T parasites which is transmitted by sandflies (Diptera; Psychodidae;

Transcription

1 J. Euk. Microbiot.. 44(5) pp by the Society of Protozoologists The Leishmania hertigi (Kinetoplastida; Trypanosomatidae) Complex and the : Their Classification and Evidence for a Neotropical - Origin of the Leishmania-Endotrypanum Clade HARRY A. NOYES,*,' BYRON A. ARANA,** MICHAEL L. CHANCE,* and RHAIZA WINGON*** *Liverpool School of Tropical Medicine, Pembroke Place, Liverpool, L3 5QA, UK, and **Centro de Investigaciones en Enfermedades Tropicales, Universidad del Valle, Guatemala. and ***Centre for Applied Entomology and Parasitology, Department of Biological Sciences, University of Keele, Staffordshire, ST5 53G, UK ABSTRACT. The relationships of the Leishmania hertigi complex and the lizard Leishmania species to the main groups of mammalian Leishmania and Endotrypanum parasites were examined. Restriction fragment length polymorphisms and sequences of small subunit ribosomal RNA genes and hybridization studies of kinetoplast DNA indicated that the L. hertigi complex was more closely related to the genus Endotrypanum than to the genus Leishmania. The lizard Leishmania species were found to be at the crown of the Leishmania tree. The data provides strong evidence for a Neotropical origin of the EndotrypanudLeishmania clade since the parasites closest to the root of the tree are all found exclusively in the Neotropics. The evolution of the Leishmania/Endotrypanurn clade in relation to the evolution of the known hosts of these parasites is discussed. Supplementary key words. Biogeography, Bradypodidae, Coendu, Cricetidae, evolution, Hystricomorpha, sloth, Trypanosoma cruzi. HE genus Leishmania is an important group of protozoan T parasites which is transmitted by sandflies (Diptera; Psychodidae; Phlebotominae) and is endemic amongst lizards and six orders of mammals. The majority of mammalian Leishmania species cause chronic cutaneous lesions in humans and parasites of the L. donovani complex may cause a visceral disease. There are believed to be approximately 400,000 cases of leishmaniasis each year amongst a population at risk of about 400 million [5]. The mammalian Leishmania species are divided into two subgenera-the L. (Viannia), consisting of nine species which are restricted to the Neotropics, and the L. (Leishmania) consisting of eleven New World species and ten Old World species [49]. Two further groups of Leishmania species have an uncertain relationship to the two named subgenera, these are the Leishmania hertigi complex and the lizard Leishmania. The L. hertigi complex consists of two species-l. hertigi and L. deanei. These parasites are restricted to Neotropical porcupines (Rodentia; Hystricomorpha; Erethizontidae; Coendu) and an unknown sandfly vector. They are not infectious to humans and only transient infections have been achieved experimentally in hamsters or mice. The L. hertigi complex species have been provisionally included in the L. (Leishmania) as both groups develop in the midgut of sandflies (suprapylarian) in the laboratory. In contrast L. (Viannia) parasites develop in the hindgut and midgut (peripylarian) and the lizard Leishmania develop in the hindgut (hypopylarian) of their respective sandfly hosts [33]. However the L. hertigi complex parasites have a nuclear DNA buoyant density which is intermediate between the ranges of the other Leishmania species and the genus Endotrypanum (Table 1). Twelve species of Leishmania have been isolated from the blood of lizards [29]. These parasites have been placed in a separate genus-the Sauroleishmania [29], principally because they infect a separate class of vertebrates and a separate genus of sandflies-sergentomyia. In contrast mammalian Leishmania species are parasites of sandflies of the genera Phlebotomus and Lutzomyia. The haematozoic lizard Leishmania are restricted to the Old World even though some host species, such as Hemidactylus sp. are found in both hemispheres [20], and some New World sandflies transmit Trypanosoma sp. between lizards. The genus Sauroleishmania has not been universally adopted [23, To whom correspondence should be addressed. Telephone: ; Fax: ; harry@liv.ac.uk Abbreviations: GB, GenBank; DNA, Kinetoplast DNA; MYA, Million Years Ago; RFLP, Restriction Fragment Length Polymorphism; SSU rfwa, Small Sub Unit Ribosomal RiboNucleic Acid Recent phylogenies of the small subunit ribosomal RNA (SSU rrna) gene of the Trypanosomatidae could not resolve the lizard parasite L. tarentolae and the mammalian Leishmania species [l 1, 381. The Sauroleishmania is the only trypanosomatid genus for which sequence data is available that cannot be resolved by this method. Three species of Leishmania are found mainly or exclusively in the gut of the lizard [29]. These species were all first described before 1925 when Wenyon's definition of Leishmania was current [55]. More recent studies have shown that these parasites are probably trypanosomatids of Diptera that can survive in the lizard gut [20, 30, 531. Therefore the lizard gut trypanosomatids will be disregarded in the following discussion of the lizard Leishmania. Phylogenies of the SSU rrna gene indicate that digenetic parasitism has developed independently on up to four occasions in the Trypanosomatidae [22]. One of these groups of digenetic parasites is the EndotrypanumlLeishmania clade which appears to be a sister group of the CrithidialLeptomonas clade of monogenetic insect parasites. We have used a number of methods in order to clarify the position of the L. hertigi complex and the lizard Leishmania within the EndotrypanumlLeishmania clade. RFLP analysis of SSU rdna has been shown to be useful for the classification of Trypanosoma cruzi strains and anuran Trypanosoma species [14, 151, we have used this method to determine the relationship of representatives of both the L. hertigi complex and the lizard Leishmania with respect to Endotrypanum sp. and the mammalian Leishmania subgenera. A partial sequence of the SSU rrna gene of L. hertigi was used to support the RFLP evidence. Hybridization studies of the variable region of kinetoplast (mitochondrial) minicircle DNA of 14 Leishmania species were also used to confirm the position of the lizard Leishmania sp. MATERIALS AND METHODS Parasites from the Liverpool cryobank were revived in semisolid blood agar and maintained in HOMEN [8]. DNA was prepared by standard methods [28]. Parasite strains used are shown in Table 1. An unnamed trypanosomatid (G755) which was isolated in Guatemala was used as an outgroup. The sequence of the most variable region of the SSU rrna gene of G755 had previously been obtained (Genbank (GB) accession number U59491) (bases 938-1,422 in the L. donovani GB:X07773 sequence). This sequence indicated that this parasite clusters between the Crithidialleptomonas clade and the EndotrypanumlLeishmania clade (Fig. 3). Therefore, of the trypanosomatids for which se-

2 5 12 J. EUK. MICROBIOL., VOL. 44, NO. 5, SEPTEMBER-OCTOBER 1997 Table 1. Parasite strains used in the present study. All strains were from the Liverpool School of Tropical Medicine cryobank with the exception of G755 which was isolated in Guatemala by Dr Arana and MHOMNEIXXLbV which was isolated in Venezuela by Dr Bonfante-Ganido. Escuela de Medicina, Universidad Centro Occidental, Lara, Venezuela. Parasites are ranked primarily in descending order of nuclear DNA buoyant density and secondarily in descending order of kinetoplast DNA buoyant density. Nuclear (ndna) and kinetoplast DNA (kdna) buoyant densities are indicated after each strain [12, 181. Where the buoyant densities of a given strain have not been determined the value of a closely related strain has been included in brackets where available. DNA buoyant density is related to GC content by the equation GC composition = (X )/0.0098, where X = DNA buoyant density. The international code for Leishmania strains consists of four elements. The first element indicates the host class and genus in the case of vertebrate hosts and the species in the case of sandfly hosts, the second the country of isolation, the third the year of isolation and the fourth element is a local strain number [3]. In the first element M Mammal, R Reptile and I Insect; HOM Homo, NYC Nyctomys, COE Coendu, BRA Bradypus, CHO, Choloepus, HEM Hemidactylus, TAR Tarentola, GYM Gymnodactylus, in the second element IR Iran, ET Ethiopia, SD Sudan, IL Israel, BZ Belize, BR Brazil, PA Panama, DZ Algeria, SE Senegal, SU Soviet Union, KE Kenya, VE Venezuela, CR Costa Rica. The fourth element corresponds to the strain number given by the isolating laboratory. However a number of strains have been described under the Liverpool School of Tropical Medicine cryobank accession numbers which begin with LV. Where this is the case the original strain number given by the isolating laboratory is followed by the Liverpool accession number in the next column. DNA buoyant densities Liverpool Species Group affiliation International code strain no. ndna DNA L. (L.) tropica L. (L.) fropica L. (L.) donovani L. (L.) major L. (L.) major L. (L.) mexicana L. (L.) amazonensis L. hoogstraali L. (V.) panamensis L. tarentolae L. tarentolae L. gymnodacryli L. adleri L. (V.) braziliensis L. (V.) guyanensis L. deanei L. hertigi E. monterogeii E. schaudinni E. schaudinni Promastigote trypanosomatid Viannia subgenus Viannia subgenus Viannia subgenus L. hertigi complex L. hertigi complex Endotrypanum sp. Endotrypanum sp. Endotrypanum sp.? MHOM/IR/89/ARD MHOMmU60/LRC-L39 MHOM/ET/67/HU3 MHOM/SD/72LV305 MHOM/IL/67/LRC-L137 MNYC/BZ/62/M379 MHOM/BR/73/M1845 RHEM/SD/63/NG-26 MHOM/PA/7 1LS94 RTARIDZI39fTarVI RTAWSW67/G10 RGYM/SU/64/Ag RLAT/KE/54/1433 MHOMNE/XXLbV MHOM/BR/75/M4 147 MCOEISR/XX/M808 MCOE/PA/65/C-8 MCHO/CR/62/A9 MBRAA'AI67LV5 8 MCHOPM68LV59 G755 LV357 LV9 LV561 LV78 LV3 1 LV414 LV 108 LV247 LV30 LV402 LV42 LV88 (1.719 ( ( ( ( (1.71G ) 1.707) ) ) ) ) quence data is available, this parasite is the most closely related to the EndotrypanumlLeishmania clade. PCR primers were designed using the personal computer programme PRIMER (Lincoln, s. E., Daly, M. J., Lander, E. s., MIT Centre for Genome Research and Whitehead Institute for Biomedical Research, Nine Cambridge Centre, Cambridge, MA 02142, USA). Primers WSSUF (GCTTGTTI'CAAGGACT- TAGCC) and WSSUR (GAAATATCGGTGAACTTTCGG) were designed to anneal to the ends of the coding regions of published sequences of the SSU rrna gene of trypanosomatids and used with the following temperature profile: 30 cycles of 94" C, 30 s; 54" C, 1 min; and 72" C, 2 min. A 561-bp sequence of the SSU RNA gene of a number of kinetoplastids was also identified (corresponding to bases 901-1,461 of L. donovani (GB:X07773 sequence) which when analyzed using the DNAPARS program in PHYLIP version 3.5~ (J. Felsenstein, Dept. of Genetics SK-50, University of Washington, Seattle, WA 98195, USA.) generated a dendrogram of the Trypanosomatidae with a topology in general agreement with published phylogenies of the whole gene 138, 391. In an alignment of the complete SSU rrna gene of C. fasciculatu (GB:X03450), Leptornonas (GB:X53914), E. rnonterogeii (GB: X5391 l), L. donovani (GB:X07773) and L. amazonensis (GB: X53912) it was found that 58 out of 96 (60%) of the informative sites were in the region between bases 938-1,422 in the L. donovani GB:X07773 sequence. Primers SSU561F TGGGA- TAACAAAGGAGCA and SSU56 1R CTGAGACTGTAACCT- CAAAGC were designed to amplify the sequence between bas- es 901 and 1,461 using 30 cycles of 94" C, 30s; 54" C, 1 min; 72" C, 2 min. The L. hertigi (LV42) and G755 PCR products of these primers were purified on S400 spin columns (Pharmacia) and ethanol precipitated before cycle sequencing, using the same forward and reverse primers, on a model 377 automated sequencing system (Applied Biosystems). DNA was PCR amplified in 20 mm (NH,),SO,; 75 mm Tris- HCl ph 9.0; 0.01% Tween; 2.0 mh4 MgC1,; 200 pm dntps; 40 Ud-l Taq polymerase; 1 pm primers. PCR amplified SSU rdna was digested with AZuI, HueIII, HhuI, Hi@, HpaII, RsaI and Sau ~1 of crude PCR product was separated on 1% agarose gels against DNA size markers (Boehringer Mannheim molecular weight marker VI). Concentration of the PCR product was determined visually by comparison with size markers. 50 ng of crude PCR product was digested in a final volume of 20 p1 with one unit of enzyme (New England Biolabs). 5 pl of each digest was loaded on 6% 29:l wlw polyacrylamide: bisacrylamide gels in a Bio-Rad Mini-Protean I1 apparatus. Gels were silver stained [6]. The variable region of kinetoplast DNA minicircles was amplified using primers 132 (ACTGGGGGTTGGTGTAAAA- TAG) [46] and LiR (TCGCAGAACGCCCCT) for 30 cycles of 94" C, 30 s and 54" C, 90 s. PCR amplified kinetoplast DNA (kdna) variable region probes were prepared using primers 132 and LiR and hybridized to dot blots of kdna amplified with the same primers [9]. SSU rdna restriction patterns were analyzed using the PHY- LIP package. The MIX parsimony programme was applied di-

3 NOYES ET AL.-EVIDENCE FOR A NEOTROPICAL ORIGIN OF LEISHMANIA 513 Table 2. Sizes of restriction fragments found after digestion of the SSU rrna gene. The complete gene was digested with HaeIII, Hinfr. HpaII, RsaI and Sau96I of the complete SSU rdna gene amplified with WSSUF and WSSUR and bases 901-2,142 (L. donovani GB: X07773) amplified with SSU561F and WSSUR with HhaI and bases 901-1,461 amplified with SSU561F and SSU561R with AM. Hinfr generated six fragments of 230, 213, 154, 122/126 (not resolved), 63 and 57-bp in all strains, these are grouped and indicated by the number "6" in the table. G755 used as outgroup, LV88 E. monterogeii, LV58 E. schaudinni, LV402 L. deanei, LV30 L. adleri, LV31 L. hoogstraali, LV247 L. gymnodactyli, LV108 and LV414 L. tarentolae, LV9 L. donovani, M379 L. mexicana, M4147 L. guyanensis. G LV LV LV LV LV LV LV LV LV M M Enzyme MWt Sau96I 1,200 1, HaeIII Hinfr 1, HpaII RsaI 1, HhaI AluI L 394- = Fig. 1. Sau96I restriction digest of the complete 16s SSU rrna gene of the species indicated. DNA was PCR amplified with primers WSSUF and WSSUR. Strain numbers are as indicated in Table 2. The sizes of fragments are shown in Table 2. Fragments of approximately 1,000-bp and larger in lanes 7-14 are partial digests. The mammalian Leishmania (lanes 12-14) are clearly distinguished from the lizard Leishmania (lanes 7-1 1) and L. deanei (lane 6) and the Endotrypanum species (lanes 3-5). This digest indicated that L. herreri (MCOE/CW 74LV344) was an Endotrypanurn species, this was confirmed by additional methods and is the subject of a separate communication [44]. rectly to the RFLP data for the presence or absence of bands. The data was also analyzed by calculating Nei's distance (equation 21 in Nei and Li [42]) for each pair of species and inserting the values into a distance matrix from which dendrograms were prepared using the FITCH programme and the UPGMA option in the NEIGHBOR programme. Bootstrap values were calculated using the SEQBOOT programme followed by MIX. RESULTS HaeIII, Hinfl, HpaII, RsaI and Sau96I restriction digests of the complete SSU rrna gene amplified with WSSUF and WSSUR were prepared from representatives of Leishmania, and Endotrypanum species. An Ah1 digest of the partial SSU rrna gene amplified with primers SSU561F and SSU561R and a HhaI digest of the products of SSU561F and WSSUR were also used since an analysis of an alignment of the available sequences of the SSU rrna gene indicated that these enzymes would only discriminate in these regions. The restriction fragments produced by these enzymes together with their sizes are shown in Table 2. The enzyme that produced fingerprints most suitable for identifying mammalian Leishmania, lizard Leishmania and Endofrypanum species was Sau96I which generated unique restriction patterns for each of these groups (Fig. I). The mammalian Leishmania species had a fragment of 452-bp absent from all other Leishmania and Endotrypanum species, although this band was present in the outgroup strain G755 this strain could be distinguished by a fragment at 500-bp which was not present in Leishmania species. The five lizard Leishmania strains had identical fingerprints with this enzyme and produced a fragment of 656-bp that was not present in the mammalian Leishmania

4 514 J. EUK. MICROBIOL., VOL. 44, NO. 5, SEPTEMBER-OCTOBER Kirnura distance = E. rnonrerogeii E. schaudinni L. deanei L. (L..) donovani L. (L.) mexicam - - L. tarentolae L. (L.) major L. (V.) guyanensis L. (L.) amazonensis - L. (L.) donovani G755 C. fasciculata L. fl.1 donovani LV9 0 L. IL.1 infanfum tptl c. L. IL.) mexicam M379 L. fl.1 amazonensis CV78 3 L. fl.i major tv fl.) major LU56?:C22. L. il.1 iropica ARD 1. il.1 tropfca iv357 L. IV.) panamensis is94 0 L. IV. j guyanensis M4147 t) L. [V.) braziliensis P2 ts L. IV.) brazifiensis LbV :; L. adlerilv30 L. hoogstraaii LV31 0 L. gymnodactyii LV247 L. tarentolae LV108 L. tarentolae LV414 Fig. 4. Dot blot of 10 ng and 1 ng of kinetoplast minicircle variable region DNA. The DNA was PCR amplified with primers 132 and LiR from the species indicated. Leishmania strains which do not have a subgeneric designation are lizard parasites. The dot blot was first probed with the 132 PCR primer which was terminal transferase labelled with 32P (Amersham) and washed in 1 X SSC, 0.1% SDS at 42" C, to determine loading. The blot was then stripped and hybridized to a L. tarenrolae (LV414) variable region kdna probe prepared by PCR using primers 132 and LiR and washed in water containing 0.1% SDS at 42" C. No hybridisation was detectable to the L. (Viunnia) strains despite relatively high loading in the case of L. guyanensis. Weak hybridisation was detectable to the L. (Leishmania) strains despite slightly lower loading in some cases such as the L. tropica strains. Hybridization to the lizard Leishmania strains was proportional to loading except in the case of the homologous L. tarenrnlae LV414 strain which gave a more intense signal. species. The L. hertigi complex species L. deanei had a unique restriction pattern with this enzyme with a fragment at approximately 1,200-bp but none at 656 or 560-bp which were found in the lizard and mammalian Leishmania strains. The Endotrypanum strains had a fragment at 1009-bp not found in any other strains tested (Table 2, Fig. 1). A dendrogram was prepared directly from the presence absence data shown in Table 2 using the MIX parsimony programme in PHYLIP (Fig. 2). L. deanei from the L. hertigi complex was found on the Endotrypanum branch of the dendrogram in 90% of bootstrap replicates when the data was analyzed with the MIX parsimony programme. An alternative analysis using the distance matrix programme FITCH placed L. deanei on the Leishmania branch (not shown), although in all other respects the topology of the FITCH dendrogram was identical to that of the MIX dendrogram. In both cases L. deanei was adjacent to the node separating Endotrypanum and Leishmania branches. L. hertigi and L. deanei were compared using HaeIII, HpaII, HhaI and RsaI and no differences were found (not shown). The lizard Leishmania species (L. tarentolae, L. gymnodacryli, L. hoogstraali and L. adleri) were at the tip of the Leishmania branch irrespective of how the data was analyzed. In order to confirm the position of the L. hertigi complex the product of PCR primers SSU561F and SSU561R was sequenced. Reliable data was obtained for the positions homologous to positions 938-1,422 in the L. donovani GB:X07773 sequence and this was deposited at Genbank (GB:U59492). Dendrograms based on this sequence together with the homologous region of a number of Leishmania species and representatives of Endotrypanum, Crithidia, Leptomonas species and T. cruzi were prepared using the DNAPARS programme (not shown) and with the DNADIST programme followed by FITCH (Fig. 3). In both these dendrograms L. hertigi was on the Endotrypanum branch and this position was supported in 9 1 % of bootstrap replicates using the parsimony method and 95% using the distance method. The Kimura distances calculated using the DNADIST programme indicated that L. hertigi was closer to E. monterogeii (Kimura distance ) than to L. (V.) guyanensis (Kimura distance ). The lizard Leishmania species L. tarentolae was closer to all the mammalian Leishmania species used (Kimura distances ) than L. hertigi was to any other Lxishmania species (Kimura distances ). Since the lizard Leishmania species appeared to be more closely related to the L. (Leishmania) species than to the L.

5 NOYES ET AL.-EVIDENCE FOR A NEOTROPICAL ORIGIN OF LEISHMANIA 515 (Viannia) species (Fig. 2), a dot blot of the variable region of kinetoplast DNA minicircles was prepared to test for cross hybridisation between the lizard and mammalian Leishmania kdna (Fig. 4). The blot was first probed with the 132 oligonucleotide primer which had been used to amplify the kdna to test for variations in loading, some variation in loading was detected but as the L. (Viannia) species L. guyanensis appeared to have a higher loading than any of the L. (Leishmania) species this was not considered an obstacle to accurate interpretation of the data. A L. tarentolae probe hybridized strongly at high stringency (0.1% SDS, 42" C) to 10 ng of all the lizard Leishmania species tested except L. hoogstraali, weakly to 10 ng of all L. (Leishmania) species tested but there was no detectable hybridisation to L. (Viannia) species. A L. hertigi probe hybridized most strongly to L. hertigi and less strongly to L. deanei. There was a very weak signal with all other Leishmania and Endotrypanum species (not shown). DISCUSSION The relationship of the lizard Leishmania to the mammalian Leishmania. The SSU rdna RFLP analysis and the pattern of cross hybridisation of the kdna probe indicated that the lizard Leishmania are more closely related to the L. (Leishmania) than to the L. (Viannia). SSU rrna sequence based phylogenies have been unable to resolve L. tarentolae from the mammalian Leishmania species with confidence [38] and (Fig. 3). The restriction enzymes used had been selected by a search of an alignment of L. tarentolae, L. donovani and E. monterogeii sequences for restriction enzyme sites that would discriminate between these species. The resulting restriction fragment data therefore contained relatively few uninformative sites, unlike the sequence data in which extensive regions of complete homology are found. The higher resolution of the RFLPs is therefore presumably a consequence of using mainly informative sites with an effective length of approximately 128-bp and the elimination of the large number of uninformative sites in the sequence data. The cross hybridization between L. tarentolae and L. (Leishmania) species kdna in the dot blot but not with L. (Viannia) species supports the results of the RFLP analysis. A recent RNA polymerase I1 phylogeny indicated that the lizard Leishmania are more closely related to the L. (Leishrnania) than to the L. (Viannia) [17]. There is also some in vivo evidence of a close relationship between mammalian and lizard Leishmania species. Humans infected with lizard parasites have produced positive Montenegro skin test reactions up to two years later. Mammalian parasites have been found to be infective to lizards, and transient infections of mammals including humans have been reported with certain lizard Leishmania strains [l, 7, 371. Only the nuclear DNA buoyant densities (Table 1) indicate that the lizard Leishmania are more closely related to the L. (Viannia) than to the L. (Leishmania). The relationship of the L. hertigi complex to the Leishmania and Endotrypanum. The L. hertigi complex was found to be most closely related to the genus Endotrypanum when the RFLP data was analyzed with the parsimony programme MIX but most closely related to the genus Leishmania when the RFLP data was analyzed with the distance matrix programme FITCH. However the partial sequence of the 561-bp region of the SSU rrna gene indicated that it clustered with Endotrypanum species in 95% of bootstrap replicates. Other ultrastructural, gene phylogeny and DNA buoyant density studies also suggest that the L. hertigi complex is more closely related to the genus Endotrypanum than to the genus Leishmania [ 17, 18, 411. Part of the definition of the genus Leishmania is that these parasites develop as amastigotes in macrophages [54]. Although prevalence of L. hertigi in porcupines may be high (SO-70%), parasites are difficult to detect except by culture and consequently the cell type parasitized is not yet known [24, 32, 581. Until the cell type parasitized can be established it will not be clear to what extent the DNA based data may conflict with the biological characteristics of these parasites. Endotrypanum parasites develop as epimastigotes or trypomastigotes in sloth (Edentata: Bradypodidae) erythrocytes and have not been reported from any other family of mammals [48]. Consequently the L. hertigi complex would appear to have a closer affinity with Leishmania than with Endotrypanurn parasites on biological grounds, since they develop as amastigotes rather than trypomastigotes or epimastigotes. Further studies of the biology of the L. hertigi complex species and of other genes will be required to determine the affiliation of these parasites with confidence. However it is clear that the L. hertigi complex appeared soon after the Endotrypanum and Leishmania genera separated, and before the lizard Leishmania species arose, an observation that has important consequences for the understanding of the evolution of the whole group. Molecular evidence for a Neotropical origin of the LeishmanialEndotrypanum clade. In the phylogenies of the LeishmanialEndotrypanum clade (Fig. 2, 3) one of the two major branches consisted exclusively of New World parasites-l. hertigi and Endotrypanum sp., whilst the other branch consisted of Leishmania parasites from both the Old World and the New World. There is only one isoenzyme based phylogeny of the mammalian Leishmania species which includes both subgenera [51, 521. Two major branches were identified, one of which contained the New World L. (Viannia) subgenus and the other the New and Old World L. (Leishmania) subgenus. In the absence of an outgroup, the L. hertigi complex was assigned to the L. (Leishmania) subgenus from which it branched off first. The L. mexicana complex, which is also restricted to the New World, branched off after the L. hertigi complex and the Old World complexes were placed at the crown of the tree. This phylogeny in which Old World parasites are found only at the tip of one of the sub-branches is consistent with a Neotropical origin of the EndotrypanumlLeishmania clade. The overall GC content for a large number of mammalian and lizard Leishmania and Endotrypanum species can be calculated from their DNA buoyant densities [12, 13, 181. The GC content of the parasites rises through the Neotropical species and then through the Old World species (Table 1). species fall into two groups, L. adleri and L. hoogstraali and some L. tarentolae have values similar to members of the L. (Viannia) subgenus at the low end of the range whilst L. agamae and other L. tarentolae have values higher than any mammalian Leishmania species. The GC composition of the gene flanking regions of a number of trypanosomatids has been estimated [2] and follows the branching order of the trypanosomatids in the phylogeny of Fernandes et al. [22]. Consequently low GC content may be the primitive state and the genus Endotrypanum and New World species of Leishmania may be closer to the primitive state of this character than the Old World Leishmania species. Biogeographical evidence for a Neotropical origin of the LeishmanialEndotrypanum clade. Sloths are important reservoir hosts of both Endotrypanum sp. and L. braziliensis, which are represented on both major branches of the Endotrypanuml Leishmania clade. It is therefore possible that ancestors of modern sloths, which are members of the only surviving Eutherian order which was indigenous to the Neotropics throughout the Tertiary (65-6 million years ago MYA) [lo, 45, 471, were early hosts of Leishmania like organisms. Marsupials were also endemic to the Neotropics throughout the Tertiary and a number

6 516 J. EUK. MICROBIOL., VOL. 44, NO. 5, SEPTEMBER-OCTOBER 1997 of supposedly monogenetic trypanosomatids can thrive in the anal scent glands of marsupials. It has therefore been proposed that development in these glands may have been a stage in the transition from monogenetic to digenetic parasitism in the Leishmania and Trypanosoma cruzi lines [ 19, 261. The hystricomorph ancestors of the tree porcupines which are hosts of L. hertigi appeared in the Neotropics in the late Eocene (54-38 MYA) at the earliest, probably from the Nearctic, and the earliest Erethizontidae appeared in the fossil record in the early Oligocene (37-25 MYA) [45]. Therefore the late Eocene is probably the earliest date when the L. hertigi complex could have separated from the ancestors of Endotrypanum and Leishmania species. It may also be significant that the Kimura distance between the two T. cruzi strains in Fig. 3 (0.0760) is greater than between Endotrypanum and Leishmania ( ) indicating that the Endotrypanum and Leishmania genera may have diverged after the two major groups of T. cruzi parasites. T. cruzi is a parasite of eight mammalian orders and reduviid bugs and is restricted to the New World [40]. T. cruzi may have evolved in South America at the end of the Cretaceous ( MYA) at the time of the major mammalian radiation and the isolation of the Neotropics from the rest of the World [16, 431. The greater distance between the two T. cruzi strains would be consistent with an early Tertiary separation of the Endotrypanum and Leishmania genera in edentate hosts, and a later still separation of the L. hertigi complex at about the time the Hystricomorpha arrived in South America in the late Eocene (54-38 MYA) [ 10, 451. Representatives of all the surviving mammalian orders that were present in South America during the Oligocene and Miocene (37-6 MYA), except the Chiroptera, are at least occasionally infected with Leishmania and some are associated with highly specialized species of Endotrypanum or Leishmania that are restricted to South America and to specific hosts [25, 34, 45, 501. The migration of Leishmania to the Old World. It has previously been suggested that the original radiation of the Leishmania species occurred from Central Asia on the grounds of the present epidemiology and the historical record [21, 361. The radiation of the Old World Leishmania species from Central Asia is consistent with an earlier arrival from the Neotropics via the Isthmus of Panama and across Beringia. However it is possible that Beringia was already too cold for sandflies by the time the Neotropics became permanently reconnected to the Nearctic by the formation of the Isthmus of Panama in the early to mid Pliocene (5-3 MYA) [56, 571. Consequently Leishmania parasites may have migrated to the Nearctic earlier via the Antilles or via the chain of islands that was the precursor of Central America. The presence of a distinctive strain of L. mexicana in the Dominican Republic which may be indigenous [27, 311, suggests that Leishmania parasites can be carried across open sea and become established in new habitats. The climate of the west coast of the Nearctic and Beringia reached a warm optimum in the Miocene about 13 MYA since which time temperatures at higher latitudes have declined although with several reversals [lo], the passage of sandflies through this region may have been possible for much of the Miocene. The genus Leishmania may therefore have arrived in Central Asia in the Miocene (24-6 MYA), possibly in cricetid rodents which are the most important hosts of Leishmania parasites in Central Asia and which themselves underwent a major radiation during the Miocene [4, 471. The lizards that share burrows with cricetids may then have become infected, leading to the development of the lizard Leishmania in Central Asia. The lizard parasites may have developed at a time when the climate was already too cold for these species to cross back through Ber- ingia although they could spread south into Africa. The presence of Sergentomyia as well as Phlebotomus sandflies in the rodent and lizard burrows may have permitted occasional infections of lizards with mammalian parasites to become an established cycle. The classification of the lizard Leishmania species. The molecular data does not support generic status for the lizard Leishmania species. It has been proposed by Saf janova that the lizard Leishmania species should be placed in the subgenus L. (Sauroleishmania) [33]. The molecular data, with the exception of the DNA buoyant densities, is consistent with the lizard parasites being a discrete group that could legitimately be accorded subgeneric status. However, if the biogeographic account of the origin of the lizard Leishmania is correct, these parasites are phylogenetically more closely related to the Old World L. (Leishmania) species than they are to the New World L. (Leishmania) species and the resurrection of the subgenus L. (Sauroleishmania) would render the subgenus L. (Leishmania) paraphy letic. Conclusions. The DNA evidence (RFLP of SSU rrna, kdna dot blots, RNA polymerase I1 phylogeny [I71 and some ndna buoyant densities) indicates that the lizard Leishmania species are at the crown of the genus Leishmania together with the L. (Leishmania) species. This topology is consistent with the present distribution of these species. Consequently, although alternative hypotheses may be tenable, the hypothesis that accounts for the most of the currently available data is that the LeislzmanialEndotrypanum clade arose in the Neotropics in the Palaeocene (65-55 MYA) or Eocene (54-38 MYA), with a subsequent migration during the Miocene (24-6 MYA) or Pliocene (5-3 MYA) to the Nearctic and then to the Old World where the lizard Leishmania species evolved. ACKNOWLEDGMENTS We gratefully acknowledge the financial support of the International Scientific Cooperation Programme of the European Union (C11-CT ). We thank D. Croan (University of Technology Sydney) for discussing his unpublished data with us and Prof H. Townson and Prof M. Molyneux for critical readings of the manuscript. LITERATURE CITED 1. Adler, S Leishmania. Adv. Parasitol., 2: Alvarez, E, Robello, C. & Vignali, M Evolution of codon usage and base contents in kinetoplastid protozoans. Mol. Bid. Evol., 11 : Anon Control of the Leishmaniases. Report of a WHO Expert Committee. World. Health. Organ. Tech. Rep. Ser., 793: Ashford, R. W. & Bettini, S Ecology and Epidemiology: Old World. In: Peters, W. & Killick-Kendrick, R. (ed.), The Leishmaniases in Biology and Medicine, Vol. 1. Academic Press, London. Pp Ashford, R. W., Desjeux, P & DeRaadt, P Estimation of population at risk of infection and number of cases of Leishmaniasis. Parasitol. Today, 8: Bassam, B. J., Caetano-Anolles, G. & Gresshoff,? M Fast and sensitive staining of DNA in polyacrylamide gels. Anal. Biochem., 196: Belova, E. M Reptiles and their importance in the epidemiology of Leishmaniasis. Bull. Wid. Hlth. Org., 44: Berens, R. L. & Marr, J. J An easily prepared defined medium for cultivation of Leishmania donovani promastigotes. J. Parasitol., 64: Bozza, M., Fernandez, O., Degrave, W. M. & Lopes, U. G Characterization of Old World Leishmania species using amplified minicircle variable regions as molecular probes. Trans. R. SOC. Trop. Med. Hyg., 89: Briggs, J. C Biogeography and plate tectonics. Developments in Palaeontology and Stratigraphy, Elsevier, Amsterdam.

7 NOYES ET AL.-EVIDENCE FOR A NEOTROPICAL ORIGIN OF LEISHMANIA Briones, M. R., Nelson, K., Beverley. S. M., Affonso, H. T., Camargo, E. I? & Floeter-Winter, L. M Leishmania rarentolae taxonomic relatedness inferred from phylogenetic analysis of the small subunit ribosomal RNA gene. Mol. Biochem. Parasitol., 53: Chance, M. L., Gardener, F'. J. & Peters, W Biochemical taxonomy of Leishmania as an ecological tool. In: Ecologie des leishmanioses CNRS. Paris. Pp Chance, M. L., Peters, W. & Shchory, L Biochemical taxonomy of Leishmania I: Observations on DNA. Ann. Trop. Med. Parasitol., 68: Clark, C. G., Martin, D. S. & Diamond, L. S Phylogenetic relationships among anuran trypanosomes as revealed by riboprinting. J. Euk. Microbiol., 42: Clark, C. G. & Pung, 0. J Host specificity of ribosomal DNA variation in sylvatic Trypanosoma cruzi from North America. Mol. Biochem. Parasitol., 66: Cox, C. B Vertebrate palaeodistributional patterns and continental drift. J. Biogeog., 1: Croan, D. & Ellis, J. In Press. Monophyletic origin of the genus Sauroleishmania. Archiv fur Protistenkunde. 18. Croft, S. L., Chance, M. L. & Gardener, F'. J Ultrastructural and biochemical characterisation of stocks of Endotrypanum. Ann. Trop. Med. Parasitol., 74: Deane, M. I? & Jansen, A. M From a mono to a digenetic life-cycle: how was the jump for flagellates of the family Trypanosomatidae. Mem. Inst. Oswaldo. Cruz., 83: Dollahon, N. R. & Janovy, J., Jr Insect flagellates from feces and gut contents of four genera of lizards. J. Parasitol., 57: Elgood, C The early history of the Baghdad boil. J. R. Asiatic SOC., Fernandes, A. I?, Nelson, K. & Beverley, S. M Evolution of nuclear ribosomal RNAs in kinetoplastid protozoa-perspectives on the age and origins of parasitism. Proc. Natl. Acad. Sci. USA, 90: Frech, G. C., Bakalara, N., Simpson, L. & Simpson, A. M In vitro RNA editing-like activity in a mitochondria1 extract from Leishmania tarentolae. EMBO Journal, 14: Herrer, A Leishmania hertigi sp. n. from the tropical porcupine, Coendou rothschildi Thomas. J. Parasitol., 57: Herrer, A., Christensen, H. A. & Beumer, R. J Reservoir hosts of cutaneous leishmaniasis among Panamanian forest mammals. Am. J. Trop. Med. Hyg., 22: Jansen, A. M., Carreira, J. C. & Deane, M. I? Infection of a mammal by monogenetic insect trypanosomatids (Kinetoplastida, Trypanosomatidae). Mem. Inst. Oswaldo. Cruz., 83: Johnson, R. N., Young, D. G., Butler, J. E & Bogaert-Diaz, H Possible determination of the vector and reservoir of leishmaniasis in the Dominican Republic. Am. J. Trop. Med. Hyg., 46: Kelly, J. M Isolation of DNA and RNA from Leishmania. In: Hyde, J. E. (ed.), Protocols in Molecular Parasitology, Vol. 21. Humana Press, Totoa, NJ. Pp Killick-Kendrick, R., Lainson, R., Rioux. J. A. & Saf'janova, V. M The taxonomy of Leishmania like parasites of reptiles. In: Rioux, J. A. (ed.) Leishmania. Taxonomie et phylogenese. Applications eco-epidemiologique IMEEE. Montpelier. Pp Killick-Kendrick, R. & Wallbanks, K. R The identity of Leishmania chamaeleonis Wenyon, Trans. R. SOC. Trop. Med. Hyg., 75: Kreutzer, R. D Genetic analysis of Leishmania mexicana populations from Texas, Latin America, and the Caribbean. Am. J. Trop. Med. Hyg.. 43: Lainson, R. & Shaw, J. J Leishmanias of neotropical porcupines: Leishmania hertigi deanei nov. subsp. Acta Amazonica., 7: Lainson, R. & Shaw, J. J Evolution, classification and geographical distribution. In: Peters, W. & Killick-Kendrick, R. (ed.), The Leishmaniases in Biology and Medicine, Vol. 1. Academic Press, London. Pp. I Lainson, R., Shaw, J. J., Silveira, E T, Desouza, A. A. A., Braga, R. R. & Ishikawa, E. A. Y The dermal leishmaniases of Brazil, with special reference to the eco-epidemiology of the disease in Amazonia. Mem. Inst. Oswaldo. Cruz., 89: Landweber, L. E & Gilbert, W RNA editing as a source of genetic variation. Nature, 363: Lysenko, A. J The distribution of leishmaniasis in the Old World. Bull. Wld. Hlth. Org., 44: Manson-Bahr, F'. E. C. & Heisch, R. B Transient infection of man with a Leishmania (L. adleri) of lizards. Ann. Trop. Med. Parasitol., 55: Marche, S., Roth, C., Phillipe, H., Dollet, M. & Baltz, T Characterisation and detection of plant trypanosomatids by sequence analysis of the small subunit ribosomal RNA gene. Mol. Biochem. Par- asitol., 71~ Maslov, D. A. & Simpson, L Evolution of parasitism in kinetoplastid protozoa. Parasitol. Today, 11: Miles, M. A Transmission cycles and heterogeneity of Trypanosoma cruzi. In: Lumsden, W. H. R. & Evans, D. A. (ed.), Biology of the Kinetoplastida, Vol. 2. Academic Press, London. Pp Molyneux, D. H Virus-like particles in Leishmania parasites. Nature, 249: Nei, M. & Li, W. H Mathematical model for studying genetic variation in terms of restriction endonucleases. Proc. Natl. Acad. Sci. USA, 76: Novacek, M. J Mammalian phylogeny: shaking the tree. Nature, 356: Noyes, H. A., Perez Camps, M. A. & Chance, M. L Leishmania herreri (Kinetoplastida; Trypanosomatidae), is more closely related to Endotrypanum (Kinetoplastida; Trypanosomatidae) than to Leishmania. Mol. Biochem. Parasitol., 80: Patterson, B. & Pascual, R The fossil mammal fauna of South America. In: Keast, A,, Erk, E & Glass, B. (ed.), Evolution, Mammals and Southern Continents. State University of New York Press, Albany. Pp Rodgers, M. R., Popper, S. J. & Wirth, D. E Amplification of kinetoplast DNA as a tool in the detection and diagnosis of Leishmania. Exp. Parasitol., 71: Savage, R. J Mammal Evolution. British Museum Natural History, London. 48. Shaw, J. J Taxonomy of the haemoflagellates of sloths and of related flagellates found in edentates; statistical evaluation of morphology; differentiation of the genus Endotrypanum and of the two species within it; trypanosomes in edentates; Trypanosoma preguci sp. n.; Trypanosoma leeuwenhoeki sp. n.; summarized descriptions of Endotrypanum schaudinni and of the three new species of haemoflagellates of sloths. In: The Haemoflagellates of Sloths. London School of Hygiene and Tropical Medicine, London. Pp Shaw, J. J Taxonomy of the Genus Leishmania: Present and future trends and their implications. Mem. Inst. Oswaldo. Cruz., 89: Shaw, J. J. & Lainson, R Ecology and epidemiology: New World. In: Peters, W. & Killick-Kendrick, R. (ed.), The Leishmaniases in Biology and Medicine, Vol. 1. Academic Press, London. Pp Thomaz-Soccol, V., Lanotte, G., Rioux, J. A,, Pratlong, E, Martini-Dumas, A. & Serres, E Monophyletic origin of the genus Leishmania Ross, Ann. Parasitol. Hum. Comp., 68: Thomaz-Soccol, V., Lanotte, G. L., Rioux, J. A,, Pratlong, E, Martini-Dumas, A. & Serres, E Phylogenetic taxonomy of New World Leishmania. Ann. Parasitol. Hum. Comp., 68: Vickerman, K The identity of L. chamaeleonis Wenyon Trans. R. SOC. Trop. Med. Hyg., 59: Vickerman, K The diversity of the kinetoplastid flagellates. In: Lumsden, W. H. R. & Evans, D. A. (ed.), Biology of the Kinetoplastida, Vol. 1. Academic Press, London. Pp Wenyon, C. M Observations on the intestinal protozoa of three Eygyptian lizards, with a note on a cell invading fungus. Parasitology, 12: Wolfe, 3. A An analysis of Neogene climates in Beringia. Palaeogeogr., Palaeoclimatol., Palaeoecol., 108: Wolfe, J. A Tertiary climatic changes at middle latitudes of western North America. Palaeogeogr., Palaeoclimatol.. Palaeoecol., 108: Zeledon, R., Ponce, C. & de Ponce, E Finding of Leishmania hertigi in the Costa Rican porcupine. J. Parasitol., 63: Received , ; accepted