Role of KNOX genes in the evolution and development of floral nectar spurs

A key question in biology is how changes in gene function or regulation produce new morphologies during evolution. The nectar spur is an evolutionarily labile structure known to influence speciation in a broad range of angiosperm taxa. Here, the genetic basis of nectar spur development, and the evolution of differences in nectar spur morphology, is investigated in Linaria vulgaris and two closely related species of orchid, the primitively longer-spurred Dactylorhiza fuchsii, and more derived short-spurred D. viridis (Orchidinae, Orchidaceae). Despite considerable morphological and phylogenetic differences, nectar spur ontogeny is fundamentally similar in each of the study species, proceeding from an abaxial bulge formed on the ventral petal relatively late in petal morphogenesis. However, spur development is progenetically curtailed in the short-spurred orchid D. viridis. In each case spur development involves class 1 KNOTTED1-like homeobox (KNOX) proteins. KNOX gene expression is not restricted to the spur-bearing petal, indicating that additional components are required to define nectar spur position, e.g. canonical ABC genes, determinants of floral zygomorphy, and additional (currently unknown) factors. However, constitutive expression of class 1 KNOX proteins in transgenic tobacco produces flowers with ectopic outgrowths on the petals, indicating that KNOX proteins alone are, to some degree, capable of inducing structures similar to nectar spurs in a heterologous host. Interestingly, KNOX gene expression is high in the ovary of all study taxa, suggesting that KNOX proteins may also have been involved in the evolution of this key angiosperm feature. Although principally involved in maintaining indeterminacy in the shoot apical meristem (SAM), members of the KNOX gene family have been co-opted in the evolution and development of compound leaves where they suppress differentiation and extend the morphogenetic potential of the leaf. A similar model is presented here to explain the role of KNOX proteins in nectar spur development. Co-option of KNOX gene expression to the maturing perianth delays cellular differentiation, facilitating the development of the nectar spur but requiring additional, unknown factors, to determine nectar spur fate. As facilitators of nectar spur development, changes in the spatio-temporal patterns of KNOX gene expression may alter the potential for nectar spur development and explain the critical length differences observed between the orchids D. fuchsii and D. viridis (and among other angiosperm taxa). Taken together, the available data indicate that KNOX genes confer a meristematic state upon plant tissues in a variety of morphogenetic contexts, making the gene family a potentially versatile tool to mediate a wide variety of evolutionary transformations.Thanks for the generous financial support of the Dept. Plant Sciences, University of Cambridge and Corpus Christi College, Cambridge