Transposon Mutagenesis of Trypanosoma brucei Identifies Glycosylation Mutants Resistant to Concanavalin A

We have engineered Trypanosoma brucei with a novel mariner transposition system that allows large populations of mutant cells to be generated and screened. As a proof of principle, we isolated and characterized two independent clones that were resistant to the cytotoxic action of concanavalin A. In both clones, the transposon had integrated into the locus encoding a homologue of human ALG12, which encodes a dolichyl-P-Man: Man(7)GlcNAc(2)-PP-dolichyl-alpha6-mannosyltransferase. Conventional knock-out of ALG12 in a wild-type background gave an identical phenotype to the mariner mutants, and biochemical analysis confirmed that they have the same defect in the N-linked oligosaccharide synthesis pathway. To our surprise, both mariner mutants were homozygous; the second allele appeared to have undergone gene conversion by the mariner-targeted allele. Subsequent experiments showed that the frequency of gene conversion at the ALG12 locus, in the absence of selection, was 0.25%. As we approach the completion of the trypanosome genome project, transposon mutagenesis provides an important addition to the repertoire of genetic tools for T. brucei.

occurs via a nonreplicative cut-and-paste mechanism, in which a staggered double-strand break at each end of the transposable element releases it from the donor DNA molecule, freeing it to be ligated into a staggered cut at the target site (19). Transposition is independent of host-specific factors (20), which has allowed mariner to be used for gene disruption in a variety of organisms, including Leishmania (21), nematodes (22) and vertebrates (23,24).
We have used the Mos1 mariner transposon from Drosophila mauritiana (25). This 1.3-kb transposon is flanked by 28-bp inverted repeats and encodes a single protein of 345 amino acids, the transposase. Mos1 integrates at a TA dinucleotide, which is duplicated upon insertion (reviewed in (17)). By expressing Mos1 transposase constitutively, from a chromosomally integrated gene, and supplying a transposable donor cassette on a single-copy autonomously replicating plasmid (26), we were able to build a system that gives a high transposition efficiency in cell populations of unlimited size, which would not be possible if the system depended upon transient transfection of the mariner components.
As a test of the system, we decided to select mutants that resisted the cytotoxic effects of Concanavalin-A (ConA), which binds to the N-glycans on the EP1 and EP3 Procyclins, and kills procyclic T. brucei via an unknown mechanism that resembles programmed cell death in metazoa (27,28). ConA has been used to isolate glycosylation mutants of T. brucei after chemical mutagenesis (29), but the lack of methods for genetic mapping prevented the mutated genes from being identified.
From a T. brucei population selected for mariner transposition events, we isolated and characterized two independent mutants, in which the transposon had disrupted a previously uncharacterized ORF at different sites. The ORF encodes a probable orthologue of the ALG12 mannosyltransferase. Mutations at the asparagine-linked glycosylation (ALG) pathway are the by guest on March 24, 2020 http://www.jbc.org/ Downloaded from origin of a number of severe inherited human diseases, known collectively as congenital disorders of glycosylation or CDG (30). Cloning of human ALG12 showed that mutation in this locus is the underlying cause of a new subtype of CDG I (31). Targeted deletion of both ALG12 alleles in wild-type T. brucei produced the same phenotype as the mariner mutants. The Nglycan precursor in wild-type T. brucei is Man 9 GlcNAc 2 -PP-dolichol, whereas ALG12 mutants make a smaller precursor, Man 7 GlcNAc 2 -PP-dolichol.

EXPERIMENTAL PROCEDURES
Constructs and Cell Lines-One copy of the Mos1 transposase coding sequence from pBluescribe M13+/Mos1 (25) was targeted to the TUBULIN locus on Chromosome I of wildtype procyclic T. brucei of the Lister 427 strain, by transfection with NotI-linearized pSgl29, which was derived by several cloning steps from a version of the TUB-targeting construct pHD309 (32) containing a puromycin-resistance gene in place of hygromycin phosphotransferase. The donor transposon cassette, pSgl33 (Fig. 1A), is a derivative of the single-copy autonomously replicating T. brucei plasmid pT13-11 (26). As an intermediate step, pBSMos1Hyg was made by cloning the hygromycin phosphotransferase gene, flanked by 5' and 3' ACTIN-derived-splicing and polyadenylation signals (the SmaI-StuI fragment of pHD309), into the SalI site in the transposase ORF of pBluescribe M13+/Mos1. The XhoI-HindIII fragment from pBSMos1Hyg was blunt-end cloned into the EcoNI site of pT13-11, to create pSgl33. ALG12 knockout constructs were made by PCR-amplifying 300 bp of genomic sequences flanking the ALG12 ORF and inserting selectable marker cassettes between them. The neomycin (pSgl39) and hygromycin (pSgl41) cassettes, both containing ACTIN-derived-splicing and polyadenylation signals,were obtained by SmaI-StuI digestions of pLew114 and pHD309 (33), respectively. Bloodstream and procyclic ALG12 null mutants were obtained by sequentially 6 transfecting wild-type T. brucei with NotI-linearized pSgl39 and pSgl41. Elimination of the ALG12 gene was confirmed by Southern hybridization. Trypanosomes were cultivated and transfected as previously described (32).
DNA Analysis-Southern analysis was carried out according to standard protocols (34).
Approximately 2 µg of genomic DNA was run on an 0.8% agarose gel and transferred to Hybond-N membranes (Amersham) by standard procedures. Probes were generated using a ConA Screening-6 x10 7 transposase-expressing procyclic trypanosomes were transfected with the mariner donor plasmid (pSgl33) and amplified for nine days in the presence of 50 µg ml -1 G418. Approximately 4 x 10 8 G418 R cells were washed and resuspended in SDM-79 containing 25 µg ml -1 of hygromycin. The cells were grown for 10 days to allow expression of mutant phenotypes. The culture (100 ml) was treated with 2 µg ml -1 of ConA (Sigma, St Louis, MO) overnight, then centrifuged (50 x g for 5 min) to eliminate large clumps of cells. One round of ConA addition, centrifugation and recovery for a few days in the absence of ConA, is considered a 'treatment'. We performed one more 'treatment' with 2 µg ml -1 of ConA, five with 5 µg ml -1 and one with 10 µg ml -1 of ConA. The entire procedure took approximately one month.
Identification of TbALG12 Gene-The hygromycin and ConA-resistant cells were cloned and the mariner integration sites were determined as described above, except that a different pair of primers was used: SG16 (CGGCACGAAACTCGACATGTTGACTGCACT, beginning 161 bp from the mariner 5' IR) and SG17 (CGAAGACGCCGAACTGCAAGCATTATTGGA, beginning 872 bp from the mariner 3' IR). The inverse PCR product was cloned into pGEM-T Easy vector (Promega) after A-tailing (performed according to the manufacturer's directions).
The cloned DNA was sequenced and analyzed as described above.

Preparation of Cell Lysates and Radiolabeling of Oligosaccharide-Lipids-Hypotonic
lysates of wild-type, ALG12 +/-, ALG12 -/-, and mariner ConA R trypanosomes were prepared as described (35), except that tunicamycin was omitted during preparation of the membranes. For labeling oligosaccharide-lipid (OSL) and other glycolipids, we used a protocol designed for GPI synthesis (35). (DuPont) and exposed to X-ray film (Amersham) at -80 ºC.
Mild Acid Hydrolysis of OSL-Dried [ 3 H]Man-labeled lipids were resuspended in 100 µl of 0.1 M HCl and heated at 100 ºC. After 10 min, the tubes were cooled to 0 °C and the released glycans separated from glycolipids. n-Butanol (100 µl) was added, and the sample was equilibrated twice with 100 µl of water-saturated n-butanol. Glycans, in the aqueous phase, were dried in a SpeedVac and dissolved in 40 % 1-propanol for TLC analysis using nbutanol/acetone/water (6:5:4 V/V/V).

MALDI-TOF-MS Analysis of
Procyclins-For mass-spectrometry, procyclins were purified from 2 x 10 8 freeze-dried parasites by sequential extraction with organic solvents (7,9). To remove GPI anchors from the procyclins, the dry butanolic extracts were treated with 25 µl of 48% aqueous hydrofluoric acid for 14 h at 0 °C. After treatment, samples were immediately frozen in dry-ice-ethanol, dried in a SpeedVac, and resuspended in 20 µl of 0.1% TFA. An aliquot (~0.5 µl; ~5 x 10 6 parasite equivalents) was co-crystallized with α-cyano-4hydroxycinnamic acid as the matrix and analyzed in a PerSeptive Biosystems Voyager Elite mass spectrometer (7). To confirm assignments, HF-treated samples (2 µl, ~2 x 10 7 parasite equivalents) were submitted to mild acid hydrolysis with 40 mM TFA for 15 min at 100 °C, and an aliquot analyzed by MALDI-TOF-MS as described above (7).

Transposition in Trypanosoma brucei-To circumvent the low transfection efficiency of T.
brucei, we started with the objective of establishing permanent cell lines that could be amplified prior to selecting for transposition, and screening or selecting for specific phenotypes. The approach that we used was to split the two components of mariner, creating a donor transposon containing a selectable marker that could only be expressed after transposition into a transcribed region of a chromosome, and providing tranposase activity in trans. Some early experiments that failed to result in transposition included attempts to mobilize a chromosomally integrated transposon cassette by placing a chromosomally integrated copy of transposase under the control of an inducible promoter. On the other hand, co-transfection of separate plasmids encoding the transposase and a donor cassette resulted in transposition into the genome, so it was clear that-in principal-mariner could function in T. brucei.
The approach that achieved our objectives was to create a cell line in which the mariner transposase gene was integrated into the TUBULIN locus on chromosome I, where it is constitutively expressed by read-through transcription, and cloning the donor element into pT13-11, an autonomously replicating single-copy episome (26), creating pSgl33 (Fig. 1A). pSgl33 is stable under G418 selection but does not confer hygromycin resistance: HYG is in the opposite orientation to the direction of transcription, which originates from the single EP-1 procyclin promoter that drives NEO expression. Transposase-expressing trypanosomes were transfected with pSgl33 and amplified in the presence of G418, which was subsequently replaced by hygromycin, and the cells were distributed into multi-well plates. Hygromycin-resistant clones were observed at a frequency of approximately 5% (240 cells were distributed per 24 wells in a typical experiment, yielding ~12 clones per plate). Southern analysis of these clones suggested that the donor element had transposed to different regions of the genome (data not shown). The mariner integration sites from several clones were recovered by inverse PCR. Sequence analysis ( Fig. 1B) showed that they all contained the mariner cassette flanked by a TA dinucleotide and sequences that were absent from the donor plasmid DNA, but were represented in the T. brucei genome database.
ConA Screening and Identification of the TbALG12 Gene-To test the potential for isolating specific mutants with this system, we chose to screen procyclic T. brucei for resistance to ConA, which was expected to identify genes involved in N-glycosylation. T. brucei mutants with reduced ConA binding had been isolated after chemical mutagenesis (29,36) but, although these mutants were biochemically characterized, the genetic defect responsible for the altered phenotype could not be identified.
The transposase-expressing cell line was transfected with donor plasmid pSgl33 and the population was amplified under G418 selection. The G418-resistant population was then washed, split, and grown either in the absence or presence of 25 µg ml -1 of hygromycin for 10 days, to allow full expression of any mutant surface glycoproteins before adding ConA. After the 6th cycle of ConA treatment, only the cells growing in the presence of hygromycin no longer agglutinated. After two more cycles of ConA treatment, analysis of ConA binding by flow cytometry confirmed the low affinity of the surviving cells for ConA. Preliminary DNA analysis (data not shown) suggested that we had a mixed population of mutants that were resistant to both hygromycin and ConA. After cloning by limiting dilution, Southern blot analysis of individual clones showed two patterns (data not shown). One clone (F and L) of each type was selected for further characterization. Once the integration site in clone F was amplified, cloned and sequenced, primers were made to amplify the flanking sequences of the targeted ORF and the ORF itself. DNA analysis ( Fig. 2A) showed that the plasmid-derived transposable cassette had integrated into the same gene in both clones. More surprising was the absence of bands characteristic of the wild-type allele in both mutants (see later). The insertion site in clone F was a TA dinucleotide in the middle of the gene, whereas the insertion in clone L had occurred in a TA located 7 bp upstream of the translation initiation codon (Fig. 2B). BLAST searching identified the target as an unannotated ORF on chromosome II.
The disrupted ORF appeared to be an orthologue of the human ALG12 gene, which encodes dolichyl-P-Man:Man 7 GlcNAc 2 -PP-dolichyl α6-mannosyltransferase (31). The predicted amino acid sequences of T. brucei, C. elegans, D. melanogaster, budding yeast, fission yeast, mouse and human ALG12 are aligned in Fig. 3. TbALG12 has 38% and 33% similarity with S. cerevisiae and human ALG12, respectively. Motif 1 (TKVEESF) is conserved only in the ALG12 family of α6-mannosyltransferases, whereas motif 2 (the sequence HKEXRF flanked by hydrophobic regions) also occurs in the α2-mannosyltransferases (37). ALG12 is a single-copy gene on chromosome II and northern analysis showed that it is expressed in both procyclic and bloodstream T. brucei (data not shown).
ConA Resistance Phenotype is due to Gene Conversion-T. brucei is diploid, so transposonmediated insertional mutagenesis was expected to result in the isolation of heterozygous cells.
However, restriction mapping ( Fig. 2A) and PFGE separation of chromosomal DNA (Fig. 4) suggested that both mariner clones F and L had undergone a gene-conversion-mediated loss of heterozygosity (LOH). Conventional targeted deletion of the ALG12 ORF from wild-type cells showed that the wild-type and ALG12 +/heterozygotes were equally sensitive to ConA: deletion of both alleles was necessary to confer resistance (data not shown).  Table 1 for mass assignments).

Mariner ConA R Mutants Express Procyclins with Altered Glycosylation-To
There is also a small amount of unglycosylated EP1-2, and small amounts of unglycosylated EP3 and EP1-1 are probably also present, but obscured by other peaks. PNGase F treatment of aqueous HF-treated procyclins yielded peaks of identical mass (m/z 8,506, 9,213 and 10,314), corresponding to the unglycosylated polypeptides, from wild-type and mariner ConA R clones ( Fig. 6). Finally, analysis of the C-terminal fragments after mild trifluoroacetic acid hydrolysis, which cleaves the EP procyclins at their mild-acid-labile Asp-Pro bonds (7), showed the same characteristic C-terminal ions from EP1-1, EP1-2 and EP3 (m/z 7,001, 5,870 and 5,191) in all samples (data not shown). Taken together, these analyses suggest that, in contrast to wild-type cells, which express glycosylated EP procyclins bearing a Hex 5 HexNAc 2 , both mariner ConA R clones express the same proteins with altered glycosylation, predominantly modified by a shorter high-mannose glycan (Hex 4 HexNAc 2 ) or a hybrid-type glycan with an extra HexHexNAc sequence that is probably a terminal N-acetyllactosamine (see below). The presence of peptides carrying a Hex 5 HexNAc 3 glycan (denoted as species B) suggests that one of the terminal αMan residues is capped with an N-acetyllactosamine (βGal-βGlcNAc), repeat that is similar to the modification found in the T. brucei ConA R clone ConA 1-1 (29,38).

Analysis of Procyclins from ALG12
To confirm this interpretation, we incubated aqueous-HF-treated ALG12 -/procyclins with a mixture of bovine testis β-galactosidase and jack bean β-hexosaminidase. Simultaneous incubation with these enzymes removes N-acetyllactosamine units (9). As shown in Fig. 7B, this treatment eliminated species B but A remain intact, confirming that only the former glycopeptides contain a terminal N-acetyllactosamine. Identical results were obtained after treatment of procyclins from both mariner ConA R clones (data not shown).

Analysis of the OSL Precursor from Mariner ConA R Mutants and ALG12 -/cells-A defect
in ALG12 implies that mariner mutants should make OSL precursor with a shorter glycan core, which, in turn, will account for the altered glycan structure on their procyclins. To characterize the OSL precursor, glycolipids were synthesized in a cell free system consisting of washed trypanosome membranes. Under these conditions, incubation with UDP-GlcNAc and GDP- (35). Although the mature OSL (Man 9 GlcNAc 2 -PP-lipid) was synthesized in the wild-type cell free system, as expected (39), and by ALG12 +/membranes (Fig. 8A, lanes 1 and 2), it was absent from the ALG12 -/and the two mariner mutant samples (lanes 3-5), where less polar species-designated 'Mutant OSL'-were observed, migrating slightly faster than PP1. Other components were identified as GPI/OSL intermediates and PP1 and P3, the well-characterized GPI anchor precursors of procyclic trypanosomes (40,41). Mutant OSL components were sensitive to mild acid hydrolysis, and their synthesis was inhibited by tunicamycin (data not shown), consistent with their proposed designation as N-glycan precursors. Based on this experiment, we concluded that the mariner and ALG12 -/mutants make a less polar oligosaccharide-PP-dolichol, whose glycan structure must be shorter than that from wild type cells.

To prove that mutant OSL has a smaller glycan, total [ 3 H]Man-labeled glycolipids from
ALG12 -/and from ALG12 +/-, which produces large amounts of native OSL (Fig. 8), were subjected to mild acid hydrolysis under conditions that do not affect GPIs. The released glycans were recovered from the aqueous phase after butanol-water partitioning, and the products were analyzed by TLC. After mild acid hydrolysis, a ladder of glycans ranging from Man 3 GlcNAc 2 up to Man 9 GlcNAc 2 were released from ALG12 +/-OSLs (Fig. 8B, left lane). The same ladder of [ 3 H]Man-labeled glycans was observed from ALG12 -/-OSLs, except that the most hydrophilic component was smaller and co-migrated with the Man 7 GlcNAc 2 standard. Thus, as predicted, disruption of ALG12 in procyclic T. brucei produces a mature OSL that has a shorter glycan, containing a Man 7 GlcNAc 2 instead of Man 9 GlcNAc 2 , as in wild-type cells (39).

DISCUSSION
Due to their ancient divergence from the lines of evolution that have provided most of the model organisms used for the study of basic biological processes, trypanosomes differ from other eukaryotes in a wide range of pathways-from gene transcription to glycolysis. Trypanosomes are also important pathogens that exert their virulence through novel mechanisms such as antigenic variation, which gives us two important motives to study them. Although trypanosomes represent the most intensively studied group of differently evolved eukaryotes, their novelty results in a greatly reduced ability to identify the functions of important genes by bioinformatics methods. To facilitate the elucidation of important processes in trypanosomes, it is vital to develop forward-genetics tools, to supplement the available reverse-genetics approaches.
Experiments in Leishmania major provided the first example of trans-kingdom transposition of the Drosophila mariner element (21), and prompted us to evaluate its potential as a genetic tool for T. brucei. Our first attempts to establish the mariner system in T. brucei were based on the construction of a cell line in which we could regulate expression of the mariner transposase, using the tet repressor system (33,42), and a separately integrated copy of a donor transposon that allowed for the selection of transposition events. This approach did not work, even when the donor was provided on a transiently transfected plasmid. Fortunately, the constitutive expression of transposase that we subsequently adopted did not lead to unstable insertions that we could detect within the time frame of our experiments. When hygromycin-resistant cell lines were cultivated for more than two months, in the absence of selection, the Southern blot pattern of the HYG gene was unchanged (data not shown). In Drosophila melanogaster, MosI-dependent mobilization of mariner transposons containing various inserts, including GFP at the SalI site is very inefficient: no mobilization was detected in 831 samples (43).
We based our decision to use ConA resistance, as a real-life test of mariner mutagenesis, on the foreknowledge that ConA-resistant mutants could be generated by chemical mutagenesis, and on the preliminary characterization of one such mutant (ConA 1-1: Paul T. Englund, personal communication). The initial biochemically identified defect in ConA 1-1 was a reduction in the conversion of polyprenol to dolichol, which was tentatively ascribed to mutation of one allele of an unidentified polyprenol reductase gene. Thus, a mariner hit in one allele of the same unidentified gene was expected to confer resistance to ConA. Subsequent work, however, showed that the dolichol-linked glycan precursor in ConA 1-1 also contained a modified glycan that would be consistent with a concomitant mutation in ALG12, although this has not been confirmed. The possibility remains that the ConA 1-1 phenotype is due to disruption of one polyprenol reductase allele, and that the consequent dolichol deficiency causes an alteration in the structure of the OSL (39).
The mariner-targeted gene in the two clones that we characterized in detail is ALG12, based initially on sequence homology and confirmed by structural analysis of the mutant glycans. The T. brucei ALG12 protein is considerably larger than it is in other organisms. It contains several insertions within the region aligned in Fig. 3, and a carboxy-terminal extension. There is no structural information that can be used to interpret these differences, at present.
Our structural data on the Procyclin N-glycans from the ALG12 -/and mariner ConA R clones are consistent with the lack of transfer of the α1-6-linked Man residue (the enzymatic product of ALG12) to the Manα1-6βManβ1-arm of the tri-mannosyl core (Fig 8C). This is also consistent with the formation of an OSL carrying a smaller glycan (Man 7 GlcNAc 2 ) by these mutant cells (Figs. 8A and B). The implications of such a defect, with regard to the structure of mature glycans linked to surface glycoproteins, have been discussed extensively (29,36,39,44).
One issue that has been clarified, from our study of ALG12 -/cells, is that upon transfer of the truncated (Man 7 GlcNAc 2 ) OSL onto nascent proteins, the Man 4 GlcNAc 2 glycan that results from the trimming of all α1-2Man residues can also be modified by addition of a terminal Nacetyllactosamine repeat. The hybrid-type glycans are responsible ConA-resistance, as they are poorly bound by ConA. It also suggests that the trypanosome UDP-GlcNAc:glycoprotein GlcNAc transferase type I, GnTI, prefers Man 4 GlcNAc 2 instead of Man 5 GlcNAc 2 (as in other organisms) as a substrate. It also may explain why procyclic trypanosomes, which are known for making exclusively high Man oligosaccharides, never modifies the Man 5 GlcNAc 2 glycan linked to procyclin molecules. Further research on the expression and substrate specificity of the trypanosome GnTI is necessary to fully clarify this issue.
RNA interference (RNAi) is another potentially powerful approach that has recently been used for large-scale genetic screening in T. brucei. Procyclic T. brucei were stably transfected with an inducible RNAi-generating vector containing a library of random genomic sequences.
The selection of ConA-resistant mutants was also used as one test of this approach, but did not result in the identification of glycosylation mutants. Instead, ConA selection led to the isolation of cells that had switched to the expression of one of the non-glycosylated Procyclin variants.
The change was shown to be due to the RNAi-induced silencing of the hexokinase gene, a totally unanticipated result (45). A screen for Tubercidin resistance also produced an unexpected result, which led directly to the demonstration of phosphoglycerate kinase as a target for tubercidin in T. brucei (46). Both of these reports emphasize the importance of applying forward genetics to T. brucei. The fact that RNAi and mariner mutagenesis identified different genes that mediated ConA resistance demonstrates the value of using alternative approaches.
Transposon mutagenesis and RNAi libraries could provide parallel and complementary approaches to forward genetics in T. brucei. RNAi is not limited by the ploidy of the organism, but it has limitations that do not apply to the mariner system. Most importantly, it is limited by transfection and chromosome-integration efficiency, which precludes its use in the bloodstream forms, and it is limited by the leakiness of the currently available vector systems, which prevent its use for genes that are very sensitive targets for RNAi. Insertional mutagenesis by electroporation of preformed transposon complexes into T. brucei procyclic forms has also been reported (47), but the efficiency of this approach precludes its general application, especially to the pathogenic stage. Transposon mutagenesis in general, however, was expected to be limited to identifying mutations where disruption of only one allele would cause a discernable phenotype.
Unexpectedly, we found that ConA selection resulted in the isolation of mutants in which both alleles of ALG12 had been disrupted, by mariner insertion at one allele followed by geneconversion-mediated LOH. That the loss of the second ALG12 allele in both cases was indeed due to gene conversion-not a second mariner hit-was inferred from the fact that the insertion was present at the same TA site in both alleles in both mutants (data not shown). In any case, a second mariner transposition should not have been possible: the plasmid supplying the selectable transposon is single-copy and rapidly lost after removal of G418. It was possible that LOH was a very low frequency event that we detected only because of the strong selection by ConA, over a significant period of time. If the intrinsic LOH frequency were low, it could be difficult, because of the size of the populations that would have to be screened, to use mariner for mutant screens when strong selection was not an option. We therefore measured the rate of LOH at the ALG12 locus in the absence of selection. To do this, we took an ALG12 +/heterozygote clone, generated by targeted gene disruption, and analyzed the proportion of cells that spontaneously lost their ability to bind fluorescein-conjugated ConA. These negative cells were sorted, immediately cloned, and PCR was used to verify that, in most of the ConA-negative clones, the second ALG12 allele had been lost. No ConA-negative clones were obtained from populations of wildtype cells. The proportion of ALG12 -/cells arising from spontaneous LOH was 0.24% and 0.25%, in two independent experiments, which was far higher than anticipated. These experiments will be reported in detail elsewhere, after they have been extended to additional loci.
LOH has also been measured in L. major (21,48), where the rate ranged from 10 -4 to 10 -6 at the DHFR-TS locus. The authors were unable to account for the large variation in frequency between experiments, but were able to demonstrate that the LOH frequency could be increased by gamma irradiation. As previously noted (48), an LOH rate at the upper end of this range could have important implications for genetic variability within this group of organisms.
Our experiments provide the first example of the use of transposon insertion to isolate specific mutations in trypanosomatids, and provide an approach that can be easily scaled up to select or screen for mutations in other pathways and processes of interest. Mariner mutagenesis will be most useful if it can identify genes involved in trypanosome infectivity and virulence. We are currently extending the system to the pathogenic bloodstream stage of T. brucei. The transposon donor cassette has been cloned into pT11-bs, a version of the pT13-11 plasmid that can be propagated in the bloodstream stage, to create pSgl35. Although the original description of these plasmids noted that pT13-11 propagated poorly in bloodstream T. brucei, in contrast to pT13-11 in procyclic forms (26), other investigators did not find this to be a problem (11), although an initial lag in growth of G418-resistant plasmid-containing cells has been observed by us and by others.
With a detectable transposition efficiency in the range of 5% (another 5% of insertions will be in the wrong orientation, with respect to the direction of transcription, and perhaps 5% will be in transcriptionally silent loci, indicating an overall transposition rate closer to 15%, comparable to the value of 23% calculated for L. major (21)), we have an efficient new tool to manipulate the T.brucei genome. We estimate the chance of mariner disrupting a single allele to be >10 -5 per cell division, based on an assumption that T. brucei contains about 5,000 genes and knowing that genes are closely packed on the chromosomes. This rate would be at least 10-fold higher than the rate of spontaneous mutation calculated for the Herpes thymidine kinase gene inserted into a VSG expression site (49), and slightly higher than the measured frequency of mariner insertion into the DHFR-TS locus in Leishmania major (21), where the spontaneous mutation frequency was in the range of 10 -6 to 10 -7 . With the end of the T. brucei genome project in sight, the challenge now lies in devising good screening techniques to help unravel the many complexities of this parasite.  This species was observed only in procyclins from wild-type and ALG12 +/cells (Fig. 5A) d Form with ethanolamine linked to the C-terminal glycine, but missing the N-glycan chain. This species was observed only in procyclins from ALG12 -/and mariner clones ( Fig. 5B and 7A). e Glycosylated polypeptide with altered N-glycan chain and an ethanolamine linked to the Cterminal glycine. This species was observed only in procyclins from ALG12 -/and mariner clones ( Fig. 5B and 7A  Manα1-2Manα1-3 Manα1-2Manα1-2Manα1-3 Manß1-4GlcNAcß1-4GlcNAc-PP-Dol