I.1. BACTERIA: PATHOGENICITY FACTORS – 53
tag. Mutants that are lost have been mutated in genes that have a function in the
pathogenic process, or which at least have a function that is needed to survive and be
retrieved in the experiment.
Ideally, for the identification of virulence, several approaches should lead to the same
gene or set of genes, and the characterisation of a gene as virulence-associated should be
based on evidence from more than one approach. Even then, the controversy between
housekeeping genes and virulence genes is not always solved. For example, the
housekeeping magnesium transport system of Salmonella typhimurium, mgtA/B, is under
PhoP/PhoQ regulation, and is activated during invasion in vitro (Smith and Maguire,
1998). One example is glutamine synthetase of Salmonella typhimurium, which is under
the regulation of ntrC (an alternative sigma factor that can be indicative for in vivo
regulation of expression) and which was identified as a virulence gene based on
phenotypic evidence, since inactivation resulted in attenuation (Klose and Mekalanos,
1997). The enzyme presumably provides glutamine to the organism while surviving in the
host, and could for that reason be considered a virulence-associated gene that enables
colonisation. Since glutamine synthetase is also present in non-pathogenic bacteria, it is
not considered a virulence gene in the comparative genetic approach. As the absence of
virulence genes in non-pathogenic bacteria receives a lot of weight in this approach,
two points need to be considered: 1) the outcome of such comparative genetics is heavily
dependent on the content of the databases used; and 2) gene function is not always
correctly predicted by comparative genetics alone. Putative virulence gene candidates
identified in this way should therefore at least be confirmed by phenotypic evidence,
despite the mentioned shortcomings of such evidence.
Trends in virulence gene identification
Due to explosive developments in genomics it is now feasible to analyse the complete
genome of bacterial pathogens by in silicio subtractive hybridisation. With the expanding
annotation of genes from genome sequences, this can lead to the identification of new
virulence genes (Field, Hood and Moxon, 1999; Frosch, Reidl and Vogel, 1998).
The annotation of these newly sequenced genes is based on sequence identity.
This identification of virulence genes by comparative genomics, based on genetic
similarity is, however, risky for several reasons.
An acceptable level of sequence conservation is seen as (indirect) evidence of
conserved function, so that the gene function of a newly sequenced gene is extrapolated
from a well-characterised analogue in another species. However, genes may have a
niche-adapted function in a particular organism, and this may have its effect on the role of
the gene product in virulence. Functional domains may not be conserved (Lee and Klevit,
2000). Therefore, sequence homology does not always predict function, and even when
there is a high degree of genetic conservation between a non-characterised gene and a
known virulence gene, the function of the gene product of the non-characterised gene as a
virulence factor should first be experimentally tested before functional homology is
assigned. Until then, the newly identified virulence gene should be annotated as
“putative”. Misannotation based on “putativism” is quite common, since it is now easier
to generate sequencing data than to experimentally prove a function of the given gene
product.
Another, diametrically opposed, pitfall of comparative genetics is that genes that
share no sequence homology can have identical functions, as is demonstrated for actA of
Listeria monocytogenes and IcsA in Shigella flexneri, whose gene products recruit host