Date Published: April 21, 2017
Publisher: Public Library of Science
Author(s): Xin Zhang, Wenzhe Liu, Xin Wang, Meng-xiang Sun.
Angiosperms distinguish themselves from gymnosperms by their ovules that are enclosed before pollination. However, how the ovules get enclosed in angiosperms remains a mystery, especially for Magnoliaceae. The only key to this mystery is finding a series of carpels transitional from fully closed with enclosed ovules to open with naked ovules. We use routine paraffin section technology, LM, SEM to document the morphology and anatomy of carpel variation in Michelia figo (Magnoliaceae). A series of carpel variations within a single flower of Michelia figo (Magnoliaceae) are documented, in which the ovules are exposed in atypical carpels. These atypical and typical carpels for the first time demonstrate clearly how the naked ovule get enclosed. Each atypical carpel, with naked ovules, clearly comprises two parts, namely, subtending foliar part and branches bearing ovules, suggesting that a typical carpel is actually an end-product of the fusion between the ovuliferous branches and subtending foliar parts. The only difference among these carpels is the extent of fusion between these two parts. This generalization is in full agreement with the molecular genetic studies on angiosperm flowers.
The crucial step in the evolution leading to angiosperms is how the formerly naked ovules become enclosed. Although the APG III system has re-shaped the understanding of angiosperms , no sound solution is available for the origin of carpels . According to the widely accepted Euanthium Theory, ovules in the ancestors of angiosperms are borne along the margins of a foliar part (so-called “megasporophyll”) and they become enclosed through longitudinal folding of such “megasporophyll”. Ovules arrangement along ventral the carpel margins in the Magnoliaceae , the assumed uniform anatomy of magnoliaceous flowers carpels , the assumed conduplicate carpels and marginal placentation seen in the Early Cretaceous Archaefructus all seem to favor this hypothesis. At the same time, this hypothesis is facing challenges from various works. For example, studies on the gynoecium of Arabidopsis thanliana indicates that the ovules can develop without carpel wall ; a so-called carpel is found derived from two primordia controlled by two sets of genes, and is supplied by two different and independent vascular systems (one for ovarian wall and one for the placenta) [8–11]; an ovule primordium demonstrates a juxtaposition of expression of REV and STM, similar to a shoot apex ; and carpels of Archaefructus are not conduplicate and do not have marginal placentation . Despite all, magnoliaceous carpels remain especially revealing in term of the nature of “carpels”. Here we document the carpel variations in Michelia figo (Magnoliaceae) to shed some light on this question. The female units in the same flower of Michelia demonstrate two extremely different (atypical and typical) organizations that are bridged by a transitional series. The ovules in atypical female units are exposed and borne on two branches (placenta) independent of the subtending foliar parts, while the ovules in typical female units appear enclosed and borne along the margins of the foliar parts. The occurrence of these two extreme carpel types in a single flower suggests that ovules are oringinally borne on branches that fuse with the margins of the subtending foliar parts in typical magnoliaceous carpels.
Both typical and atypical female units were observed in the gynoecium of Michelia figo, and all these female units can constitute a seamless transitional series (Fig 1d and 1f–1r). Some of these atypical and typical female units occur in a single flower (Fig 1f–1j and 1m–1q). The common feature of the atypical female units is that their ovules are naked and borne on branches (placenta) (Fig 1b–1i and 1m–1q), and each of female units comprises a placenta and a subtending foliar part (Fig 1c, 1d, 1f and 1m–1r). The placenta is clearly separated from the subtending foliar part, and the ovules are on the termini of ovuliferous branches, not along the margins of the subtending foliar parts (Fig 1c, 1d, 1f, 1m and 1n). As the adnation between the placenta and subtending foliar part proceeds, two ovuliferous branches separate from each other and each independently fuses with a margin of the subtending foliar part, resulting in a marginal-appearing placentation (Fig 1h, 1i, 1o and 1p). As the foliar parts enrolls, a ventral suture appears in the adaxial-distal first and extends proximally, while the ovules become increasingly enclosed (Fig 1h–1j and 1o–1q). In typical female units, the foliar parts are completely closed, and the ovules are fully enclosed and appear attached to the margins of the foliar parts (Fig 1k and 1r). Sometimes the foliar parts may have completed its closure but failed to enclose the ovules (Fig 1l). Anatomically, the dorsal and ventral vascular bundles of the fruits are derived from the same vascular bundle, and they all are collateral and amphicribral, respectively, in in the fruits of Michelia figo (Fig 2a–2e). The ovules are supplied by amphicribral placenta bundles (Fig 2a, 2b and 2e).
Placenta and foliar part in female units of angiosperms are two parts supplied by two vascular bundles of different organizations, derived from two former primordia controlled by two different sets of genes. Placenta is supplied by amphicribral vascular bundles, while the ovarian walls are supplied by collateral bundles. This generalization has been confirmed by various independent studies. These observations prompt a re-evaluating on the nature and provenance of angiosperm “carpels”.