Fitness Reduction as Side Effect of
Selection
Genes often simultaneously affect several
functional aspects of an organism, a phenom-
enon called pleiotropy. This comes as no sur-
prise to physiologists who are aware of the
intricate network of metabolic machinery by
which several functions may interact, or to
medical geneticists who often observe that a
disease phenotype associated with a single
hereditary factor involves the malfunctioning
of different organs. Pleiotropy is also fre-
quently apparent in so-called major muta-
tions: mutations with a big effect on one
particular trait of the organism almost always
display ‘side-effects’ on other traits.
Well-studied examples are mutations caus-
ing resistance to antibiotics, pesticides or other
toxic compounds. For example, we recently
isolated spontaneous Aspergillus nidulans
mutants resistant to the fungicide fludioxonil.
When tested on the standard medium without
the fungicide, these resistant strains showed a
10 to 30% decrease in growth rate compared to
the sensitive wild-type strain (Schoustra,
unpublished results). Similarly, Björkman et al.
(1998) isolated spontaneous Salmonella
typhimuriummutants that were resistant to the
antibiotics rifampicin and naladixic acid. In
subsequent competition experiments against
the wild-type strain after injection of the bacte-
ria into mice that were not treated with the
antibiotics, the mutants appeared to be at a
general disadvantage. Therefore, the resis-
tance mutations caused a reduction in viru-
lence. Nevertheless, these mutations will tend
to become common in an environment con-
taining the toxin because of the protection they
provide, despite their negative side-effects.
Do Deleterious Genes Disappear as a
Result of Selection?
Of particular interest in the context of resis-
tance mutations is the fate of resistant strains
or genotypes after termination of the appli-
cation of the relevant antibiotic or pesticide.
On the basis of naïve reasoning one might
expect that the now deleterious allele would
quickly disappear, being selected against
because of its negative side-effects. However,
there is increasing evidence that this scenario
is not very likely. A more probable course of
events involves the selection of mutations
that take away the negative side-effects of
the resistance, thus enabling these strains to
enhance their fitness while retaining their
resistance. In a more general sense, the
experiments discussed below suggest that
evolution will rarely be reversible. Therefore
the view that when a population adapted to
one environment returns to a previous envi-
ronment, evolution will (re)produce the orig-
inal genotypic state, is unlikely to be correct.
A convincing illustration is provided by
the study of Björkman et al.(1998) on antibi-
otic-resistant mutants of Salmonella
typhimurium. They selected seven mutant
strains, resistant to streptomycin, rifampicin,
or nalidixic acid. From these, six had lost
their virulence in mice, due to apparent side-
effects of the resistance mutations. The
mutant strains were then allowed to grow in
mice in the absence of antibiotics and sam-
ples were regularly examined for restoration
of virulence. After several growth cycles all
mutant strains showed restored virulence. In
one case, full restoration of virulence
appeared to result from a true reversion, i.e.
a precise back-mutation to the sensitive state.
But all others had retained their resistance.
In these cases, the restored virulence had
resulted from so-called compensatory muta-
tions – mutations at sites other than that of
the resistance mutation and apparently tak-
ing away some of the side-effects of the resis-
tance mutation that had caused the reduced
virulence. The authors inferred from their
results that reduction in the use of antibiotics
might not result in the disappearance of the
resistant bacteria already present. That this
phenomenon is not restricted to prokaryotes
is shown by recent work in our laboratory.
We have observed compensatory evolution
in fungicide-resistant A. nidulans strains,
restoring growth rate to the original level of
the parental sensitive strain while retaining
the resistance (S.E. Schoustra, unpublished).
At least one study has demonstrated a
similar phenomenon in insects. McKenzie et
al.(1993) studied the establishment of resis-
tance to diazinon in the sheep blowfly Lucilla90 R.F. Hoekstra