of [ 3 H]GlcN-PI (15,32) and some labeled downstream GPI products formed from [ 3 H]GlcN-PI at the expense of limiting amounts of dolichol-P-Man in the washed membranes (19) (Fig.  7A, lane 1). However, labeling of the cell-free system prepared from non-induced (ϪTet) cells revealed a build up of [ 3 H]Glc-NAc-PI (Fig. 7A, lane 2), consistent with significantly lower GlcNAc-PI de-N-acetylase activity in those membranes. The presence of some [ 3 H]GlcN-PI and labeled downstream GPI products in (lane 2) shows that there is still some GlcNAc-PI de-N-acetylase activity in these cells, explaining the viability of the cells at the point of harvest.
The same cell-free systems, and one prepared from wild-type cells, were labeled with GDP-[ 3 H]Man in the presence and absence of UDP-GlcNAc (in the presence of DTT to stimulate UDP-GlcNAc:PI ␣-GlcNAc-transferase activity) or synthetic GlcNAc-PI or GlcN-PI (in the presence of N-ethylmaleimide to inhibit UDP-GlcNAc:PI ␣-GlcNAc-transferase activity (49)). In the first set of experiments (Fig. 7B), the addition of synthetic GlcNAc-PI to the cell-free systems from wild-type and tetracycline-induced (ϩTet) conditional null mutant cells lead to similar levels of 3 H-mannosylated GPI products, from Man 1 GlcN-PI to glycolipid AЈ, whereas more dolichol-P-[ 3 H]Man and less 3 H-mannosylated Man 1 GlcN-PI to glycolipid AЈ were observed with the cell-free systems from the noninduced (ϪTet) conditional null mutant cells. These results are also consistent with lower GlcNAc-PI de-N-acetylase activity in the non-induced membranes. The control experiment, using GlcN-PI instead of GlcNAc-PI, produced an interesting result.
Because the processing of GlcN-PI does not require prior de-Nacetylation, we expected the products of the wild-type and induced (ϩTet) and non-induced (ϪTet) cell-free systems to be similar. We observed, however, that although the level of labeling was indeed comparable, the non-induced cell-free system produced a more complex band pattern than the wild-type and induced cell-free systems. To investigate this further, additional experiments were performed comparing the addition of GlcN-PI with the addition of UDP-GlcNAc and with no additions (Fig. 7C). As expected, in the absence of GlcN-PI and UDP-GlcNAc, the wild-type and induced (ϩTet) cell-free systems produced few 3 H-mannosylated products from the processing of the limiting amounts of GPI intermediates in these membranes (Fig. 7C, lanes 3 and 4). However, the non-induced (ϪTet) cell-free system produced abundant 3 H-mannosylated GPI intermediates, up to and including glycolipid AЈ and glycolipid (Fig. 7C, lane 9). Furthermore, the levels of these products were not increased by the addition of UDP-GlcNAc (Fig. 7C, lane 8), whereas the addition of UDP-GlcNAc greatly stimulated the formation of 3 H-mannosylated GPI intermediates in the wild-type and induced (ϩTet) cell-free systems (Fig.  7C, lanes 2 and 5). We interpret these results to mean that the non-induced (ϪTet) cells accumulate significant amounts of endogenous GlcNAc-PI in their endoplasmic reticulum membranes because of significantly reduced GlcNAc-PI de-Nacetylase activity, such that cell-free system prepared from these cells harbors a significant pool of endogenous GlcNAc-PI. Thus, when GDP-[ 3 H]Man is added to the cell-free system, endogenous GlcN-PI generated by the action of residual de-Nacetylase activity on this pool provides the necessary substrate for the formation of 3 H-mannosylated GPI intermediates. These 3 H-mannosylated GPI intermediates are based on endogenous GlcN-PI that contain predominantly stearic acid at the sn-1 position and a mixture of C18-C22 fatty acids at sn-2 (49) and, therefore, have slightly higher R F values than those 3 H-mannosylated GPI intermediates made from exogenous synthetic sn-1,2-diplamitoylglycerol-containing GlcN-PI. Thus, the combination of 3 H-mannosylated GPI intermediates made from endogenous and exogenous GlcN-PI leads to the complex band pattern in Fig. 7B (lane 6 compared with lanes 4 and 5) and Fig. 7C (lane 7 compared with lanes 1 and 6).
Regardless of the complexities of the results described above, it is clear that the GlcNAc-PI de-N-acetylase deficient membranes of the conditional null mutant grown for 48 h without tetracycline are perfectly capable of synthesizing GlcNAc-PI and of processing GlcN-PI to later GPI intermediates. This suggests that neither the multicomponent UDP-GlcNAc:PI ␣1-6 GlcNAc transferase (50) nor the downstream mannosyltransferases and ethanolamine phosphate transferases are significantly affected by a reduction in GlcNAc-PI de-N-acetylase T. brucei genomic DNA was digested with the restriction enzymes indicated and subjected to Southern blotting with the TbGPI12 probe (see "Experimental Procedures"). A restriction map of the TbGPI12 ORF predicted from the DNA sequence is shown above the blot. The results indicate that TbGPI12 is a single-copy gene per haploid genome.
level. Because the expression levels of tightly associated subunits of GPI biosynthesis complexes do affect each other (50,51), this result may suggest that the T. brucei GlcNAc-PI de-N-acetylase does not tightly associate with the upstream GlcNAc-transferase or the downstream ␣-mannosyltransferase. On the other hand, the apparent substrate channelling between the de-N-acetylase and the downstream ␣-mannosyltransferase (43) suggests some spatial proximity of these enzymes. Further experiments are required to determine whether the T. brucei de-N-acetylase is associated with other components of GPI biosynthesis or whether, like the mammalian enzyme (34), it behaves as a free-standing protein.

CONCLUSIONS
Two kinetoplastid parasite GlcNAc-PI de-N-acetylase gene homologues were readily identified and cloned, thanks to the ongoing T. brucei and L. major genome sequencing projects, and their functionality was confirmed by complementation. The TbGPI12 gene has been shown here to be a single-copy gene that is essential for the disease-causing bloodstream form of T. brucei, thus validating this particular enzyme as a potential drug target for the development of therapeutic agents against African sleeping sickness and, possibly, against related parasitic diseases, such as the leishmaniases and Chagas' disease. The genetic validation of GlcNAc-PI de-N-acetylase as an antitrypanosomal target is particularly significant because there has been considerable biochemical characterization of this enzyme (32)(33)(34)(35)(36)(37)(38). Attractive features of this potential drug FIG. 4. Construction and characterization of a conditional TbGPI12-null mutant. A, scheme of the targeted replacement of one TbGPI12 allele with PAC, the introduction of an ectopic tetracycline-inducible copy of myc-tagged TbGPI12 into the rDNA locus, and replacement of the second TbGPI12 allele with HYG. The cells used for the aforementioned transformations stably express T7 polymerase and tetracycline repressor protein (TetR) under G418 selection (39). B, aliquots of DNA (5 g) were digested with PstI (which gives conveniently sized TbGPI12-containing fragments) and Southern blotted with the TbGPI12 probe. The DNA was from wild-type target include: (i) relatively low identity and similarity (36 and 54%) between the T. brucei and human enzyme peptide sequences; (ii) significant differences between the substrate specificities of the T. brucei and human enzymes (37,38); and (iii) the recent synthesis of two potent (IC 50 , 8 nM) parasite-specific suicide substrate inhibitors of the enzyme (38).