The ecology of Escherichia coli - Escherichia coli: Pathotypes and Principles of Pathogenesis - page 12

1975 ; Yates, 2007 ; Ingle et al., 2011 ). However, 1980s' studies in the tropics
revealed high E. coli densities in environments where there were no obvious
sources of fecal contamination ( Carrillo et al., 1985 ; Fujioka et al., 1988 ;
Rivera et al., 1988 ). High E. coli counts were then observed in subtropical
environments ( Solo-Gabriele et al., 2000 ) and then in a variety of temperate
environments ( Alm et al., 2003 ; Haack et al., 2003 ; Whitman et al., 2003 ).
These studies indicated that there are environments where fecal inputs are
only one factor determining E. coli cell densities and, overall, E. coli rates
of cell division can exceed cell death rates. Subsequently, population genetic
studies revealed that similar or identical E. coli genotypes can be recovered
at multiple locations, and at different seasons, from a range of environments
( Gordon et al., 2002 ; Walk et al., 2007 ; Ratajczak et al., 2010 ). Indeed one
study of a river system in Canada found that six genotypes (0.08% of all
genotypes observed) represented 28% of the over 21 000 isolates characterized
( Lyautey et al., 2010 ). Another study found what appear to be season-specific
genotypes ( Jang et al., 2011 ). Thus, it seems while these E. coli may have
originally been host-adapted, they have evolved to the point that their persis-
tence in the external environment may have little to do with fecal inputs.
An Australian study has identified an E. coli strain that appears to be entirely
free-living. This strain (Bloom B1-001) is frequently responsible for elevated
coliform counts (>10 000/100 ml), in recreational lakes and water reservoirs
( Power et al., 2005 ). As many as 2-3 of the 'bloom' events may occur in a
particular water body each year and such bloom events are observed over many
years in the same water body. Near identical variants of this bloom strain have
been isolated from multiple water bodies across southeastern Australia. The
bloom strain can be regularly isolated from water samples at different times
of the year in the absence of elevated counts. The strain has also never been
detected in the feces of any vertebrate. In Lake Burley Griffin, Canberra, this
strain can increase its density over 100 000-fold in less than a week. For these
densities to be achieved as a result of fecal contamination would require that the
lake directly receive all of the feces produced by every resident of Canberra for
over a week. This would also require that every person in Canberra harbored
this E. coli strain.
Genetic analysis of this strain (B1-001) has revealed that, based on its core
genome, it is most closely related to strains of Shigella boydii , S. sonnei , and
S. flexerni (unpublished data). It also has a relatively small genome of 4.6 mb,
more like the genome size of Shigella strains than typical E. coli . Phenotypi-
cally it also resembles Shigella more so than other E. coli . However, the strain
lacks most, if not all, virulence factors associated with Shigella . Thus we have
two groups of strains that, phylogenetically, appear to be very closely related.
Yet one represents what appear to be free-living variants of E. coli , while the
other represents strains that are facultative intracellular parasites. Since there
are no data to suggest that Shigella and other EIEC can be carried asymp-
tomatically by anything other than humans in developed countries, could the