| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
 |
ABSTRACT |
 TOP ABSTRACT INTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
|
INTRODUCTION |
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
Many studies have been carried out on surface antigenic markers (i.e., merozoite surface protein-1 [MSP-1] and MSP-2, circumsporozoite surface protein [CSP], and glutamate-rich protein [GLURP]),8 which are a priori under strong selection due to host immune responses.9 However, these studies involved selected markers, rather than neutral ones (e.g., microsatellite loci), which are more appropriate for investigating population structure.10 Recently, 800 microsatellite markers were described for the P. falciparum genome.11 Few studies to date have used these genomic markers to investigate the population genetics of P. falciparum.7,12 A recent study of P. falciparum genetic diversity based on 12 microsatellite loci from 465 blood samples collected in different areas (Africa, South America, and Southeast Asia) described the genetic diversity observed at the level of local populations and on a global scale.7 According to these investigators, P. falciparum exhibits a range of population structures, characterized by strong linkage disequilibrium, low diversity, and extensive population differentiation in low transmission regions, while at high levels of transmission linkage disequilibrium and population differentiation are low, while genetic diversity is high.7
In the present study, we analyzed blood samples collected from individuals living in Pointe Noire, Republic of the Congo and characterized P. falciparum isolates at 28 microsatellite loci. We 1) quantified the level of genetic diversity and analyzed the genetic structure of this local population of P. falciparum; 2) compared its diversity with the species-wide diversity of P. falciparum represented by a cosmopolitan sample of 21 isolates from different continents; and 3) compared its diversity and genetic structure with samples from other countries located in a globally hyperendemic area (Democratic Republic of the Congo, Uganda, and Zimbabwe13). This last comparison is interesting because 10 of the 12 microsatellite markers used by Anderson and others7 are a subset of our 28 markers.13 Despite several characteristics common to all samples, i.e., same markers, all samples situated in areas of high transmission, all samples exhibiting high genetic diversity, our analysis showed that our sample from the Republic of the Congo, similar to their samples from Zimbabwe in contrast to their samples from the Democratic Republic of the Congo and Uganda, had significant levels of linkage disequilibrium. We discuss possible interpretations of this discrepancy and their implications in malaria epidemiology.
|
MATERIALS AND METHODS |
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
Ethical clearance. The Republic of the Congo blood-samples analyzed were collected by the "Centre Médical Sanitaire d’Elf-Aquitaine" in Pointe Noire. This clinic is accessible to all patients and is not restricted to Elf-Acquitaine employees. All patients provided informed consent before donating blood samples to be used in this study. This study was reviewed and approved by the biomedical official of the Centre Médical Sanitaire d’Elf-Aquitaine, in agreement with international ethical guidelines for biomedical research involving human subjects. The legal representatives were informed about the objectives of the study and were included only after providing consent. The other isolates were provided by the Biomedical Committee of Malariology (Caracas, Venezuela), the Laboratoire Central de Virologie (Geneva, Switzerland), the Department of Medical Microbiology (Nijmejen, The Netherlands), and the Centers for Diseases Control and Prevention (Atlanta, GA).
Microsatellite alleles and polymerase chain reaction.
We used 28 microsatellite loci distributed throughout the P. falciparum genome that were chosen from the described microsatellite markers (Table 1
).11 The amplified microsatellite loci were perfect, compound, and interrupted with di-, tri-, hexa-, or nona-nucleotide repeats.
|
Data analyses. Genetic polymorphism was measured by the number of alleles per locus (A) and Nei’s unbiased expected heterozygosity (He) adapted to haploid data using F-STAT, version 1.2.18 Differences in expected heterozygosities were tested by an exact Wilcoxon rank sum test using the S-PLUS software.19
Genotypic linkage disequilibrium (LD = non random association of genotypes occurring at different loci) was tested by the exact probability test performed using GENEPOP software, version 3.2d.20 The null hypothesis is that genotypes at one locus are independent from genotypes at the other locus. This test computed unbiased estimates by randomization (4,000,000 iterations) and by the Markov-chain method for the exact probabilities of random association for all contingency tables corresponding to all possible pairs of loci within each population.
The previous method measures the degree of association between pairs of loci. Haubold and others21 elaborated on another method, first proposed by Brown and others,22 that measured non-random association among all screened loci. This method is implemented by the LIAN 3.0 software.23 This software tests the null hypothesis of statistical independence of alleles at all loci. It also computes the distribution of the number of loci at which pairs of haplotypes within a population differ, and then calculates the variance, VD, of these pairwise differences. The sample variance is then compared with the variance expected under linkage equilibrium, VE. A distribution of VE is generated by Monte Carlo simulations, and its percentiles provide 95% confidence intervals.22 We performed 10,000 iterations to generate this distribution. The output file gives VD and VE –values, as well as a standardized index of association (IAS = (VD/VE - 1)/(1 - r), where r is the number of loci,24 a measure of haplotype-wide linkage and the 95% confidence limits determined by Monte Carlo simulations (LMC).23
For the measures of polymorphism (A and He) for multilocus linkage disequilibrium and assignment methods, we considered the Republic of the Congo sample and a cosmopolitan sample of all other isolates pooled, giving a total of 32 and 18 isolates, respectively.
|
RESULTS |
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
Allelic distribution and heterozygosities.
The twenty-eight microsatellite loci surveyed were polymorphic in all isolates (Table 1
). The number of alleles observed per locus ranged from three (L04 locus) to 15 (L03 locus). The total number of alleles detected was 236 for the 28 loci. In the sample from the Republic of the Congo, 217 alleles were scored. They represented 92% of all detected alleles. The mean ± SD number of alleles per locus (A ± SD) were 7.75 ± 2.62 for the Republic of the Congo sample and 4.79 ± 1.52 for the cosmopolitan sample.
Unbiased expected heterozygosity (He ± SD) was high in the Republic of the Congo sample (He = 0.786 ± 0.111) and comparable with that found elsewhere in samples from Uganda, Zimbabwe, and the Democratic Republic of the Congo.7 The expected heterozygosity in the cosmopolitan sample was 0.687 ± 0.109.
Linkage disequilibrium. Statistical tests for LD were conducted for all pairs of microsatellite loci; 28 and 201 of the 378 possible tests showed significant results (P < 0.05) in the Republic of the Congo and the cosmopolitan samples, respectively.
None of the significant pairwise associations involved loci located on the same chromosome. Thus, these loci are only statistically linked, not physically linked.
The LD estimated by the LIAN program was also significant in the Republic of the Congo sample (observed mismatch variance VD = 5.5, expected mismatch variance VE = 4.3, standardized index of association IAS = 0.01, simulated 5% critical value LMC = 4.9, P = 9.10-4) and in the cosmopolitan sample (VD = 30.1, VE = 5.3, IAS = 0.17, LMC = 6.5, P = 10-4).
The extensive LD detected by both methods in the cosmopolitan sample can be explained by a Wahlund effect, i.e., the mixing of several populations with different genotypic frequencies. The much lower, but significant, linkage disequilibrium detected in the Republic of the Congo sample will be subsequently discussed.
|
DISCUSSION |
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
). The number of alleles per locus was similar between the two studies, and no significant differences were detected between pairwise unbiased expected heterozygosities (P = 0.272, by the exact Wilcoxon rank sum test). It is also worth mentioning that we found no multiple genotypes in the sample from the Republic of the Congo. If we restrict the analysis to the 10 loci common to the previous study7 and our study, this remains true. The absence of multiple genotypes is strong evidence against a clonal structure in the population from the Republic of Congo.
|
After their investigation on P. falciparum population genetic structures, Anderson and others concluded that "The microsatellite data reveal a spectrum of population structures within a single pathogen species. Strong LD, low genetic diversity and high levels of geographical variation are observed in regions of low transmission, while random association among loci, high genetic diversity, and minimal geographical differentiation are observed in regions of Africa, where transmission is intense".7 The pattern exhibited by the sample from the Republic of the Congo surveyed here clearly does not fit this categorization. The discrepancy between our results and those of Anderson and others cannot be explained by a difference in the markers used in the two studies. Indeed, we found significant linkage disequilibria in our sample from the Republic of the Congo even when we considered only the 10 loci shared by the two studies presented in Table 2 (GENEPOP: 3,000,000 iterations; P value of the exact binomial test = 0.024 and LIAN: VD = 2.07, VE = 1.67, IAS = 0.03, LMC = 1.87, P = 0.001).
There are two possible explanations for this discrepancy: selfing and a Wahlund effect. Selfing could explain our results either because our sampling method may have biased sampling in favor of Plasmodium lineages that have undergone at least one generation of self-fertilization or because some amount of selfing, capable of generating small but significant levels of linkage disequilibrium, occurs even in areas of high transmission. A bias could arise because of our sampling method. Indeed, to avoid problems related to genotyping strains in the presence of multiple infections, we chose to use only isolates that after culture displayed a single genotype (see Anderson and others for an alternative method).7 Thus, it could be argued that we used only malaria lineages that resulted from at least one self-fertilization event. However, we do not believe that this method induces a bias for two reasons. First, selfing does not generate linkage disequilibrium but maintains it, i.e., if the whole population is in linkage equilibrium, examining only lineages resulting from a single self-fertilization event will not reveal significant LD. Second, Anderson and others7 showed that their results with respect to LD remained unchanged whether they considered multiple infections or only single infections. Thus, for both a priori and a posteriori reasons, the bias hypothesis cannot explain our results.
The Wahlund effect could also explain our results because although samples were collected at the same site, our results cannot rule out the presence of local population genetic structures for P. falciparum in the Republic of the Congo. Indeed, we have no information on the residence of the patients treated at the clinic where our samples were collected; the patients could live in areas harboring genetically differentiated Plasmodium populations.
Both explanations imply that at some spatial level within the Republic of the Congo P. falciparum mating is not random. Interestingly, one of the populations in Zimbabwe7 exhibited a similar pattern; while situated in a high transmission area, both a priori and based on the percentage of multiple infections and mean number of clones per individual host, this population shows significant levels of linkage disequilibrium, despite a high genetic diversity. In discussing the case of the Zimbabwe sample, Anderson and others7 examined the two explanations mentioned earlier: selfing and a Wahlund effect. The latter was an excellent a priori candidate, since the Zimbabwe sample was actually composed of two sub-samples. However, the two sub-samples were not significantly differentiated, and significant linkage disequilibria were found within both samples. This situation could also arise either from selfing, despite high transmission, or from a Wahlund effect. Indeed, a Wahlund effect could still explain the pattern revealed by the Zimbabwe sample despite the absence of genetic differentiation between the two sub-samples if the two clinics from which the two sub-samples originate receive with equal probability patients from areas harboring genetically differentiated P. falciparum populations. We do not know how plausible this is for the Zimbabwe sample.7 However, this possibility underlines the importance of knowing precisely the geographic origin of samples collected from human hosts to infer the population genetic micro-structure of P. falciparum populations.
At the local population scale, the opportunity for random mating depends on the possibility of several Plasmodium strains (sensu Hastings and Wedgwood-Oppenheim25) to co-infect a mosquito, which depends mainly on parasite diversity within infected human individuals. This, in turn, depends on the parasite density in the area under consideration. Thus, linkage disequilibrium observed in P. falciparum populations seems to be due mainly self-fertilization in regions where only a few different multilocus genotypes are present,25 which leads to a situation considered as genetic clonality.26 How large can the proportion of selfing be in an area of high transmission, and can it generate small but significant levels of linkage disequilibrium?
The proportion of selfing in a P. falciparum population will be equal to the probability of single infections in mosquitoes plus the product of the probability of multiple infections of mosquitoes by the proportion of selfed oocysts in multiply infected mosquitoes. Assuming random mating within mosquitoes,26 this latter quantity will be equal to the sum over all haplotypes of the squares of haplotype frequencies. For example, if two P. falciparum haplotypes infect a mosquito at equal frequencies, on average half of the oocysts will be outcrossed and half will be selfed. Deviations from equal frequencies would lead to lower frequencies of outcrossed oocysts. The data of Taylor27 from Tanzania, a highly endemic area, provide the basis for a rough estimate of the selfing rate. Using polymorphism data of surface proteins, Taylor reported that approximately 30% of the infected mosquitoes carried a single infection, while the mean number of Plasmodium genotypes carried by multiply infected mosquitoes was 2.38 (this latter number is inferred from her table). Assuming that all co-infecting strains are at equal frequencies (thus yielding a minimal selfing rate estimate), and that multiple infections consist of either two or three different genotypes, we obtain a selfing rate of 0.61 corresponding to Fis = 0.44. The selfing rate estimate could be overestimated because we assumed that all co-infecting strains were at equal frequencies, and underestimated because surface proteins polymorphism could have lower resolution than microsatellites,28 thus underestimating the number of coinfecting strains. Whatever its weaknesses, however, these are the only relevant data available. It is at present unclear on theoretical grounds whether this amount of selfing (0.6) can lead to the low but significant LD observed in our sample and in the Zimbabwe sample7 (see the discussion of Vitalis and Couvet).29
The evidence presented here and the Zimbabwe sample7 and the arguments provided to explain it show that even in highly endemic areas and despite high genetic diversity, small deviations from random mating, either through weak population differentiation of partial selfing, may exist. Thus, the generalization of Anderson and others7 should be slightly modified. However, we do believe that the major differences raised by them between areas of high and low endemicity remain valid. Nevertheless, the small deviations from panmixia observed in areas of high endemicity show that our understanding of the population biology of this parasite is still incomplete in such areas. Further advances will probably come from more detailed studies of P. falciparum sampled from mosquitoes.
Received November 24, 2001. Accepted for publication October 2, 2002.
Acknowledgments: We thank Dr. Philip Agnew for suggestions and linguistic assistance, and Dr. Thierry De Meeus for helpful advice and criticism. We are also grateful to Dr. Ludmel Urdaneta (Universidad de Carabobo, Centro de Investigaciones Biomedicas, Estado Avagua, Venezuela) for providing isolates from Venezuela. We also thank two anonymous referees for reviewing and improving the manuscript.
Financial support: This work was supported by the Centre National de la Recherche Scientifique (CNRS) and the Institut de Recherche pour le Développement (IRD).
Authors’ addresses: P. Durand, Y. Michalakis, S. Cestier, B. Oury, M. C. Leclerc, M. Tibayrenc, and F. Renaud, Institut de Recherche pour le Développement, Centre d’Etudes sur le Polymorphisme des Microorganismes, Unité de Rocherche (UMR) Centre National de la Recherche Scientifique-Institut de Recherche pour le Développement (CNRS-IRD) 9926, Unité de Recherche (UR) IRD 062, 911 Avenue Agropolis, BP5045, 34032 Montpellier Cedex 1, France, Telephone: 33-4-67-41-63-33, Fax: 33-4-67-41-62-99, E-mail: durand{at}mpl.ird.fr
|
REFERENCES |
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
This article has been cited by other articles:
|
|
 |
|
  D. Zhong, Y. Afrane, A. Githeko, Z. Yang, L. Cui, D. M. Menge, E. A. Temu, and G. Yan Plasmodium falciparum Genetic Diversity in Western Kenya Highlands Am J Trop Med Hyg, December 1, 2007; 77(6): 1043 - 1050. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 Z. Annan, P. Durand, F. J. Ayala, C. Arnathau, P. Awono-Ambene, F. Simard, F. G. Razakandrainibe, J. C. Koella, D. Fontenille, and F. Renaud Population genetic structure of Plasmodium falciparum in the two main African vectors, Anopheles gambiae and Anopheles funestus PNAS, May 8, 2007; 104(19): 7987 - 7992. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
H. BOGREAU, F. RENAUD, H. BOUCHIBA, P. DURAND, S.-B. ASSI, M.-C. HENRY, E. GARNOTEL, B. PRADINES, T. FUSAI, B. WADE, et al. GENETIC DIVERSITY AND STRUCTURE OF AFRICAN PLASMODIUM FALCIPARUM POPULATIONS IN URBAN AND RURAL AREAS Am J Trop Med Hyg, June 1, 2006; 74(6): 953 - 959. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
F. G. Razakandrainibe, P. Durand, J. C. Koella, T. De Meeus, F. Rousset, F. J. Ayala, and F. Renaud "Clonal" population structure of the malaria agent Plasmodium falciparum in high-infection regions PNAS, November 29, 2005; 102(48): 17388 - 17393. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
M. C. Leclerc, P. Durand, C. Gauthier, S. Patot, N. Billotte, M. Menegon, C. Severini, F. J. Ayala, and F. Renaud From The Cover: Meager genetic variability of the human malaria agent Plasmodium vivax PNAS, October 5, 2004; 101(40): 14455 - 14460. [Abstract] [Full Text] [PDF]
|
|
| |||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
