Experimental tsetse flies were from the Bristol laboratory colony of Glossina morsitans morsitans originally from Zimbabwe. Flies were kept at 25°C and 70% relative humidity, and fed on sterile defibrinated horse blood via a silicone membrane; bloodmeals for infected flies were supplemented with 2.5% w/v bovine serum albumen (Sigma A4503) 9 and 1 mM dATP 10. Male flies were used for experiments, being given the infective bloodmeal for their first feed 24-48 hours post-eclosion. The infective bloodmeal contained approximately equal numbers of bloodstream form (BSF) trypanosomes of each strain (approximately 8 × 10⁶ trypanosomes ml^-1) in sterile horse blood, or procyclics in washed red blood cells (approximately 10⁷ trypanosomes per ml of packed cells), supplemented with 60 mM N-acetylglucosamine 11 or 10 mM L-glutathione 12 to increase infection rates.

The trypanosome clones used were 1738 (T. b. brucei MOVS/KE/70/1738 13 and J10 (T. b. brucei MCRO/ZM/73/J10 CLONE 1 14) transfected with either a gene for green fluorescent protein (GFP) or monomeric red fluorescent protein (mRFP) 15 as described previously 8. A selectable marker gene, conferring resistance to the antibiotic hygromycin (Hyg) or phleomycin (Ble), was located downstream of the fluorescent protein gene in each plasmid construct. Constructs were targeted to the non-transcribed spacer of the ribosomal RNA locus. The clones used for mating were designated as follows according to fluorescence; the antibiotic resistance gene is given in brackets: 1738 GFP (Hyg), 1738 RFP (Ble), J10 GFP (Ble), J10 RFP (Hyg). The intraclonal crosses are referred to as 1738 RGFP and J10 RGFP. A yellow fluorescent hybrid clone, SG3 clone 7, was also transmitted through tsetse. This hybrid clone originated from a cross of J10 RFP (Hyg) and 1738 GFP (Hyg) and had a DNA content similar to that of the parental trypanosomes, 2N 8. Procyclic forms of SG3 clone 7 were transmitted through tsetse to obtain BSF, which were subsequently retransmitted; procyclic clones were derived from both BSF populations after transformation in vitro.

Whole tsetse alimentary tracts, from the proventriculus to the hindgut, were dissected at various timepoints after feeding (21-54 days) in a drop of phosphate buffered saline (PBS) and scored for the presence of trypanosomes. A subset of these flies was used to quantify the proportion of mixed infections in the midgut by fluorescence microscopy; for the 1738 RGFP cross, trypanosomes were counted in midguts from a sample of flies dissected 21 to 32 days post infected feed, as described by 11. Whole salivary glands from flies were dissected into a drop of PBS and viewed as wet mounts under bright field illumination to search for trypanosomes. Positive glands were then viewed by fluorescence microscopy using a DMRB microscope (Leica) equipped with a Colour Coolview camera (Photonic Science) and ImagePro Plus software (Media Cybernetics). Glands found to contain a mixed population of trypanosomes were digitally imaged before inoculation into mice. Trypanosomes from the first wave of parasitaemia were transformed to procyclics in vitro (see below).

BSF trypanosomes from mice were transformed to procyclics in Cunningham's medium 16 supplemented with 10% v/v heat-inactivated foetal calf serum, 5 μg/ml hemin and 10 μg/ml gentamycin (complete medium = CM) at 27°C and tested for drug resistance to hygromycin and/or phleomycin at 50 μg ml^-1 and 1 μg ml^-1 respectively in CM in a microtitre plate format. Plates were generally examined every 1 or 2 days for 2 weeks and individual wells passaged as necessary. Clones from transformed procyclics were obtained by limiting dilution in 96 well plates.

Genomic DNA samples were prepared from approximately 5 × 10⁷ washed procyclics using a spin column DNA purification kit (Qiagen). Samples for pulsed field gel (PFG) electrophoresis were prepared by lysing and deproteinising trypanosomes in situ in agarose blocks 17. PFG electrophoresis, blotting and hybridization were carried out essentially as described previously 8. Chromosomes were separated using a Biorad CHEF-DR III with a 2 phase program (Block 1: switch time 1800 s, voltage 2 V/cm, angle 106°, 15 hours; Block 2: switch time 300-900 s, voltage 3 V/cm, angle 106°, 50 hours) using 0.5 × TBE buffer and 0.9% agarose gels; gels were stained overnight by submersion in electrophoresis buffer containing ethidium bromide (2 μg/ml). Blots were hybridized with the following P³²-labelled DNA fragments: GFP and mRFP genes from the plasmid constructs used for transfection; β-tubulin from cDNA plasmid clone 18; 18S rRNA, trypanothione synthetase (TS; chromosome II), paraflagellar rod protein (PFR1, chromosome III) genes from T. brucei genomic DNA 19. Microsatellite analysis was carried out using eight chromosome-specific primer sets - see Tables 3, 4, 520. Primer sets for each locus were as follows: II-PLC 5' CAACGACGTTGGAAGAGTGTGAAC, 5' CCACTGACCTTTCATTTGATCGCTTTC; III-2 5' GGTGGAATGGAAGATCAGTT, 5' GTTGGAATTGTTGTTGCTGT; IV-TB4/7 5' CCGTCACACGCCATGCACGATATG, 5' CCGTTCAGTGTGCATGTTTCAC; VI-TB6/1 5' CATGATGCGGAACACATGACC, 5' CTCAATAGTGCAAGTAGTCATAC; VIII-TB8/9 5' CCAAATATGCGATTAGTTTCC, 5' TGTTTATGTGGAAGGAAATGAA; IX-TB9/6 5' AAGTGTGAGGAGTTGTTGT, 5' CACCCCTTTCATCAACATCAT; X-TB10/17 5' CTCGCACGTTCGGATTTATGTCCG, 5' GCGACTTGTGACTTGCCTTTCTTC; XI-53 5' CGTGTGTCTTGTATATCTTCT, 5' TGAATAAACAAAACATGAAACGAC. Products were resolved by electrophoresis in 1 × TAE buffer through 3-5% Metaphor agarose (Cambrex) gels.

DNA contents of individual clones were measured by flow cytometry of fixed procyclics stained with propidium iodide as described 8. Measurements were standardized by reference to the DNA content of T. b. brucei strain 427, arbitrarily assigned a value of 1.00; procyclics of strain 427 were included in every batch of samples run.

For the cross 1738 RGFP, a total of 461 flies were co-infected with approximately equal numbers of BSF trypanosomes of 1738 RFP and 1738 GFP and dissected in batches 21-54 days later. The overall midgut infection rate was 43% (197/461); in a subset of these fly midguts examined by fluorescence microscopy, the majority contained a mixture of red and green fluorescent trypanosomes (78/88 = 89%), with 8% (7/88) having only green and 3% (3/88) only red fluorescent trypanosomes. Counts of trypanosomes in midguts dissected between 21 and 32 days after infection showed no significant difference in numbers of green (1.54 × 10⁵ ± 2.75 × 10⁴) and red fluorescent trypanosomes (1.61 × 10⁵ ± 1.90 × 10⁴) (t = 0.247, df = 18, P = 0.808, n = 19, Paired-samples t-test). There was thus no detectable difference in the ability of these two recombinant clones of strain 1738 to establish infection in the fly midgut.

There were a total of 22 salivary gland infections, giving a transmission index of 11% (22/197). In seven flies, the salivary glands contained a mixed infection of red and green fluorescent trypanosomes, while the remainder had trypanosomes of only a single colour (Table 1). Small numbers of yellow trypanosomes were seen in the salivary glands of three of the seven flies with a mixed salivary gland infection (Table 2; Fig 1). This contrasts with the abundance of yellow fluorescent trypanosomes typically observed in salivary glands containing both 1738 GFP and J10 RFP in the previous interclonal cross 8.

Metacyclics from the seven flies with mixed salivary gland infections were transformed to BSF in rodents and then into procyclic populations in vitro to facilitate drug selection and cloning. No yellow fluorescent trypanosomes were observed in any of the procyclic populations, with a mixture of red and green fluorescent trypanosomes evident in only two populations (Fly 6 and 7; Table 2). Growth with antibiotics showed drug resistance consistent with the observed fluorescence of the procyclic populations, but no evidence of double-drug resistant trypanosomes (Table 2). Clones were derived from populations 1, 6 and 7, all of which originated from salivary glands with red, green and yellow fluorescent trypanosomes (Table 2). Of 20 clones analysed, four from Fly 6 and 7 lacked fluorescence, while the rest were either red or green (Table 3). These clones were genotyped by microsatellite analysis of eight chromosome-specific loci (Table 3). While all clones from Fly 1 and 6 matched the 1738 genotype, five clones from Fly 7 showed homozygosity at loci where 1738 was heterozygous, indicating that they are recombinant (Table 3; Fig 2A and 2B). In each case, homozygosity was demonstrated at 3-4 independent loci, making it unlikely that these results can be explained by gene conversion.

This interpretation was supported by analysis of molecular karyotypes by PFG electrophoresis. Inspection of the ethidium-stained gel revealed small differences among clones from Fly 7, while clones from Fly 1 and 6 were indistinguishable from 1738 GFP or 1738 RFP; Fig 3A shows a comparison of clones from Fly 6 and 7. The variability of Fly 7 clones was confirmed by hybridization of the PFG blot with a probe for 18S rDNA (Fig 3B), which is carried on several different chromosomes (rDNA arrays on chromosomes II, III and VII 19), and a series of chromosome-specific probes. For the PFR1 locus on chromosome III, where the two homologues of 1738 differ substantially in size (~1.7 and 2.2 Mb), only a single band was evident in Fly 7 clones 1, 2 and 6 (Fig 3C). As both alleles c and d were detected in these clones for microsatellite locus III-2 (Table 3), two chromosome III homologues of the same size are present; moreover, as allele c is normally located on the smaller of the two chromosome III homologues in strain 1738, but is now on a large homologue, chromosomal recombination has occurred. Further evidence of chromosomal recombination is the presence of two copies of the mRFP gene in Fly 7 clone 2, while only a single copy is present in 1738 RFP and other red fluorescent clones (Fig 3D and 3E). The mRFP gene was targeted to integrate into the non-transcribed spacer region of one of the rDNA arrays, and co-localization with a probe for the trypanothione synthetase (TS) gene showed the recombinant array to be on chromosome II (Fig 3F). While the chromosome II homologues of strain 1738 are approximately equal in size (~1.2 Mb), one homologue is much larger (~1.8 Mb) in Fly 7 clone 2 and both homologues carry the mRFP gene (Fig 3E and 3F).

Measurement of DNA contents by flow cytometry relative to strain 427 gave values of 0.91 for 1738 RFP and 1738 GFP, and 0.90 - 0.97 for the clones analysed from Fly 1 and 6 (Table 3). Four clones from Fly 7 also fell within or just outside this range, but clones 1 and 3 had substantially higher values of 1.17 and 1.21 respectively (Table 3). While there was evidence from microsatellite and karyotype data that Fly 7 clone 1 was recombinant, clone 3 showed no difference to the profile of 1738 except lack of fluorescence (Table 3).

Taken together, the microsatellite and karyotype data identify five clones (1, 2, 4, 5, 6) from Fly 7 as products of intraclonal mating of strain 1738. The increased DNA content of Fly 7 clone 3 also indicates that it is a recombinant. Clones 4 and 5 were identical on all criteria (Table 3), hence there are five individual genotypes from Fly 7 that result from intraclonal mating. This contrasts with the absolute fidelity of transmission of 1738 RFP and 1738 GFP as demonstrated by the recovery of unaltered clones from Fly 1 and 6.

For the cross J10 RGFP, a total of 407 flies were co-infected with approximately equal numbers of BSF trypanosomes of J10 RFP and J10 GFP and dissected in batches 33-41 days later. The overall midgut infection rate was 64% (259/407) and the majority of midguts in a subset examined by fluorescence microscopy contained a mixture of red and green fluorescent trypanosomes (51/62, 82%), with 16% (10/62) having only red and 2% (1/62) only green fluorescent trypanosomes. There were 13 salivary gland infections, giving a transmission index of 5% (13/259). For 11 of these flies, trypanosomes of single colour fluorescence were detected in both salivary glands (Table 1) and these were not analysed further. Of the two remaining flies, Fly 1 had one salivary gland packed with yellow fluorescent trypanosomes; the colour of fluorescence was verified using narrow bandwidth red and green filters (Fig 4); the other gland contained only red trypanosomes. Although we assume that yellow fluorescent metacyclics were present in the salivary gland, disappointingly no yellow fluorescent trypanosomes were observed in the procyclic population derived from the glands after mouse passage, which comprised red fluorescent trypanosomes only, and no double drug resistant trypanosomes were recovered by antibiotic selection (Table 2); BSF were not checked for fluorescence, as the fluorescent reporter genes are driven by a procyclin promotor, which is down-regulated in BSF. Fly 2 had a mixed infection of red and green fluorescent trypanosomes in both glands, but no yellow fluorescent trypanosomes were noticed, nor were double drug resistant trypanosomes recovered after antibiotic selection (Table 2).

Clones were obtained from Fly 1 and 2 procyclic populations prior to drug selection; all eight clones from Fly 1 showed red fluorescence, while those from Fly 2 showed either red or green fluorescence (Table 4). Analysis of molecular karyotypes did not reveal any differences from J10 (not shown). However, several clones showed homozygosity for the microsatellite locus on chromosome IX whereas J10 was heterozygous (Fig 2C; Table 4). A difference between J10 GFP and J10 RFP was detected for the microsatellite locus on chromosome VIII, while all clones had ab alleles like J10 GFP, regardless of colour of fluorescence (Table 4); we assume J10 RFP became homozygous during the selection and fly transmission involved in derivation of the lineage.

Measurement of DNA contents by flow cytometry relative to strain 427 gave values of 0.96 for J10 RFP and J10 GFP, but all clones from Fly 1 had greatly increased DNA contents, ranging from 1.48 to 1.76 (Table 4). Based on the haploid DNA content of J10 (0.48), these values indicate triploid cells. Fly 2 clone 1 also had a high DNA content, while the other clones measured had similar values to J10 (0.93 - 1.05; Table 4). The raised DNA contents of some clones from Fly 1 and 2, coupled with reduction to homozygosity at one microsatellite locus, suggest that these clones are products of intraclonal mating. This evidence is strengthened by the striking observation of widespread yellow fluorescence in one salivary gland of Fly 1 (Fig 4). This deserves comment, as such predominant yellow fluorescence has never been observed even in salivary glands from red/green crosses. It is possible that trypanosome fusion took place on a large scale in this gland for some reason, but this implies that large numbers of red and green fluorescent trypanosomes were already present; on the other hand, if the gland was colonized by a small population of yellow fluorescent trypanosomes, this implies that there was an actively dividing population, which does not square with the subsequent failure to recover any viable yellow trypanosomes.

SG3 clone 7 is a hybrid clone derived from the 1738 GFP × J10 RFP cross; it expresses both mRFP and GFP genes and therefore appears yellow by fluorescence microscopy 8. Fly transmission of this clone was carried out twice in succession, first using SG3 clone 7 procyclics in the infective bloodmeal. Three flies with infected salivary glands were obtained, but the trypanosomes were not of uniform yellow fluorescence as expected, instead showing a range of fluorescent colours including green, red and yellow (Fig 5); most surprisingly, one fly had one salivary gland of the pair containing only red fluorescent trypanosomes. The infected glands from these three flies were used to obtain BSF populations, which were transformed to procyclics in vitro; no green fluorescent trypanosomes were seen in this population and the red and yellow fluorescent clones obtained were analysed by microsatellite and karyotype analysis (Table 5, Transmission 1). No differences were detected in the genotypes of any of the clones except for the absence of the GFP gene in the red fluorescent clones (Fig 6B). The GFP gene was on the smaller homologue of chromosome VII in the yellow fluorescent clones; hybridization with a probe for 18S rDNA showed both chromosome VII homologues present and unchanged in size in all clones (not shown). The most likely explanation of these observations is that the GFP gene was removed by gene conversion, because it sits in a tandem array of rDNA repeats. However, measurement of DNA contents of two red and two yellow clones showed a range of values: while yellow clones, 7A1 and 2, both gave values of about 0.9 and are probably diploid, the two red clones had either a very high (1.58, 7D1) or low (0.78, 7D4) value (Table 5). Previously SG3 clone 7 was found to have a DNA content similar to that of its parents, 1738 GFP and J10 RFP, consistent with diploidy 8.

The two yellow clones (7A1 and 7A2) were re-transmitted through flies. Of 39 flies with infected salivary glands, all except one had salivary glands with trypanosomes of red, green or yellow fluorescence as on the first transmission of the parent clone, SG3 clone 7; the exception was one fly that had only red trypanosomes in both glands. Salivary glands from six flies, representing transmission of clone 7A1 (7A1-1 to 4) and 7A2 (7A2-1 and 2) were used to obtain bloodstream form populations, which were transformed to procyclics in vitro. As before, no green fluorescent trypanosomes were observed in the population and only yellow or red clones were obtained for microsatellite and karyotype analysis (Table 5, Transmission 2). These clones were again indistinguishable by molecular karyotype (not shown) and most were identical by microsatellite analysis, but 7A1-4 clone 1 had become homozygous at two previously heterozygous loci and it also showed a high DNA content (1.42, Table 5). A second clone from this fly (7A1-4 clone 2) also showed homozygosity at a single microsatellite locus, but was lost prior to measurement of DNA content.

One interpretation of these results is that, like its parents 1738 and J10, SG3 clone 7 also undergoes intraclonal mating, leading to the occurrence of pure red and green fluorescent trypanosomes as well as yellow ones in the salivary glands; it is an open question whether the red and green trypanosomes are haploid intermediates or diploid progeny. As most salivary glands had this appearance, SG3 clone 7 may have a higher frequency of intraclonal mating than either parental strain, but this may also reflect the fact that these events are easy to detect in a yellow clone that carries genes for both red and green fluorescence. If so, it is puzzling that these high levels of intraclonal mating did not feed through to recovery of large numbers of recombinant progeny. Like the predominance of yellow fluorescent trypanosomes in one salivary gland for J10, the observation of salivary glands filled with red fluorescent trypanosomes in sequential transmissions of SG3 clone 7 is difficult to explain. Do these uniform populations arise by founder effects 21, or are they evidence of widespread recombination? In considering these alternatives, it is noteworthy that red and green fluorescent trypanosomes were also observed in small numbers in the midguts of infected flies.