The genetic factors underlying Francisella pathogenicity and lifestyle are largely unknown, although recent studies have revealed several genes required for intracellular survival of the bacterium in murine and human systems. Sequencing and comparison of representative strains of each subspecies, such as Schu4, LVS, and OSU18, should further reveal the genes that impart greater virulence to type A strains in humans and should also help define candidate attenuation alleles for vaccine development. The most striking distinction between the type A and B subspecies is the level of genomic rearrangement that has taken place within them. Optical map and gene organization data for two type A strains, ATCC 6223 and Schu4, and two type B strains from different continents, OSU18 (North America) and LVS (Asia), reinforce the observation that while type B strains maintain their syntenic order, type A strains are highly rearranged compared to type B strains and to each other. It is noteworthy that the tnp genes for ISFtu2 may be functional in type A strains but not in type B strains. Genome rearrangements may then be stimulated in type A strains but frozen in type B strains as a consequence.
Despite these differences in gene order, overall gene content in both subspecies is highly conserved. The primary difference in gene content between the strains is that while OSU18 and Schu4 have over 300 and 200 pseudogenes, respectively, only 98 of the pseudogenes are shared. Therefore, each strain expresses over 100 unique genes that may influence its individual pathogenicity level and distribution. The FPI-encoded protein PdpD is one protein that is expressed in type A strains and is disrupted in type B isolates. A murine infection study showed that pdpD
is required for intramacrophage growth of F. tularensis
). Therefore, pseudogenes of pdpD
(both of which have been studied in F. tularensis
) and phospholipase D are at least partially responsible for the differences between type A pathogenicity and type B pathogenicity. Validation of virulence phenotypes for the other unique type A and type B alleles requires further study and would be assisted by comparison of additional Francisella
The worldwide distribution and limited variation of type B F. tularensis
strains indicate that this subspecies emerged recently and spread rapidly (5
) compared to the genetically diverse, North America-restricted type A subspecies. The two clades of the type A subspecies have been associated with specific vector species in the United States, while type B isolates have not followed any such pattern (5
). The widespread transmission of the type B subspecies suggests that among the differences that distinguish it from the type A subspecies are alleles that facilitate type B subspecies dissemination. These differences may be caused by genes uniquely expressed or upregulated in OSU18 that confer to type B strains the ability to be transmitted with fewer host barriers or nutritional requirements. Alternatively, perhaps the genes that prohibit widespread dissemination of type A strains are disrupted or are regulated differently in type B strains, thus relaxing a growth restriction and allowing type B strains to spread. Candidate alleles like the alleles encoding the OSU18 chitinase gene, a pseudogene in Schu4, may enable type B strains to survive in more diversified natural reservoirs (e.g., a greater number of tick subspecies). Further studies of Francisella
and its natural reservoirs are needed to better understand the global distribution of the various subspecies.
The relatively low level of diversity among type B strains facilitates a search for candidate attenuation alleles in LVS. The 374 genes showing OSU18/LVS variation represent the primary pool of candidate attenuation alleles. This population can be narrowed to 209 nonconservative mutations which may cause differences in virulence or attenuation. Further type A and B sequences should narrow this list further as polymorphisms naturally present in virulent Francisella populations can be identified and eliminated from the attenuation candidate list. Studies are under way to identify which of the candidate alleles are, in fact, attenuating. Related studies, not directly involving LVS, have assisted in this process. For example, the FTT0918/FTT0919 fusion mutation found in a spontaneous, attenuated type A mutant is also present in LVS, but the FTT0918 and FTT0919 homologs are intact in OSU18. Therefore, inactivation of FTT0918 is probably one factor that contributes to LVS attenuation. Other factors will be studied as more Francisella genomes become available and as the genetic tools for Francisella continue to improve.