Review Tansley insight Why we need more non-seed plant models Author for correspondence: Stefan A. Rensing1,2 Stefan A. Rensing 1 2 Tel: +49 6421 28 21940 Faculty of Biology, University of Marburg, Karl-von-Frisch-Str. 8, 35043 Marburg, Germany; BIOSS Biological Signalling Studies, Email: stefan.rensing@biologie. University of Freiburg, Sch€anzlestraße 18, 79104 Freiburg, Germany uni-marburg.de Received: 30 October 2016 Accepted: 18 December 2016 Contents Summary 1 V. What do we need? 4 I. Introduction 1 VI. Conclusions 5 II. Evo-devo: inference of how plants evolved 2 Acknowledgements 5 III. We need more diversity 2 References 5 IV. Genomes are necessary, but not sufficient 3 Summary New Phytologist (2017) Out of a hundred sequenced and published land plant genomes, four are not of flowering plants. doi: 10.1111/nph.14464 This severely skewed taxonomic sampling hinders our comprehension of land plant evolution at large. Moreover, most genetically accessible model species are flowering plants as well. If we are Key words: Charophyta, evolution, fern, to gain a deeper understanding of how plants evolved and still evolve, and which of their hornwort, liverwort, moss, Streptophyta. developmental patterns are ancestral or derived, we need to study a more diverse set of plants. Here, I thus argue that we need to sequence genomes of so far neglected lineages, and that we need to develop more non-seed plant model species. revealed much, the exact branching order and evolution of the I. Introduction nonbilaterian lineages is still disputed (Lanna, 2015). Research on animals has for a long time relied on a number of The first (small) plant genome to be sequenced was of THE traditional model organisms, such as mouse, fruit fly, zebrafish or model plant, the weed Arabidopsis thaliana (c. 130 Mbp), worm – all bilaterian Metazoa. One of the reasons for this palette of followed by the c. 500 Mbp (average sized) rice genome. organisms was the need to have models for human diseases via Number three was a tree, poplar, and number four the moss orthologue studies. In addition to these species, the genomes of Physcomitrella patens. Together with the genome of the which were sequenced early on, the last decade has seen the unicellular green alga Chlamydomonas reinhardtii, comparative sequencing of many animal genomes due to their informative genomics allowed us to infer when and how many of the major phylogenetic position, enabling evolutionary developmental (evo- molecular adaptations of plant evolution had taken place devo) studies and inference of ancestral states. For example, (Rensing et al., 2008). Since then, many plant genomes have comparative genomics revealed that the interaction of home- been sequenced; however, most of them are angiosperms. odomain (HD) transcription factors (TF) of the HOX and TALE Although those cover the majority of extant plant taxa, there are subfamilies necessary for patterning probably evolved in the last other species-rich lineages as well, in particular ferns, mosses common ancestor of the Eumetazoa (Hudry et al., 2014). Although and liverworts. Similar to animal research, it is important to many sister lineages of bilaterian animals were sequenced and cover the less species-rich, often neglected lineages as well. Ó 2017 The Author New Phytologist (2017) 1 New Phytologist Ó 2017 New Phytologist Trust www.newphytologist.com New 2 Review Tansley insight Phytologist genomes are of flowering plants (Fig. 1). Thus, our knowledge of II. Evo-devo: inference of how plants evolved how land plants evolved is severely biased, akin to looking through a An important change that occurred early in land plant evolution distorting lens. Fortunately, the past few years have seen attempts to was the establishment of a multicellular diploid ‘generation’, the close the huge gaps in land plant phylogeny by genome sequencing, sporophyte, and hence the plant-specific alternation of generations. but projects have not yet been started for all lineages (Fig. 1). Probably, the sporophyte evolved from dormant zygotes of Arabidopsis is on the extreme fringes of the plant morphospace charophytes by intercalation of mitoses (Lee et al., 2008), leading (Diaz et al., 2016), and Physcomitrella is not typical for the to the plant version of embryogenesis (see Rensing, 2016, for majority of mosses. Evo-devo and comparative genomics review). Control genes of haploid (gametophytic) and diploid approaches using only a small set of organisms carry with them a (sporophytic) generations have been determined by evo-devo high risk of misinterpretation, because individual species are used as studies employing the model moss P. patens (e.g. Sakakibara et al., representative for whole clades. In the following I pick three 2013, 2014), in which both multicellular generations are exper- examples to outline how we can profit from more sequenced model imentally tractable. More recently, genes involved in the alterna- organisms. tion of generations that act at the level of the egg cell or zygote have been identified in the liverwort Marchantia polymorpha (Rovekamp Conifers et al., 2016) and in P. patens (Horst et al., 2016). More and more evidence is emerging that the control of similar structures occurs by Together with angiosperms, gymnosperm trees (in particular orthologues, regardless of the generation in which they are conifers) are an important hallmark of past and current terrestrial expressed. For example, orthologous transcription factors control habitats. Conifers are interesting in terms of socio-economy the cellular protrusion of structures such as rhizoids and root hairs because the oldest and largest individuals known are among them. (Proust et al., 2015), and conducting cells are regulated by the same Moreover, they are an evolutionary enigma because a small number control genes in mosses and flowering plants (Xu et al., 2014). of species covers about half of the land masses. Two thirds of the c. These examples underline the importance of having data from a 1000 gymnosperm species are conifers (Christenhusz et al., 2011). diverse set of model organisms that aid evo-devo studies and Although the average angiosperm genome is 588 Mbp large, the inference of ancestral traits. Having the genome of these organisms average gymnosperm genome size is 15.48 Gbp (Kirst et al., 2003), available greatly aids such studies, and the sequencing of these although most of them are not known to be polyploid – the inflated genomes is now feasible due to much reduced sequencing costs. genome size is thus probably mainly due to high transposon activity. Maybe the longevity of the conifers is rooted in different III. We need more diversity genome structure and function. To date, we do not know much about somatic mutations and epigenetic acclimation of conifers Currently, Phytozome 12 (https://phytozome.jgi.doe.gov) lists 64 (Avramidou et al., 2015). Yet, if we consider a long-living tree as an genomes of plants and algae, and more are available elsewhere (yet assemblage of genetically divergent branches or sections (Burian many are version 1 draft genomes, the completeness of which is et al., 2016), it appears feasible that mechanisms of acclimation and hard to estimate). The problem is that > 95% of the available plant of generation of divergent seed banks exist even in individual trees. (flowering plants) Magnoliophyta Done Asterids [15] Under production Rosids [59] No project yet Basal eudicotyledons [3] Spermatophyta (seed plants) Liliopsida (monocotyledons) [22] (vascular plants) Embryophyta (land plants) Basal angiosperms [1] Tracheophyta Conifers [1] [6] Gnetales Cycadales Streptophyta Ginkgo Monilophytes (ferns) [3] Lycopodiophyta (club mosses) [1] Fig. 1 The plant tree of life. Schematic Bryopsida (mosses) [1] [4] representation of Streptophyta, rooted on the Marchantiophyta (liverworts) [2] branch leading to other eukaryotic groups Anthocerotophyta (hornworts) [1] harbouring plastids. The colour code shows species for which the genome has been Charophyta ZCC [several] sequenced and published (green), is under way (orange), or for which there is no project Charophyta KCM [1] [several] yet (purple). Clades or grades for which the branching order is unclear are shown as Chlorophyta (green algae) [10] multifurcating. Numbers in square brackets Prasinophyta [4] show for how many genomes sequencing is Rhodophyta [3], heterokonts [7], Haptophyta [1], Cryptophyta [1], currently underway, if any. Brackets to the Chlorarachniophyta [1], Dinophyta, Glaucophyta [1] right mark taxonomic groups. New Phytologist (2017) Ó 2017 The Author www.newphytologist.com New Phytologist Ó 2017 New Phytologist Trust New Phytologist Tansley insight Review 3 Due to the large genome sizes, that are also found, for example, in step to land, including an altered cell wall, and land plants evolved ferns, the first genome (of Norway spruce) was only recently after this transition (Harholt et al., 2016). Although Zygnematales sequenced and published (Nystedt et al., 2013), but several more are the closest relatives of land plants, they were apparently are works in progress (http://pinegenome.org/). Yet, there is more secondarily reduced through the course of evolution (Delwiche & to the gymnosperms than conifers – and the age old question of Cooper, 2015), leading, for example, to their peculiar form of which of them is sister to the flowering plants is still unresolved sexual reproduction – conjugation. Charales, however, are mor- (Wickett et al., 2014). Genomic sequences of thus-far neglected phologically