would experience a high degree of co-inheritance
with thefirst two enzymes of the pathway. Such a
degree of genetic linkage is also observed in the
biosynthetic pathways of DIBOA in maize, which
spans a distance of 6 cM (Frey et al. 1997 ), and in
the avenacin pathway in oat where a locus for
glucosylation is at 3.6 cM from the core gene
cluster (Qi et al. 2004 ). The presence of an
uncharacterized cytochrome P450 gene in the
region betweenIFS1andIFS2should also be
mentioned, as a yet unknown functional role in
isoflavonoid metabolism cannot be excluded.
14.8 Gene Cluster Formation
It was noted that the terpenoid biosynthetic
pathways for avenacin and thalianol, and by
implication their biosynthetic gene clusters, had
independently evolved from primary metabolism
in monocot and dicot species (Field and Osbourn
2008 ). Also, the three gene clusters for the bio-
synthesis of cyanogenic glucosides in the gen-
omes of cassava, sorghum, and L. japonicus
evolved independently (Takos et al. 2011 ). This
suggests that there is a general evolutionary
mechanism promoting this remarkable genome
organization. We recently proposed that a simple
but comprehensive explanation for gene cluster
formation can be based on a genetic principle
first described by Ronald A. Fisher in his 1930
book“The Genetical Theory of Natural Selec-
tion” (Fisher 1930 ; Takos and Rook 2012 ).
Under certain conditions interacting genetic loci
on the same chromosome will tend to reduce the
recombination frequency between them, result-
ing in ever closer genetic linkage. The situation
under which this occurs requires that alternative
gene or allele combinations are maintained
within the genome of a species. Each combina-
tion of beneficially interacting genes provides a
selective advantage, but the effect of each com-
bination is also counteracted by the effect of the
alternative gene combinations. Although gene
translocations are rare and their direction ran-
dom, the emergence of a modified chromosome
with a reduced recombination frequency between
beneficially interacting loci will replace the ori-
ginal one in a population by contributing more of
thefitter genotypes to each subsequent genera-
tion. A repeated process of gene translocation
and selection for reduced recombination leads to
ever closer physical linkage over evolutionary
time. Such gene translocations leading to the
formation of eukaryotic gene clusters have been
documented in fungi (Wong and Wolfe 2005 ;
Proctor et al. 2009 ; Slot and Rokas 2010 ). We
have argued that the Fisher model for selection
for reduced recombination applies to the evolu-
tion of sex chromosomes in animals and to gene
clusters forflower dimorphism and self-incom-
patibility loci in plants (Takos and Rook 2012 ).
Sex chromosomes evolve under sexually antag-
onistic selection by selecting for reduced
recombination between a sex-determining locus
and loci with alleles that are beneficial for the
corresponding sex (Bergero and Charlesworth
2008 ). Self-incompatibility gene clusters in plant
sexual reproduction minimally consist of a
receptor and its corresponding protein ligand.
Such self-incompatibility loci are subjected to
frequency-dependent selection, a form of bal-
ancing selection, as the rare genotypes in a
population have a reproductive advantage. The
biosynthetic pathways for many plant chemical
defense compounds are also under antagonistic
selection pressures, or phrased in population
genetic terms under balancing selection withfit-
ness varying in time and space. For example,
cyanogenic glucosides in Phaseolus lunatus
provide protection against herbivores but make
the plant more susceptible to fungal infection
(Ballhorn et al. 2010 ). Consequently, chemical
defense polymorphisms, including the presence/
absence of a biosynthetic pathway, are dynami-
cally maintained in natural populations. Such
chemical defense polymorphisms represent the
competing beneficial allele combinations that
promote gene cluster formation. The study of such
natural variation in plant chemical defense on the
genomic and ecological level, for whichL. japo-
nicusprovides a very suitable model system, will
further contribute to our understanding of how
ecological interactions shape eukaryotic genomes.14 Plant-Specialized Metabolism and Its Genomic Organization... 159