Why don’t mosses flower?

MADS-box and homeobox genes encode transcription factors that are involved in diverse developmental processes. Members of both these large families are present in plants, animals and fungi. Many of the genes which control the vegetative and reproductive development of flowering plants are members of two well defined subclades of the families – the MIKC (MADS-box) and KNOX (homeobox) genes. In two recent publications, Neil Ashton and coworkers report that MIKC-type and KNOX-type genes are also present in the moss Physcomitrella patens (Champagne & Ashton, see pp. 23–36 in this issue; Krogan & Ashton, 2000). This implies that the last common ancestor of mosses and vascular plants, which existed more than 400 million years ago (MYA), already contained homologues of genes that control flower development, generate shoot apical meristems, and are involved in many other ontogenetic processes in extant flowering plants. Functional characterization of these moss genes promises deep insights into the phylogeny of plant developmental control genes and its contribution to the evolution of land plant body plans. ‘Explaining the mechanisms that generated the enormous diversity of plant and animal body plans is still one of the greatest biological challenges’ During evolution, complex multicellular organisms originated several times independently from unicellular ancestors. Striking examples abound within the animals (metazoa) and land plants (embryophytes – such as mosses, ferns and seed plants (Fig. 1)). Explaining the mechanisms that generated the enormous diversity of plant and animal body plans is still a challenge. One of the most promising current approaches to solve the mysteries of macroevolution is evolutionary developmental genetics (‘evo-devo’). Evo-devo assumes that there is a close interrelationship between developmental and evolutionary processes (Gilbert et al., 1996). This is so because even the most complex organisms are produced anew in each generation by developmental processes that usually start with a single fertilized egg-cell. In multicellular organisms, evolution of form is thus always the evolution of developmental processes. Since development is largely under genetic control, changes in developmental control genes may be a major aspect of evolutionary changes in morphology (Gilbert et al., 1996; Theißen & Saedler, 1995; Theißen et al., 2000). Understanding the phylogeny of these genes may thus contribute significantly to understanding the evolution of plant and animal form. The ancestry of some clades of transcription factor genes in the evolution of land plants. A phylogenetic tree of some major taxa of land plants is shown; the topology of the tree is according to recent publications ( Kenrick & Crane, 1997 ; Nickrent et al., 2000 ). Angiosperms (flowering plants) together with the gymnosperms constitute the seed plants. These together with the ferns (including allies such as horsetails) are the vascular plants. The ages (in million years after (MYA)) given at three nodes of the tree are estimates based on fossil or molecular evidence. At the root of the tree a minimal set of MADS-box and homeobox genes that was probably present in the last common ancestor of mosses and vascular plants is shown. These were precursors of the many highly specialized MIKC and KNOX genes we know from extant flowering plants. ‘≥ 1’ symbolizes that the last common ancestor of mosses and vascular plants already had at least one gene of that type. In recent years it has been discovered that many of the key developmental control genes of plants and animals are members of multigene families that encode transcription factors; the homeobox and MADS-box families are arguably the most important for understanding the interrelationships between development and evolution (Gehring, 1992; Theißen et al., 1996, 2000). The defining characteristic of all MADS-box genes is the presence of a highly conserved, approx. 180-nucleotide-long DNA sequence, the MADS-box, which encodes the DNA-binding domain of MADS-domain transcription factors (Theißen et al., 1996, 2000). Similarly, the homeobox genes are defined by the presence of a approx. 180-nucleotide-long DNA sequence, the homeobox, which also encodes a DNA-binding domain, the homeodomain (Gehring, 1992). The MADS-domain and the homeodomain, however, have very different structures and are phylogenetically unrelated. Outside the DNA-binding domains, MADS-domain and homeodomain proteins are structurally diverse. However, like many other eukaryotic transcription factors, they have a modular structural organization, which provides the basis for further subdivision. Most plant MADS-domain proteins known belong to a single clade with a conserved structural organization, the MIKC-type domain structure. It comprises (from the N- to C-terminus) a MADS (M-) domain, which is the major determinant of DNA-binding, but which also performs dimerization and accessory factor binding functions; an intervening (I-) domain, which constitutes a key molecular determinant for the selective formation of DNA-binding dimers; a keratin-like (K-) domain, which promotes protein dimerization; and a C-terminal (C-) domain, which is involved in transcriptional activation, or in the formation of ternary or quaternary protein complexes (Riechmann & Meyerowitz, 1997; Becker et al., 2000; Theißen et al., 2000). By phylogenetic reconstruction, MIKC-type genes can be further subdivided into a considerable number of defined gene clades. Most clade members share highly related functions and similar expression patterns during vegetative or reproductive growth (Theißen et al., 1996, 2000). Plant homeodomain proteins can be classified into four distinctive groups, of which the KNOX proteins are of special interest. KNOX proteins are composed of (again from the N- to C-terminus) a nonconserved region that may contain transcriptional activation domains; a conserved KNOX domain, parts of which may also function in transcriptional activation; a GSE-domain rich in glycine (G), serine (S) and glutamate (E) residues, which may be involved in determining the longevity of the KNOX proteins; an ELK-domain rich in glutamate (E), leucine (L) and lysine (K) residues that may be involved in protein–protein interactions; the homeodomain responsible for DNA-binding; and a short, nonconserved C-terminal region. By phylogenetic reconstruction, KNOX genes can be subdivided into two different clades, termed class 1 and class 2 genes (Reiser et al., 2000). Many of the homeobox genes of animals act as homeotic selector genes, meaning that they are involved in differentiating different body regions from each other, especially along the anterior–posterior axis of the developing animal. They do so by encoding transcription factors that bind to regulatory DNA sequences of their ‘target genes’ and either activate or repress these genes as appropriate for the development of each segment of the animal. These ‘target genes’, or genes further downstream, function as ‘realizator genes’, meaning that they encode the enzymes and structural proteins needed for each body segment of the animal to develop its characteristic identity. Plants also have many homeobox genes, but these seem not to function as homeotic selectors (Meyerowitz, 1997). The KNOX genes are the most well characterized plant homeobox genes. In flowering plants such as maize and Arabidopsis, class 1 KNOX genes are needed for the maintenance of cells in different spatial domains in an undifferentiated, undetermined state. This general function appears to be required for the generation and maintenance of shoot apical meristems (SAMs) and inflorescence meristems (IMs). In line with this, class 1 KNOX genes are expressed in SAMs and IMs, but (with interesting exceptions) not in lateral organ primordia (Reiser et al., 2000). Class 2 KNOX genes exhibit expression patterns that are less restricted and more diverse than those of the class 1 genes. Up to now, no function has been assigned to class 2 genes. The role of homeotic selector genes is taken over in flowering plants by some MIKC-type genes. Especially well known is the role of some of these genes in the specification of floral organ identity (Weigel & Meyerowitz, 1994; Theißen, 2001; Theißen & Saedler, 1999, 2001). The different ‘floral organ identity genes’ (or ‘floral homeotic genes’) are generally expressed in those organs whose identity they specify, such as sepals and petals, or stamens and carpels. However, angiosperm MADS-box genes are also involved in many other developmental processes inside and outside the flower (Theißen et al., 2000). Interestingly, like the homeobox genes of plants, the MADS-box genes of animals also do not function as homeotic selector genes, but have other functions in signal transduction, muscle development and a number of other ontogenetic events (Theißen et al., 2000). Since MADS-box and homeobox genes play such important roles in flowering plant development, understanding the phylogeny of MADS-box and homeobox genes may strongly improve our understanding of the origin and evolution of flowering plant architecture. For example, SAM function and floral organ identity are totally depending on the activity of some homeobox or MADS-box genes, respectively. The origin of some of the most important architectural features of flowering plants therefore can only be understood in the context of homeobox and MADS-box gene phylogeny. Thus the question arises as to when and how the genes arose during evolution that today control the development of SAMs or of floral organs. To answer this, the phylogeny of homeobox and MADS-box genes has to be reconstructed and superimposed on the phylogeny of land plants. Towards that goal, MADS-box and homeobox genes have to be studied in phylogenetically informative taxa, including all major groups of land plants. Extant land plants (embryophytes) are probably a monophyletic group and comprise bryophytes (liverworts, hornworts, mosses) and vascular plants (Kenrick & Crane, 1997). The vascular plants range from relatively simple clubmosses (lycophytes) and ferns to complex seed plants (spermatophytes), comprising gymnosperms and angiosperms (Kenrick & Crane, 1997; Nickrent et al., 2000). Bryophytes are probably not monophyletic, but which of the bryophyte lineages is the most basal is still controversial (Qiu & Palmer, 1999; Nickrent et al., 2000). A widely held view places liverworts as the first branch of the land plant tree, whereas some studies indicate that hornworts are the earliest land plants, with mosses and liverworts jointly forming the second deepest lineage (Kenrick & Crane, 1997; Nickrent et al., 2000). In any case, mosses have a phylogenetic position near the base of extant land plants (Fig. 1) and are therefore crucial for our understanding of the evolution of land plants and their genes. The publications from Neil Ashton's lab (Krogan & Ashton, 2000; Champagne & Ashton, see pp. 23–36 in this issue) report that at least two different MIKC-type MADS-box genes and at least four different KNOX-type homeobox genes are present in Physcomitrella. This provides immediate insights into the evolution of key developmental control genes of land plants. Moreover, the novel findings may open the door to comparative functional studies and thus further our understanding of the genetic mechanisms behind the evolution of land plant architecture and the interrelationships between development, genes and evolution in green plants. The isolated moss genes have the same domain structure as their flowering plant counterparts: thus an N-terminal region, a KNOX-domain, a GSE-domain, an ELK-domain, a homeodomain, and a C-terminal region could be identified in the conceptual PhyscomitrellaKNOX proteins. Also, the Physcomitrella MIKC proteins have MADS-, intervening, keratin-like and C-terminal domains (Krogan & Ashton, 2000) as in their relatives from flowering plants (Theißen et al., 2000), gymnosperms (Becker et al., 2000) and ferns (Münster et al., 1997). In line with this there is a high degree of conservation of intron number and positions between flowering plant MIKC and KNOX genes and the corresponding moss loci. This strongly supports the view that all the MIKC genes are homologous – that is, derived from a common ancestral gene from which the extant moss, fern and seed plant genes have been generated by speciation events (generating orthologous genes), gene duplications (generating paralogous genes), sequence diversification, and fixation. Analogous conclusions hold for the KNOX genes. Domain shuffling was clearly of minor importance during the evolution of the KNOX and MIKC genes. Phylogenetic reconstructions revealed the presence of both class 1 and class 2 KNOX genes in Physcomitrella. This suggests that the gene duplication and sequence diversification events that created both KNOX gene subfamilies occurred before the separation of the moss lineage from the lineage that led to extant vascular plants (Fig. 1), assumed to have occurred approx. 450 MYA. In contrast, no MIKC gene duplication has been reported that predates the separation of the moss and vascular plant lineages. However, representatives of a novel type of MIKC genes, for which LAMB1 from the clubmoss (lycophyte) Lycopodium (Svensson et al., 2000) is the only published member so far, have now also been isolated from Physcomitrella (K. Henschel et al. unpublished data). Thus the complexity of the MADS-box gene family in Physcomitrella is higher than previously assumed (Krogan & Ashton, 2000). The lycophyte and moss genes have structurally deviant I-, K- and C-regions and are therefore termed MI*K*C* genes here. The novel findings suggest that a gene duplication followed by structural changes in at least one of the duplicates has resulted in the existence of two different types of plant MADS-box genes, MI*K*C* and MIKC genes, and that these events predate the separation of the moss and vascular plant lineages, which occurred about 450 million years ago. In summary, the phylogenetic reconstructions are compatible with the assumption that there was a relatively small pool of MIKC and KNOX genes present in the last common ancestor of mosses and vascular plants. A reconstruction of the minimal set of genes that was present lists at least one representative of each, MIKC, MI*K*C*, class 1 and class 2 KNOX genes (Fig. 1). Descendants of that pool of genes have been generated by gene duplications and sequence diversifications. It seems that most (if not all) of the gene duplications that generated the relatively high number of MIKC and KNOX genes in extant flowering plants took place after the split between the moss and vascular plant lineages. Whole clades of MIKC and KNOX genes may therefore be orthologous to corresponding moss genes, but orthology relationships between single moss and flowering plant genes may not exist. Thus mosses certainly have homologues, but probably not orthologues of those genes that control, for example, floral organ identity or SAM function in flowering plants. The similarities in the domain structures between moss and flowering plant MIKC and, likewise, KNOX proteins, suggest that all ontogenetic functions of MIKC and KNOX genes, irrespective of whether they are found in mosses, ferns or seed plants, are based on very similar physico- or biochemical properties, such as capacities for sequence-specific DNA-binding or protein–protein interaction. The findings of Ashton and coworkers are of special interest because Phycomitrella is the moss model system (Cove & Knight, 1993; Reski, 1999), and should seriously facilitate attempts to determine the function of the genes of interest. Not only that Physcomitrella can be stably transformed and is amenable to many other techniques of molecular genetics, but it is actually the only land plant so far in which gene knock-out and allele replacement approaches via homologous recombination are directly accessible to study plant development (Schaefer & Zryd, 1997). The phenotypes of PhyscomitrellaMIKC and KNOX gene knock-outs may give us a clue about the ancestral functions of genes whose extant relatives are now involved in very special and important ontogenetic processes such as generating SAMs and flowers. Since flowers are the hallmark of the flowering plants, and derivatives of the SAM give rise to the whole flowering plant part which is above ground, the developmental processes governed by KNOX and MIKC genes are of crucial importance for the architecture of the flowering plant sporophyte, which is very different from the structure of both the moss sporophyte and the dominating moss gametophyte. Thus by studying the functions of MIKC and KNOX genes in Physcomitrella and comparing them to those of their flowering plant homologues, fascinating new insights into the evolution of developmental control gene functions may be obtained. Why don’t mosses flower? It turns out that like ferns (Münster et al., 1997), mosses have homologues, but probably not orthologues, of floral homeotic genes (Krogan & Ashton, 2000) – but one should not be satisfied with such an answer. We are just at the beginning of understanding the evolution of developmental control genes and morphological features, and which complex relationships exist between both kinds of evolutionary processes. The work by Ashton and coworkers is one of a number of starting points that may turn out to be revealing.