Hox
genes control the body plan of animals along the anterior-posterior axis (Figure 1). Their expression patterns
and protein function determine the identity of each body segment, and changes
in their functions are often responsible for the variations in the number and
structure of body segments in vertebrates.
 |
|
Figure 1. The spatial expression
zones of the Hox genes along the antero-posterior axis of a fly (a) and a mouse
(b). Note the correspondence with the chromosomal order of the genes. This is
referred to as colinearity. (From Carroll, S. et al. 2005, From DNA
to Diversity, Blackwell.)
|
Among
many vertebrates, snakes are fascinating models for studying body plan
evolution: they have more vertebrae than other vertebrates and have ribs even
in the posterior parts of their trunk, where most vertebrates have rib-less
vertebrae. It has been known in mice that Hox6 is expressed in the anterior
trunk and Hox10 in the posterior trunk (Figure
2). The former is an activator of rib formation, and the latter a
suppressor. When mouse Hox10 was ectopically expressed in the anterior trunk,
rib development was inhibited. Interestingly in snakes, Hox6 expression is
extended to the posterior trunk and overlaps with Hox10 expression (Figure 2). Because Hox10 function is dominant over Hox6
function in the rib development in mice, it was hypothesized that snake Hox10
lost its function to repress rib formation.
To test
this hypothesis, Guerreriro and colleagues (1) examined a gene regulatory
network involving Hox10 in a recent
PNAS paper. The authors hypothesized that loss of gene expression or protein
function of Hox10 in snakes might be
responsible for the posterior extension of ribs. However, expression patterns
of Hox10 between mice and snakes were
similar, and the snake Hox10 protein could still repress rib development in
transgenic mice, rejecting the hypothesis.
 |
|
Figure 2. Loss of Hox10
rib-suppressing activity in the trunk of the snake. In mice suppression of Myf5/6
in the posterior part of the body is responsible for rib-less phenotype.
However, in snakes Hox10 cannot suppress Myf5/6
resulting in rib-bearing vertebrae in the trunk. (From
Woltering JM, 2012, Current Genomcis, 13(4))
|
If
it is not Hox10, then are the downstream
target genes of Hox10 responsible for
the phenotypic difference? The authors examined several known downstream target
genes of Hox10. Interestingly, they
found a single base-pair difference in a highly conserved enhancer region of
the snake Myogenic factor 5 (Myf5), a transcription factor gene that
controls skeletal muscle development. In mice, this enhancer is repressed by Hox10 in the posterior trunk and is
activated by Hox6
and Pax3 in the anterior trunk. When
the snake mutation was introduced into the mouse Myf5 enhancer, ectopic reporter expression in the posterior trunk
was observed, which is consistent with the posterior extension of ribs in
snakes.
Is
this change specific to the snake lineage? The authors' survey of the enhancer
sequences from a variety of vertebrate species discovered independent
occurrences of the same mutation in at least two other lineages, elephant/hyrax/manatee and tenrec. Because the former
three species also have longer rib-cages than other mammals, it is possible
that the Myf5 enhancer is a hotspot (2) for morphological evolution.
Note that the title is a bit misleading, because the causal mutation is fixed
in the snake lineage, and therefore is not a polymorphism.
Although
the findings look clear-cut, it was puzzling that the single base-pair mutation
in the snake enhancer is within the binding site of both Hox6 (activator) and Hox10
(repressor). How did the snake enhancer lose only repression, but not both activation
and repression? The answer was quite unexpected: although Hox6 can bind directly
on the enhancer in mice, Hox6 can also indirectly interact with the enhancer
via interaction with Pax3, whose binding is not affected by the mutation.
Because Hox10 does not interact with Pax3, Hox10's interaction with the snake
enhancer is significantly reduced resulting in ectopic expression of Myf5. This result explains why snake Myf5 enhancer lost its repression by Hox10, yet maintained activation by Hox6.
 |
|
Figure 3. Model for rib growth vs.
repression in mice and snakes. In mice, Hox6, an activator of Myf5, is expressed in the anterior
trunk, and Hox10, a repressor of Myf5,
in the posterior trunk. In snakes, Hox6 is expressed in both anterior and
posterior trunks and Hox10 only in the poster trunk. However, neither Hox6 nor
Hox10 can bind directly to the snake Myf5
enhancer due to a single-nucleotide change within the enhancer. Unexpectedly in
snakes, Hox6 still can activate Myf5
with the help of Pax3, whose binding to the enhancer is independent of the
mutation. The newly discovered mechanism of Hox6/Pax3 is also conserved in
mice. (From Mansfield JH, 2013, PNAS)
|
The
present study highlights the power of studying evolution in the framework of
gene regulatory networks: it not only provided genetic clues for morphological
differences, but also deepened our understanding on the underlying molecular
mechanisms. Is functional variation within conserved genes the main underlying
causes of morphological evolution? We need more of these studies to answer this
question.
References
1. Guerreiro I, et al. (2013) Role of a polymorphism in a Hox/Pax-responsive enhancer in the evolution of the vertebrate spine. Proceedings of the National Academy of
Sciences 110:10682-10686.
2. Stern DL (2011) Evolution, development, & the
predictable genome (Roberts and Co. Publishers).
|
