Diptera: Psychodidae - Semantic Scholar

MORPHOLOGY, SYSTEMATICS, EVOLUTION

The Taxonomic Status of Genetically Divergent Populations of Lutzomyia longipalpis (Diptera: Psychodidae) Based on the Distribution of Mitochondrial and Isozyme Variation ˜ ANGO,3 DOUGLAS NORRIS,2 JAZZMIN ARRIVILLAGA,1 JOHN- PAUL MUTEBI,2 HERMES PIN 4 1 BRUCE ALEXANDER, M. DORA FELICIANGELI, AND GREGORY C. LANZARO5

J. Med. Entomol. 40(5): 615Ð627 (2003)

ABSTRACT The sand ßy, Lutzomyia longipalpis (Lutz & Neiva) reputedly is a complex of cryptic species; however, there is currently no consensus as to the number of species in the complex or their geographic distributions. We conducted phylogenetic analyses of 31 populations from throughout the species range, using seven isozyme loci and genes in the mitochondrial genome. Analyses of these two independent sets of markers were largely concordant and revealed four distinct clades that support the existence of four species. The four clades have distinct geographic ranges: (1) Brazil (Species A ⫽ Lu. longipalpis sensu stricto), (2) Laran (Species B ⫽ Lu. pseudolongipalpis), (3) cis-Andean (Species C), and (4) trans-Andean (Species D). The cis-Andean clade may be subdivided further into two groups, one in Colombia and one in northwestern Venezuela, but their taxonomic status remains unresolved. Knowledge that Lu. longipalpis is a complex of species may ultimately shed light on anomalies in the epidemiology of visceral leishmaniasis in the New World. KEY WORDS Lutzomyia longipalpis, species complex, isozymes, mitochondrial DNA, phylogeny

THE SAND FLY Lutzomyia longipalpis (Lutz and Neiva 1912) sensu lato is the primary vector of Leishmania chagasi the etiological agent of zoonotic visceral leishmaniasis (ZVL) in South and Central America. It has a wide geographical range that extends from the Yucatan Peninsula in Mexico to northern Argentina and Paraguay (Young and Duncan 1994), including all the countries of Central America except Belize and most of those of tropical South America east of the Andes. There are no published records of Lu. longipalpis from Guyana, Surinam, or French Guiana. In addition, there are no records of this species from the PaciÞc Coast of Colombia, Ecuador, Peru, or Chile. Although Lu. longipalpis has one of the widest geographical ranges of any species in the genus, its closest relatives (members of the subgenus Lutzomyia, series longipalpis) are almost entirely restricted to Brazil, where they are associated with calcareous rocks and caves (Williams 1999). 1 Universidad de Carabobo, BIOMED, Centro Nacional de Referencia de Flebotomos, La Morita, Maracay, Venezuela, Apdo. 4873 (e-mail: [email protected]). 2 University of Texas Medical Branch, Department of Pathology, Galveston, Texas 77555Ð 0609. 3 Universidad Central de Venezuela, Instituto de Zoologia Tropical, Laboratory Biologia y Ecologia de Poblaciones, Caracas, Venezuela. 4 Laborato ´ rio de Entomologia Me´ dica, Centro de Pesquisas Rene´ Rachou, Fundaç aõ Oswaldo Cruz, Av. Augusto de Lima 1715, 30190002 Belo Horizonte-MG, Brazil. 5 University of California, Department Entomology, Briggs Hall, Davis, California, USA, 95616 Ð 8579.

This species was Þrst described as Phlebotomus longipalpis by Lutz and Neiva (1912) from specimens collected from Saõ Paulo and near Cordisburgo in the state of Minas Gerais, Brazil. The original description is incomplete, no holotype was designated, and the original series of specimens have been lost. The species was redescribed by Franç a in 1920, using specimens from Mangaratiba-Quixada´ (Ceara´), Saõ Paulo, and Minas Gerais in Brazil as well as “Rives du Parana´” Paraguay. Molecular and biochemical studies of Lu. longipalpis have resulted in the emergence of three schools of thought regarding its taxonomic status: (1) that it is a complex of morphologically similar, allopatric species (Lanzaro et al. 1993, Lanzaro and Warburg 1995, Mutebi et al., 1998, 1999, Arrivillaga et al. 2000), (2) that it is a complex of species, some of which may occur sympatrically, isolation being maintained by the production of chemically distinct terpenoid pheromones (Ward et al. 1983, 1988, Hamilton et al. 1996, Lampo et al. 1999), and (3) that it is a single heterogeneous species with a wide geographical distribution (Bonnefoy et al. 1986, Mukhopadhay et al. 1998, Munstermann et al. 1998). Based on isoenzyme comparisons and cross-mating experiments using laboratory-reared sand ßies from Costa Rica, Colombia and Brazil, Lanzaro et al. (1993) suggested that Lu. longipalpis was a complex of at least three allopatric species. Subsequent isozyme studies

0022-2585/03/0615Ð0627$04.00/0 䉷 2003 Entomological Society of America

Colombia Bucaramanga Palo Gordo Neiva Durania Costa Rica Brasilito Liberia Honduras Rancho Grande Orocuina Tiger Island San Francisco del Coray Pavana San Juan Bautista Los Guatales Nicaragua Pochomil Las Huertas Venezuela Curarigua Cojedes Trujillo El Paso Brazil Pakaraima Santarem Salvaterra Baturite Jacobina Lapinha caves Fortaleza Sobral Camara Bacabal Itapipoca Montes Claros 58.4 (10.4) 40.3 (7.6) 11.3 (1.4) 27.9 (2.9) 71.3 (11.2) 71.4 (11.1) 41.3 (10.4) 47.7 (7.7) 38.7 (7.3) 36.3 (7.3) 18.1 (2.1) 18.9 (2.3) 28.1 (4.6) 18.1 (2.1) 49.0 (8.8) 15.1 (1.8) 15.0 (1.6) 15.4 (1.7) 16.4 (1.6) 61.6 (3.8) 54.3 (5.9) 52.3 (2.8) 65.3 (2.8) 63.7 (3.1) 59.9 (4.9) 60.3 (5.0) 62.9 (4.7) 59.1 (3.2) 60.6 (4.4) 62.1 (5.2)

10⬚28⬘ N, 85⬚18⬘ W 10⬚28⬘ N, 85⬚18⬘ W

14⬚02⬘ N, 87⬚41⬘ W 13⬚29⬘ N, 87⬚06⬘ W 13⬚17⬘ N, 87⬚37⬘ W 13⬚40⬘ N, 87⬚30⬘ W 13⬚25⬘ N, 87⬚20⬘ W 13⬚34⬘ N, 87⬚16⬘ W 13⬚21⬘ N, 87⬚37⬘ W

11⬚45⬘ N, 86⬚30⬘ W 13⬚02⬘ N, 86⬚34⬘ W

09⬚59⬘ N, 69⬚55⬘ W 09⬚40⬘ N, 68⬚40⬘ W 09⬚25⬘ N, 70⬚30⬘ W 09⬚59⬘ N, 69⬚55⬘ W

05⬚30⬘ N, 60⬚40⬘ W 02⬚26⬘ S, 54⬚41⬘ W 00⬚46⬘ S, 48⬚31⬘ W 04⬚20⬘ S, 38⬚53⬘ W 11⬚11⬘ S, 40⬚31⬘ W 19⬚03⬘ S, 43⬚57⬘ W 03⬚43⬘ S, 38⬚30⬘ W 03⬚42⬘ S, 40⬚21⬘ W 00⬚46⬘ S, 48⬚31⬘ W 00⬚46⬘ S, 48⬚31⬘ W 03⬚30⬘ S, 39⬚35⬘ W 16⬚43⬘ S, 43⬚52⬘ W

Isozyme n (S.E.)

07⬚08⬘ N, 73⬚09⬘ W 07⬚08⬘ N, 73⬚09⬘ W 02⬚56⬘ N, 75⬚18⬘ W 08⬚08⬘ N, 73⬚09⬘ W

Location

n ⫽ 12 (F ⫽ 12) n ⫽ 14 (E ⫽ 14) n ⫽ 16 (K ⫽ 16) n ⫽ 9 (A ⫽ 9) -

n ⫽ 16 (A ⫽ 2, B ⫽ 1, C ⫽ 1, D ⫽ 12) n ⫽ 14 (G ⫽ 14) n ⫽ 20 (D ⫽ 20)

-

n ⫽ 16 (G ⫽ 16) -

n ⫽ 17 (A ⫽ 17) n ⫽ 17 (A ⫽ 17) n ⫽ 16 (A ⫽ 16) n ⫽ 16 (A ⫽ 16) -

n ⫽ 12 (A ⫽ 10, B ⫽ 2) n ⫽ 12 (A ⫽ 12) -

-

n ⫽ 12 (A ⫽ 12) -

n ⫽ 12 (A ⫽ 12)

n ⫽ 12 (A ⫽ 14) -

n ⫽ 14 (A ⫽ 14)

16S region n (haplotype freq.)

JOURNAL OF MEDICAL ENTOMOLOGY

n ⫽ 28 (L ⫽ 25, M ⫽ 3) n ⫽ 13 (H ⫽ 12, I ⫽ 1) n ⫽ 34 (G ⫽ 34) n ⫽ 32 (K ⫽ 32) n ⫽ 16 (J ⫽ 16) n ⫽ 35 (N ⫽ 35) -

n ⫽ 39 (A ⫽ 36, B ⫽ 2, C ⫽ 1) n ⫽ 32 (E ⫽ 30, F ⫽ 2) n ⫽ 29 (A ⫽ 26, B ⫽ 1, D ⫽ 2)

-

n ⫽ 25 (S ⫽ 25) -

n ⫽ 13 (F ⫽ 13)

n ⫽ 10 (A ⫽ 10) -

n ⫽ 12 (Q ⫽ 12) n ⫽ 39 (R ⫽ 39)

n ⫽ 14 (A ⫽ 14)

12S region n (haplotype freq.)

n ⫽ 38 (O ⫽ 35, P ⫽ 3)

COI region n (haplotype freq.)

Collection sites, sample sizes for isozyme and mitochondrial analyses and haplotype frequencies for 31 populations of Lu. longipalpis

Population

Table 1.

616 Vol. 40, no. 5

September 2003

ARRIVILLAGA ET AL.: MOLECULAR-BIOCHEMICAL ANALYSIS OF THE Lu. longipalpis

617

Fig. 1. Collection sites for the 31 populations Lu. longipalpis included in this study.

compared wild-caught specimens from different localities within Colombia (Lanzaro et al. 1997), Brazil (Mutebi et al. 1999), and Central America (Mutebi et al. 1998), the results of which provided additional support for the hypothesis that one species occurs in each of the three regions. Arrivillaga et al. (1999a,b) and Arrivillaga and Feliciangeli (2000) presented evidence for two cryptic species within the Lu. longipalpis complex in Venezuela, based on two diagnostic isozyme loci and morphometric analysis of larval mouthpart characters. Lampo et al. (1999) reported the occurrence of two species occurring sympatrically at Curarigua (El Paso), Venezuela, based on the same two isozyme loci. One of these species showed similarities to both Brazilian and Colombian populations, whereas the other appeared to be indigenous to Venezuela. Although the patchy distribution of American visceral leishmaniasis and different manifestations of Le. chagasi infection may be explained by reasons other than the involvement of several species within a Lu. longipalpis complex, the public health importance of the disease is such that the true identity of the vector in each focus needs to be determined. Visceral leishmaniasis has been reported from all countries where Lu. longipalpis is present, with the exception of Panama (WHO 1990). An atypical, nodular form of cutaneous leishmaniasis caused by Le. chagasi has been reported in some areas of Central America (Zeledon et al., 1989; Ponce et al., 1991, Belli et al., 1999). Failure of the parasite to visceralize in many patients in Cen-

tral America may be linked to lower concentrations of the vasodilator maxadilan in the saliva of Lu. longipalpis from this region (Warburg et al., 1994, Yin et al., 2000). Simultaneous analyses of genetic markers in the nuclear and mitochondrial genomes can provide useful insights into the evolutionary history and taxonomic status of populations. In the current study we investigated geographical variation in the cytochrome c oxidase I gene (COI), as well as the 12S and 16S ribosomal genes all within the mitochondrial genome. In addition, we studied several enzyme encoding nuclear genes with the goal of clarifying the phylogenetic and taxonomic relationships among populations of Lu. longipalpis.

Materials and Methods Sand Fly Collections. Specimens of Lu. longipalpis were collected from 31 different localities in six countries (Table 1; Fig. 1). Flies were aspirated directly from domestic animals (horses, cows, pigs, and chickens) or from surfaces in and around their enclosures. In some places these collections were augmented with collections from CDC light traps. Three other sand ßy species were analyzed as outgroups: Phlebotomus papatasi (Scopoli) from a colony originating from Israel, Lutzomyia evansi (Nun˜ ez Tovar) collected in Costa Rica and Lutzomyia gomezi (Nitzulescu) collected in Venezuela.

618


mtDNA Variation Polymerase Chain Reaction (PCR) Amplification. Genomic DNA was extracted and puriÞed from frozen single sand ßies using previously described methods (Bender et al. 1983). Three target DNA fragments were ampliÞed: COI, 12S and 16S mtDNAs. AmpliÞcation reactions were carried out in 25 ␮l containing 0.5 U of Taq polymerase (Biolase), one X enzyme buffer with MgCl2 (2.5 ␮l/reaction), 100 nmol dNTP, and 3 ␮l DNA as template. AmpliÞcation and sequencing of puriÞed products were achieved using the following primers: COI region (Kambhapati and Smith 1995) TGATCAAATTTATAAT (⫹) and GGTAAAATT AAAATATAAACTTC (⫺); 12S region (Simon et al. 1994) TACTATGTTACGACTTAT (⫹) and AAACTAGGATTAGATACCC (⫺); 16S region (Simon et al. 1994) TTACGCTGTTA TCCTAA (⫹) and CACCTGTTTAACAAAAACAT (⫺). SSCP and Sequencing Analyses. Polymorphism was detected by single strand conformation polymorphism (SSCP) analysis for each specimen (12Ð39 individuals) per population for 13 populations (Table 1; Fig. 1). Gels for SSCP were composed of 0.6X TBE (1X TBE ⫽ 53 mM Trizma base, 53 mM boric acid, 1.5 mM EDTA ph 8.0), 7.9% acrylamide, and 0.21% N,N⬘-methyllenebisacrylamide. Following PCR ampliÞcation 7 ␮l of product was removed to a 500 ␮l tube containing 9 ␮l of denaturing loading mix (20 mM NaOH, 90% formamide, 0.05% bromophenol blue, and 0.05% xylene cyanol). The tube was tapped to mix the contents, spun brießy in a microcentrifuge, heated to 95⬚C for 3 min in a thermal cycler and plunged into ice for at least 10 min. From this cooled mixture, 16 ␮l was loaded into the SSCP gel. Electrophoresis was carried out on vertical slab gels (16 ⫻ 18 cm) that were run in circulating electrode buffer (0.5X TBE) refrigerated at 4⬚C. Current was maintained at 20 mA until the xylene cyanol dye had migrated to the bottom of the gel. Gels were then silver stained, dried, scored, and photographed. Polymorphism was detected by SSCP in the following way: Preliminary studies veriÞed the reproducibility of SSCP migration patterns. This information was used to develop an optimal protocol. SSCP was Þrst used to identify unique haplotypes within single populations. These were then used to create haplotype proÞles for each population screened. Population proÞles were compared by rerunning putative haplotypes, from all populations, on a single gel. Haplotypes were veriÞed by direct cycle sequencing on an ABI-377 automated sequencer (Perkin-Elmer). Sequence Analyses. Sequences for all COI, 12S and 16S mitochondrial haplotypes initially were aligned using Clustal V (Higgins et al. 1992) and manually adjusted for obvious misalignments. Gaps were treated as missing data. All phylogenetic analyses were computed using PAUP 4.0b2 (Swofford 1999). Neighbor-joining (NJ) analysis (Saito and Nei 1997) was computed under the assumptions of the Kimura-2parameter model (K2P). Maximum parsimony (MP) analysis was computed with equally weighted and

Vol. 40, no. 5

reweighted characters (Farris 1983, Goloboff 1993) based on the consistency index (Farris 1969) by heuristic search using the tree bisection-reconnection (TBR) branch-swapping algorithm with 100 random stepwise additions. This adjusts the weight of characters (nucleotides) based on their Þt to the most parsimonious tree after deleting uninformative characters (Carpenter 1996). ConÞdence values were estimated by bootstrapping (1,000 replications for COI and 100 for 12S and 16S). The nucleotide composition, pairwise base differences, and genetic distance (K2P) were determined with PAUP 4.Ob2. Isozyme Variation Isozyme Analysis. Sand ßies from 30 populations (Table 1; Fig. 1) were assayed for genetic variation at seven isozyme loci using standard starch gel electrophoresis. Individual sand ßies were homogenized in 8 ␮l of distilled water. The homogenate was fractionated by electrophoresis on 12.5% (wt:vol) horizontal starch gels. The seven enzymes assayed were: hexokinase (E. C. 2.7.1.1), isocitrate dehydrogenase- one (E. C. 1.1.1.42), malic acid dehydrogenase-1 (E.C. 1.1.1.37), malic acid dehydrogenase- two (E. C. 1.1.1.37), malic enzyme-1 (E. C. 1.1.140), phosphoglucoisomerase (E. C. 5.3.1.9), and alpha-trehalase (E. C. 3. 2.1.28). Running buffers and histochemical staining methods were as described by Lanzaro et al. (1993). Differences in migration distances of bands at the same locus were designated as different alleles, and multiple bands at the same locus in an individual were assumed to be heterozygous. Loci were designated with positive or negative codes depending on whether they migrated to the anode or cathode. For each locus, the most frequent electromorph was given an rf value of 1.00. Sand ßies from three laboratory stocks [i.e., Melgar (Colombia), Liberia (Costa Rica) and Lapinha Caves (Brazil)] used by Lanzaro et al. (1993) were included on each gel to aid in identiÞcation of alleles. Population Genetic Analysis. Individual ßy genotypes were organized into text input Þles and populations were analyzed for genetic variability, compliance to Hardy-Weinberg expectations, and genetic relatedness using the computer program, BIOSYS-1. (Swofford and Selander 1989). Allele frequencies and NeiÕs genetic distance were calculated. Phenetic analysis was completed using a cluster analysis employing the unweighted pair-group method (unweighted pairgroup method with arithmetic average) of NeiÕs genetic distance (1978). Isozyme Phylogenetic Analyses. Raw input Þles for the BIOSYS-1 program were converted to PHYLIP 3.5c version format (Phylogeny Inference Package; Felsenstein, 1995) and the continuous character maximum likelihood (CONTML) program was performed for gene frequencies (all alleles at each locus). Lu. evansi was used as an outgroup, global rearrangements were allowed, and the input order of species (OTUÕs ⫽ populations in this case) were randomized. Genetic distances were computed using the GENDIST program (PHYLIP 3.5c) assuming that each locus con-

September 2003


619

Fig. 2. SSCP gel illustrating mtDNA COI haplotypes from four phylogenetic species in the Lu. longipalpis complex. (1): size ladder, (2): Curarigua 1, (3): Curarigua 10, (4): Curarigua 3, (5): Curarigua 6, (6): Curarigua 2, (7): Trujillo 12, (8): Trujillo 10, (9): Lapinha 51, (10): Lapinha 55, (11): Lapinha 53, (12): Lapinha 27, (13): Lapinha 24, (14): Lapinha 58, (15): open, (16): Baturite 38, (17): leader, (18): Lu. gomezi, (19): Lu. whitmani, (20): Ph. papatasi, (21): Liberia 60, (22): Liberia 61, (23): Liberia 66, (24): Liberia 63, (25): Bucaramanga 90, (26): Bucaramanga 87, (27): Huila, (28): open, (29): ladder, (30): Bucaramanga 86.

tained all alleles, and Nei distances were calculated from the original data set (allele frequency). A search for the best hypothetical phylogenetic tree was performed using the FITCH program (PHYLIP 3.5c) using the distances obtained from GENDIST as input; negative branches were not allowed, and Lu. evansi was Þxed as reference group (outgroup). Global rearrangements were allowed and the input order of OTUÕs were randomized with Þve jumbles during randomization. Further analysis was done with the KITSCH program (PHYLIP 3.5c) which uses the same additive model employed in FITCH but assumes an evolutionary clock. Results mtDNA Variation Variation at the COI Locus. Within-population haplotype variation was low, although variation among haplotypes between populations was high. This indicated that variation at the COI locus is a useful genetic marker to distinguish cryptic species in the Lu. longipalpis complex (Fig. 2). We identiÞed 19 haplotypes, based on COI sequence (Table 1; Fig. 3). Seven haplotypes (G, J, K, N, Q, R, S) were unique to single populations. Four populations contained two haplotypes, not shared with any other (E, F, H, I, L, M, O, P). Two populations, separated by 8 km (Curarigua and El Paso), contained four haplotypes, three were shared (A, B, and D) and one occurred in only one population (C).

Values for K2P genetic distances indicated the presence of four population groups with low intra-group versus high inter-group distances (Table 2). Phylogenetic analysis of COI sequences revealed four well resolved clades, corresponding to the population groups based on K2P genetic distances (Fig. 4). The G ⫹ C content did not vary signiÞcantly among haplotypes (16 Ð19%, P ⫽ 0.992). In general, transversions favored divergence among clades, whereas transitions favored divergence within clades. Phylogenetic trees based on both NJ (not shown) and MP (Fig. 4) had identical topologies, with high bootstrap scores and large nucleotide divergence among clades (10.91Ð9.47%). The strict consensus tree, based on 131 informative characters and derived from three equally parsimonious trees, illustrated the relationships among four strongly supported clades: Laran (Venezuela), cis-Andean (Colombia/Venezuela/Brazil) trans-Andean (Central America/Colombia/Venezuela), and Brazilian (Brazil). Mitochondrial 12S and 16S Ribosomal Variation. A 395 bp fragment of the 12S gene was studied by SSCP analyses from a total of 155 individual ßies from 11 populations. Of these, 14 were sequenced, representing eight haplotypes (Table 1). The NJ analysis indicated that all populations are contained within a single monophyletic clade (76% bootstrap); however, the tree topology indicates some phylogenetic structuring with two groups supported by low bootstrap values (34% and 64%). However, the MP strict consensus analysis did not support the NJ results (internal boot-

620


Vol. 40, no. 5

Fig. 3. Unique mtDNA haplotypes found for the CO I gene fragment from 13 populations of Lu. longipalpis. Each haplotype was assigned a letter designation (AÐS). Haplotype frequencies for each population are presented in Table 1. Sequence for a ⬇495 bp fragment of the CO I gene is presented. (- indicates gaps indels). Abbreviations designating haplotypes refer to site of origin as follows: jacobina ⫽ Jacobina, Brazil; Cu ⫽ Curarigua, Venezuela; La ⫽ Lapinha Caves, Brazil; Tru ⫽ Trujillo, Venezuela; Pa ⫽ Pavana, Honduras; Li ⫽ Liberia, Costa Rica; Bu ⫽ Bucaramanga, Colombia; baturite ⫽ Baturite´ , Brazil; ne ⫽ Huila, Brazil; ro ⫽ Pacaraima, Brazil; santa ⫽ Santarem, Brazil; salva ⫽ Salvaterra, Brazil. Numbers following some names indicate that more than one haplotype was recovered from that site.

September 2003


621

Table 2. Kimura-2- parameter (K2P) genetic distances among four clades based on cythocrome c oxidase I from 13 populations of Lu. longipalpis. The number of populations for each clade are designated N. Standard error in parenthesis

Laran (N ⫽ 2) cis-Andes (N ⫽ 4) trans-Andes (N ⫽ 4) Brazil (N ⫽ 5)

Laran

cis-Andes

trans-Andes

Brazil

0.013 (0.006) Ñ Ñ Ñ

0.125 (0.02) 0.055 (0.022) Ñ Ñ

0.129 (0.006) 0.130 (0.016) 0.077 (0.029) Ñ

0.123 (0.003) 0.101 (0.012) 0.101 (0.01) 0.0324 (0.01)

strapping ⬍50%), indicating no divergence among populations based on 107 informative characters (trees not presented). SSCP analysis of a 360 bp fragment of the 16S mtDNA locus was conducted on 140 ßies from 10 populations. Sequence obtained for 11 individuals revealed only two haplotypes (Table 1). MP and NJ analyses, employing 25 informative characters, produced phylogenetic trees with no signiÞcant bootstrap

values, indicating no divergence at this locus among the populations studied (trees not presented). COI versus 12S/16S Based Phylogenies. Phylogenetic analysis of mtDNA ribosomal sequences revealed a high degree of homogeneity among populations of Lu. longipalpis collected over an extensive geographical area. In contrast, phylogenetic analysis of COI sequence, conducted on the same set of populations, revealed signiÞcant levels of divergence. This

Fig. 4. Strict Consensus tree obtained by MP analysis based on 131 mtDNA COI informative characters. Numbers on branches are bootstrap values based on 1,000 replicates. Labels at the terminus of each branch provide the name of the population. The speciÞc haplotype is indicated in parentheses by an alphabetical designation corresponding to those in Fig. 4 and in Table 1. Lu. evansi represents the in group and Ph. papatasi the outgroup.

622


Vol. 40, no. 5

Table 3. Genetic distance (Nei 1978) based on seven isozyme loci, within and between species in the Lu. longipalpis species complex. Range in parentheses Species

No. of populations

Species B (Laran)

Species B Species A Species C1 Species C2 Species D

2 11 11 2 4

0.004 (0.004Ð0.004) .325 (0.276Ð0.391) 0.487 (0.438Ð0.539) 0.415 (0.345Ð0.493) 0.466 (0.370Ð0.547)

Species A (Brazilian) 0.003 (0.000Ð0.015) 0.150 (0.133Ð0.166) 0.147 (0.116Ð0.175) 0.145 (0.113Ð0.168)

difference may be the consequence of a slower rate of evolution of the ribosomal region. Isozyme Variation Genetic distance coefÞcients, D (Nei 1978), based on analysis of seven enzyme encoding loci revealed high levels of genetic differentiation among population groups originating from Þve regions; Colombia, Central America, Brazil and two areas in Venezuela (Table 3). Cluster analysis was conducted on a matrix of D values for all pairwise comparisons among 30 populations employing the unweighted pair group algorithms using arithmetic averages (unweighted pairgroup method with arithmetic average) of Sneath and Sokal (1973). In the resulting dendrogram (Fig. 5) populations in each of the Þve geographic regions clustered together. Levels of genetic distance among the Þve groups (D ⫽ 0.12Ð 0.49) were high and con-

Species C1 (cis-Andean: Colombian)

Species C2 (cis-Andean: Venezuelan)

0.000 (0.000Ð0.001) 0.141 (0.107Ð0.174) 0.133 (0.084Ð0.165)

0.015 (0.015Ð0.015) 0.121 (0.073Ð0.172)

Species D (trans-Andean)

0.017 (0.003Ð0.030)

sistent with values previously reported for sibling species in the genus Lutzomyia (Lanzaro and Warburg 1995). Within regions, values for D were consistently below 0.02, typical for levels of divergence between local populations within a single insect species. The total FST for all populations for all loci was high 0.62, indicating genetic sub-structuring among populations. Effective migration rates (Nem) among all 30 Þeld collected populations were extremely low, Nem ⫽ 0.15, indicating restricted gene ßow among populations. Several enzyme loci diagnostic for the Þve population groups identiÞed in this study included Mdh-2, Gpi, and Hk (Table 4). Gels illustrating Mdh-2 and Gpi zymograms are presented in Fig. 6. Isozyme Phylogeny CONTML Analysis. A single rooted tree was obtained (Fig. 7A) from 5,547 trees analyzed based on

Fig. 5. A unweighted pair-group method with arithmetic average dendrogram constructed from values for pairwise genetic distance (Nei 1978) for 30 Þeld populations of Lu. longipalpis in Colombia, Venezuela, Central America and Brazil. Labels on branches indicate the geographic region of origin for each population group.

September 2003


623

Table 4. Allele frequencies for 3 isozyme loci that discriminated 5 clades from field-collected Lu. longipalpis in South and Central America. Frequencies of the most common alleles are highlighted in bold print. Clades in this table are based on cluster analysis of Nei’s genetic distance D (Fig. 5) Clade Locus Mdh-2 (N) A B C D Gpi (N) A B C D E F G H I J Hk (N) A B C

cis-Andean (Colombia)

trans-Andean

Brazilian

cis-Andean (Venezuela)

(226) 0.998 0.002 0 0

(714) 0.996 0.004 0 0

(791) 0.994 0.004 0.001 0

(40) 0.900 0.063 0 0.038

(40) 0.087 0.063 0 0.850

(162) 0.892 0.022 0.040 0.022 0.003 0.003 0.012 0 0 0.006

(523) 0.004 0.973 0.005 0 0.001 0.007 0 0 0.011 0

(641) 0.004 0.011 0.013 0.002 0 0.007 0 0.001 0.962 0

(40) 0.063 0.025 0.863 0.013 0 0 0.038 0 0 0

(39) 0.026 0.013 0.154 0.026 0.026 0 0 0 0.756 0

(116) 0.897 0.078 0.026

(336) 0.991 0.007 0.001

(742) 0.993 0.004 0.003

(34) 0.956 0.015 0.029

(39) 0.051 0 0.949

Laran

Fig. 6. (A) A starch gel stained for malic acid dehydrogenase-2 (Mdh-2, E.C. 1.1.1.37). Samples include specimens from four populations in Venezuela. Samples from Trujillo and Cojedes represent species C and carry the A allele at this locus. Samples from El Paso and Curarigua represent species B and carry the B allele. CO and BZ are colony controls from Melgar, Colombia and Lapinha Caves, Brazil respectively. (B) A starch gel stained for the enzyme phosphoglucoisomerase (Gpi, E. C. 5.3.1.9). The gel contains 30 Lu. longipalpis individuals from eight populations. Bucaramanga and Palo Gordo are in Colombia and represent species C1; Liberia and Brasilito are in Costa Rica and represent species D; Salvaterra and Fortaleza are in Brazil and represent species A; and Curarigua is in Venezuela and represents species B. There are six alleles illustrated on this gel. The A allele is the most common allele in species C; the B allele is the most common in species A and B; and the I allele is most common in species D. The F, S1 and S2 alleles are less common.


Fig. 7. Phylogenetic trees illustrating relationships among 30 populations of Lu. longipalpis. (A) Tree based on CONTML analysis based on allele frequencies at seven loci. (B) Tree based on KITSCH analysis based on genetic distance (Nei 1978). Lu. evansi was designated as outgroup in both analyses. Methods are described in detail in the text.

624 Vol. 40, no. 5

September 2003


625

Fig. 8. Geographical distribution of species in the Lu. longipalpis complex. (A) Distribution of species based on unweighted pair-group method with arithmetic average cluster analysis of genetic distance (Nei 1978) computed from allozyme frequencies (seven loci) among 30 populations (Fig. 5). (B) Distribution of species based on phylogenetic analysis of isozyme data [same data as in (A), Fig. 7B]. (C) Distribution of species based on phylogenetic (maximum parsimony) of sequence for the mitochondrial cytochrome c oxidase I locus recovered from 13 populations of Lu. longipalpis (Fig. 4).

allozyme frequencies. The tree topology shows four independent, monophyletic clades: cis-AndeanÐ Colombia represents an ancestral, Andean gene pool, closely related to three recently evolved and monophyletic internal groups: Brazilian, Laran (northwest Venezuelan populations) and trans-AndeanÐ Central American populations. The two monophyletic lineages (Brazilian and trans-Andean) contain complex internal branching (Fig. 7A). The Brazilian populations consist of two internal groups. The most recently evolved in-group is represented by two Venezuelan populations that make up the Laran Clade. However, the internal “paraphyletic” grouping indicates a different divergence time for these groups. The two internal Brazilian groups contain populations that do not have well-deÞned geographical distributions, indicating extensive gene ßow between them. The transAndean Clade contains a cluster of closely related Central American populations and appears to represent an ancestral sister-group of the cis-Andean Clade. This arrangement indicates extensive gene ßow among populations in Central America.

KITSCH Analysis This analysis revealed four monophyletic clades (Fig. 7B). The most ancestral sister groups are the Brazilian and Laran clades. The most recently diverged group is represented by the trans-Andean Clade, recently diverged from the cis-Andean Clade. Analyses of Isozyme Variation versus Phylogenetic Analysis of Isozymes. Different topologies were obtained from the phylogenetic analyses performed (CONTML and KITSCH). However, similar conclusions regarding the relationships among populations and their taxonomic status can be drawn from the two trees illustrated in Fig. 7. Both analyses revealed four clades that may be interpreted as phylogenetic species. The phenogram, based on unweighted pair-group method with arithmetic average cluster analysis of genetic distance (Fig. 5) clearly resolved the same groups, although it further split the cis-Andean Clade into two groups (Colombian and Venezuelan). Discussion mtDNA and isozyme variation were used as complementary tools for the identiÞcation of species

626


boundaries among members of the Lu. longipalpis species complex. Results using the two approaches were largely concordant. We recognized four cryptic species in what we considered the Lu. longipalpis species complex: Species A, the Brazilian Clade, is represented by 11 populations sampled throughout Brazil; Species B, the Laran Clade, is represented by populations from El Paso and Curaruiga (La Rinconada) in northwest Venezuela; Species C, the cis-Andean Clade, is represented by populations in Colombia (Bucaramanga, Palo Gordo, Neiva, Durania), Venezuela (Trujillo and Cojedes), and northern Brazil (Pacaraima); and Species D, the trans-Andean Clade is represented by 11 populations from various parts of Central America. The distribution of our species A includes the areas from which Lu. longipalpis was originally described and should be considered to represent Lutzomyia longipalpis sensu stricto. Our species B has been formally described as a new species, Lutzomyia pseudolongipalpis Arrivillaga and Feliciangeli (Arrivillaga and Feliciangeli 2001). Formal descriptions including assignment of new scientiÞc names are currently underway for species C and D. There are some inconsistencies in the outcome of our analyses. Phenetic analysis (unweighted pairgroup method with arithmetic average) based on genetic distance (DNei) computed from allozyme frequencies grouped the 30 populations into Þve rather than four groups (Fig. 5). The cis-Andean Clade, species C was split. We designated species C1 to include specimens collected from Bucaramanga, Palo Gordo, Durania, and Neiva in Colombia and species C2 included populations from Trujillo and Cojedes in Venezula 5). This arrangement was based largely on three discriminating markers. In the isozyme-based phylogenetic analysis the differences because of autapomorphic characters, such as MDH- two and HK, provide no phylogenetic information among populations. The GPI character is a synapomorphy and is the most informative character in the MP analysis (data not shown). Consequently the phylogenetic approach resolved four groups, rather than Þve (Fig. 7B). However, examination of Fig. 7B revealed that C1 and C2 are split and their placement indicates that the two groups were recently diverged. This arrangement is not supported by the COI mitochondrial sequence data. For the present, we are suggesting that four species, rather than Þve, be considered until this discrepancy is resolved. The mitochondrial phylogeny indicates spatial overlap in the distributions of species C and D that is not evident in the analyses based on allozymes. Acknowledgments In Brazil, we thank Italo Sherlock, of FIOCRUZ, Salvador; Raimundo Nonato de Souza and Irineu Guerreiro de Souza of FIOCRUZ, Fortaleza; Irolanda da Rocha Barata, Evaldo Carneiro das Chagas and Lourdes Maria Garcez Silveira of Instituto Evandro Chagas, Belem; Elizabeth Rangel, Claudio Meneses and Diamar C. Pinto of the Entomology Department FIOCRUZ, Rio de Janeiro for Þeld and technical as-

Vol. 40, no. 5

sistance. We are greatly indebted to the authorities at FIOCRUZ in Salvador, Fortaleza, and Rio de Janeiro as well as the Instituto Evandro Chagas, Belem for providing us with transportation. We are grateful to Francisco Freitas, Herberto Carvalho Dantas Filho, Leonidas Monte Batista, Manoel Caetano Duarte, and Osvaldo Carlos Silva da Silveira for their Þeld assistance in Santarem. We also thank Paulo Calaç a and Evangley Rios for Þeld assistance in Montes Claros. In Venezuela we thank Milagros Oviedo of Universidad de Los Andes for the Þeld collaboration in Trujillo localities, Freddy Arias, Arturo Bravo, Maria Martinez, Florencio Mendoza, and Carlos Mendez of the Centro Nacional de Referencia de Taxonomõá de Flebo´ tomos for technical assistance, Yadira Rangel for assistance with isozyme analysis and Juan Carlos Navarro for recommendations for the phylogenetic analyses. In Costa Rica we thank Marco Herrero and Julio Rojas of Universidad Nacional in Heredia; in Honduras we thank Carlos Ponce of the Central Laboratory, Ministry of Health, Tegucigalpa. In Nicaragua we would like to acknowledge the assistance of Alejandro Belli and Sonia Valle of the Centro Nacional de Diagnostico y Referencia, Ministero de Salud, Managua. Finally, we thank Bill Sweeney and Scott Marriott of UTMB, Galveston, for technical support. This research was supported, in part, by a grant from the John D. and Catherine T. MacArthur Foundation program entitled “Molecular Biology of Parasite Vectors,” by Grant No. AI39540 from the National Institutes of Health to G.C.L and by funds from Conselho Nacional de Desenvolvimento Cientõ´Þco e Tecnolo´ gico (CNPq), Fundaç aõ Nacional de Sau´ de (FUNASA) to Bruce Alexander and a fellowship from TDR-WHO to J. C. A. Travel to Roraima, Brazil was funded in part by Fundaç aõ Instituto Oswaldo Cruz

References Cited Arrivillaga, J. C., and M. D. Feliciangeli. 2000. Diferenciacio´ n morfolo´ gica larval entre poblaciones de Lutzomyia longipalpis (Lutz & Neiva, 1912) de Venezuela (Diptera: Psychodidae. Bol. Entomol. Venez. 15: 229 Ð234. Arrivillaga, J. C., and M. D. Feliciangeli. 2001. Lutzomyia pseudolongipalpis: the Þrst new species within the longipalpis (Diptera: Psychodidae: Phlebotominae) complex from La Rinconada, Curarigua, Lara State, Venezuela. J. Med. Entomol. 38: 783Ð790. Arrivillaga, J. C., J. C. Navarro, and M. D. Feliciangeli. 1999a. Morfologia y quetotaxia del tagma cefalico de Lutzomyia (Diptera: Psychodidae). Bol. Entomol. Venez. 14: 1Ð12. Arrivillaga, J. C., Y. Rangel, M. D. Feliciangeli, and G. Lanzaro. 1999b. Morphogenetic studies of Lutzomyia longipalpis (Psychodidae: Phlebotominae). J. Am. Mosq. Contro. Assoc. 15: 410 Ð 411. Arrivillaga, J. C., Y. Rangel, M. Oviedo, and M. D. Feliciangeli. 2000a. Genetic Divergence among Venezuelan populations of Lutzomyia longipalpis (Dipteras: Psychodidae: Phlebotominae). J. Med. Entomol. 37: 325Ð330. Arrivillaga, J. C., Y. Rangel, M. Oviedo, and M. D. Feliciangeli. 2000b. Correlated morphological and genetic diversity among Lutzomyia longipalpis (Diptera: Psychodidae) collections in Venezuela. J. Am. Mosq. Cont. Assoc. 16: 171Ð174. Belli, A., D. Garcia, X. Palacios, et al. 1999. Widespread atypical cutaneous leishmaniasis caused by Leishmani (L.) chagasi in Nicaragua. Am. J. Trop. Med. Hyg. 61: 380 Ð385. Bender, W., P. Spierer, and D. S. Hogness. 1983. Chromosomal walking and jumping to isolate DNA from the ace and rosy loci and the bithorax complex in Drosophila melanogaster. J. Mol. Bio. 168: 17Ð33.

September 2003


Bonnefoy, S. M., F. Tibayrenc, J. P. Le Pont, P. Dujardin, P. Desjeux, and F. J. Ayala. 1986. An isozymic study of Lutzomyia longipalpis (Diptera: Psychodidae), the vector of visceral leishmaniasis in the “ Yungas” (Bolivia). Cah. O.R.S.T. O. M. Ser. Entomol. Med. Parasitol. 24: 213Ð217. Carpenter, J. M. 1996. Uninformative bootstrapping. Cladistics 12: 177Ð181. Farris, J. S. 1969. A successive approximations approach to character weighting. Syst. Zool. 26: 269 Ð276. Farris, J. S. 1983. The logical basis of phylogenetic analysis, pp. 1Ð36. In N. I. Platnick and V. A. Funk [eds.], Advances in Cladistics. Columbia University Press, New York. Felsenstein, J. 1995. Phylip 3.5c version. Phylogenetic inference package䉷. Copyright 1986 Ð1995 by Joseph Felsenstein and the University of Washington. Franç a, C. 1920. Observations sur le genre Phlebotomus. II.ÑPhlebotomus du Nouveau Monde (Phlebotomes du Bresil et du Paraguay). Bull. Soc. Portug. Sci. Nat. 8: 215Ð236. Goloboff, P. A. 1993. Estimating character weights during tree search. Cladistics 9: 83Ð91. Hamilton, J.G.C., R. D. Ward, M. J. Dougherty, et al. 1996. Comparison of the sex-pheromone components of Lutzomyia longipalpis from areas of visceral and atypical cutaneous leishmaniasis in Honduras and Costa Rica. Ann. Trop. Med. Parasitol. 90: 533Ð541. Higgins, D. G., A. J. Bleasby, and R. Fuchs. 1992. Clustal V improved software for multiple sequence alignment. Comp. Appl. Biol. 8: 189 Ð191. Kambhampati, S., and P. T. Smith. 1995. PCR primers for the ampliÞcation of four insect mitochondrial gene fragments. Insect. Mol. Biol. 4: 223Ð236. Lampo, M., D. Torgerson, L. M. Marquez, M. Rinaldi, C. Z. Garcia, and A. Arab. 1999. Occurence of sibling species of Lutzomyia longipalpis (Diptera: Psychodidae) in Venezuela: Þrst evidence from reproductively oslated sympatric populations. Am. J. Trop. Med. Hyg. 61: 1004 Ð1009. Lanzaro, G. C., K. Ostrovska, M. Herrero, P. G. Lawyer, and A. Warburg. 1993. Lutzomyia longipalpis is a species complex: Genetic divergence and interspeciÞc hybrid sterility among three populations. Am. J. Trop. Med. Hyg. 48: 839 Ð 847. Lanzaro, G. C., and A. Warburg. 1995. Genetic variability in phlebotomine sandßies: possible implications for leishmaniasis epidemiology. Parasitol. Today 11: 151Ð154. Lanzaro, G. C., B. Alexander, J. P. Mutebi, J. Montoya-Lerma J. Montoya-Lerma, and Warburg, A. 1997. Genetic variation among natural and laboratory colony populations of Lutzomyia longipalpis from Colombia. Mem. Oswaldo Cruz. 93: 65Ð70. Mukhopadhyay, J. K., K. Ghosh, A. Azevedo, E. Rangel, and L. E. Munstermann. 1998. Genetic polymorphism of morphological and biochemical characters in Natal Brazil populations of Lutzomyia longipalpis (Diptera: Psychodidae). J. Am. Mosq. Cont. Assoc. 14: 277Ð282. Munstermann, L.E.M., A. C. Morrison, C. Ferro, R. Pardo, and M. Torres. 1998. Genetic structure of local populations of Lutzomyia longipalpis (Diptera: Psychodidae) in Central Colombia. J. Med. Entomol. 35: 82Ð 89. Mutebi, J. P., E. Rowton, M. Herrero, et al. 1998. Genetic variability among populations of the sandßy Lutzomyia (Lutzomyia) longipalpis (Diptera: Psychodidae) from Central America. J. Med. Entomol. 35(2): 169 Ð174. Mutebi, J. P., B. Alexander, I. Sherlock, et al. 1999. Breeding structure of the sand ßy Lutzomyia longipalpis (Lutz & Neiva) in Brazil. Am. J. Trop. Med. Hyg. 61: 149 Ð157.

627

Nei, M. 1978. Estimation of average heterozygosity and genetic distance from a small number of individuals. Genetics 89: 583Ð590. Ponce, C., E. Ponce, A. Morrison, et al. 1991. Leishmania donovani chagasi: new clinical variant of cutaneous leishmaniasis in Honduras. Lancet 33: 67Ð70. Saito, N., and M. Nei. 1997. The neighbor-joining method a new method for reconstructing phylogenetic trees. Mol. Bio. Evol. 4: 406 Ð 425. Simon, C., F. Frati, A. Beckenbach, B. Crespi, H. Liu, and P. Flook. 1994. Evolution weighting and phylogenetic utility of mitochondrial gene sequences and compilation of conserved polymerase chain reaction primers. Ann. Entomol. Soc. Am. 87: 651Ð700. Sneath, P.H.A., and R. R. Sokal. 1973. Numerical taxonomy. W. H. Freeman, San Francisco. Steiner, W.W.M., and D. J. Joslyn. 1979. Electrophoretic techniques for the genetic study of mosquitoes. Mosq. News 39: 35Ð54. Swofford, D. L. 1999. Phylogenetic Analysis Using Parsimony (PAUP), version 4.0b. 2. Sinauer, Inc. Suderland, MA. Swofford, D. L., and R. B. Selander. 1989. BIOSYS-1. A computer program for the analysis of allelic variation in population genetics and biochemical systematics. Illinois Natural History Survey, Urbana, IL. Warburg, A., E. Saravia, G. C. Lanzaro, R. G. Titus, and F. Neva. 1994. Saliva of Lutzomyia longipalpis sibling species differs in its composition and capacity to enhance leishmaniasis. Phil. Trans. R. Soc. Lond. B. 345: 223Ð270. Ward, R. D., A. L. Ribeiro, P. D. Ready, and A. Murtagh. 1983. Reproductive isolation between different forms of Lutzomyia longipalpis (Lutz & Neiva) (Diptera: Psychodidae), the vector of Leishmania donovani chagasi Cunha & Chagas and its signiÞcance to kala-azar distribution in South America. Mem. Inst. Oswaldo Cruz. 78: 269 Ð280. Ward, R. D., A. Phillips, B. Burnet, and C. Marcondes. 1988. The Lutzomyia longipalpis complex reproduction and distribution, pp. 257Ð269. In Biosystematics of Haematophagous Insects. Service. M. W., 37th edition. Clarendon Press, Oxford, England. World Health Organization. 1990. Leishmaniasis Control. Technical report series No. 793. WHO. Geneva. Williams, P. 1999. Patterns in the geographical distribution of members of the genus Lutzomyia Franç a (Diptera: Psychodidae-Phlebotominae), pp. 455Ð502. In J. F. Burger [ed.], Contributions to the Knowledge of Diptera. A Collection of Articles on Diptera Commemorating the Life and Work of Graham B. Fairchild. Mem. Entomol., International 14, Associated Publishers, Gainesville, FL. Yin, H., Norris, D. E., and Lanzaro, G. C. 2000. Sibling species in the Lutzomyia longipalpis complex differ in levels of mRNA expression for the salivary peptide, maxadilan. Insect Mol. Biol. 9: 309 Ð314. Young, D. M., and M. A. Duncan. 1994. Guide to the identiÞcation and geographic distribution of Lutzomyia sandßies in Mexico, The West Indies, Central and South America (Diptera: Psychodidae) Mem Am. Entomol. Inst. No. 54. Associated Publishers, Gainesville. FL. Zeledon, R., H. Hidalgo, A. Viquez, and A. Urbina. 1989. Atypical cutaneous leishmaniasis in a semiarid region of north-west Costa Rica. Trans. R. Soc. Trop. Med. Hyg. 83: 789. Received for publication 5 September 2001; accepted 5 December 2002.