How to spot a flower


(Author for correspondence: tel +44 1223 333938;


The great diversity of floral form has always fascinated biologists, and many studies have sought to explain this diversity by focussing on the interactions of pollinators with different floral morphologies in a wide range of species. By contrast, our understanding of the developmental genetic mechanisms through which floral diversity is generated is mostly limited to a handful of model species. Antirrhinum majus and Petunia hybrida have proved excellent models in which to explore the generation of floral symmetry (Luo et al., 1996), flower colour (Quattrocchio et al., 1998, 1999; Albert et al., 2011), petal texture (Noda et al., 1994; Perez-Rodriguez et al., 2005) and venation patterning (Schwinn et al., 2006). However, many other aspects of floral diversity merit closer examination, and advances in sequencing and gene expression analysis now allow the study of a wider range of traits in non-model organisms. For example, in this issue of New Phytologist, Martins et al. (pp 958–969) present an exciting analysis of petal spot development in Clarkia gracilis (Onagraceae), a native North American species.

‘Petal spots are intimately linked to plant fitness through their influence on pollinator-mediated pollen transfer and resultant seed set.’

Petal spots are groups of pigmented cells that occur in a defined region or regions of the petal, and are sometimes associated with elaborated epidermal cell morphologies. They are widely distributed among many angiosperm families (Fig. 1), and their effects on pollinators have been studied in several systems, including two different Clarkia species (Jones, 1996; Moeller, 2005; Eckhart et al., 2006). The petal spots of Clarkia gracilis were the subject of a classical genetic study in the 1980s, which demonstrated that the presence and position of the spot were controlled by two independent loci (Gottlieb & Ford, 1988). Martins et al. show that Clarkia petal spots develop through differential regulation of the enzymes involved in anthocyanin synthesis. Essentially, expression of the gene encoding F3′H (flavonoid 3′-hydroxylase) occurs throughout the petal early in development, but production of the resulting red/purple cyanidin/peonidin pigments is limited to the spot area by the spatially restricted expression of DFR2 (encoding dihydroflavonol reductase) (Fig. 2). Martins et al. isolated an additional duplicate copy of DFR, DFR1, that is expressed later in petal development throughout the petal. This allows pigment to be produced everywhere in the petal, but the concurrent late expression of the gene encoding F3′5′H (flavonoid 3′,5′-hydroxylase) means that this pigment is pale pink malvidin. Spot formation is therefore, in part, the product of a duplication at the DFR locus. This was followed by the acquisition of a distinct spatio-temporal expression pattern for each of the duplicated gene copies, combined with temporal regulation of the genes encoding F3′H and F3′5′H.

Figure 1.

Examples of flowering plant families with species that exhibit petal spots. (a) Nemophila maculata (Boraginaceae); (b) Gorteria diffusa ‘Soeb’ (Asteraceae); (c) Cistus cyprius (Cistaceae); (d) Digitalis purpurea (Plantaginaceae); (e) Pelargonium echinatum (Geraniaceae); (f) Sparaxis elegans (Iridaceae); (g) Linum grandiflorum ‘Bright eyes’ (Linaceae); (h) Dendrobium nobile (Orchidaceae).

Figure 2.

Petal spot development in Clarkia gracilis is the result of spatially restricted expression of DFR2 early in petal development, when F3′H expression leads to the production of cyanidin in the petal spot. Later expression of DFR1 and F3′5′H throughout the petal leads to production of malvidin in the rest of the petal. DHK, dihydrokaempferol; DHQ, dihydroquercetin; DFR, dihydroflavonol reductase; DHM, dihydromyricetin; F3′H, flavonoid 3′-hydroxylase; F3′5′H, flavonoid 3′,5′-hydroxylase.

Studies of the genus Clarkia in native habitats have demonstrated that insect pollinators can visually distinguish between spotted and non-spotted flowers. Populations of Clarkia xantiana ssp. xantiana are polymorphic for the presence and number of red petal spots and pollinated by the specialist bee species Hesperapis regularis (Melittidae), Lasiglossum pullilabre (Halictidae) and Ceratina sequoia (Apidae), which all show frequency dependent behaviour in selecting between spotted and non-spotted flowers (Moeller, 2005; Eckhart et al., 2006). By contrast, the spotted form of Clarkia gracilis was shown to set up to 32% more seed than the non-spotted form when grown in a mixed plot (Jones, 1996). These data emphasize the importance of extrinsic factors such as pollinators in understanding the evolutionary processes leading to the formation of petal spots, in addition to understanding genetic regulation. Petal spots are intimately linked to plant fitness through their influence on pollinator-mediated pollen transfer and resultant seed set. Consequently these traits are the target of natural selection, and, by influencing pollinators, may play an important role in establishing species boundaries. Thus, petal spots offer an important opportunity to link natural genetic variation underlying diverse spot phenotypes with the fitness of this phenotypic variation for pollinator attraction. As work continues in diverse species it will become interesting to compare the findings from the Clarkia system with parallel experimental systems such as that of the South African daisy Gorteria diffusa. Gorteria diffusa has raised black petal spots on its ray florets, exists as several floral variants, and has been subjected to a detailed morphological analysis (Thomas et al., 2009). The G. diffusa spot is surprisingly complex, consisting of different cell types and localized pigment deposition. As with Clarkia, pollination data suggest a complex interaction between these petal spots and their pollinator, in this case the small bee-fly Megapalpus capensis (Ellis & Johnson, 2010).

The paper by Martins et al. emphasizes the importance of two key themes in evolutionary biology – transcriptional regulation and gene duplication. Promoters are modular entities, comprising a composite assembly of distinct regulatory elements, which allow transcription of a single gene to be activated at different times and places to exert different functions. For example, in the common pea Pisum sativum, the single copy gene UNIFOLIATA is expressed at the margin of growing leaves, where it controls the formation of characteristic compound leaves, but is also later activated in the floral meristem to trigger flower development (Hofer et al., 1997). Gene duplication can also contribute to this process of functional differentiation through neo- or sub-functionalization. The coding sequences of paralogous copies can diverge and create proteins with distinct properties. Alternatively, the expression of the two copies, now piloted by distinct regulatory regions, can change so that each duplicate ends up with its own unique spatio-temporal expression pattern. While several studies, including that of Martins et al., have revealed the existence of such diverging expression patterns between duplicates, pinning down why and how the two copies are expressed differently is still challenging, even in a post-genomic era. Obtaining whole genome and transcriptome sequences for thousands of individuals has become relatively straightforward, yet it remains difficult to reconstruct transcriptional networks. While conserved protein sequences are easily identified, transcriptional regulatory modules are much more challenging to predict (Farnham, 2009). Thus, our ability to identify regulatory elements and, among them, those accounting for the acquisition of new expression patterns, is one of the next big challenges in evolutionary developmental biology.

By contrast, uncovering the transcriptional regulation responsible for differential expression patterns of pigmentation seems entirely feasible given our detailed understanding of anthocyanin synthetic pathways, and the availability of next-generation sequencing technologies. Anthocyanin synthesis is regulated by a complex comprising R2R3 MYB, bHLH (basic-helix-loop-helix) and WD40 proteins, which are a source of phenotypic variation and have been an essential component in generating plant epidermal cell diversity. While the bHLH and WD40 proteins controlling pigment synthesis are also involved in the production of compounds fulfilling various roles in vegetative tissues, the expression of R2R3 MYB transcription factors is, by contrast, highly tissue- or even cell-type specific: in Petunia hybrida, four R2R3 MYB genes have been shown to regulate different aspects of anthocyanin expression in the flower (Quattrocchio et al., 1998, 1999; Albert et al., 2011). Thus, R2R3 MYB regulators appear to be promising candidates to explain the diversity of spatio-temporal expression patterns of anthocyanin biosynthetic genes observed in Clarkia. In addition to these candidate gene approaches, next generation sequencing technologies now allow the retrieval of differentially expressed regulators on a much greater scale. These technologies can be used to generate transcriptional profiles from tissue-specific developmental stages, to identify genes that regulate part of a biosynthetic pathway or development of a morphological trait, and to reveal the existence of duplicated genes, like those found by Martins et al.

In conclusion, Martins et al. have not only developed a promising system for future research into petal spot biology, but also present exciting insights that are broadly applicable to the development of many morphological traits in non-model systems. In particular, it will be interesting to determine the extent to which the genetic mechanisms identified here are able to explain sub-specific or inter-specific variation in spot morphology within the Clarkia genus as a whole. More broadly, it will also be important to compare genetic regulation of petal spot development in Clarkia with that found in the petal spots of diverse angiosperm lineages. The Clarkia system is an excellent example of the use of natural variation coupled with laboratory-based experimental techniques to explore the genetic regulation of morphological traits with quantifiable impacts on fitness. We hope to see many more such studies in coming years.