PMCCPMCCPMCC

Search tips
Search criteria 

Advanced

 
Logo of tropmedLink to Publisher's site
 
Am J Trop Med Hyg. 2010 January; 82(1): 156–164.
PMCID: PMC2803528

Evidence for Pleistocene Population Divergence and Expansion of Anopheles albimanus in Southern Central America

Abstract

The micro-geographic structure of Anopheles albimanus was studied in southern Central America using partial sequences of the mtDNA cytochrome oxidase subunit I gene (COI). Analysis of molecular variance supported significant genetic structure between populations from Costa Rica and western Panama versus those from central-eastern Panama (ΦCT = 0.33), whereas the within group divergence was shallow and statistically insignificant (ΦST = 0.08). Furthermore, a statistical parsimony network depicted three divergent groups of haplotypes that were not evenly distributed across the study area. Our findings are in partial agreement with previous studies, yet they do not support physical barriers to gene flow or contemporary isolation by distance in this region. Instead, three co-occurring groups of An. albimanus may be the result of multiple introductions, most likely caused by historical fragmentation and subsequent secondary contact. In addition, the molecular signature of population expansion of An. albimanus was detected in central-eastern Panama approximately 22,000 years ago (95% confidence interval [CI] 10,183–38,169). We hypothesize that the population structure of An. albimanus, as determined by our COI locus analysis, is the result of late Pleistocene climatic changes in northern South America.

Introduction

Malaria prevalence has decreased overall during the last decade in Central America.1 However, Panama suffered a major outbreak from 2001 to 2005, which represented 60% of the total number of malaria cases reported in the past 35 years.2 The main vector in Panama is Anopheles albimanus; nevertheless, six other Anopheles species involved in malaria transmission elsewhere in the Americas have been recorded in Panama.3 Malaria is currently under control, but an influx of non-immune travelers into endemic areas, the presence of drug-resistant Plasmodium falciparum,4 and the recent discovery of Anopheles darlingi in eastern Panama,5 are expected to worsen the situation. Studies on the population structure of malaria vectors are important to determine the extent of genetic exchange among populations and to predict the spread of genes conferring insecticide resistance.6,7 In theory, focal control interventions against mosquitoes would only be effective if mosquito immigration from outside the target area is low. In addition, such studies may provide key information on lineage divergence, discrete populations, and timing of demographic phenomena, all of which can influence the involvement of vectors in malaria transmission.810

Anopheles albimanus is a malaria vector in the subgenus Nyssorhynchus with considerable ecological adaptability and broad, mostly coastal, geographic distribution in the Neotropics.11 Generally, An. albimanus is crepuscular, zoophilic, exophagic, exophilic, and seasonally abundant.12 However, heterogeneity in environmental factors, host availability, and vector control strategies across the species range may lead to population differentiation. Although several studies have shown that An. albimanus is a single taxon,1316 there is evidence for significant intraspecies variation in host-feeding behavior, Plasmodium susceptibility, longevity, and insecticide resistance.1719 These differences may affect the ability to transmit Plasmodium sp. and the response to vector control strategies. For example, in Belize, An. albimanus is regarded as a secondary vector because it seldom enters human habitations, and prefers to feed on domestic animals.19,20 However, domestic animals are rarely encountered in some areas of Central America, where indigenous people reside and malaria is endemic. In these situations, high numbers of An. albimanus and exposure to mosquito bites caused by poor housing construction may enhance its importance as a vector.3,5

Previous research on the population genetics of An. albimanus using mitochondrial DNA (mtDNA) ND5 sequences found low to moderate differentiation between Central and South American populations, attributed mainly to isolation by distance (IBD).21 In contrast, analysis of microsatellite loci using the same sample localities showed a small effect of distance, but uncovered differences in allele frequencies between Costa Rica and eastern Panama.22 The discrepancy in allele frequencies may be attributed either to the mountain range that crosses Costa Rica and western Panama, which may act as a barrier to gene flow, or to population contraction in Panama caused by intense insecticide use.22 Nonetheless, the analysis of only four populations in Costa Rica and Panama in the latter studies may have produced inaccurate results for the IBD analysis, and demographic history (e.g., stability of the population size) and natural selection (e.g., neutrality), which are known to influence inferences of population structure in Anopheles mosquitoes,8,10 were not explored in these studies.

Intensive chemical and physical larval control (e.g., source reduction) was carried out in Panama during and after the construction of the Panama Canal (1904–1914). Likewise, Panama has used nationwide residual insecticide spraying (IRS) with DDT, propoxur, fenitrothion, deltamethrin, and cyflutrin since 1957 as the main control method for malaria vectors. To date, little is known about the impact of these control strategies on the population structure of An. albimanus and other malaria vectors. The mtDNA has proven to be extremely useful in identifying the population structure and demographic history in malaria vectors because of its lack of recombination and high levels of variability, and it has been used extensively to study anopheline mosquitoes in the Neotropics, allowing for comparison across species and geographic areas.8,2125 In this study, we use partial sequences of the mitochondrial gene, cytochrome oxidase subunit I (COI), and extensive sampling throughout southern Central America, to address the following questions: 1) Is there genetic differentiation in An. albimanus at this micro-geographic scale?; and if so 2) What are the main factors driving this differentiation (e.g., geographic distance, physical barriers to dispersal, natural selection, and/or demographic phenomena)?

Materials and Methods

Mosquito collection.

Females of An. albimanus were collected outdoors by human landing catches and using Centers for Disease Control (CDC) miniature light traps (model 2848, BioQuip Products, Inc., Rancho Dominguez, CA) from 16 localities spanning roughly 1,100 km of a northwestern-southeastern transect across southern Central America (Figure 1; Table 1). The landing catch protocol was reviewed and approved by a Human Ethics Review Board at McGill University, in Canada and the Instituto Commemorativo Gorgas de Estudios de la Salud (ICGES), in Panama City. All mosquitoes were killed with chloroform, identified to species level in the field with the morphological key of Wilkerson and Strickman,26 and stored in desiccant within 24 hours of capture. Samples were hand carried to the United States and stored at −80°C at the Griffin Laboratory until molecular work was performed.

Figure 1.
Geographic distribution of the COI haplotypes in Central America. The letters correspond to the haplotype(s) observed for each locality. The circled numbers correspond to the localities in Table 1, positioned on the map according to the longitude and ...
Table 1
Anopheles albimanus collection localities and haplotype frequency*

Molecular procedures.

Eight to 25 mosquitoes were analyzed from each locality and samples from different days, collectors, and sampling methods were randomly included in the DNA analysis. We followed the procedures used by Mirabello and Conn25 for DNA isolation and polymerase chain reaction (PCR) amplification. We used the primers: UAE3/UAE10 (5′ TATAGCATTCCCACGAATAAATAA-3′/5′ TCCAATGCACTAATCTGCCATATTA-3′)27 to amplify a 1,300-bp fragment of the COI gene. The PCR products were cleaned using CentriSpin 40 columns (Princeton Separations, Freehold, NJ), and sequencing was carried out by the Wadsworth Center Genomics Core Facility (New York State Department of Health). Sequences were unambiguously aligned with Sequencher 3.0 (Gene Codes Corp, Ann Arbor, MI), grouped together by site, and trimmed in PAUP, version 4.0.28 Sequence fragments of 824 bp (only the overlap between the two primers), spanning positions 1873–2698 of the mitochondrial genome of An. gambiae, were used to determine unique haplotypes in MacClade, version 3.0.29 The DNA sequences were checked for stop codons and confirmed as that of COI by translation into amino acid sequences using Anopheles quadrimaculatus mtDNA genetic code in MEGA 4.0.30

Structure of genetic variation.

The program TCS version 1.1231 was used to calculate the number of mutational steps necessary to link any two haplotypes at the 95% confidence level (CI), as determined in ParsProb 1.1.32 A statistical parsimony network represented genealogical relationships among An. albimanus haplotypes, and homoplasy was resolved based on the algorithm estimation rules described in Crandall and Templeton33 and summarized in Uthicke and Benzie.34 Analysis of molecular variance (AMOVA) was carried out in Arlequin 3.11.35 First, we grouped the sequences of An. albimanus as western (1–9) and central–eastern localities (10–16) (Figure 1; Table 1) to test for congruence with previous studies.21,22 Second, we grouped the sequences from localities in the Pacific (1, 2, 3, 8, 9, 10, 11, and 15) and the Atlantic coasts (4, 5, 6, 7, 12, 13, 14, and 16) (Figure 1; Table 1) to test for genetic divergence caused by potential ecological differences (e.g., precipitation and/or tide regimen). Hierarchical AMOVA tests using both haplotype frequencies and the number of mutations between haplotypes (Φ) partitioned the genetic variance within and among populations within groups (ΦSC, ΦST) and among groups (ΦCT).

The Mantel analysis was used to test the null hypothesis of the independence of the geographic and genetic distance by a pairwise matrix of linearized genetic distances, estimated by FST, and ln geographic distance among all populations. We also assessed the effect of the Central American Cordillera on genetic differentiation by testing whether genetic distance between populations on either the Atlantic or the Pacific coast correlates better with geographic distance than with direct-line distances. This procedure was carried out using a third matrix that included a categorical variable in the partial Mantel test.36 The significance of the Mantel test was determined by permutations of N = 10,000 through a non-parametric procedure implemented in the IBD web service version 3.15.37

Demographic history.

Neutrality tests of Tajima's D,38 Fu and Li's F* and D*,39 were used to test for strict neutrality (e.g., the majority of mutations are selectively neutral), whereas Fu's FS40 and R241 were used to test for population size stability.42 These statistics detect deviations from the pattern of polymorphism expected from a neutral model of evolution in a similar fashion, but use the information in the sample differently. Fu's FS40 and R241 are among the most powerful tests to detect population expansion and genetic hitchhiking, whereas Tajima's D,38 Fu and Li's F* and D*,39 are the most effective tests to detect background selection. Neutrality tests were calculated in DnaSP, version 4.50.0243 for each population and for the haplotype groups defined in the parsimony network. Significance of neutrality tests was assessed by 10,000 coalescence simulations in DnaSP, version 4.50.02.43 In addition, haplotype and nucleotide diversities44 were computed in Arlequin 3.1135 for each population and for the groups in the parsimony network.

The mismatch distribution was performed to distinguish between an exponentially growing population depicting a smooth unimodal distribution and a stable population that has reached mutation drift-equilibrium depicting a multimodal, or erratic distribution, caused by stochastic lineage loss.45 To quantify the smoothness of the mismatch distribution,46 we calculated the raggedness (r) statistic in DnaSP, 4.50.02.43 The mismatch distribution was calculated in Arlequin 3.1135 for the entire data set and for the haplotype groups in the parsimony network. Net divergence (Da)44 among populations from western (1–9) and central-eastern Panama (10–16) (Figure 1; Table 1) was estimated in MEGA 4.030 using the Tamura and Nei model.47 Standard errors (SE) were estimated by the bootstrap method using 10,000 replicates. Divergence time (T) between groups of populations was estimated as (Da = 2 μtD) where 2 μ is the divergence rate,8 and the 95% CIs for the divergence times were calculated as ±1.96 SE of the net distances.

Results

In total, 265 female mosquitoes (Table 1) from 16 localities were sequenced for an unambiguous alignment of 824 bp (both directions). Sequences were AT-rich (70.01% A and T bases) as encountered in the COI gene of other anopheline species.8,25 There were no nonfunctional genes (e.g., pseudogenes) as shown by the absence of stop codons, the prevalence of synonymous substitutions, low pairwise divergence, and clear electropherograms. Sequences were deposited into GenBank (accession nos.: FJ516463–FJ516553). The total number of segregating sites was 85, with 85% transitions and 15% transversions, the majority of which were silent substitutions. Ninety-one haplotypes were detected, 27 (29.6%) were shared among localities, and 64 (70.3%) were singletons (Figures 1 and and22).

Figure 2.
Statistical parsimony network of 91 COI haplotypes of Anopheles albimanus. The letters correspond to the haplotypes observed for the 16 localities in Table 1. The solid black dots represent missing haplotypes. Dotted lines depict 7 mutational steps between ...

Structure of genetic variation.

The statistical parsimony network illustrates the relationship among haplotypes of An. albimanus. All haplotypes differed by less than 12 mutational steps, so they could be connected parsimoniously. Three groups, each separated by 7 mutational steps, suggested substantial haplotype partition (Figure 2). Group A was star-shaped with short branches and an excess of singleton mutations, the signal of a demographic expansion, background selection, or selective sweep.40,44,48 This group contained haplotype A1, which appears to be the most ancestral owing to its high frequency (75/265 sequences), wide geographic distribution (12/16 localities), and interior position in the network.48 In addition, 82.4% of the haplotypes in group A were distributed exclusively in central-eastern Panama (Figure 1; Table 1). In contrast, group B had longer branches, more missing haplotypes, fewer singleton mutations, and 90.2% of its haplotypes were encountered exclusively in Costa Rica and western Panama (Figure 1; Table 1). Group C included only 9.8% of the total haplotypes in the parsimony network, which is star-shaped with short branches and only two missing haplotypes (Figure 2). Similar to group B, 88.8% of the haplotypes in group C were restricted to Costa Rica and western Panama (Figure 1; Table 1). These three haplotype groups represent a substantial genetic division in the data set, showing roughly 0.8% sequence divergence and an average number of 7.5 nucleotide differences (k) among them.

Hierarchical AMOVA tests attributed 33.99% (P = 0.0005) of the total genetic variance to the western versus eastern group comparison, thus partially supporting previous findings,21,22 whereas the Pacific versus Atlantic group comparison accounted only for 6.93% (P = 0.711) of the total variance and was not statistically significant (Table 2). Furthermore, the Mantel analysis to test for IBD among all populations was statistically insignificant (R2 = 0.0214, P = 0.577) suggesting little input of direct-line geographic distance as an explanatory factor of the observed spatial pattern of genetic diversity. The correlation coefficient using a third categorical variable to test the effect of geographic distance on the genetic distance after controlling for the effect of the Central American Cordillera was not statistically significant (R2 = 0.0714, P = 0.377).

Table 2
Comparative results of a hierarchical analysis of molecular variance (AMOVA), using Φ statistics from haplotype frequencies and genetic divergence among populations of Anopheles albimanus. (a) Costa Rica and western Panama (localities 1–9) ...

Demographic history.

Overall values of haplotype and nucleotide diversities were higher (ranging from 0.77 to 0.98 and 0.008 to 0.01, respectively) for populations in Costa Rica and western Panama than for populations in central-eastern Panama (ranging from 0.45 to 0.96 and 0.001 to 0.006, respectively) (Table 3). By group, nucleotide diversity was highest in B, followed by lower and similar values for A and C (Table 4). The mismatch distributions for the entire data set (all populations) and for B alone depicted multimodal curves, and did not fit a model of sudden population expansion, thus suggesting a stationary population at equilibrium. In contrast, A and C were unimodal and fit the model of sudden population expansion (Figure 3). In support of these findings, the mismatch distribution for A was strongly biased toward low divergence values with a mode of three nucleotide changes, suggesting a relatively recent expansion event (Figure 3). The average number of nucleotide differences (k) for A, B, and C was 1.31, 4.83, and 1.10, respectively; suggesting higher diversity and perhaps equilibrium only in B. The raggedness statistic for the entire data set and for A and B was low and not significant providing further support for population expansion. In contrast, C showed a higher and significant value for the raggedness statistic, therefore rejecting the model of sudden population expansion (Figure 3).

Figure 3.
Mismatch distribution for groups A, B, and C of Anopheles albimanus. Bars are observed distribution, and the dashed line shows the distribution simulated under a model of sudden expansion.
Table 3
Summary statistics for polymorphisms of Anopheles albimanus*
Table 4
Summary statistics for polymorphisms and neutrality tests of Anopheles albimanus

Neutrality tests of Tajima's D,38 Fu and Li's F* and D*,39 and Fu's FS40 showed negative and significant values for the entire data set, and for A (Table 4), thus allowing the rejection of either strict neutrality or population size stability, because of either background selection, selective sweep, or demographic expansion. The R2 test41 was also low and significant for A further supporting the departure from equilibrium most likely because of demographic expansion (Table 4). The time since expansion for A was calculated using t = τ/2u, where u = μ × number of base pairs sequenced × generation time, and μ is the mutation rate.44,45 The Drosophila mutation rate of 10−8/site/year49 and 10 generations/year8 were used in this calculation. The estimate of τ, from the raggedness calculation, is 3.39 for A. Therefore, the time since expansion for An. albimanus in central-eastern Panama is approximately 22,000 years ago (95% CI 10,183–38,169), in the late Pleistocene.

Discussion

Pattern of genetic structure.

The role of An. albimanus as a primary malaria vector of Plasmodium vivax and P. falciparum has been historically acknowledged; yet, its capacity to transmit malaria parasites varies across its range.3,19,20 Furthermore, evidence for cryptic species in An. albimanus have been unsubstantiated so far1316 and population differentiation seems to be low overall.19,20 The shallow population structure of An. albimanus in the Americas may be attributed to long-term effective population size caused by high dispersal capability, opportunistic feeding behavior, and ecological plasticity. Nevertheless, this study shows considerable geographic structure for An. albimanus across Costa Rica and the Isthmus of Panama, with a genetic discontinuity within less than 300 km.

Our results partially support previous findings on the population structure of An. albimanus using the mtDNA ND5 gene and microsatellites, which depicted differences in allele and haplotype frequencies between Costa Rica and eastern Panama.21,22 The authors in these studies invoked isolation by distance, physical barriers to gene flow, and/or population contraction in Panama resulting from intense insecticide use as the likely causes of population differentiation. However, neither IBD nor barriers to dispersal across Costa Rica and Panama were supported by our results. The correlation coefficient and the Mantel test remained low and statistically insignificant despite several analyses, and 27 (29.6%) haplotypes were shared among populations and 11 (13.5%) were shared between the Atlantic and Pacific coasts of Costa Rica and western Panama where the mountain range reaches its maximum altitude (roughly 3,000 m). These findings indicate that the Central American Cordillera has not restricted the dispersal of An. albimanus in the past. They also show that the spatial pattern of haplotype diversity depicted by the COI gene is not caused by restricted gene flow as a result of geographic distance.

Mountain ranges (cordilleras) have been identified as barriers to gene flow for some American malaria vectors. The Green Mountains in the eastern United States are believed to restrict dispersal of Anopheles punctipennis, and some lineages of Anopheles nuneztovari s.l. were partially separated by the eastern Andean Cordillera.24,50 Because these two Anopheles species have specific ecological requirements, and limited geographic distributions, mountain ranges are likely to restrict their dispersal. In contrast, An. albimanus is capable of flying up to 32 km, and has been found as high as 1,941 m12,18; thus, low mountain ranges are unlikely to play a substantial role in its geographic structure. A positive and significant relationship between geographic and genetic distances was reported for An. nuneztovari s.l. and An. darlingi in South America using nuclear markers: the white gene,7,9 and microsatellites.7,51,52 However, in most of these cases the correlation was weak, and not detected when different lineages and geographic regions were analyzed separately.7,9,50 Furthermore, IBD was not detected in An. nuneztovari s.l. with analysis of mtDNA restriction fragment length polymorphisms (RFLPs)53 or in An. darlingi using the COI gene.25 Given that neither An. darlingi nor An. nuneztovari s.l. are at mutation-drift equilibrium,9 misleading IBD may be the result of historical fragmentation, initial colonization pathway mimicking the IBD pattern, or inappropriate sampling.54

Mosquito control during the construction of the Panama Canal was carried out in the former Canal Zone and few control measures were undertaken in the surrounding areas. However, we observed reduced diversity in several localities 95 to 410 km from the former Canal Zone (Figure 1). Even though IRS is efficient in controlling endophilic Anopheles species, the exophagic and exophilic behavior displayed by An. albimanus in Central America19,20 make the efficacy of IRS uncertain even in highly susceptible populations. Moreover, avoidance behavioral responses against DDT exhibited by Panamanian populations of An. albimanus may have lessened the physical contact between this species and sprayed surfaces.55 For the previous reasons, we believe that intensive mosquito control alone does not explain the lower diversity and different haplotype composition seen in localities of central-eastern Panama as hypothesized by Molina-Cruz and others.22

Nevertheless, a lower mutation rate of the COI gene indicates that this marker is not as sensitive as microsatellites to detect more recent demographic events.22 However, the larger effective population size and longer coalescent time of microsatellites (e.g., nuclear genome), suggest that population contraction in Panama because of mosquito control is unlikely, as insecticide treatment is a recent phenomenon and microsatellites are lagging indicators of population processes.56 The general agreement of previous microsatellite and mtDNA ND5 gene studies,21,22 and rather similar results in our COI gene analysis, could indicate that the processes causing genetic differentiation in An. albimanus are genome-wide effects, and that they occurred at a much earlier time frame than mosquito control in Panama. The high prevalence and wide distribution of An. albimanus during the last 35 years in Panama also argue against a recent bottleneck caused by insecticide treatment.3,5

Demographic history.

The unimodal and statistically insignificant mismatch distribution under the demographic expansion model and 4 of 5 negative and significant neutrality tests in group A, indicate that An. albimanus is not at equilibrium, because of a past bottleneck, background selection, population expansion, or selective sweep in central-eastern Panama. The neutrality tests cannot distinguish between demographic phenomena and selection as the causes of departure from equilibrium. Therefore, we performed the McDonald-Kreitman neutrality test (MK) in DnaSP, version 4.50.0243 using all the sequences of An. albimanus from central-eastern Panama and 52 COI sequences of An. triannulatus from the same localities as an outgroup (unpublished data). The MK considers the evolution of a protein-coding gene in two closely related species in which nucleotide differences can be classified as either synonymous or non-synonymous, and it was used to investigate the influence of natural selection on the COI sequences. The MK test yielded a statistically insignificant P value of 0.870 in the Fisher's exact test; therefore, the departure from equilibrium observed in An. albimanus from eastern Panama is not a result of directional selection, but more likely to an expansion.

The observed haplotype partition seems to be more related to the demography of An. albimanus and a complex history of dispersal events throughout southern Central America. The genetic signature of population historical processes can remain for long periods of time, obscuring the true geographical population structure that may exist.57 In general, populations from Costa Rica and western Panama showed higher diversity, with a decreasing trend in the nucleotide diversity from Costa Rica toward eastern Panama. Although this may be interpreted as a single founder event followed by sequential population bottlenecks in the colonizing front, and range expansion from Costa Rica into eastern Panama, an alternate interpretation is the co-occurrence and haplotype mixing of three maternal lineages (A, B, and C) in populations of Costa Rica and western Panama, and the predominance of only A in populations of central-eastern Panama. Furthermore, B and C depicted a decreasing trend in haplotype diversity from Costa Rica toward eastern Panama, whereas A also decreased in haplotype diversity, but in the opposite direction (Figure 1). This pattern may represent historical geographic fragmentation in eastern Panama and a subsequent re-invasion by two different source populations (B and C) from the north and one source population from Colombia (A). The fact that old and predominant haplotypes from Costa Rica and western Panama gave rise to haplotypes found in eastern Panama and vice versa, further supports a pattern of recent regional secondary contact.

Taken together, these findings suggest a discontinuous colonization pattern of An. albimanus throughout Costa Rica and the Isthmus of Panama as the most likely explanation for the observed distribution of genetic diversity. The dispersal of different groups of mosquitoes (multiple introductions) could explain three co-occurring and genetically differentiated maternal lineages in Costa Rica and Panama, which may have arrived at different times, with a separate demographic history and therefore an already significant genetic structure between them.58 The divergence time (tD) from the net nucleotide substitution per site (Da = 0.004 ± 0.001, 95% CI) between western and central-eastern populations of An. albimanus placed the time since divergence around 200,000 years ago (235,000 ± 165,000, 95% CI) in the late Pleistocene.

Anopheles albimanus is believed to have originated in the Caribbean islands and subsequently invaded the American continent, but no definitive information on the colonization path exists.22 Group B seems to be the oldest owing to its long branches and missing haplotypes in the parsimony network, and higher nucleotide diversity, which may support a north-south colonization path throughout Costa Rica and Panama; however, this disagrees with the coalescent theory, which predicts that group A is the oldest because of haplotype A1.48 Nevertheless, given the non-equilibrium scenario and likely past geographic fragmentation of An. albimanus in the present study, it is possible that the COI gene does not have enough resolution to infer the original colonization pathway of An. albimanus throughout southern Central America. Because of the haploid nature and maternal mode of inheritance, mtDNA genes have 4-fold smaller effective population size than nuclear markers, and therefore the COI gene may have been affected by more recent demographic phenomena during the Pleistocene, thus erasing more ancient signals of colonization events.

Pollen data from marine records of the Panama Basin and the Pacific slope of the northern Andes show cold and humid conditions between 39,410 and 28,120 years ago with moderately high sea levels and less precipitation. In contrast, the coldest and driest conditions and the lowest sea levels occurred between 28,120 and 14,500 years ago.59 These environmental changes extend through the population expansion time proposed for An. albimanus in this study. Anopheles albimanus is highly opportunistic, able to develop in a wide variety of breeding habitats18; however, it is more abundant in coastal plains, close to the shore, and at less than 100 m.11 Therefore, a combination of Pleistocene climatic oscillation, most likely involving temperature, precipitation, and sea level changes in eastern Panama, seems to be a plausible explanation for the population expansion of An. albimanus.

The entire data set also concurs with population expansion, yet B and C showed a less clear signal. These two groups are distributed mostly in Costa Rica and western Panama where mountains and upland plains might have served as a refuge providing alternative and more stable breeding sites during glacial periods. It is noteworthy that haplotype A1 was not recorded from the Atlantic coast of Costa Rica and western Panama. Pleistocene glaciations exposed most of the Gulf of Panama, located in the narrowest strip of land of central Panama,60 and this may have facilitated re-invasion of this expanding haplotype through the Pacific coast, whereas re-invasion through the Atlantic coast might have been more irregular.

Conclusions

In the present study, the COI gene uncovered significant genetic structure at a micro-geographic scale. The genetic differentiation in An. albimanus appears to be caused by Pleistocene historical fragmentation, re-colonization, and population expansion, rather than to contemporary geographic structure, insecticide treatment, and selective pressures on the COI locus. Other studies carried out in fresh water fishes, the pseudoscorpion Cordylochernes scorpioides and the tungara frog Physalaemus pustulosus, have supported a complex biogeographical history in the Isthmus of Panama where significant genetic differentiation is seen at short distances, and waves of colonization, extinctions, and re-invasions appear to be the main forces shaping the distribution of genetic diversity.6163 Our results have to be viewed with caution as they may represent divergence only in the mitochondrial genome and therefore a nuclear marker will be required to test these findings. In addition, more research will be needed to determine whether these three maternal lineages may have different potential roles in malaria transmission.

Acknowledgments

We are grateful to Lisa Mirabello from the Division of Cancer Epidemiology and Genetics, Bethesda, (US) and Jorge Motta, Director, Instituto Conmemorativo Gorgas de Estudios de la Salud, Panama, for technical assistance and logistical support. We thank Luis Guillermo Chaverri Sánchez and Alvaro Herrera Villalobos from Instituto Nacional de Biodiversidad de Costa Rica (INBio) for helping us to obtain the required collecting and export permits and for fieldwork guidance in Costa Rica. We thank Rolando Torres Cosme (Departamento de Control de Vectores del Ministerio de Salud, Panama) and Jordane Roy Leblanc (McGill University) for fieldwork collaboration. We also thank Oris Sanjur, Maribel Gonzales, and Grethel Grajales from Smithsonian Tropical Research Institute (STRI), who provided guidance in the laboratory. We are grateful to Harilaos Lessios and Matthew Miller from STRI for their valuable suggestions on an earlier draft of this manuscript.

Footnotes

Financial support: The Secretariat for Science, Technology and Innovation of Panama (SENACYT) through a scholarship and research grant COL08-066 awarded to Jose Loaiza, partially financed this research. Additional financial support was provided by the Smithsonian Tropical Research Institute (STRI), the Organization of American States (OAS), Natural Sciences and Engineering Research Council of Canada (NSERC), and National Institutes of Health (NIH) grant (AI) R0154139-02 to JEC. Research at the Institute of Parasitology is supported by a regroupement stratégique from Fonds Québécois de la recherche sur la nature et les technologies (FQRNT).

Authors' addresses: Jose R. Loaiza and Marilyn E. Scott, Department of Natural Resource Sciences, McGill University, Ste. Anne de Bellevue, Quebec, Canada, E-mails: ac.lligcm.liam@aziaol.esoj and ac.lligcm@ttocs.nyliram. Eldredge Bermingham, Smithsonian Tropical Research Institute, Balboa Ancon, Republic of Panama, E-mail: ude.is@mahgnimreb. Jose Rovira, Departamento de Entomologia Medica, Instituto Conmemorativo Gorgas de Estudios de la Salud, Ciudad de Panama, E-mail: ap.bog.sagrog@arivorj. Jan E. Conn, Griffin Laboratory, The Wadsworth Center, New York State Department of Health, Slingerlands, NY, E-mail: gro.htrowsdaw@nnocj.

References

1. Pan American Health Organization . Regional Strategic Plan for Malaria in the Americas 2006–2010. Washington, DC: Pan American Health Organization; 2006.
2. Ministerio Nacional de Salud de Panamá . Análisis de la Situación de Malaria en Panamá. Boletín Epidemiológico; 2007. http://www.minsa.gob.pa Available at. Accessed May 4, 2009.
3. Loaiza JR, Bermingham E, Scott ME, Rovira JR, Conn JE. Species composition and distribution of adult Anopheles (Diptera: Culicidae) in Panama. J Med Entomol. 2008;45:841–851. [PMC free article] [PubMed]
4. Calzada EJ, Samudio F, Bayard V, Obaldia N, De Mosca IB, Pascale JM. Revising antimalarial drug policy in Central America: experience of Panama. Trans R Soc Trop Med Hyg. 2009;102:694–698. [PubMed]
5. Loaiza J, Scott M, Bermingham E, Rovira J, Conn J. Short report: Anopheles darlingi in Panama. Am J Trop Med Hyg. 2009;81:23–26. [PMC free article] [PubMed]
6. Lehmann T, Licht M, Elissa N, Maega BT, Chimumbwa JM, Watsenga FT, Wondji CS, Simard F, Hawley WA. Population structure of Anopheles gambiae in Africa. J Hered. 2003;94:133–147. [PubMed]
7. Mirabello L, Vineis JH, Yanoviak SP, Scarpassa VM, Povoa MM, Padilla N, Nicole LA, Conn JE. Microsatellite data suggest significant population structure and differentiation within the malaria vector Anopheles darlingi in Central and South America. BMC Ecol. 2008;8:3. [PMC free article] [PubMed]
8. Walton C, Handley JM, Tun-Lin W. Population structure and population history of Anopheles dirus mosquitoes in Southeast Asia. Mol Biol Evol. 2000;17:962–974. [PubMed]
9. Mirabello L. Molecular Population Genetics of the Malaria Vector Anopheles darlingi throughout Central and South America using Mitochondrial, Nuclear and Microsatellites Markers. Albany, NY: University at Albany; 2007. Ph.D. Thesis.
10. O'Loughlin SM, Okabayashi T, Honda M, Kitazoe Y, Kishino H, Somboon P, Sochantha T, Nambanya S, Saikia PK, Dev V, Walton C. Complex population history of two Anopheles dirus mosquito species in south Asia suggests the influence of Pleistocene climate change rather than human-mediated effects. J Evol Biol. 2008;21:1555–1569. [PubMed]
11. Faran ME. Mosquito studies (Diptera: Culicidae) XXXIV. A revision of the Albimanus section of the subgenus Nyssorhynchus of Anopheles, contrib. Am Entomol Inst. 1980;15:1–215.
12. Breeland SG. Studies on the ecology of Anopheles albimanus. Am J Trop Med Hyg. 1972;21:751–754. [PubMed]
13. Hobbs J, Sexton J, St. Jean Y, Jacques J. The biting and resting behavior of Anopheles albimanus in northern Haiti. J Am Mosq Control Assoc. 1986;2:150–153. [PubMed]
14. Beach RF, Mills D, Collins FH. Structure of ribosomal DNA in Anopheles albimanus (Diptera: Culicidae) Ann Entomol Soc Am. 1989;81:641–648.
15. Narang SK, Seawright JA, Suarez MF. Genetic structure of natural populations of Anopheles albimanus in Colombia. J Am Mosq Control Assoc. 1991;7:337–345. [PubMed]
16. De Merida AM, De Mata MP, Molina E, Porter CH, Black WC. Variation in ribosomal DNA intergenic spacers among populations of Anopheles albimanus in South and Central America. Am J Trop Med Hyg. 1995;53:469–477. [PubMed]
17. Collins WE, Skinner JC, Warren M, Richardson B. Studies on human malaria in Aotus monkeys. VII. Comparative infectivity of two strains of Plasmodium vivax to Anopheles freeborni, An. maculatus, and four strains of An. albimanus. J Parasitol. 1976;62:190–194. [PubMed]
18. Frederickson EC. Bionomics and Control of Anopheles albimanus. Washington, DC: Pan American Health Organization, Pan American Sanitary Bureau Regional Office of the World Health Organization; 1993.
19. Grieco JP, Achee NL, Roberts DR, Andre RG. Comparative susceptibility of three species of Anopheles from Belize, Central America, to Plasmodium falciparum (Nf-54) J Am Mosq Control Assoc. 2005;21:279–290. [PubMed]
20. Achee NA, Achee N, Grieco JP, Andre RG, Rejmankova E, Roberts DR. A mark release-recapture study to define the flight behaviours of Anopheles vestitipennis and An. albimanus in Belize, Central America. J Am Mosq Control Assoc. 2007;3:276–282. [PubMed]
21. De Merida AM, Palmieri M, Yurrita M, Molina A, Molina E, Black WC. Mitochondrial DNA variation among Anopheles albimanus populations. Am J Trop Med Hyg. 1999;6:230–239. [PubMed]
22. Molina-Cruz A, De Merida AM, Mills K, Rodriguez F, Schoua C, Yurrita MM, Molina E, Palmieri M, William CB. Gene flow among Anopheles albimanus populations in Central America, South America, and the Caribbean assessed by microsatellites and mitochondrial DNA. Am J Trop Med Hyg. 2004;71:350–359. [PubMed]
23. Conn JE, Mirabello L. The biogeography and population genetics of neotropical vector species. Heredity. 2007;99:245–256. [PubMed]
24. Fairley TL, Renaud TM, Conn JE. Effects of local geographic barriers and latitude on population structure in Anopheles punctipennis (Diptera: Culicidae) J Med Entomol. 2000;37:754–760. [PubMed]
25. Mirabello L, Conn JE. Molecular population genetics of the malaria vector Anopheles darlingi in Central and South America. Heredity. 2006;96:311–321. [PubMed]
26. Wilkerson RC, Strickman D. Illustrated key to the female anopheline mosquitoes of Central America and Mexico. J Am Mosq Control Assoc. 1990;6:7–34. [PubMed]
27. Lunt DH, Zhang DX, Szymura JM, Hewitt GM. The insect cytochrome oxidase I gene: evolutionary patterns and conserved primers for phylogenetic studies. Insect Mol Biol. 1996;5:153–165. [PubMed]
28. Swofford DL. PAUP (Phylogenetic Analysis Using Parsimony) and Other Methods. Sunderland, MA: Sinauer Associates; 2003. Version 4.
29. Maddison WP, Maddison DR. MacClade: Analysis of Phylogeny and Character Evolution. Sunderland, MA: Sinauer Associates; 1997. Version 3.07.
30. Tamura K, Dudley J, Nei M, Kumar S. MEGA 4: molecular evolutionary genetics analysis (MEGA) software version 4.0. Mol Biol Evol. 2007;24:1596–1599. [PubMed]
31. Clement M, Posada D, Crandall KA. TCS: a computer program to estimate gene genealogies. Mol Ecol. 2000;9:1657–1659. [PubMed]
32. Posada D, Crandall KA, Templeton AR. GeoDis: a program for the cladistic nested analysis of the geographical distribution of genetic haplotypes. Mol Ecol. 2000;9:487–488. [PubMed]
33. Crandall KA, Templeton AR. Empirical tests of some predictions from coalescent theory with applications to intraspecific phylogeny reconstruction. Genetics. 1993;134:959–969. [PubMed]
34. Uthicke S, Benzie JAH. Gene flow and population history in high dispersal marine invertebrates: mitochondrial DNA analysis of Holothuria nobilis (Echinodermata: Holoturoidea) populations from the Indo-Pacific. Mol Ecol. 2003;12:2635–2648. [PubMed]
35. Excoffier L, Laval G, Schmeider S. Arlequin (version 3.0): an integrated software package for population genetic data analysis. Evol Bioinform Online. 2005;1:47–50. [PMC free article] [PubMed]
36. Mantel N. The detection of disease clustering and a generalized regression approach. Cancer Res. 1967;27:209–220. [PubMed]
37. Jensen J, Bohonak AJ, Kelley ST. Isolation by distance, web service. BMC Genet. 2005;6:13. [PMC free article] [PubMed]
38. Tajima F. Statistical method for testing the neutral mutation hypothesis by DNA polymorphisms. Genetics. 1989;123:585–595. [PubMed]
39. Fu YX, Li WH. Statistical tests of neutrality of mutations. Genetics. 1993;133:693–709. [PubMed]
40. Fu YX. Statistical tests of neutrality of mutations against population growth, hitchhiking and background selection. Genetics. 1997;147:915–925. [PubMed]
41. Ramos-Onsins SE, Rozas J. Statistical properties of new neutrality tests against population growth. Mol Biol Evol. 2002;19:2092–2100. [PubMed]
42. Kimura M. The Neutral Theory of Molecular Evolution. Cambridge, England: Cambridge University Press; 1983.
43. Rozas J, Sanchez-Del Rio JC, Messeguer X, Rozas R. DnaSP, DNA polymorphism analyses by the coalescence and other methods. Bioinformatics. 2003;19:2496–2497. [PubMed]
44. Nei M. Molecular Evolutionary Genetics. New York: Columbia University Press; 1987.
45. Harpending HC, Sherry ST, Rogers AR, Stoneking M. The genetic structure of ancient human populations. Curr Anthropol. 1993;34:483–496.
46. Rogers AR. Genetic evidence for a Pleistocene population explosion. Evolution. 1995;49:608–615.
47. Tamura K, Nei M. Estimation of the number of nucleotide substitutions in the control region of mitochondrial DNA in humans and chimpanzees. Mol Biol Evol. 1993;10:512–526. [PubMed]
48. Castelloe J, Templeton AR. Root probabilities for intraspecific gene trees under neutral coalescent theory. Mol Phyl Evol. 1994;3:102–113. [PubMed]
49. Powell JR, Caccone A, Amato GD, Yoon C. Rate of nucleotide substitution in Drosophila mitochondrial DNA and nuclear DNA are similar. Proc Natl Acad Sci USA. 1986;83:9090–9093. [PubMed]
50. Mirabello L, Conn JE. Population analysis using the nuclear white gene detects Pliocene/Pleistocene lineage divergence within Anopheles nuneztovari in South America. Med Vet Entomol. 2008;22:109–119. [PubMed]
51. Conn JE, Vineis JH, Bollback JP, Onyabe DY, Wilkerson RC, Póvoa MM. Population structure of the malaria vector Anopheles darlingi in a malaria-endemic region of eastern Amazonian Brazil. Am J Trop Med Hyg. 2006;74:798–806. [PubMed]
52. Scarpassa VM, Conn JE. Population genetic structure of the major malaria vector Anopheles darlingi (Diptera: Culicidae) from the Brazilian Amazon, using microsatellite markers. Mem Inst Oswaldo Cruz. 2007;102:319–327. [PubMed]
53. Conn J, Mitchell SE, Cockburn AF. Mitochondrial DNA analysis of the neotropical malaria vector Anopheles nuneztovari. Genome. 1998;41:313–327.
54. Telles MP, Diniz-Filho JA. Multiple Mantel test and isolation by distance, taking into account long-term historical divergence. Genet Mol Res. 2005;4:742–748. [PubMed]
55. Trapido H. Modified response of Anopheles albimanus to DDT residual house spraying in Panama. Am J Trop Med Hyg. 1952;15:853–861. [PubMed]
56. Zink RM, Barrowclough GF. Mitochondrial DNA under siege in avian phylogeography. Mol Ecol. 2008;17:2107–2121. [PubMed]
57. Templeton AR. Nested clade analyses of phylogeographic data: testing hypotheses about gene flow and population history. Mol Ecol. 1998;7:381–397. [PubMed]
58. Kolbe JJ, Glor RE, Schettino LR, Lara AD, Larson A, Losos JB. Genetic variation increases during biological invasion by a Cuban lizard. Nature. 2004;431:177–181. [PubMed]
59. Gonzáles C, Urrego LE, Martínez JI. Late quaternary vegetation and climate change in the Panama basin: palynological evidence from marine cores ODP 677 and TR 163-38. PALAEO. 2006;234:62–80.
60. Golik A. History of the Holocene transgression in the Gulf of Panama. J Geol. 1968;76:497–507.
61. Bermingham E, Martin AP. Comparative mtDNA phylogeography of neotropical fresh water fishes: testing shared history to infer the evolutionary landscape of lower Central America. Mol Ecol. 1998;7:499–517. [PubMed]
62. Zeh JA, Zeh DW, Bonilla MM. Phylogeography of the harlequin beetle-riding pseudoscorpion and the rise of the Isthmus of Panama. Mol Ecol. 2003;12:2759–2769. [PubMed]
63. Weigt LA, Crawford AJ, Stanley Rand A, Ryan M. Biogeography of the tungara frog, Physalaemus pustulosus: a molecular perspective. Mol Ecol. 2005;14:3857–3876. [PubMed]

Articles from The American Journal of Tropical Medicine and Hygiene are provided here courtesy of The American Society of Tropical Medicine and Hygiene