- Top of page
- Materials and Methods
Objective To describe the genetic diversity of Plasmodium vivax isolates from different areas in the Brazilian Amazon using 11 polymorphic microsatellites and to evaluate the correlation between microsatellite variation and repeat array length.
Methods Microsatellites with variable repeat units and array lengths were selected using in silico search of the P. vivax genome. We designed primers and amplified the selected loci in DNA obtained from patients with P. vivax acute infections.
Results Positive correlation between repeat array length and microsatellite variation was detected independently of the size of repeat unit (di, tri, or tetranucleotide). We used these markers to describe the genetic variability of P. vivax isolates from four geographic regions of the Brazilian Amazon. Substantial variability was observed among P. vivax isolates within populations, concurrent with high levels of multiple-clone infections and high linkage disequilibrium. Overall, structured populations were observed with moderate to high genetic differentiation.
Conclusion The markers studied are useful tools for assessing population structure of P. vivax, as demonstrated for Brazilian populations and for searching for evidence of recent selection events associated with different phenotypes, such as drug resistance.
Loci microsatellites: Détermination de la variabilité génétique du Plasmodium vivax
Objectifs: Décrire la diversité génétique des isolats de P. vivax provenant de différentes régions de l’Amazonie brésilienne en utilisant 11 microsatellites polymorphes et évaluer la corrélation entre la variation des microsatellites et la taille des réseaux répétitifs.
Méthodes: Les microsatellites avec des unités répétitives et des tailles de réseaux variables ont été sélectionnés à partir d’une recherche in silico du génome de P. vivax. Nous avons conçu des amorces et amplifié les loci sélectionnés à partir d’ADN provenant de patients avec des infections aiguës àP. vivax.
Résultats: Une corrélation positive entre la taille des réseaux répétitifs et la variation des microsatellites a été détectée indépendamment de la taille de l’unité répétitive (di, tri ou tétranucléotide). Nous avons utilisé ces marqueurs pour décrire la variabilité génétique des isolats de P. vivax de quatre régions géographiques de l’Amazonie brésilienne. Une variabilité importante a été observée chez les isolats de P. vivax au sein des populations en parallèle avec des taux élevés d’infections à clones multiples et un déséquilibre de lien important. Dans l’ensemble, des populations structurées ont été observées avec une différenciation génétique modérée àélevée.
Conclusion: Les marqueurs étudiés sont des outils utiles pour évaluer la structure des populations de P. vivax, comme démontré pour les populations brésilienne et pour la recherche de preuves d’événements récents de sélection, associés à des phénotypes différents, tels que la résistance aux médicaments.
Loci microsatelitales: Determinación de la variabilidad genética de Plasmodium vivax
Objetivos: Describir la diversidad genética de aislados de P. vivax de diferentes áreas en la Amazonía Brasilera utilizando 11 microsatélites polimórficos, y evaluar la correlación entre la variabilidad microsatelital y la longitud de los fragmentos repetitivos.
Métodos: Se seleccionaron microsatélites con unidades repetitivas y longitudes variables utilizando una búsqueda in silico del genoma de P. vivax. Se diseñaron cebadores y se amplificaron los loci seleccionados en ADN obtenido de pacientes con infecciones agudas de P. vivax.
Resultados: Se detectó una correlación positiva entre la longitud de los fragmentos repetitivos y la variación microsatelital independientemente del tamaño de la unidad repetitiva (di, tri o tetranucleótido). Hemos utilizado estos marcadores para describir la variabilidad genética de aislados de P. vivax provenientes de cuatro regiones geográficas de la Amazonía brasilera. Se observó una variabilidad sustancial entre los aislados de P. vivax dentro de las poblaciones, junto con altos niveles de infección con múltiples clones y un alto desequilibro de ligamiento. En general, se observaron poblaciones estructuradas con una diferenciación genética moderada a alta.
Conclusión: Los marcadores estudiados fueron herramientas útiles para evaluar la estructura poblacional de P. vivax, tal y como se ha demostrado para poblaciones brasileras, así como para buscar evidencia de eventos recientes de selección asociados a diferentes fenotipos, tales como la resistencia a medicamentos.
- Top of page
- Materials and Methods
More than two billion people are at risk of malaria infection, and approximately 300–500 million clinical cases are reported annually (Hay et al. 2004). Malaria primarily afflicts populations in tropical and subtropical regions of the world, where the temperature and rainfall are suitable for the development of malaria-causing parasites in Anopheles mosquitoes (Greenwood et al. 2008). Plasmodium vivax is the most globally widespread species causing human malaria. Within an endemic area, parasite genetic diversity is frequently observed, and many field parasites circulate within the human host (De Souza-Neiras et al. 2007). The ability to distinguish between isolates and populations is a requirement for understanding the local and global epidemiology of P. vivax, with a practical significance to implement control strategies.
Compared with the more virulent parasite Plasmodium falciparum, knowledge about the genetic variability of P. vivax is limited, in part because of difficulties of its in vitro maintenance. Although P. vivax surface antigens show high genetic diversity, this is likely because of the combined effects of population history of the parasite and natural selection acting on these loci (Escalante et al. 2004). On the other hand, analyses of allelic variation at multiple independent loci such as microsatellites, that more likely are neutral or nearly neutral, provide a most effective way to assess the population structure of P. vivax (Karunaweera et al. 2008). Leclerc et al. (2004) reported low microsatellite variability: only 1 of 13 identified microsatellites was polymorphic, and 9 were completely monomorphic among the eight P. vivax populations analysed. Imwong et al. (2006) designed primers for 11 dinucleotide microsatellites that showed high allelic diversity and interpreted the difference in their results compared to those of Leclerc et al. (2004) as because of the dependence of microsatellite variation on the repeat array length. Long arrays would be more diverse than short arrays because slippage mutations become exponentially more common with an increase in array length. Recently, 14 polymorphic microsatellites were described and used to analyse genetic and geographic variability in P. vivax isolates from Brazil (only Acre State), Vietnam, and Sri Lanka (Karunaweera et al. 2007, 2008).
The aim of this work is determining the pattern of genetic diversity and population structure of P. vivax isolates from four malaria endemic areas of the Brazilian Amazon. For this purpose, we characterize 11 polymorphic microsatellites with different repeat units (di-, tri-, and tetranucleotides) and lengths of repeat arrays (9–49 number of repeats in Sal-1 P. vivax strain). Six of these loci are identified for the first time in this study. We also elucidate the association between microsatellite variation and repeat array length.
- Top of page
- Materials and Methods
Twenty-two microsatellites were selected from the P. vivax genome to design primers. Eleven of them did not amplify well or were not polymorphic in a preliminary PCR (data not shown). The other 11 polymorphic microsatellite loci were distributed across the P. vivax genome, occurring in 8 of 14 chromosomes, most of them in non-coding regions (Table 2). Only four microsatellites were lying in coding regions, three of them in hypothetical proteins. Six microsatellite loci were newly identified and the other five have been described by Carlton et al. (2008) when the P. vivax genome was sequenced. Fifty-three isolates from four geographic areas in the Brazilian Amazon were genotyped for the 11 microsatellites. The number of alleles varies across loci from 4 (PvMS1) to 31 (PvMS11), with an average of 12 alleles (Table 3). A positive correlation was found between the number of repeat units and the number of alleles (Figure 2). anova analysis revealed no statistical difference in the number of alleles between di-, tri-, or tetranucleotide repeats (P = 0.619).
Table 2. Microsatellite characteristics: chromosome location, repeat unit, primers identities
|Microsatellite||Chromosome*||Repeat unit†||Identity of forward primer‡||Identity of reverse primer‡|
|PvMS1||12||GT9||Adapter-related protein complex 1 gamma 2 subunit, putative (ID:5475130)||Adapter-related protein complex 1 gamma 2 subunit, putative (ID:5475130)|
|PvMS2||3||CA12||Hypothetical protein, (ID:5472022)||Hypothetical protein, (ID:5472022)|
|PvMS3||8||TA18||2318 bp at 5′ side: hypothetical protein (ID:5473614)||6942 bp at 3′ side: hypothetical protein (ID:5473615)|
|PvMS4||6||TA21||2868 bp at 5′ side: citrate synthase, mitochondrial precursor (ID:5471402)||27 bp at 3′ side: pre-mRNA splicing factor (ID:5471403)|
|PvMS5||3||CAT10||Hypothetical protein (ID:5472646)||Hypothetical protein (ID:5472646)|
|PvMS6||14||TGA19||Hypothetical protein (ID:5473303)||Hypothetical protein (ID:5473303)|
|PvMS7||2||TAA22||4023 bp at 5′ side: Small_Subunit_rRNA (ID:5471924)||1008 bp at 3′ side: ABC transporter (ID:5471923)|
|PvMS8||13||TGTA7||166 bp at 5′ side: hypothetical protein (ID:5476138)||1942 bp at 3′ side: nuclear transport factor 2 (ID:5476139)|
|PvMS9||6||CATA11||3735 bp at 5′ side: hypothetical protein (ID:5471301)||5124 bp at 3′ side: hypothetical protein (ID5471328)|
|PvMS10||5||TATG14||2540 bp at 5′ side: hypothetical protein (ID:5474410)||13 bp at 3′ side: steroid dehydrogenase kik-i (ID:5474411)|
|PvMS11||8||CATA49||3161 bp at 5′ side: hypothetical protein (ID:5473784)||190 bp at 3′ side: hypothetical protein (ID:5373785)|
Table 3. Locus characteristics: frequency of alleles, range of array length, numbers of alleles, and Heterozygosity expected (HE)
|Allele||Allele frequency on each locus|
| 1||15||21|| 2||12|| 2|| 2|| 4|| 2|| 3||60|| 2|
| 2|| 8||15||24|| 1|| 7|| 2||26|| 2||16|| 4|| 2|
| 3||48||33||13||12|| 4|| 4||28|| 2|| 3|| 5|| 2|
| 4||28|| 6|| 4||14||14|| 7|| 9||33||10||12|| 3|
| 5|| ||13||15||14||14||31|| 4|| 2||28|| 4|| 3|
| 6|| || 6|| 7||14|| 5|| 4||12||24||36|| 5|| 2|
| 7|| || 4|| 4|| 3||23|| 2|| 2||18|| 3|| 2|| 3|
| 8|| || 2||11|| 4|| 2|| 2||11|| 4|| || 4|| 7|
| 9|| || ||17|| 4|| 2|| 4|| 2||11|| || 4|| 3|
|10|| || || 2|| 1||18|| 2|| 2|| || || 2|| 2|
|11|| || || || 3|| 2|| 7|| 2|| || || || 2|
|12|| || || || 4|| 7|| 5|| || || || || 5|
|13|| || || || 9|| || 5|| || || || || 7|
|14|| || || || 1|| || 7|| || || || || 7|
|15|| || || || 1|| || 2|| || || || || 2|
|16|| || || || || || 2|| || || || || 5|
|17|| || || || || || 4|| || || || || 2|
|18|| || || || || || 2|| || || || || 7|
|19|| || || || || || 7|| || || || || 2|
|20|| || || || || || || || || || || 2|
|21|| || || || || || || || || || || 2|
|22|| || || || || || || || || || || 3|
|23|| || || || || || || || || || || 3|
|24|| || || || || || || || || || || 3|
|25|| || || || || || || || || || || 5|
|26|| || || || || || || || || || || 5|
|27|| || || || || || || || || || || 2|
|28|| || || || || || || || || || || 2|
|29|| || || || || || || || || || || 2|
|30|| || || || || || || || || || || 3|
|31|| || || || || || || || || || || 2|
|Alleles†|| 4|| 8||10||15||12||19||11|| 9|| 7||10||31|
| H E || 0.775|| 0.487|| 0.797|| 0.815|| 0.788|| 0.521|| 0.782|| 0.758|| 0.609|| 0.636|| 0.922|
Figure 2. Correlation analysis between the number of repeat units of microsatellite loci and the number of alleles detected among all Plasmodium vivax Brazilian isolates. Correlation coefficient of Pearson r = 0.899, P = 0.0002.
Download figure to PowerPoint
Diversity indexes in the studied populations range from 0.57 (Amapa) to 0.77 (Amazonas) (0.712 mean HE) (Table 4). We found 51 eleven-loci haplotypes, most of them unique. Only Amapa possessed two haplotypes, each one appearing twice. Multiple infections were frequently observed (57% patients), with higher levels in Amazonas and Rondonia (Table 4). When we considered as evidence of multiple infections the presence of more than one peak for two or more loci, this number dropped to 32%. All populations showed several rare and few common alleles, as expected from neutrality (Figure 3). By assuming the two-phase mutation model and using the Bottleneck software, we determined that all the populations fit the null hypothesis of constant population size (data not shown). However, this result should be interpreted with caution because 11 loci may not have enough power to detect recent or weak reductions in population size.
Table 4. Microsatellite-based analysis of the genetic diversity (expected heterogeneity –HE), linkage disequilibrium (), and genetic distances (FST) between populations of Plasmodium vivax from Brazilian geographic regions with different risk of malaria infection
|Populations||Risk of malaria infection†||(HE) ± SD||MI (%)‡|| || F ST |
|AP||Low||0.57 ± 0.27||45||0.4228*||–|| || |
|RO||High||0.75 ± 0.13||69||0.0770*||0.2370*||–|| |
|PA||High||0.74 ± 0.13||40||0.1706*||0.1732*||0.1115*||–|
|AM||Medium||0.77 ± 0.16||73||0.0344||0.2352*||0.0681*||0.1355*|
|Average||–||0.71 ± 0.18||57||–||–||–||–|
Figure 3. Allele frequency distribution for each of studied population: Amapa (AP), Amazonas (AM), Para (PA), and Mato Grosso (MT).
Download figure to PowerPoint
The FST values ranged from 0.068 to 0.237, with the highest values involving the Amapa population (Table 4). The lowest intrapopulation diversity and highest interpopulation differentiation of the Amapa population is consistent with the highest linkage disequilibrium observed in this population. Significant LD was found in 3 of 4 geographic regions (Table 4). The highest value observed in Amapa was maintained even when only unique haplotypes were analysed ( = 0.3483, P < 0.05), excluding a recent clonal expansion as an explanation for this result (Ferreira et al. 2007). We also analysed the pattern of linkage disequilibrium between pair of markers. None marker pair showed significant LD in all populations, only one pair (PvMS1-PvMS8) showed LD in three populations, 16 pairs of markers showed LD in two populations, and 21 pairs in one population (Figure 4). As only one of three pairs of markers located in the same chromosome presented LD in two populations, and significant LD was observed between markers in different chromosomes, inbreeding or recent admixture, instead of physical linkage is the most feasible explanation for the overall LD observed.
Figure 4. Linkage disequilibrium analyses for pair of loci in each studied population. Each box represents the level of LD for a pair of loci. Black box represents a pairs of loci in linkage disequilibrium. The level of significance was defined as 0.05. These analyses were performed using Arlequin version 3.1 software.
Download figure to PowerPoint
The population structure was evaluated using the Bayesian model-based approach implemented in Structure software, with the most likely K = 4. In the Figure 1, we represent in the circles the proportion of these four inferred genetic subpopulations (represented by colors) for each of the geographic populations. Each geographic population showed heterogeneity with one distinct predominant genetic subpopulation. One of these genetic subpopulations (yellow) was predominant in Amazonas population and was the second predominant in all the other three populations. This pattern of diversity likely results from few parasites with their origin clearly assigned to a single genetic subpopulation and most of them with mixed ancestry (Figure 5).
Figure 5. Population structure of Plasmodium vivax inferred from microsatellite typing of 53 isolates using the STRUCTURE program. Each individual parasite is represented by a vertical bar, which is partitioned into K (4) colored segments, representing genetic subpopulations. Isolates are organized by geographical origin.
Download figure to PowerPoint
- Top of page
- Materials and Methods
In this work, we described 11 microsatellites to analyse the genetic structure of Amazonian P. vivax populations, as well as the influence of the repeat unit and array length in microsatellite variability of this species. The Brazilian Amazon is a large region with malaria epidemiological characteristics that are geographically variable. For these reasons, the Brazilian Amazon is a particularly interesting region for population structure analysis.
Seventeen of the 42 microsatellite loci from the panel described so far as polymorphic in field populations of P. vivax, including loci described here, were also described by Carlton et al. (2008). These microsatellite loci were not homogeneously distributed among the genome, concentrating in chromosomes 3, 6, and 8. Microsatellites described by Karunaweera et al. (2007) were lying principally in hypothetical proteins encoding genes and microsatellite loci described here and by Imwong et al. (2006, 2007) were lying mainly in no coding regions, these loci are more likely near-neutral, and therefore, constitute a convenient choice to study the genetic and population structure of P. vivax. The major difference between these microsatellites was the repeat unit: only dinucleotides (Leclerc et al. 2004; Imwong et al. 2006); tri- and tetranucleotides (most of them composite and/or imperfect) (Karunaweera et al. 2007); or variable repeat units (Imwong et al. 2007). Here, we describe di-, tri-, and tetranucleotide repeat units, all of them perfect microsatellites.
We observed a substantive variation in the number of alleles per locus and a positive correlation between the number of alleles and the number of repeat units. This finding is consistent with the fact that microsatellite variation is dependent on the length of repeat arrays because of an exponential increase in the amount of DNA polymerase slippage that is observed with increasing repeat numbers (Imwong et al. 2006; Russell et al. 2006). Our results corroborate the studies performed by Imwong et al. (2006) demonstrating the importance of microsatellite array length, extending the findings for tri- and tetranucleotide repeat loci.
Using our 11 microsatellites, we showed that isolates from distinct geographic areas in Brazilian Amazon display extensive genetic diversity, and we illustrated that frequent multiple-clone infections concur with a high LD. A similar profile was also demonstrated for the Acre population, another Brazilian state (Ferreira et al. 2007), and for a Sri Lankan population using a different set of markers (Karunaweera et al. 2008). Ferreira et al. (2007) showed a slightly lower mean of alleles/locus, but similar He, consistently high LD and high multiplicity of infections. In contrast, regions of low transmission of P. falciparum are associated with strong LD, low genetic diversity, and low levels of multiple-clone infections (Anderson et al. 2000). Thus, P. vivax appears to be more diverse and more frequently associated with high multiple-clone infections than the sympatric P. falciparum isolates (Ferreira et al. 2007). Significant deviations from random association among loci suggest that the rate of recombination may be sufficiently low relative to mutation, such that LD is maintained. Alternatively, the populations may result from admixture with a genetically divergent parasite population, and insufficient time has passed for recombination to homogenize these two populations (Anderson et al. 2000). Our data support the latter hypothesis because genetic mixture was observed among populations and in parasite populations from South America. Moreover, multilocus linkage disequilibrium has not been breaking down by multiple-clone infection because of differential gametocytes production between variants with temporal expansion of particular variants, or frequent strand slippage in mitotic replication of parasites generating diversity without affecting LD (Anderson et al. 2000).
One of our populations, the Amapa, was the most homogenous and differentiated. This population displayed the highest LD globally and between pairs of locus and intermediate percentage of multiple-clone infections. We have some hypotheses for these findings that are not mutually exclusive. First of all, this population may have been established more recently than other populations, and therefore did not have time to accumulate new mutational events. However, no evidence for a bottleneck or founder effect was revealed by our analysis using Bottleneck software. Secondly, Amapa state has a very low API value; when the transmission rate is low, the chances of recombination between different parasites are consequently low. Moreover, because only few Amapa-infected individuals have multiple infections, mosquitoes must bite different persons to gain access to more than one parasite variants, which reduce the number of recombination events. As a consequence, linkage disequilibrium increases in the population. The third hypothesis postulates the presence of a different vector in this area. The major vector in Brazil is Anopheles darlingi, although there are several reports demonstrating Anopheles marajoara as the major vector in this state (Conn et al. 2002). Particular variants could be better adapted to this vector, because coadaptation of vectors and parasites has been proposed to explain parasite adaptation on local hosts (Joy et al. 2008).
In conclusion, we demonstrated a high diversity within and between Amazon parasite populations. We also showed that the Amapa population is the most divergent, likely because of many peculiarities associated with this state, such as a low API and multiple malaria vectors. Finally, this work presents molecular markers that could be used to increase our overall understanding of the local malaria situation, which is important in defining regional strategies for disease control.