Behavior is
the most evolutionarily flexible phenotype, and can discriminate
between closely related species in the absence of any
obvious morphological changes. Behavioral characters are
obviously the endpoint of the expression of many genes,
but differences between species may be encoded by single
genes. A clear example of this comes from the study of
the period (per) gene in Drosophila. This gene
encodes a critical component of the circadian 24h clock,
as well as determining the periodicity of an ultradian
60s cycle in the D. melanogaster male's courtship
song. Transfer of the period gene from D. simulans
to D. melanogaster period mutants, generates host
flies which sing with the species-specific 40s lovesong
cycle characteristic of D. simulans males.
The species-specificity
for the lovesong cycle has furthermore been mapped to
a small repetitive region within the central portion of
the per coding sequence (Wheeler et al 1991). Consequently,
in this example, a coding change, rather than a regulatory
change, in a single gene determines all of the species-specificity
in behavior between these two species. The per gene also
determines species-specific patterns of circadian locomotor
activity, in that the D. pseudoobscura per gene
can transfer the pseudoobscura pattern to D. melanogaster
per-mutant hosts (Petersen et al 1988). Chimaeric genes
between these two species show that the N-terminal half
of the per coding sequence encodes the species specific
locomotor pattern. The corresponding experiments with
locomotor patterns have also been performed with the per
gene of Musca domestica, the housefly. Again, the
per gene by itself is entirely responsible for the species-specific
differences in locomotor behavior, and the sequences determining
these profiles are found in the N-terminal part of the
PER product (Piccin et al, in prep).
Finally, the
nonA gene in Drosophila also contributes species-specific
information to the male lovesong. Mutations in this gene
cause both visual and lovesong defects, particularly in
the pulse component of the song in D. melanogaster.
Interestingly, the mutant song is reminiscent of the normal
pattern of D. virilis songs. The D virilis nonA
gene was isolated and transformed into D. melanogaster
nonA mutants. The gene fully restored the mutants' defective
visual behaviour, revealing that rescue was robust, but
the lovesong showed some of the characteristics of D.
virilis pulses. Statistical analysis clearly revealed
differences between D. melanogaster songs and those
of the virilis nonA transformants. However, unlike
per, only a small proportion of the species specific variance
in the song was transferred along with virilis
nonA. Conseqeuntly, nonA is one of probably a number of
song genes which contribute to the species specific patterns
in song signals. In conclusion, the molecular basis of
species-specific behavior has been dissected revealing
that in two phenotypes, lovesong rhythms and circadian
locomotor profiles, species differences are dependent
solely on the per gene. In the case of song pulses, the
nonA gene is one of probably several genes which each
contribute a small proportion of the species-specific
behavioral variation. The latter finding conforms with
the traditional view of speciation, which assumes that
species differences occur by the accumulation of substitutions
at many loci.
