16. Evolution of Reproductive Organs in Vascular Plants
251
the AGL12, AGL15, and AGL17 groups are likely to have diversified approximately 340 million years ago, indicating that these genes originated in a seed plant
lineage after the branching of ferns from the Trimerophytophyta, approximately
400 million years ago. Therefore, at least nine (AGL2, AGL6, API, AGL14, AG,
AP3-PI, AGL12, AGL15, and AGL17) groups had already diversified in the common ancestor of gymnosperms and angiosperms (Fig. 3).
The A-, B-, and C-function genes include members of API, AP3-PI, and AG
groups, respectively, and manage development of sepals, petals, stamens, and
gynoeciums in angiosperms. It is intriguing that orthologous genes of API have
not been reported from gymnosperms. Members of the API group have indispensable roles in petal and sepal development. The lack of these organs in gymnosperms may be related to the lack of API orthologs in the gymnosperm lineage
(Fig. 3). In Arabidopsis and Antirrhinum, the loss of function mutants of genes in
the AP J group form the flowers composed of only stamens and carpels, which are
partly similar to gymnosperm reproductive organs. It should be noted that the mutant flowers are not completely the same as gymnosperm flowers (Mandel et-al.
1992). For example, the apJ mutant flower is hermaphroditic and different from
the monoecious strobilli of gymnosperms. The apJ leal double loss-of-function mutant
in Arabidopsis loses any floral organs (Kempin et al. 1995). These results suggest
that other genes are involved in the evolution of flowers in addition to the A-function gene.
6 Possible Changes Needed for Flower Evolution
Based on the gene tree (Fig. 2) and the discussion above, all of A-, B- and Cfunction genes are likely to have had already diverged in the most recent common
ancestor of gymnosperms and angiosperms. It is odd that the common ancestor did
not have floral organs notwithstanding that it had all the orthologs of the angiosperm
A-, B- and C-function genes. The two most probable explanations are (1) that the
ancestral A-, B- and C- function genes of the common ancestor regulated downstream genes in different ways from the present A-, B- and C- function genes of
angiosperms and/or (2) that the A-, B- and C- function genes themselves were
regulated differently in the common ancestor from that in angiosperms. The former
possibility should be investigated by experiments in which angiosperm B- or Cfunction genes are swapped for their gymnosperm orthologs. If the gymnosperm
genes can complement the angiosperm B- or C-function genes in the angiosperm
loss-of-function mutants (e.g., apetala3, pistilata, and agamous mutants of
Arahidopsis), the down-stream genes regulated by the B- or C-function genes should
be similar between angiosperms and living gymnosperms, and this will demonstrate that their most recent common ancestor possibly had similar genetic cascades. Recently Rutledge et al. (1998) and Tandre et al. (1998) revealed that ectopic expression of the gymnosperm AG ortholog converted sepals to carpels and
petals to stamens, suggesting that the gymnosperm AG ortholog can control the
251
the AGL12, AGL15, and AGL17 groups are likely to have diversified approximately 340 million years ago, indicating that these genes originated in a seed plant
lineage after the branching of ferns from the Trimerophytophyta, approximately
400 million years ago. Therefore, at least nine (AGL2, AGL6, API, AGL14, AG,
AP3-PI, AGL12, AGL15, and AGL17) groups had already diversified in the common ancestor of gymnosperms and angiosperms (Fig. 3).
The A-, B-, and C-function genes include members of API, AP3-PI, and AG
groups, respectively, and manage development of sepals, petals, stamens, and
gynoeciums in angiosperms. It is intriguing that orthologous genes of API have
not been reported from gymnosperms. Members of the API group have indispensable roles in petal and sepal development. The lack of these organs in gymnosperms may be related to the lack of API orthologs in the gymnosperm lineage
(Fig. 3). In Arabidopsis and Antirrhinum, the loss of function mutants of genes in
the AP J group form the flowers composed of only stamens and carpels, which are
partly similar to gymnosperm reproductive organs. It should be noted that the mutant flowers are not completely the same as gymnosperm flowers (Mandel et-al.
1992). For example, the apJ mutant flower is hermaphroditic and different from
the monoecious strobilli of gymnosperms. The apJ leal double loss-of-function mutant
in Arabidopsis loses any floral organs (Kempin et al. 1995). These results suggest
that other genes are involved in the evolution of flowers in addition to the A-function gene.
6 Possible Changes Needed for Flower Evolution
Based on the gene tree (Fig. 2) and the discussion above, all of A-, B- and Cfunction genes are likely to have had already diverged in the most recent common
ancestor of gymnosperms and angiosperms. It is odd that the common ancestor did
not have floral organs notwithstanding that it had all the orthologs of the angiosperm
A-, B- and C-function genes. The two most probable explanations are (1) that the
ancestral A-, B- and C- function genes of the common ancestor regulated downstream genes in different ways from the present A-, B- and C- function genes of
angiosperms and/or (2) that the A-, B- and C- function genes themselves were
regulated differently in the common ancestor from that in angiosperms. The former
possibility should be investigated by experiments in which angiosperm B- or Cfunction genes are swapped for their gymnosperm orthologs. If the gymnosperm
genes can complement the angiosperm B- or C-function genes in the angiosperm
loss-of-function mutants (e.g., apetala3, pistilata, and agamous mutants of
Arahidopsis), the down-stream genes regulated by the B- or C-function genes should
be similar between angiosperms and living gymnosperms, and this will demonstrate that their most recent common ancestor possibly had similar genetic cascades. Recently Rutledge et al. (1998) and Tandre et al. (1998) revealed that ectopic expression of the gymnosperm AG ortholog converted sepals to carpels and
petals to stamens, suggesting that the gymnosperm AG ortholog can control the