Kinetoplast DNA minicircles are inherited from both parents in genetic hybrids of Trypanosoma brucei

Molecular and Biochemical Parasitology, 42 (1990) 45-54

45

Elsevier MOLBIO 01370

Kinetoplast D N A minicircles are inherited from both

parents in genetic hybrids of Trypanosoma brucei W e n d y Gibson and Lisa Garside Department of Pathology, University of Bristol Veterinary School, Langford, Bristol, U.K (Received 8 January 1990; accepted 28 March 1990)

We have examined the inheritance of kinetoplast DNA (kDNA) m genetic crosses of trypanosomes. In 2 independent crosses of

Trypanosoma brucei spp. trypanosomes, the kDNA maxicircles which carry the genes for mitochondrial biogenesis, were inherited from one parent only, as already found by other workers. However, the other component of kDNA, the minicircles, were inherited from both parents. This was demonstrated by Southern analysis using cloned minicircle probes. The inheritance of kDNA is therefore not uniparental. Our data point to fusion of the parental kinetoplast DNA networks during genetic exchange, with gradual loss of one or other parental maxicircle type due to random segregation of maxicircles at subsequent mitotic divisions. We infer that the first event of genetic exchange is fusion of parental trypanosomes (either haploid or diploid), followed at some point by fusion of the parental mitochondria. Key words: Kinetoplast DNA; Mitochondrial DNA; Trypanosome; Trypanosoma brucet; Genetic exchange

Introduction There is now abundant evidence that trypanosomes of the Trypanosoma brucei group can undergo genetic exchange during their cycle of development in the insect vector [1--4], but where in the tsetse fly genetic exchange takes place and by what mechanism are questions still to be answered. No haploid stage in the trypanosome lifecycle has yet been substantiated [5,6], although the demonstration that alleles segregate and recombine points to a meiotic division at some stage [3,4]. With regard to the kinetoplast DNA (kDNA), it has been shown by analysis of polymorphic restriction sites that the kDNA maxicircles of the hybrid progeny can be inherited from either parent, with the implied assumption that this applies to the whole of the kinetoplast [ 2 4 ] . Correspondence address: Wendy Gibson, Department of Pathology, University of Bristol Veterinary School, Langford, Bristol, BSI8 7DU, U.K.

Abbreviations: kDNA, kinetoplast DNA; SSC, standard saline

However, the trypanosome kinetoplast contains 2 types of DNA circles, maxi- and minicircles (for reviews, see refs. 7-9). Maxicircles are equivalent to the mitochondrial DNA of other eukaryotes and code for ribosomal RNAs and proteins necessary for mitochondrial biogenesis. In T. brucei spp. there are 50-100 maxicircles per kinetoplast, which are homogeneous in sequence. By contrast, the minicircles have no known function, although they have been shown to code for a small RNA [10], and are heterogeneous in sequence. There are 10000 or so minicircles per kinetoplast, which fall into an estimated 300 sequence classes according to analysis of renaturation kinetics. Each trypanosome stock has a unique set of minicircles, as judged from restriction enzyme digestion [11]. How this high level of heterogeneity is maintained is unknown. One suggestion is that genetic exchange, involving mitochondrial fusion, is responsible [12]. To test this hypothesis, we have examined both the maxi- and minicircles of hybrid trypanosomes from 2 genetic crosses.

citrate; SDS, sodium dodecyl sulphate; VSG, variant surface glycoprotein; RFLP, restriction fragment length polymorphism. 0166-6851/90/$03.50 © Elsevier Science Publishers B.V. (Biomedical Division)

46 Materials and Methods

Trvpanosomes. Hybrid trypanosome clones were derived from 2 separate genetic crosses: (1) T. b. rhodesiense 058 (MHOM/ZM/74/58 (CLONE B) and T. b. brucei 196 (MSUS/CI/78/TSW 196 (CLONE A)); clones from this cross are designated G1-G9 and W1-W5 [4]. (2) T. b. brucei 196 and T. b. brucei J l 0 (MCRO/ZM/73/J10 (CLONE A)); clones from this cross are designated Y1, Y2, X1-4 and Z1-8 (Gibson and Garside, unpublished results). Preparation of DNA. Kinetoplast DNA was extracted using a modification of the shearing method of Simpson and Berliner [13,14]. Several preparations of both bloodstream and procyclic culture form trypanosomes were made for 058 and 196, while J l 0 and all other clones were examined as bloodstream form preparations. Total DNA was prepared by standard methods [15].

Electrophoresis, blotting and hybridisation. DNAs were restricted and size-fractionated using 0.7% agarose gels to separate maxicircle fragments and 1.5-1.8% gels to separate minicircle fragments. unless stated otherwise in figure legends. Gels were blotted in the standard way [16]. Nitrocellulose filters were hybridised with oligo-labelled DNA probes [ 17] as previously described [18] and washed to a stringency of 0.1 × SSC, 0.1% SDS or 0.01 x SSC, 0.1% SDS at 65°C as indicated in the figure legends (1 x SSC = 0.15 M NaCI, 0.015 M Na-citrate, pH 7).

Cloning. Kinetoplast DNA of 058 was restricted with AluI and that of 196 with PstI/HincII. Fragments were ligated into appropriately cut plasmid vectors and cloned into Escherichia coli HB101. Colonies were screened with labelled kDNA from 058 or 196; all positive colonies hybridised with both probes. Plasmids were isolated and checked for inserts of 1 kb or less. Inserts were labelled and hybridised with blots of minicircle DNA as above. Results

Comparison of kDNA maxi- and minicircles by

gel electrophoresis. Hybrid trypanosome clones were derived from 2 separate genetic crosses (for details, see Materials and Methods). Clones G2-G4, W3 and W5 from the 058/196 cross are all hybrid progeny clones, as assessed by analysis of molecular karyotype and RFLPs in housekeeping and VSG genes; clone G1 is also hybrid, but appears to have exchanged only kDNA, having nuclear DNA polymorphisms identical to parent 058. but kDNA maxicircles of parent 196 [4]. For the J10/196 cross, clones X2, X4 and Z6 are hybrid as judged by molecular karyotype and RFLPs in housekeeping and VSG genes; clones Y I and Y2 were identical to parent J10 in every character examined (Gibson and Garside, unpublished results). The kDNA maxicircles of 058 are larger than those of 196 and have an extra EcoRI site [4]. Similarly, the maxicircles of 196 and J10 can easily be distinguished by differences in Cfol sites [14]. These maxicircle polymorphisms are shown in Fig. 1. Each hybrid clone had maxicircles of only one parental type, as demonstrated by gel electrophoresis and also hybridisation for the 058/196 cross [4]. The minicircles were examined by gel electrophoresis of Hinfl or TaqI digests of purified kDNA (Fig. 2). Both these enzymes cut most minicircles at least once [11], yielding a complex pattern of fragments which can be viewed as a fingerprint of the minicircle network. These fingerprints appear to be relatively stable. Fig 2B shows that TaqI digests of 3 different preparations of parent 196 kDNA are identical. Similarly, TaqI digests of kDNA from parent 058 trypanosomes were identical whether the preparations were of bloodstream form trypanosomes before or after fly transmission, or procyclic culture form. However, the hybrid clones from this cross and also the J10/196 cross have different minicircle digests from both their parents and each other, as can be seen by studying Figs. 2A and B. For example, note that the Hinfl digests of X2 and Z6 minicircles are clearly different from those of their parents, J10 and 196, in the flanking lanes (Fig. 2A). Neither hybrid clone has the minicircle digest expected for its maxicircle type. The maxicircle type of each clone is obvious from the pattern of bands larger than linearised minicircles, which

47

progeny

progeny 058

196

G1

G2

G3

G4

J10

W5 B

A

Y1

Y2

Q!~ ~

X2

~~

X4

Z6

196

~

~

e l

a-

'1

br

I

C-

d0-

mc

mq=

196

I

b

I

058

I

b

I

I

a

d

I •

I

c

I

196

III c

I

c

I

JlO

Idlel

I

a

II

b

I

a

II

b

I

I

2 kb Fig. 1. Maxicircle polymorphisms. Panel A shows an ethidium-stained gel with EcoRI digests of purified kDNA from parents 058 and 196 and 5 hybrid progeny clones. 058-type maxicircles: G3, G4 and W5; 196-type maxicircles: G1 and G2. Panel B shows a similar gel with Cfol digests of parents Jl0 and 196 with 5 progeny clones, of which 3 are hybrid (X2, X4 and Z6) and 2 are identical to J10 (YI, Y2). Some of the digests show streaking due to undissolved material and the Y2 lane is very weak. J10-type maxicircles: YI, Y2, X2; 196-type maxicircles: X4, Z6. Sketch maps below the photographs indicate the positions of EcoRI (A) and Cfol (B) sites on the linearised maxicircles of the parental trypanosomes, together w~th positions of fragments a--e. mc, linearised minicircles, size 1 kb. are maxicircle fragments - there are clearly HinfI p o l y m o r p h i s m s which distinguish the maxicircles o f J10 and 196 (Fig. 2A). Similarly, in Fig. 2B, the TaqI minicircle digests o f hybrid clones G3 and W 5 differ f r o m the flanking parental digests o f 058 and 196 k D N A . The possible contribution o f maxicircle fragments smaller than 1 kb to fingerprint differences can be disregarded, since the hybrid clones do not

share the parental minicircle network appropriate to their maxicircle type. There are 2 possible explanations for the above: either the hybrid clones have inherited minicircles f r o m both parents thus producing a new pattern o f minicircle fragments, or minicircles from a single parent have been reassorted during genetic exchange changing the relative numbers o f minicircles belonging to each sequence class. To

L~

C/96L

L/96L

96L

SM

~0 Q

E]

-OLU

8~0 •

V

"OUJ

]l-r'

9Z

CX 96L

961- g M

~9

1.9 SM

t~9 £ 0

8~0

49

distinguish these possibilities, we hybridised minicircle blots with cloned minicircle fragments.

Comparison of minicircles by hybridisation. Minicircle fragments from 058 or 196 were cloned and colonies screened with labelled kDNA from 058 or 196; all positive colonies hybridised with both probes. Six plasmids with inserts of 1 kb or less were selected at random: three derived from 196 kDNA (pTK196-9, insert 300 bp; pTK19611, insert 500 bp; pTK196-14, insert 500 bp), and three from 058 kDNA (pTK058-6, pTK058-66, pTK058-82, all inserts 1 kb). Inserts from all 6 probes were hybridised with blots of minicircles digested with HinfI or HinclI. Fig. 3 shows three of the resultant autoradiographs. Each of the probes hybridises with many minicircle fragments in the 3 parental trypanosomes and hybrid clones, highlighting the differences observed above by simple gel electrophoresis. Probe 14 from plasmid pTK196-14 gives the overall impression of identical hybridisation in all lanes except 058 and J10 (panel A; G4 lane appears weak because there is less DNA; see Fig. 2A), i.e., all hybrid clones contain 196 minicircles, whether they have 196 maxicircles or not. Similarly for probe 9 from plasmid pTK1969, hybrid clones possess certain fragments which are seen in the 196 lanes, but not the 058 or J10 lanes (Fig. 3B, arrows). On the other hand, certain fragments seen in the 058 lanes with probes 9 and 66 are also visible in hybrid clones G1-4, W3 and W5 (Fig. 3B and C, arrowheads), indicating the presence of 058 as well as 196 minicircles in these hybrid clones. Similarly, certain fragments seen in the corresponding J10 lanes are present in clones X2 and Z6. Moreover, the intensity of hybridisation with probe 66 in lanes J 1 0 , X2 and Z6 is markedly greater than that in the 196 lanes (Fig. 3 panel C),

again indicating the presence of J10 minicircles in these hybrid clones. These results were confirmed using other blots of HinfI, HinclI or TaqI digests of kDNA or total DNA from the same set of trypanosome clones hybridised with various minicircle DNA probes. Thus we conclude that the hybrid clones from 2 different crosses have hybrid kDNA minicircle networks, although each has inherited the maxicircles of one parent only.

Fusion of whole minicircle networks? If parental kDNA networks simply fused during genetic exchange, mixing up all the minicircles, the outcome would be progeny with similar networks containing roughly 50:50 minicircles from each parent. Fig. 4 shows that this is not the case. Hinfl and TaqI digests of minicircles from hybrid clones G2 and G3 give a different pattern of hybridisation to simple mixtures of minicircles from 058 and 196, the 2 parental trypanosomes. In addition, the TaqI digests of G2 and G3 minicircles have a prominent extra fragment not seen in the parental digests (Fig. 4, arrowed). Comparison of the minicircle networks of clones G2 and G3 shows them to be indistinguishable; this is obvious from Fig. 4, although somewhat disguised by intensity differences in Figs. 2 and 3. Yet G2 has 196-type maxicircles and G3 has 058type maxicircles (Fig. 1; also checked on samples used for Fig. 4). These hybrid clones were previously found to have an identical molecular karyotype and to share an extensive range of nuclear DNA polymorphisms [4]. Since the chance of finding unrelated trypanosomes with an identical recombinant nucleus at numerous loci is small, it was assumed that clones G2 and G3 most probably arose from a single predecessor with a hybrid nucleus and both types of maxicircle. Now it seems likely that this predecessor also had a hy-

Fig. 2. Minicircle polymorphisms. Photographs show ethidium-stained race gels of (A) Hinfl and (B) Taql digests of purified kDNA from the clones indicated. Gels were 2-2.5% agarose run at 4-5 V c m - i. Maxicircle type is indicated by symbols as follows: circle, 058, triangle, 196, star, J10. mc, linearised minicircles, size 1 kb. Bands greater than l kb in size derive from maxicircles, and the most right-hand 4 lanes of panel A reveal a Hinfl polymorphism in the J l 0 maxicircle, All kDNAs were prepared from bloodstream form trypanosomes, except for lane 8 in panel A (196 procyclics) and lanes 1, 2 and 8-11 in panel B (058 procyclics, 2 different preparations of 196 procyclics respectively). In panel B, the second of each pair of digests had twice the enzyme concentration of the first. Arrows indicate prominent fragments which only occur in certain lanes. Lane 3 of panel A is underloaded.

50

058

A

A :~ G3 G4 W5 G1 G2 W3 196 196 X2 Z6 J10

Q

!

B

mo- p

•

•

•

•

•

•

•

•

058 G3 G4 W5 G1 G2 W3 196 196 X2 Z6 J10

mc-

¢:

Probe

Probe

14

C

• 058

• G3

• G4

Probe

66

• W5

•

•

•

A

•

9

*

•

,

G1 G2 W3 196 196 X2 Z6 J10

mc-

Fig. 3. Hybridisation of Hinfl blot m Fig. 2A with 3 cloned minicircle fragments as indicated. Pv,st-hybridisational washes were to 0.1 × SSC at 65°C. Each probe was removed by washing with 0.1 M NaOH, before the next hybridisation. Nomenclature as Fig. 2A, except arrows indicate 196 minicircle fragments present in hybrid clones, but not 058 or Jl0, and arrowheads indicate 058 minicircle fragments present m hybrid clones, but not 196. G4 lane shows only weak hybridisation because there is less DNA in the lane (see Fig. 2A). Bands greater in size than linearised minicircles (mc = l kb) are probably mlnicircle ohgomers.

51 058 75% G2 G3 50% G2 G3 25% 196

058 75% G2 G3 50% G2 G3 25% 196

058 75% G2 G3 50% G2 G3 25% 196

~

o

e

mc-

4-"

Probe 9

Hinfl

Probe 9

Taql

4,-

Probe 66

Taql

Fig. 4. Comparison of mixtures of parental kDNAs with kDNA of hybrid clones. Purified kDNAs of 058 and 196 were digested separately with Hinfl or TaqI, and then mixed in the proportions shown; 75% refers to a 3:1 ratio of 058 to 196 kDNA, and so on. Other lanes are pure digests of the kDNAs indicated. Electrophoresis was through 1.5% agarose at 4 V cm -1 . Blots were hybridised with the probes indicated; post-hybridisational washes were to 0.01 x SSC at 65°C. mc, linearised minicircles, 1 kb; bands above this size are probably minicircle oligomers. Arrows indicate non-parental TaqI fragment in hybrid clones G2 and G3.

brid minicircle network, since the minicircle networks of the derived clones G2 and G3 are indistinguishable. Presumably, subsequent cell division rapidly resolved the mixture of maxicircles, producing otherwise identical hybrid clones. Discussion

Our results show clearly that hybrid trypanosomes inherit kDNA minicircles from both parents, but maxicircles from one parent only. How does this come about? The most likely possibility is that kDNA networks from the 2 parents simply fuse during genetic exchange, mixing up the minicircles and also the maxicircles. Subsequent mitotic divisions could yield networks of approximately the same minicircle composition, while at the same time resolving the maxicircles into a homogeneous set. As discussed by Birky [19] in relation to the inheritance of organelle DNA, the non-stringent replication and

partitioning of a relatively small number (100) of DNA molecules of 2 types leads to the eventual loss of one type, i.e., uniparental inheritance. For example, in baker's yeast, Saccharomyces cerevisiae, the mitochondrial DNA of heteroplasmic progeny becomes homogeneous in about 20 generations [19]. In the case of the trypanosome hybrids, at least 40 generations elapsed after genetic exchange before the kDNA of progeny clones was examined, giving ample time for the estimated 50-100 maxicircles to segregate. Similar stochastic mechanisms would not have the effect of homogenisation on the minicircles because of their large number and variety. An alternative possibility to explain the uniparental inheritance of maxicircles but biparental inheritance of minicircles, is that parental kDNA networks remain separate, but pick up minicircles released from the network of the other parent. If this were the case, then hybrid clones G2 and G3, which are identical except for maxicircle type and

52 probably derive from the same predecessor, would be expected to have very different minicircle networks. In fact, the minicircle networks of G2 and G3 are indistinguishable, and the most likely option remains kinetoplast fusion. Whatever the actual mechanism for generating hybrid minicircle networks, the first prerequisite is fusion of mitochondria preceded by trypanosome fusion. The possibility that minicircles recombine during genetic exchange has not been addressed by these experiments. If this occurred, new types of minicircle would be generated, which is one explanation for the non-parental TaqI fragment seen in hybrid clones G2 and G3 in Fig. 4B and C (arrowed). However, it is equally possible that the visibility of this fragment in the hybrid clones results from a simple increase in the relative amount of this particular minicircle in the hybrid relative to parental kDNA. Previous exhaustive analysis of the nuclear DNA of clone G1 showed it to be identical to parent 058 in every character [4]. However, it has inherited maxicircles from parent 196 and the minicircle network is also hybrid. Somehow this aberrant clone recombined only its kDNA. Three facts can now be gleaned concerning the genesis of Gl: the original 058 parent fused with a 196 trypanosome; either the 196 parent had no nucleus or this was lost after trypanosome fusion; the 058 nucleus underwent only mitotic, not meiotic divisions. In a previous study of kDNA restriction site polymorphisms in T. brucei spp. [14], T. brucei J10 along with 3 other stocks was found to have maxicircles of a distinctive type. This seemed to confirm the separateness of these stocks from mainstream T. brucei, as they had already been designated 'the kiboko group' on the basis of characteristic isoenzymes [22]. All isolates of this group originated from 'big game' areas of East Africa and thus appeared to be associated with a wild animal-tsetse fly transmission cycle. However, since J10 mated readily with TSW 196, a West African T. b. brucei, there can be no doubt that the kiboko group is part of the main interbreeding continuum of T. brucei spp. As suggested previously [12], the factor which maintains the extreme heterogeneity of minicircles in T. brucei spp. is likely to be genetic exchange.

The degree of minicircle heterogeneity may serve as an indicator of the frequency of genetic exchange in kinetoplastid species. At one end of the scale is T. brucei with a proven mating system and extremely heterogeneous minicircles. At the other is T. evansi, a close relative of T. brucei, which is barred from genetic exchange by its inability to undergo cyclical transmission in the tsetse fly; T. evansi has homogeneous minicircles [10]. Midway on the scale is T. cruzi, which has intermediate minicircle heterogeneity and is considered to have given up genetic exchange [9,20]. On this basis, T. b. gambiense, T. congolense and T. mega with their extremely heterogeneous minicircles [9,21], would be expected to undergo genetic exchange.

Acknowledgements This work was supported by the Medical Research Council. We are indebted to Dr. Michael Miles for encouraging us to take another look at minicircles. We thank Dr. Paul Englund for helpful discussion and Prof. Pier Borst for constructive criticism of the manuscript.

References 1 Jenm, L., Marti, S., Schweizer, J., Betschart, B., Lepage, R. W.F., Wells, J.M., Tait, A.. Pamdavoine, P., Pays, E. and Stelnert, M. (1986) Hybrid formation between African trypanosomes during cychcal transmission. Nature 322, 173-175. 2 Sternberg, J., Tait, A., Haley, S., Wells, J.M., Lepage, R.W.F., Schweizer,J. and Jenni, L. (1988) Gene exchange in African trypanosomes: characterisation of a new hybrid genotype. Mol. Biochem. Parasitol. 27, 191-200. 3 Sternberg, J., Turner, C.M.R., Wells, J.M., RanfordCartwrlght, L.C., Lepage, R.W.F. and Tait, A. (1989) Gene exchange in African trypanosomes: frequency and allellc segregation. Mol. Biochem. Parasitol. 34, 269-280. 4 Gibson, W.C. (1989) Analysis of a genetic cross between Trypanosoma brucei rhodesiense and T. b. brucei. Parasitology 99, 391-402. 5 Shapiro, S.Z., Naessens, J., Liesegang, B., Moloo, S.K. and Magondu, J. (1984) Analysis by flow cytometry of DNA synthesis during the life cycle of African trypanosomes. Acta Trop 41, 313-323. 6 Kooy, R.F., Hirumi, H., Moloo, S.K., Nantulya, V.M., Dukes.. P., Van der Linden, P.M., Duijndam, W.A.L., Janse, C.J. and Overdulve,J.P. (1989) Evidencefor diploidy in metacyclic forms of African trypanosomes. Proc. Natl. Acad. Sci USA 86, 5469-5472. 7 Borst, P. and Hoeijmakers,J.H.J, (1979) Kinetoplast DNA. Plasmid 2, 20-40.

53 8 Stuart, K. (1983) Kinetoplast DNA, mitochondrial DNA with a difference. Mol. Biochem. Parasitol, 9, 93-104. 9 Simpson, L. (1986) Kinetoplast DNA in trypanosomid flagellates. Int. Rev. Cytol. 99, 119-179. l0 Rohrer, S.P., Michelotti, E,F., Torri, A1.F. and Hadjuk, S.L. (1987) Transcription of kinetoplast DNA minicircles. Cell 49, 625-632. 11 Borst, P., Fase-Fowler, F. and Gibson, W.C. (1981) Quantitation of genetic differences between Trypanosoma brucei gambiense, rhodestense, and brucei by restriction enzyme analysis of kinetoplast DNA. Mol. Biochem. Parasitol. 3, 117-131. 12 Borst, P., Fase-Fowler, F. and Gibson, W.C. (1987) Kinetoplast DNA of Trypanosoma evansi. Mol. Biochem. Parasttol. 23, 31-38. 13 Simpson, L. and Berliner, J. (1974) Isolation of kDNA from Leishmania tarentolae in the form of a network. J. Protozool. 21, 382-393. 14 Gibson, W., Borst, P. and Fase-Fowler, F. (1985) Further analysis of intra-specific variation in Trypanosoma brucei using restriction site polymorphisms in the maxi-circle of kinetoplast DNA. Mol. Biochem. Parasitol. 15, 21-36. 15 Van der Ploeg, L.H.T., Bernards, A., Rijsewijk, F. and Borst, P. (1982) Characterisation of the DNA duplicationtransposition that controls the expression of two genes for variant surface glycoproteins in Trypanosoma brucel.

Nucleic Acids Res. 10, 593-609. 16 Southern, E.M. (1975) Detection of specific sequences among DNA fragments separated by gel electrophoresis. J. Mol. Biol. 98, 503-517. 17 Feinberg, A.P. and Vogelstein, B. (1983) A technique for radiolabelling DNA restriction endonuclease fragments to high specific activity. Anal. Biochem. 132, 6-13. 18 Gibson, W.C., Dukes, P. and Gashumba, J.K. (1988) Species specific DNA probes for the identification of trypanosomes in tsetse. Parasitology 97, 63-73. 19 Birky, C.W. (1983) Relaxed cellular controls and organelle heredity. Science 222, 468-475. 20 Tibayrenc, M., Ward, P., Moya, A. and Ayala, F.J. (1986) Natural populations of Trypanosoma cruzi, the agent of Chagas disease, have a complex multiclonal structure. Proc. Natl. Acad. Sci. USA 83, 115-119. 21 Borst, P., Fase-Fowler, F., Weijers, P.J. Barry, J.D., Tetley, L. and Vickerman, K. (1985) Kinetoplast DNA from Trypanosoma vivax and Trypanosoma congolense. Mol. Biochem. Parasitol. 15, 129-142. 22 Gibson, W.C., Marshall, T.F. de C. and Godfrey, D.G. (1980) Numerical analysis of enzyme polymorphism: a new approach to the epidemiology and taxonomy of trypanosomes of the subgenus Trypanozoon. Adv. Parasitol. 18, 175-246.