Geographic differentiation in Neisseria meningitidis Daniel J K

Geographic differentiation in Neisseria meningitidis Daniel J. K. A. 1 Wilson , 1 Jolley , R. H. 2 Claus , 1 Urwin , F. M. C. J. 2 Hessler , 1 Maiden , M. N. 2 Frosch 1 Mc. Carthy , and U. 2 Vogel 1. Peter Medawar Building for Pathogen Research, University of Oxford 2. Institute for Hygiene and Microbiology, University of Würzburg 1. Introduction Composition and prevalence of different sequence types (STs) is known to vary with geographic location in populations of carried Neisseria meningitidis. Such geographic differentiation may be the result of gene flow restricted by distance, or may be maintained by natural selection, perhaps in response to the genetic make-up of the host population. Therefore geographic differentiation has the potential to inform us about the rate at which types circulate in the population, and what selection pressures shape the pathogen population. 2. Isolate collection We studied 822 isolates obtained from healthy carriers in Bavaria, Germany, an area that spans approximately 200 miles (320 km) from its north west to its south east corner. Isolates were collected from 17 principal locations. 11 of these were schools whose pupils’ ages ranged from 3 to 25. The other 6 were military barracks (prefixed ‘BW’ in Figure 1) whose recruits’ ages ranged from 17 to 26. Isolates were taken from recruits within two weeks of their arrival at the military barracks. Multilocus sequence typing (MLST) was carried out on all isolates. Figure 1 shows the relative prevalence of clonal complexes found in the study (sequence types that could not be assigned to a clonal complex are excluded). The pie charts differ considerably between sampling locations, indicating that the composition of sequence types varies across sampling locations. 4. Results 20 miles ST-41/44 complex/Lineage 3 ST-23 complex/Cluster A 3 ST-53 complex ST-35 complex ST-198 complex ST-32 complex/ET-5 complex ST-22 complex ST-162 complex ST-254 complex ST-92 complex ST-269 complex ST-11 complex/ET-37 complex ST-18 complex ST-116 complex ST-364 complex ST-376 complex ST-8 complex/Cluster A 4 ST-106 complex ST-334 complex 3. Analysis of geographic differentiation Some variation in prevalence is expected simply from the sampling process. To test the null hypothesis Figure 1. Clonal complex prevalence in Bavaria, Germany. Map © 2001 Microsoft Corp. that isolates collected from any location are a random sample of the whole population, we performed Analysis of Molecular Variance (AMOVA) and Mantel tests. AMOVA estimates FST, which is the proportion of genetic variation that can be attributed to differences between locations, relative to the population at large. The Mantel test estimates rho, which is the correlation between genetic distance and geographic distance. Both test for values significantly different from zero by permutation. AMOVA and the Mantel test can investigate geographic differentiation at three levels: at the nucleotide sequence level (which we concatenated across loci), at the allelic profile level, and at the sequence type (ST) level. Because the allelic profile throws away some information, it emphasizes differences between sequences, so we expect more significant results. This applies to STs even more so. Table 1 shows that there is significant geographic differentiation between Bavarian schools, but not between military barracks. The effect is significant at the nucleotide, allele and ST level. Table 2 shows that genetic distance and geographic distance are positively correlated for schools, but there is no correlation for military barracks. The correlation for schools is significant at the allele and ST level. 5. Conclusions The results are consistent with genetic isolation by distance. Schools, whose catchment areas are small, show a pattern of local differentiation, whereas military barracks, whose catchment areas cover the entire region, do not exhibit geographic differentiation. While the differentiation for schools is significant, its effect is small (FST ~1%) indicating that the differences have arisen recently. The results show that transmission rates within Bavaria are not sufficiently high to eradicate the association between particular STs and locations, but a sample of isolates taken from a single location would be almost as diverse as a sample of isolates collected from the whole population. This work was supported by Deutsche Forschungsgemeinschaft, the German Federal Ministry of Education and Research, and the Senator Kurt und Inge Schuster-Stiftung. The Bavarian Government and the German Armed Forces are gratefully acknowledged for their assistance. DJW is a BBSRC research student, NMc. C is a Wellcome Trust Clinical Training Fellow and MCJM is a Wellcome Trust Senior Research Fellow.