| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
 |
ABSTRACT |
 TOP ABSTRACT INTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
|
INTRODUCTION |
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
With the important exception of Lu. pseudolongipalpis (Arrivillaga & Feliciangeli 2001),15 morphological characters have not consistently provided useful characteristics for distinguishing Lu. longipalpis sibling species. In this context, Mangabeira first reported that Lu. longipalpis (Phlebotomus longipalpis) was either highly polymorphic or more than one species.16 He also proposed that the north and northeastern Brazilian populations of Lu. longipalpis that differ in the number of pairs of pale patches (spots) on the abdomen of males could be a separate species.16 However, it has been subsequently demonstrated that some populations whose males were characterized by different spot morphology could interbreed while Lu. longipalpis from different areas but with similar spot phenotype did not always interbreed.17 In Lu. longipalpis, the spots are the visual manifestation of glandular areas underlying the paler cuticle surface that release sex pheromones that have been shown to influence the sexual behavior of Lu. longipalpis.18,19 Four male sex pheromone chemotypes have been identified in Brazilian populations: (S)-9-methyl-germacrene-B (9MGB), 3-methyl-
-himachalene (3MAH) together with -himachalene (AH), cembrene diterpene (CEMB), and a fourth cembrene isomer.19–24 Male copulation/courtship songs are strikingly distinct between populations in which males have different pheromone chemotypes.18,25,26 Furthermore, populations with distinct pheromone types and copulation love songs have shown sequence differences for the speciation27 period gene.28,29 Since love songs and male attractant pheromones are used in maintaining speciation by providing pre-zygotic mating barriers,27 their genes are likely to be significant components in the genetic structuring of the Lu. longipalpis group.
Genetic evidence for a Lu. longipalpis species complex has been reported over large distances, usually between countries, but the taxonomic status of Lu. longipalpis within a particular geographic region remains more controversial.15,27–42 Among Venezuelan populations, differences in genetic and morphological characteristics have resulted in the first formal description of a new species from the Lu. longipalpis species complex, Lu. pseudolongipalpis Arrivillaga & Feliciangeli 2001.15 In contrast, variation at allozyme loci between populations of Lu. longipalpis from different areas within Brazil has been interpreted as consistent with a single but genetically heterogeneous species.30,32,35–40 More recently, the distribution of mitochondrial DNA haplotypes in Central and South American Lu. longipalpis led the investigators to conclude that although there are several Lutzomyia clades, there is probably only a single, recently established Brazilian clade.38,39 The results of these studies conflict with those of breeding experiments,17,19 pheromone analyses,19–23 and those on the pattern of genetic variation in the periodicity and cacophony genes between Brazilian populations of Lu. longipalpis, which together have provided convincing evidence for the presence of separate clades in Brazil.28,29,41–43 Furthermore, for at least one Brazilian location in the northeastern state of Ceará (Sobral), there is cumulative unambiguous evidence (lack of cross-mating in the laboratory, distinct pheromones and copulation songs, genetic variation) for reproductive isolation between two sympatric populations that can best be described as bona fide species,17,19,29,42 although the contemporary level of gene flow needs to be reassessed to ascertain whether speciation is relatively recent.43 Part of the apparent misunderstanding about the phylogeographic structure of this taxon appears to have arisen from differential mutation rates among the genetic markers used and from studies largely conducted without considering pre- and post-cross-mating barriers.
Accurate epidemiologic data is required for efficient control of any disease. Therefore, establishing correct taxonomic distinctions between putative sibling species of Lu. longipalpis is likely to contribute to the basic foundation upon which a successful management strategy to control New World visceral leishmaniasis may be built. Furthermore, the degree of gene flow between Lu. longipalpis siblings is likely to affect the spread of insecticide resistance genes, as has been suggested for malaria associated mosquitoes in sub-Saharan Africa.44
The objective of this study was to use variation at micro-satellite loci to investigate genetic differences between populations of Lu. longipalpis from South America (Venezuela and Brazil), with particular emphasis on the male sex pheromone type. We have also included samples from two taxonomically recognized species to provide a context to the pattern and strength of genetic differences between Lu. longipalpis siblings from Brazil and Venezuela with contrasting sex pheromones.
|
MATERIALS AND METHODS |
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
and Figure 1); three of these localities (Guayabita, Lu. pseudolongipalpis at La Rinconada, Curarigua, and Sobral) were sampled twice. The Brazilian populations of Jacobina, Lapinha, Marajó, Natal, and Lu. cruzi were laboratory-established colonies derived from isofemale lines from females collected at the sites indicated in Table 1 and reared in the laboratory as described.45 The Jacobina colony was used at approximately 120 generations whereas the Lapinha, Marajó, Natal, and Lu. cruzi colonies were 100, 20, 16, and 3 generations, respectively.
|
|
). The Venezuelan Lu. longipalpis were collected using Centers of Disease Control (CDC ) light traps on December 9, 1997 and again on July 21, 2001 at Guayabita in Aragua State; on July 12, 2001 in El Layero and August 1, 2001 in Altagracia in Guárico State; and on January 18–25, 2001 in Las Cabreras on Margarita Island in Nueva Esparta State. Lutzomyia pseudolongipalpis Arrivillaga and Feliciangeli 2001 samples included males caught in CDC light traps on December 11, 1997 in La Rinconada, Curarigua in Lara State, and a smaller field collection of females collected at the same original site using CDC light traps in October 2000. The pheromone type at each location has been identified for all but two of the samples (Altagracia and Las Cabreras).
Sand fly specimens were preserved in 98% (v/v) ethanol until DNA extraction.
Genotyping of sand flies. DNA was extracted from decapitated sand flies, homogenized in 100 µL of lysis buffer (80 mM NaCl, 172 mM sucrose, 130 mM Tris-HCl, pH 8.0, 50 mM EDTA, 0.5% [w/v] sodium dodecyl sulfate), followed by incubation at 65°C for 30 minutes and protein precipitation with potassium acetate.46,47 The DNA was precipitated with ethanol and resuspended in 30 µL of sterile water. One microliter of DNA was used for a polymerase chain reaction (PCR) in a 10-µL reaction volume that contained 75 mM Tris-HCl, pH 8.9, 20 mM (NH4)2SO4, 0.01% [v/v] Tween 20, 0.2 mM of each dNTP, 3.0 mM MgCl2, 100 ng of each primer, and 0.25 units of Taq polymerase (ABgene, Epsom, United Kingdom). We genotyped all sand flies at five dinucleotide microsatellite loci, LIST6-002, LIST6-004, LIST6-006, LIST6-012, and LIST6-029, which were isolated from the genome of a Lu. longipalpis colony derived from Jacobina, Brazil.47 Thermal cycling conditions of loci are described elsewhere.47 For allele visualization, the forward primers flanking the microsatellite loci LIST6-002, LIST6-004, LIST6-006, and LIST6-012 were 5' labeled (Invitrogen Ltd., Paisley, United Kingdom) with D2, D4, D4, and D3 fluorescent dyes, respectively. The forward primer flanking LIST6-029 was 5' labeled with IRD700 (MWG Biotech UK Ltd., Inc., Milton Keynes, United Kingdom). The PCR products from LIST6-002, LIST6-004, LIST6-006, and LIST6-012 were pooled along with a 400-basepair (D1) size standard (Beckman Coulter, Fullerton, CA) and separated by capillary electrophoresis through a denaturing polyacrylamide gel on a Ceq2000XL automated sequencer (Beckman Coulter). Alleles were sized using the cubic model of analysis in the Ceq2000XL fragment analysis software (Beckman Coulter). Genotypes at the LIST6-029 locus were obtained by separating alleles on a 6% (w/v) denaturing polyacrylamide gel on a LiCor4200 apparatus (MWG Biotech UK Ltd.) and sized alongside a microSTEP~24a ladder (Microzone Ltd., Haywards Heath, United Kingdom).
Data analysis. Deviation from genotypic linkage equilibrium among all locus-pair combinations was estimated.48 Observed and expected genotype frequencies at each locus were tested for deviation from Hardy-Weinberg equilibrium (HWE) conditions for every sample (different sample dates from the same location were tested separately) using the HWE test available in ARLEQUIN VERSION 2.001.49 Genetic diversity was characterized from the number of alleles (Na), observed heterozygosity (Ho), expected heterozygosity (He), and Wright’s inbreeding coefficient (f ).50 All statistics were calculated for each locus and each sample separately, and also across all loci for Na and f. Na, Ho, and He were calculated using the program POPULATIONS version 1.2.28.51
The extent of genetic differentiation between samples was investigated by calculating fixation indices based on an infinite allele model because allele sizes were not consistent with a stepwise mutational model and FST-based approaches are considered to be more robust when less than 20 loci are used.52 Genetic differentiation between samples was assessed from the unbiased estimator (
) of Weir and Cockerham53 of the FST of Wright50 calculated using ARLEQUIN version 2.001.49 The significance of the estimates of genetic differentiation among samples was assessed with 10,000 permutations of genotypes. We first estimated the extent of genetic differences between sand flies collected from the same location but on different sampling dates, and then estimated pairwise genetic differentiation between all locations but with samples from the same location and different collection dates pooled (to increase the sample size at each site).
Isolation by distance genetic structure was examined from the correlation of pairwise estimates of genetic differentiation between samples (samples from the same location but different sample dates were treated separately) against the corresponding logarithms of geographic distance (log km) separating the populations.54 ARLEQUIN version 2.00149 was used for this test, with genetic distance defined as /(1 – ).55 A Mantel test (10,000 permutations) was used to assess the significance of any correlation between the genetic and geographic distances using the permutation procedure implemented by ARLEQUIN version 2.001.49 Partial correlations were also undertaken for the 9MGB and the CEMB samples separately (there were too few 3MAH samples for a meaningful analysis). A neighbor-joining tree55 of pairwise estimates of between pooled samples was constructed using POPULATIONS version 1.2.28.51 The robustness of the tree was evaluated from 500 bootstrap resamplings across individuals.
|
RESULTS |
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
) yielded 89 alleles across all 5 microsatellite loci. Summary characteristics for all samples and microsatellite loci are provided in Tables 1 and 2, respectively. The number of alleles per locus varied from 11 at LIST6-029 to 23 at LIST6-004 over all individuals, with between 1 and 13 alleles per sampled population. The mean number of alleles per sampled population and over all loci ranged between 2.2 and 9.2, although it was never greater than 6 in any colony-raised sample (Table 2). The number of alleles was correlated with sample size (Pearson product-moment correlation [
], r = 0.631, degrees of freedom [df] = 15, P = 0.009) and this relationship was significant only for the wild-caught sand fly samples ( wild-caught, r = 0.771, df = 10, P = 0.005; colony r = 0.677, df = 4, P = 0.209) when the samples were analyzed separately according to their origin. Despite the apparent effects of sample size and origin upon the number of alleles, there were no differences overall in either the number of alleles, expected heterozygosity, or observed heterozygosity at any locus between wild-caught and colony sand fly samples (P > 0.05 for all tests, by Mann-Whitney U test for all tests) except for Ho at LIST6-029 (P = 0.002, by Mann-Whitney U test).
|
= 0.05, k = 10), which for all but one case (Jacobina LIST6-002) was due to a deficit of heterozygotes (Table 2). At least one significant heterozygote deficit was observed at each locus, although almost half (10 of 23) of the deviations from HWE were observed at LIST6-029. Three (Lu. pseudolongipalpis 1997, Lapinha, and Lu. cruzi) of the 16 samples demonstrated no deviations from expected HWE conditions (Table 2). The inbreeding coefficient varied considerably from –0.471 to 0.737 at individual microsatellite loci and between –0.048 and 0.397 over all loci (Table 2). Linkage disequilibrium between loci was significant (P < 0.05) for 18 out of the 145 possible pairwise tests (15 tests were not made because loci were monomorphic), although only 8 of these comparisons were significant (P < 0.05) after a within-sample (k = 10) sequential Bonferroni correction.56 The Natal sample, demonstrated a relatively disproportionate amount of linkage disequilibrium (3 significant locus comparisons, data not shown), but because the majority of samples were in linkage equilibrium, all loci were retained for the data analyses.
Temporal genetic differentiation.
Only Lu. pseudolongipalpis showed significant genetic differences between samples collected on different sampling dates (Sobral CEMB, = 0.0018, P = 0.472; Sobral 9MGB, = 0.0049, P = 0.281; Guayabita, = 0.0219, P = 0.052; Lu. pseudolongipalpis, = 0.160, P < 0.001). Since this difference was likely to be a statistical effect of the small number of individuals collected from La Rinconada, Curarigua (see Discussion), genotypes from the same locality were pooled for all samples for the analyses of spatial genetic differentiation.
Spatial genetic differentiation.
Genetic differentiation between samples was generally high, with values ranging from 0.024 (Las Cabreras-El Layero) to 0.558 (Marajó Lu. pseudolongipalpis); the pairwise estimates of were significant (P < 0.05) for all comparisons. All estimates of (P < 0.05) involving Lu. cruzi and Lu. pseudolongipalpis were greater than 0.176 and 0.388, respectively. Genetic differentiation between the two samples with different sex pheromones at Sobral was as great (or higher) as many of the comparisons between samples that are separated by considerable distances (Table 3).
|
/1 – ) and the geographic distance (log km) between sample localities (r = 0.31) for all samples, its effect was rather weak with distance, accounting for less than 10% of the variation in genetic differences among samples (Figure 2a and Table 4). This relatively explanatory capability of geographic distance is a result of large genetic differences between samples with different pheromones at small and intermediate geographic distances (Figure 2). There was a marked increase in the strength of isolation by distance genetic structure when the samples with CEMB and 9MGB pheromones were analyzed separately, with geographic separation accounting for 97% and 54%, respectively, of the genetic variation between samples (Table 4). The correlation between geographic distance and genetic differentiation was significant only for the 9MGB samples (P = 0.0028) after a sequential Bonferroni correction ( = 0.05, k = 3), most likely because the small numbers of CEMB samples available to examine this relationship decreased the statistical power of this test.
|
|
showed five groups: Lu. pseudolongipalpis by itself (3MAH); a well-supported (mostly 9MGB) clade containing Altagracia, Guayabita, El Layero, and Las Cabreras from Venezuela; Lu. cruzi by itself (a 9MGB chemotype); a well-supported and distinct CEMB clade that consists of the Marajó, Sobral (CEMB), and Natal samples; and a group containing the remaining samples (Sobral 9MGB, Jacobina 3MAH plus AH, and Lapinha MGB) from Brazil. Phylogeographic trees based upon other genetic distance metrics produced identical tree topologies. |
DISCUSSION |
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
Microsatellite loci have become the mainstay of modern population genetics,57 but it is relevant that two processes may reduce genetic variability and therefore affect the relationships between our samples. First, longer (and typically more polymorphic) alleles may be restricted to the focal species (used to isolate the microsatellites).58 We found no evidence for this ascertainment bias because although there was significant variation (different sampling dates were pooled) in allele length among samples (df = 11, P < 0.001 for all loci, by Kruskal-Wallis test), the size ranges of alleles were significantly longer in the Jacobina sample at only two loci and generally overlapped between most samples. Second, this study like many others used sand flies from experimental colonies that may have lost genetic variation through a founder effect. The influence of colonization upon genetic variability and its potential effect upon understanding the genetic relationships of Lu. longipalpis has been questioned, with genetic differences noted for some59 but not all60,61 comparisons between field and colony samples. We could not explicitly test for a loss of variation (there are too few loci for a meaningful bottleneck analysis), but noted that although we observed fewer alleles in the laboratory populations, there were no statistically robust differences in genetic variability between field and colony samples. Furthermore, although Mukhopadhyay and others59 reported a reduction in genetic variability in laboratory-reared Lu. longipalpis from Lapinha compared with its field counterpart, the two sampled populations clustered together but separately from other Brazilian populations in a genetic distance-based dendogram.59 We believe that although there was undoubtedly a reduction of genetic variation in the colonized material, relative to the field sampled sand fly populations, the magnitude of this reduction did not appreciably affect the phylogeographic substructuring shown in Figure 2. A similar phylogenetic tree to that shown in Figure 3 was obtained when the two long-term colony populations (Lapinha and Jacobina) were excluded from the analysis. This indicates that the effect of colonization was negligible in this instance. However, the results presented here would be strengthened by the use of more loci and additional field-caught sample sets.
|
is due to differences in allele frequency (probably because of small sample sizes and sampling error) at a single locus (LIST6-002), rather than different alleles per se. Furthermore, the phylogenetic structure based upon samples from individual years was identical to that with pooled groups and displayed 100% bootstrap support for the two Lu. pseudolongipalpis samples on their own distinct branch. Although fine-scale genetic differences between temporal collections of the same population might exist, the genetic stability found in this study allowed us to increase sample size through pooling collections made on different dates in the same location.
Geographic distance between Lu. longipalpis populations was a relatively poor predictor of the level of genetic differentiation among populations when all samples were considered regardless of male sex pheromone chemotype (Figure 2a and Table 4). We note that not only are there genetic differences between Brazilian Lu. longipalpis per se, but that the type of sex pheromone shapes the genetic structure of this sand fly taxon. This conclusion is apparent from the contrasting patterns of isolation by distance with and without regard to the sex pheromone (Figure 2) and the phylogeographic clustering by pheromone, particularly the striking separation of the sympatric Sobral samples between two Brazilian species or clades that are based on different chemotypes (Figure 3). Three independent studies29,42,43 showed strong genetic differentiation between sympatric populations with 9MGB and CEMB pheromones at Sobral, although the magnitude of these differences in one of the studies was lower at Sobral than among other members of the Lutzomyia species complex.43 Although two previous reports described genetic differentiation between a sample from Sobral and other Brazilian populations, because neither study distinguished between pheromone types, it is difficult to interpret their data with respect to speciation processes relevant to Lu. longipalpis.31,61
The smaller genetic differences among Venezuelan samples (except Lu. pseudolongipalpis) compared with those from Brazil reflects the variation in geographic scale between samples (Table 3 and Figure 3). To what extent the genetic divergence between Venezuelan samples represents restricted gene flow or the presence of sibling species is unclear without breeding experiments. All sex pheromone-typed Venezuelan populations (with the exception of Lu. pseudolongipalpis) were of the wide spread 9MGB type (also in Honduras, Guatemala, Colombia, and Bolivia, as well as the Brazilian Lapinha population).19,62,63 The collection sites for the Venezuelan study collections (El Layero, Las Cabreras, Guayabita, and La Rinconada in Venezuela) were located in tropical dry forest type of vegetation. The role of the Chivacoa Gap in isolating the Laran Clade (Lu. pseudolongipalpis) from other Lu. longipalpis populations has been previously discussed,38 although isolation is not complete since Lu. pseudolongipalpis (3MAH) was found in sympatry with a 9MGB population at El Paso, only 8 km from La Rinconada, Curarigua.33 The Brazilian CEMB populations were located in drier sites along the northeastern region (Marajó and S
o Luis Islands, Natal, Joo Pessoa, Las Jaibas-Maceió, Alagoas, and Santarem).19,24,63–65 The Jacobina population (the only known 3MAH/AH population) lies in a narrow valley between two mountain ranges east of the Chapada Diamantina Ridge and well separated from the nearest sex pheromone-typed populations. However, at Sobral (a dry caatinga region in northeastern Brazil), two genetically differentiated 9MGB and CEMB populations share the same ecologic habitat.19,42 It would be interesting to estimate the relative contribution of pre-mating and post-mating barriers to gene flow between these sympatric populations.
This albeit incomplete contemporary geographic distribution of the pheromone chemotypes and the weak correlation between genetic and geographic distances described begs an explanation. The simplest one, based on ecologic distribution, would assume 9MGB as the ancestral chemotype in Lu. longipalpis across the savannahs of South America, followed by subsequent speciation to either 3MAH/AH or CEMB. Forest refuges arose in the dry climatic periods of the cyclical oscillations of cool versus warm ~100,000-year intervals characteristic of the Pleistocene and Recent epochs of the Quaternary period and have been shown to foster increased rates of speciation and biodiversity.66
Given the size of Brazil, the presence of sibling species in this region should not be surprising. However, the controversy regarding the taxonomy of Lu. longipalpis generally results from the pattern of genetic differentiation among Brazilian samples, which has been interpreted as representing members of the same biologic species or clade.12,30,32,35–40 However, genetic divergence in the per intron and in the cac IVS6 intron among the populations from Jacobina, Lapinha, and Natal supports the occurrence of sibling species or near species in Brazil.28,29,41,42 Overall, the pattern and magnitude of genetic differences between Brazilian sand fly samples reported here is consistent with other studies, importantly including male songs and sex pheromones that indicate that Lu. longipalpis exists as four or five cryptic species or clades within Brazil. These species (siblings, clades, or races) appear to be sibling 1, cembrene pheromone, and burst-type copulation song in northeastern Natal, Marajó, and Sobral CEMB populations; sibling 2, 9MGB pheromone and song pattern I pulse copulation song, at Lapinha Cave; sibling 3, 9MGB pheromone and song pattern II pulse copulation song at Sobral; sibling 4, (3MAH plus AH) pheromone and polycyclic pulse copulation song at Jacobina; and a possible sibling 5, cembrene isomer pheromone at Jaibas in Minas Gerais State.16–24 The present number of known siblings in the South American subcontinent would be six, including the 3MAH-producing Lu. pseudolongipalpis Arrivillaga & Feliciangeli 2001 from La Rinconada and El Paso in the Lara State of Venezuela.15 Current genetic data have been interpreted as consistent with a single species of Lu. longipalpis in Central America, but given the presence of several cryptic species in Venezuela and Brazil, we believe that these data should be treated with caution.10–14,30–32,67 Our data supports only one (Lu. pseudolongipalpis) of the four (cis and trans Andean, Lu. pseudolongipalpis or Laran, and Brazilian) clades suggested by isoenzyme and mitochondrial analyses.38,39 These analyses were interpreted to support a single clade in Brazil, despite a population collected on the Brazilian side of the Roraima formation that clustered with the trans-Andean clade and not within the Brazilian clade.38,39 In these reports, it was concluded that the large genetic distance between the four clades supports vicariance (separation or division of a group of organisms by a geographic barrier) as the most significant force in their evolution.38,39 As shown in Figure 2, our data for six Brazilian Lu. longipalpis populations contests this conclusion.
Furthermore, most studies have overlooked the distinct CEMB pheromone clade with regard to its effect on the patterns and processes of genetic evolution in Brazilian sand fly populations. For example, an absence of isolation by distance genetic structure has been reported,36,37,40 although three of the populations examined (Lapinha, Jacobina, and Santarém) produced different sex pheromones.19 Where there are sufficient samples, as observed for the 9MGB samples in Figure 3, we demonstrated a stronger pattern of isolation by distance (the Venezuelan and the Brazilian clades) between populations with similar sex pheromones than that observed between all populations. Although genetic differences between Brazilian samples are expected to arise through geographic separation within a particular sex pheromone clade, speciation processes (i.e., pre-zygotic barriers) other than vicariance are clearly shaping evolution within the Lu. longipalpis complex. From the pattern of sample clustering (Figure 3), we suspect that males from Las Cabreras and possibly Altagracia possess 9MGB or a related sex pheromone.
Lutzomyia pseudolongipalpis is as distinct from the main Venezuelan clade of Lu. longipalpis as is Lu. cruzi from the Brazilian Lu. longipalpis samples (Figure 3). Female Lu. cruzi (a 9MGB population) and Lu. longipalpis are indistinguishable from each other, both transmit L. chagasi/infantum to humans, and have been found in sympatry in Mato Grosso do Sul State in Brazil.68,69
Genetic differences between Lu. pseudolongipalpis and other Venezuelan sand fly populations have been reported,33,34 and our data support the recent taxonomic redescripton of Lu. longipalpis from La Rinconada, Curarigua.15 It is significant that Lu. pseudolongipalpis does not cluster with the other 3MAH sample (Jacobina, Brazil). Although this could largely be an effect of geographic separation, genetic differences are also expected from the male sex pheromone chemistry of these samples because in contrast to Lu. pseudolongipalpis, males from the Jacobina Lu. longipalpis population produce the unmethylated -himachalene (AH) in addition to 3MAH (Table 1).62 The Jacobina (3MAH plus AH) population mapped within the Brazilian 9MGB cluster in the phylogenetic tree in Figure 3. This agrees with previous per gene intron analyses, which indicated either introgression or the presence of common ancestral polymorphisms between these two homosesquiterpene-producing populations.28
To conclude, it is important to recognize that genetic differences between allopatric populations are in most instances insufficient evidence for the presence of cryptic speciation. Our immediate aim was not to define the taxonomic status of Lu. longipalpis, but to reinforce the importance of sex pheromone type (and other behavioral traits) with respect to the pattern of genetic differentiation) in Lutzomyia sp., and emphasize that if these traits are ignored then phylogeographic patterns attributed to geography will be confounded by speciation due to putative mate-recognition systems. The identification of candidate mating barriers and their genetic determinants is a prerequisite to progress our understanding of the mechanisms that drive speciation within Lu. longipalpis. Considerable progress in this area is being achieved in the Drosophila genus and in the Anopheles gambiae complex, whose genomes have been sequenced.27,44,70
Investigating the epidemiology of leishmaniasis in the light of these new taxonomic distinctions would seem an appropriate step towards controlling this disease. It is interesting that Brazil has most of the cases of visceral leishmaniasis in the New World and that of the 3,000–4,000 cases of visceral leishmaniasis reported in Brazil annually during 1999–2000, more than 90% were contributed by the northeastern states of Piauí, Maranhão, Bahia, and Ceará,4–6,9 and the northern states of Pará, Rio Grande do Norte, and Roraima,7,8 which is precisely the area occupied by one of the cembrene isomer producing siblings (CEMB clade in Figure 3) of Lu. longipalpis.
Received May 27, 2004. Accepted for publication May 23, 2005.
Acknowledgments: We thank Jose Wellington and Lindenbergh Souza (Fundação Nacional de Saúde, Ceará, Brazil) and Arturo Bravo (BIOMED, Maracay, Venezuela) for technical assistance in the field. We are also grateful to Emma Jones for sand fly decapitations.
Financial support: This study was supported by the Wellcome Trust (London, United Kingdom), Universidad de Carabobo (Maracay, Venezuela) (project CDCH-FCS-2001-005), Coordenação de Aperfeiçoamento de Pessoal de Nivel Superior (Brazil), and Filho de Amparo à Pesquisa do Estado do Rio de Janeiro (Brazil).
* Address correspondence to Rhayza D. C. Maingon, School of Life Sciences, Keele University, Staffordshire, ST5 5BG, United Kingdom. E-mail: r.d.c.maingon{at}biol.keele.ac.uk 
Authors’ addresses: Phillip C. Watts and Stephen J. Kemp, School of Biological Sciences, Liverpool University, Crown Street, Liverpool, L69 7ZB, United Kingdom, Telephone: 44-151-795-4512, Fax: 44-151-795-4408, E-mails: p.c.watts{at}liv.ac.uk and s.j.kemp{at}liv.ac.uk. J. Gordon C. Hamilton, School of Life Sciences, Keele University, Staffordshire, ST5 5BG, United Kingdom, Telephone: 44-178-258-3026, Fax: 44-178-258-3516, E-mail: bia28{at}keele.ac.uk. Richard D. Ward, School of Life Sciences, Keele University, Staffordshire, ST5 5BG, United Kingdom, Telephone: 44-178-258-3417, Fax: 44-178-258-3516, E-mail: bia40{at}keele.ac.uk. Harry A. Noyes, School of Biological Sciences, Liverpool University, Crown Street, Liverpool, L69 7ZB, United Kingdom, Telephone: 44-151-794-7334, Fax: 44-151-795-4408, E-mail: harry{at}liv.ac.uk. Nataly A. Souza, Departamento de Entomologia, Oswaldo Cruz Institute, Av. Brasil 4365, Manguinhos, CEP 21045-900, Rio de Janeiro, Brazil, Telephone: 55-21-598-4320, Fax: 55-21-257-34468, E-mail: souzana{at}ioc.fiocruz.br. M. Dora Feliciangeli, Centro Nacional de Referencia de Flebotomos, BIOMED, Universidad de Carabobo, Nucleo Aragua, Maracay, Apartado 4873, Venezuela, Telephone and Fax: 58-243-242-0559, E-mail: mdora{at}telcel.net.ve. Reginaldo Brazil, Departamento de Bioquimica e Biologia Molecular, Laboratorio de Bioquimica e Fisiologia de Insetos, Av. Brasil 4365, Rio de Janeiro 21 045-900, Brazil, E-mail: rpbrazil{at}ioc.fiocruz.br. Rhayza D. C. Maingon, School of Life Sciences, Keele University, Staffordshire, ST5 5BG, United Kingdom, Telephone: 44-178-258-4219, Fax: 44-178-258-3516, E-mail: r.d.c.maingon{at}biol.keele.ac.uk.
Reprint requests: Rhayza D. C. Maingon, School of Life Sciences, Keele University, Staffordshire, ST5 5BG, United Kingdom.
|
REFERENCES |
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
-himachalene; sex pheromone of Lutzomyia longipalpis (Diptera:Psychodidae) from Jacobina, Brazil. J Chem Ecol 22: 2331–2340.
-himachalene; sex pheromone of Lutzomyia longipalpis (Diptera:Psychodidae) from Jacobina, Brazil. J Chem Ecol 222335–2336.
-himachalene is confirmed, and the relative stereochemistry defined, by synthesis as the sex pheromone of the sandfly Lutzomyia longipalpis from Jacobina, Brazil. Chem Comm 3-Methyl--himachalene; sex pheromone of Lutzomyia longipalpis (Diptera:Psychodidae) from Jacobina 355–356.
| ||||||||||||||||||||
, 
, 
, and • indicate that males release (S)-9-methyl-germacrene-B, cembrene, 3-methyl-
indicate that the type of sex pheromone is unknown.
= cembrene; • = 3-methyl-
,