PFGE with two discriminatory rare-cutting enzymes NheI and SpeI) was used to identify groups of closely related strains in 134 isolates of serogroup X N. meningitidis from countries in Africa, Europe, and North America. All but 3 of 102 isolates from Africa had similar PFGE patterns (Figure 1, clonal grouping X-I). In contrast, 19 of 32 isolates from Europe and North America had distinct PFGE patterns (Figure 2) that differed from those of the African isolates. However, among the latter 32 strains, similar PFGE patterns were observed for 13 isolates from the United Kingdom, Germany, the Netherlands, and the United States (Figure 2, clonal grouping X-II).

Forty-one isolates, each representing a distinct PFGE pattern, were analyzed by MLST. For bacteria from which multiple isolates with a similar PFGE pattern had been detected, we tested at least one representative from each year and country of isolation. Together with other data in the MLST WEB site (http://www.mlst.net), 39 distinct sequence types (STs) have been found in 50 serogroup X meningococci. The general structure of a neighbor-joining tree of allelic differences resembles a bush, with little phylogenetic structure (Figure 3). However, isolates with similar PFGE patterns were assigned to closely related STs. All 29 clonal grouping X-I isolates analyzed by MLST were in STs ST181, ST182, or ST751 (Figure 3), which differ by one to three of the seven gene fragments (Table). Similarly, all five clonal grouping X-II isolates were in STs 24 and 750, which differ by one of the seven gene fragments (Table). The three unusual African isolates (strain designations D87, D91, and D93) were in ST188, which is very distinct from STs of clonal grouping X-I (Figure 3). These results show that numerous serogroup X isolates from Africa and nearly half the serogroup X isolates from Europe and North America belong to two clonal groupings, while other serogroup X isolates from Europe or North America are quite diverse.

The general population structure of N. meningitidis is panmictic as a result of the frequent import of alleles from unrelated Neisseriae (20,23). Furthermore, several MLST studies have demonstrated that meningococci from healthy carriers are highly diverse (16,26). Phylogenetic trees of different housekeeping genes from N. meningitidis are no more congruent with each other than with random trees (27). Our results for sequence typing of housekeeping genes of serogroup X meningococci also fit this pattern. Phylogenetic analysis of allele differences resulted in a bushlike tree that does not seem to contain any deep phylogenetic information. However, two clonal groupings were found in this otherwise panmictic group of bacteria. The same isolates were assigned to both clonal groupings by two independent methods, MLST and PFGE, indicating that these assignments reflect real genetic relationships and do not depend on the methods used.

Similar concordant genetic relationships were discerned in epidemic serogroup A N. meningitidis by multilocus enzyme electrophoresis (MLEE), random amplification of polymorphic DNA (RAPD), and MLST; population genetic analyses confirmed that the population structure of these bacteria is clonal (28). Concordant groupings were also discerned by MLEE and MLST among the so-called hypervirulent serogroup B and C isolates of the ET-5 complex, ET-37 complex, lineage III, and cluster A4 (16). Although their apparent clonality may reflect an epidemic population structure (23), this possibility has been excluded for epidemic serogroup A meningococci (28). Therefore, multiple clonal groupings exist in N. meningitidis, even though the population structure of most of the species is panmictic.

The population structure of subgroup III serogroup A meningococci seems to represent continual, sequential replacement of fit genotypes by related variants during periods of several years to decades (19). In subgroup III, nine genoclouds, each consisting of a frequent genotype plus its rarer, less fit variants, have been identified during 3 decades of pandemic spread. Our PFGE data for clonal grouping I of serogroup X suggest that clonal grouping X-I also has a genocloud structure. Two sets of PFGE variants (group Ia and Ib), which might each represent a genocloud, were detected in different countries (Mali, Chad, and Niger; and Burkina Faso and Ghana, respectively). Additional analyses of polymorphic genes are necessary to clarify the uniformity of these groups and to test the similarity of their population structure compared with that of subgroup III.

Both serogroup X clonal groupings described here were isolated over decades, on multiple occasions, and from diverse locations. Clonal grouping X-I (1970-2000) was isolated from different countries in West Africa, and clonal grouping X-II (1986-1999) was isolated from Europe and North America. For clonal grouping X-I in Ghana, the disease rate in healthy carriers was estimated to be 3/10,000 (12). Clonal grouping X-1 is thus of considerably lower virulence than serogroup B ET-5 complex bacteria (disease/carrier rate of 2,100/10,000 [29]) or serogroup A subgroup III bacteria during a postepidemic period in a vaccinated population (100/10,000 [9]).

The relationship between bacterial fitness and clonality has not yet been investigated extensively in natural isolates. Variation in virulence between bacterial genotypes leads to more uniformity in disease isolates than in carriage organism in Streptococcus pneumoniae (30) and Staphylococcus aureus (31). However, our data suggest that the clonal structure of certain meningococcal genotypes need not reflect virulence but rather is associated with genotypes that are particularly fit at colonizing the nasopharynx and spreading from person to person.

Although clonal grouping X-I bacteria are less virulent than serogroup A and B meningococci, they are still pathogenic. Most strains described here were isolated from asymptomatic carriers or patients with rare endemic cases. However, group Ia caused a meningitis outbreak with >60 cases in 1997 in Niger (11). Group Ib caused a smaller outbreak in 2000 in Ghana (12). These results suggest that X-I meningococci may even be capable of causing epidemics. Meningococci are naturally transformable, and horizontal DNA transfer is frequent in these bacteria (20–22). Meningococcal carriage is usually low in interepidemic periods in Africa (1,12,32,33), offering less opportunity for horizontal genetic exchange, which could account for the low genetic variability in serogroup X meningococci in Africa.

For more than a decade, many countries in the African meningitis belt have vaccinated extensively with A/C polysaccharide vaccines (34). Recently, mass vaccination with conjugated serogroup C vaccines has been implemented in the United Kingdom, and strong initial protection has been reported (35). However, if effective, these vaccines may well select for the spread of bacteria for which they are not protective (36), including unusual causes of disease such as serogroups Y, W135, and X. Capsule switching due to DNA transformation has been documented (37,38), and effective vaccination against serogroups A and C may select for capsule switch variants of fit genotypes expressing a capsular polysaccharide not included in the vaccination program. The recent outbreaks after the 2000 Hajj pilgrimage, caused by W135 ET-37 complex meningococci (39,40), may reflect exactly such selection. These findings support the development of comprehensive conjugate vaccines that include capsular polysaccharides from formerly rare causes of disease such as serogroup X.