The fruit is a key morphological innovation promoting explosive radiation of angiosperms. In addition, fruit provide an essential source of vitamins, proteins, fibers, and carbon-hydrates in the human diet. Because of natural selection and adaptive evolution, fruit characters such as color, shape, size and texture, have been greatly diversified in angiosperms (Figure 1). In many cases, variation in the fruit character is adaptive to specific dispersal strategies. In the past decades, our understanding of fruit development and the underlying genetic networks has expanded through extensive studies in mouse ear cress (Arabidopsis thaliana) and tomato (Solanum lycopersicum), which are models of dry fruits and fleshy fruits, respectively. However, it remains a fundamental question in biology how fruit diversity evolves at the genetic level.
Figure 1. The diversity of fruits in angiosperms
Brassicaceae plants are mainly found in the north temperate zone with 350 genera and 4000 species. Brassicaceae are also known as Cruciferae since flowers of this family have their four sepals and petals arranged in a cross. In contrast to this consistent floral arrangement, the Brassicaceae family exhibits an extraordinary diversity in fruit shape with individual species forming cylindrical, disc-formed, spherical or heart-shaped structures (Figure 2), thus providing an ideal system to trace the developmental genetics and evolutionary mechanism underlying fruit shape diversification.
Figure 2. Fruit shape diversity in Brassicaceae
Fruits from the Capsella genus develop valves that are extended at the apical end, giving them a heart-shaped appearance. This unique fruit shape is gradually established from an ovate-shaped gynoecium after pollination. Using time-lapse live imaging, we conducted an organ-wide cell behavior analysis on fruit shape morphogenesis and found that the establishment of the heart-like morphology is controlled by both anisotropic growth and dynamic changes in cellular growth and division (Figure 3). At the onset of fruit development (i.e. upon fertilization), cell divisions are widespread in the fruit valves (seed pod walls). Subsequently, cells with a high level of anisotropic growth at the apical part of the fruit contribute to the heart shape by pushing the upper part of the valve outward. Meanwhile, in these anisotropic cells, cell division activities decreased significantly, and active cell divisions were observed at the edge of the valve tips. This complementary pattern of cell division and anisotropic growth generates heterogenicity in cell size along the apical-basal axis.
To get insight into the molecular basis underlying cell growth and division during fruit development, we further compared transcriptomic datasets of the valve tips at the single cell level between two consecutive and critical developmental stages of the heart. Using known maker genes, we enriched the epidermal cells and reconstructed the developmental trajectories. This analysis showed that expression of genes involved in cell divisions are significantly reduced as development proceeds, which is consistent with the live-imaging analysis. Furthermore, when we treated the gynoecium with a cell-cycle inhibitor immediately prior to fertilization, valve tip growth was significantly suppressed. These data indicate that heart-shape fruit development in Capsella requires the maintenance of cell division in apical part of the fruit.
Figure 3. Developmental and evolutionary mechanism of fruit shape diversification
In 2019, we reported that heart-shaped fruit development depends on localized auxin synthesis in the valve tips. Here we tried to understand the connection between auxin maxima and cell division in the valve tips. During plant development, the overlap between auxin maxima and cell divisions are found cells with pluripotent activities, such as shoot apical meristem (SAM), later root primordia and leaflet primordia. We therefore checked the expression of maker genes involved in the stem cell maintenance in the SAM. All the genes we checked display a conserved expression in the SAM as observed in Arabidopsis. To our surprise, the class-I KNOX transcription factor, CrSTM, which controls plant growth and development by maintaining stem cell identity in the meristem, is also expressed in the fruit. CrSTM expression is localized in the region with higher cell division activities, especially at the apical edge of the valves.
Ectopic gene expression is associated with regulatory mutations in promoter regions, which may disrupt the repressive elements that suppress gene expression or generate new cis-elements direct gene expression at new position. In order to identify the cis-elements responsible for the fruit-specific expression of CrSTM, we generated a series of promoter deletions and found a 93-bp segment around 1,000 bp upstream from the transcription start site contains the information that are required for the CrSTM expression in the valve tips. In this region, we identified a 6-bp element that known as the STM-binding site in Arabidopsis. Indeed, subsequent molecular and biochemical analysis shows CrSTM could bind to and activate its own expression thus providing a forward feedback loop to sustain its own expression in the fruit valves. The contribution of CrSTM on fruit morphology was evidenced by several gene editing lines targeting the coding sequences and the STM-binding site. The initiation of CrSTM expression in the fruits depends on two redundant Auxin Response Factors (ARF), CrARF6 and CrARF8, who binding to the CrSTM promoter activate CrSTM expression.
Arabidopsis is a close relative of Capsella, which develops a long cylindrical silique and lacks the expression of the STM (AtSTM) gene in the fruits. A close-up inspection on the AtSTM promoter found the putative STM-binding site has two interspaced base pair mutations, which possibly abolishes its function. We next engineered the AtSTM promoter by including the STM-binding site in situ and generated GFP-tagged transgenic lines. Intriguingly, the fruits of these lines no longer adopt cylindrical shape because of over-proliferation of epidermal cells consequent upon ectopic expression of AtSTM and cell cycle associated genes. Therefore, modifying the expression of STM-orthologous gene has the potential to change fruit shape by hijacking the cell division pathway.
We next tried to explore the prevalence of the STM-based mechanism by sampling multiple Brassicaceae species with novel fruit shapes. We found that the STM-binding site has evolved multiple times independently, and therefore representing a molecular convergence. Remarkably, we observed a perfect correlation between species harboring the STM-binding site and those undergoing a shape change from the gynoecium to the mature fruit. We hypothesize this could be due to post-fertilization cell divisions in the fruits caused – at least in part – by STM expression and autoregulation (Figure 3).
This study uncovered an evolutionary mechanism underlying fruit shape diversification in Brassicaceae. We hypothesize that the origin of the STM-binding site by evolutionary tinkering may represent a particularly efficient mechanism for generating morphological diversity.