I.1. BACTERIA: PATHOGENICITY FACTORS – 51
Like Koch’s postulates, the “molecular Koch’s postulates” cannot always be applied
rigidly. If the virulence phenotype is multifactorial, as will usually be the case, the gene
products identified as virulence factors may either be a “classical” virulence factor or an
accessory factor that is essential for expression of the phenotype, but not directly
involved in it. As an example: the fimbriae, hairlike surface structures, that are virulence
factors of uropathogenic Escherichia coli strains carry an adhesin molecule at their tip
that performs the directly virulence related task of adherence to epithelial cells of the
host. They can, however, only efficiently perform this task when carried at the tip of the
fimbriae that are composed of other protein molecules that lack the adhesive property.
The gene identified as a virulence factor may not even be a structural gene, coding for a
gene product, but may have a regulatory function in the expression of the structural gene.
In the literature there is a tendency to describe all genes that pass the tests described in the
molecular Koch’s postulates as virulence genes. This approach has resulted in the
identification as “virulence genes” genes that are not directly involved in virulence as
such, but are indispensable for the expression of the virulent phenotype because they are
required in some way for correct expression of virulence genes. In fact, the molecular
approach may detect a whole spectrum of “virulence genes” ranging from “true”
virulence genes to genes encoding “house-keeping enzymes” that through some remote
mechanism influence the virulence phenotype. This may indicate a need for a more
restrictive definition of virulence genes than simply genes that are detected in virulence
assays.
A definition of bacterial virulence should enable the discrimination between “true”
virulence genes that are directly associated with the virulent phenotype, and accessory
genes, that will also be identified as virulence genes by screening methodologies that rely
on gene inactivation resulting in attenuation of virulence. A well-known example of a
housekeeping gene identified as a virulence factor is the gene aroA (as well as other
“housekeeping” genes; see Uzzau et al., 2005), inactivation of which results in
attenuation of pathogenicity. The aroA gene, however, is involved in aromatic amino acid
biosynthesis, and as such is present in both pathogens and non-pathogens and is not
considered a virulence gene. This is easily understood in the case of aroA, but when the
gene product has no known housekeeping function, such genes would be described in the
literature as virulence genes. The problem is where to draw the line in the continuum
between “virulence-associated genes” and “housekeeping genes”. In order to exclude
housekeeping genes from the set of “virulence” genes, the requisite is often added to
Falkow’s molecular postulates that virulence genes should not be expressed outside the
host. However, this would exclude certain well-characterised virulence genes, for
instance lipopolysaccharide (LPS)-producing enzymes are expressed under all
circumstances, and yet LPS is a generally accepted virulence factor. Moreover, lack of
expression outside the host may be a reflection of the applied culture conditions.
In conclusion, the border between virulence-associated genes and housekeeping genes
cannot be sharply defined.
Molecular approaches to identify virulence genes
Three basic approaches are used to identify virulence genes: genetic methods to
obtain phenotypic evidence for virulence, methods based on the proposed
immunogenicity of virulence factors for immunological evidence and comparative
genetic methods that yield additional indirect evidence.