Survey
* Your assessment is very important for improving the workof artificial intelligence, which forms the content of this project
* Your assessment is very important for improving the workof artificial intelligence, which forms the content of this project
Journal of Experimental Botany, Vol. 52, Roots Special Issue, pp. 413±417, March 2001 Evolution and genetics of root hair stripes in the root epidermis Liam Dolan1 and Silvia Costa Department of Cell Biology, John Innes Centre, Norwich NR4 7UH, UK Received 10 June 2000; Accepted 9 September 2000 Abstract Root hair pattern develops in a number of different ways in angiosperm. Cells in the epidermis of some species undergo asymmetric cell divisions to form a smaller daughter cell from which a hair grows, and a larger cell that forms a non-hair epidermal cell. In other species any cell in the epidermis can form a root hair. Hair cells are arranged in files along the Arabidopsis root, located in the gaps between underlying cortical cell files. Epidermal cells overlying a single cortical cell file develop as non-hair epidermal cells. Genetic analysis has identified a transcription factor cascade required for the formation of this pattern. WEREWOLF (WER) and GLABRA2 (GL2) are required for the formation of non-hair epidermal cells while CAPRICE (CPC) is required for hair cell development. Recent analyses of the pattern of epidermal cells among the angiosperms indicate that this striped pattern of cell organization evolved from non-striped ancestors independently in a number of diverse evolutionary lineages. The genetic basis for the evolution of epidermal pattern in angiosperms may now be examined. Key words: Root epidermis, root hair stripes, evolution, genetics, epidermal pattern. Cellular patterning diversity in the root epidermis The root epidermis of most angiosperms is composed of hair cells and non-hair cells that develop in de®ned patterns. Hairs are tip-growing extensions of epidermal cells that play a variety of functions including anchorage, water absorption, nutrient uptake etc. Some species have lost the ability to make root hairs while in other species 1 every cell in the epidermis forms a hair. Exceptional root hairs are found in the Commelinaceae (which includes Tradescantia) where they may originate in the cortex (Pinkerton, 1936). Hairs may fail to develop in older roots of many species with the development of symbiotic mycorrhizae, or their formation may be inhibited by environmental factors. For example, the roots of Elodea canadensis form hairs when in physical contact with a soil substrate, but roots are hairless when the plant is free ¯oating (Cormack, 1937). Nevertheless, hairs are generally epidermal in origin and their patterning re¯ects their mode of development. The patterns of cellular organization in the root epidermis have been described (Leavitt, 1904; Cormack, 1947; Clowes, 2000). The main types are summarized here. Alternate patterns resulting from asymmetric cell divisions Asymmetric cell division in an epidermal cell gives rise to a large cell (atrichoblast) that develops into a hairless epidermal cell and a shorter `specialized' cell that forms a root hair (trichoblast). This pattern of development is widespread among monocot taxa but restricted to a small group of dicots, the paleoherbs (such as water lillies), which recent DNA-based phylogenies have shown to be closely related to the monocots (Chase et al., 1993). Among the monocots there are at least two distinct modes of development associated with asymmetric cell division. In the ®rst case, the daughter cell nearest the meristem forms the root hair (Vd in the Clowes, 2000, notation). Root hairs of the Alismataceae, Hydrocharitaceae, Araceae, Commelinaceae, Typhaceae, Zingiberaceae, Haemodoraceae, and Pontederiaceae develop in this way. In the second mode, the daughter cell furthest from the meristem (Vp) forms a hair cell. The latter pattern is found among the Restionaceae, Juncaceae, To whom correspondence should be addressed. Fax: q44 1603 456844. E-mail: [email protected] ß Society for Experimental Biology 2001 414 Dolan and Costa Cyperaceae, and Poaceae. These families constitute a major derived clade within the monocots (Chase et al., 1995). It is therefore possible that the Vp asymmetric mode of epidermal development arose once in a common ancestor to this group. Examination of epidermal pattern in key groups can be used to test this hypothesis. Random pattern Root hairs can develop in epidermal cells in any position, relative to the underlying cortical cells, and morphologically distinguishable trichoblasts do not form. This pattern of hair cell development is prevalent among the dicots and is found in many monocot taxa. The proportion of cells that develop root hairs depends on environmental factors (Cormack, 1947). Hairs may develop on every epidermal cell, no cells or on a subset of cells (Cormack, 1935; Clowes, 2000). Striped pattern Plants with the striped pattern develop hairs in cell ®les interspersed with ®les of non-hair cells (Fig. 1). Cell ®les (T in Fig. 2) overlying anticlinal cortical cell walls (ACCWs) form root hairs and cells overlying periclinal cortical cell walls (PCCW) (A in Fig. 2) form non-hair epidermal cells. The cells over the ACCWs are shorter and less vacuolated than cells overlying the PCCW because of their slightly shorter cell cycle time. This difference in cell size between the two cell types is visible in the meristem and maintained through the mature region of the root (Fig. 3). This pattern was ®rst described for members of the Brassicaceae (Cormack, 1935; BuÈnning, 1951). It has recently been described in other families including the Capparaceae, Resedaceae, Caryophylaceae, Portulacaceae, Aizoaceae, Salicaceae, Euphorbiaceae, Boraginaceae, Hydrophyllaceae, and Acanthaceae (Clowes, 2000). Interestingly, Onagraceae and Urticaceae contain species with striped and non-striped epidermal patterns (Clowes, 2000). Evolution of the striped pattern Analysis of the pattern of epidermal cells among diverse groups of angiosperms indicates that the striped pattern of hair cell organization evolved independently in a number of lineages (Clowes, 2000) (Fig. 4). The striped pattern evolved at least once within the Capparales (the order that includes the Brassicaceae) after the emergence of the Tropaeolaceae. Within the Capparales, the Brassicaceae, Capparaceae, and Resedaceae exhibit the striped pattern. The Capparaceae and Resedaceae are sister groups of the Brassicaceae and the three taxa form a monophyletic group (Rodman et al., 1998). The striped pattern has also been observed in the Limnanthaceae, but the epidermal cell patterns of groups more derived than the Limnanthaceae such as Fig. 1. Cellular organization of the Arabidopsis root. Scanning electron micrograph showing the organization of epidermal cell types in the epidermis. Orange cells are atrichoblasts and non-hair cells. Blue cells are trichoblasts and root hairs. Fig. 2. Schematic representation of a transverse section through a root in the meristematic zone, showing the position of trichoblasts (T) and atrichoblasts (A) relative to underlying cortical cells. Yellow indicates the position of lateral root cap cells and blue cells are cortical cells. Fig. 3. Meristematic cellular organization of wild-type and mutant roots. (A) Wild-type root with ®les of shorter trichoblasts in the ACCW position (arrowhead) and atrichoblasts in the PCCW position (arrow). (B) Cellular organization of the epidermis of the root of a plant homozygous for the cpc mutation showing that cells are more atrichoblast-like in morphology. (C) Cellular organization of the epidermis of the root of a plant homozygous for the wer mutation showing that cells are more trichoblast-like in morphology. The arrowhead indicates the location of the cell ®le located in the ACCW position and the arrow indicates the location of cells in the PCCW position. Cells were stained in propidium iodide which ¯uorescently stains the intercellular spaces, and imaged with a confocal microscope. Images are presented in reverse contrast to enhance resolution. Root hair pattern in angiosperms 415 Fig. 4. The distribution of species with striped (S) and non-striped (NS) epidermis in the Capparales. The epidermal pattern has not been determined in the Tovariaceae. The phylogeny is based on Rodman et al. (Rodman et al., 1998). Branch lengths are not indicative of distance. the Gyrostemonaceae, Tovariaceae, Pentadiplandraceae, Koeberliniaceae, Bataceae, and Salvadoraceae, have not yet been described. Root hairs can form in any position in the Tropaeolaceae, and this pattern of development is therefore considered ancestral. The simplest explanation is that the striped pattern evolved once among the Capparales in a taxon ancestral to Limnanthaceae but more derived than the Tropaeolaceae. Nevertheless, the possibility cannot yet be ruled out that the striped pattern arose more than once in the Capparales. Characterization of root hair development in other taxa within the Capparales is required to distinguish between single and multiple origin models. This phylogenetic analysis suggests that the derived, striped pattern evolved from an ancestral non-striped state among the Capparales and independently in a number of other dicot families. Alternatively, it is possible that the striped pattern is ancestral and was progressively lost in many clades. The prevalence of the random pattern throughout the whole of the ¯owering plants (monocots and dicots) would suggest that the random patterning is the ancestral condition and it is more parsimonious to suggest that the striped pattern has arisen independently in many plant groups. The development of more reliable phylogenies for these groups, and further characterization of the organization of root epidermal cells in key groups identi®ed by these phylogenies, will be instructive in distinguishing between these alternatives. If the striped pattern evolved a number of times, independently, it will be instructive to determine if the same regulatory genes were involved in morphological change in each case. The characterization of genes required for the development of pattern in the Arabidopsis epidermis is providing useful tools to begin such an analysis. Cellular organization of the Arabidopsis root epidermisÐa model system The Arabidopsis root epidermis consists of 16±24 cell ®les and is derived from a ring of 16 initials that also gives rise to lateral root cap cells (Figs 1, 2, 3A; Dolan et al., 1993). Variation in the number of cell ®les in the epidermis occurs as a result of rare longitudinal anticlinal divisions that take place in trichoblasts (Berger et al., 1998a, b). At the end of the meristematic zone (when cell division ceases), lateral root cap cells die and the epidermis emerges at the root surface. The epidermal cells undergo rapid elongation and initiate hairs when elongation (in the direction of the long axis of the root) ceases. Root hairs develop from trichoblasts that are arranged in ®les overlying the ACCWs (Fig. 2). At maturity hair cells are shorter than the adjacent non-hair epidermal cells and this difference in length can be traced back to the meristem where the two cell types can be easily distinguished. Laser microsurgical experiments indicate that positional information directs cell fate in the epidermis (Berger et al., 1998a). It is likely that this information is in place by the torpedo stage of embryogenesis and maintained in the developing meristem during postembryonic growth of the root. A clonal analysis of epidermal development shows that the positional information may be located in the cell wall, indicating that protoplast±cell wall interactions are necessary for the establishment of cell pattern in the root (Berger et al., 1998a). The molecular basis of this information remains to be de®ned. A cascade of transcription factors regulated the development of epidermal pattern Genetic analysis of epidermal development in Arabidopsis has identi®ed genes required for the development of the characteristic striped pattern of hair cell development. To date, a cascade of transcriptional regulators has been identi®ed that speci®es the identities of cells in the epidermis. CPC (CAPRICE) and WER (WEREWOLF) are the earliest acting genes in this pathway and both are required for cell speci®c transcription of GLABRA2 (GL2), which encodes a homeodomain protein expressed in atrichoblasts required for the development of non-hair 416 Dolan and Costa cells (Di Cristina et al., 1996; Masucci et al., 1996; Wada et al., 1997; Lee and Schiefelbein, 1999). Plants homozygous for loss of function mutations in WER have a hairy phenotype, i.e. all epidermal cells develop root hairs, suggesting that WER is a positive regulator of non-hair cell development. Epidermal cells in the meristem of plants homozygous for wer are indistinguishable morphologicallyÐthere are no clearly differentiated trichoblasts and atrichoblasts (Fig. 3C), indicating that WER activity is required for the repression of hair cell identity early, in the meristem, before root hairs have formed. The WER protein is a member of the MYB family of transcriptional regulators, suggesting that WER is required for the transcription of genes involved in non-hair cell development and is expressed in non-hair cells. CPC, on the other hand, mutates to a hairlessu decreased hair cell density phenotype. The differences between atrichoblasts and trichoblasts are reduced in plants homozygous for cpc mutation (Fig. 3B). This suggests that CPC is either a positive regulator of hair cell development or a negative regulator of non-hair cell development, i.e. it promotes the development of root hair cells. CPC is also a member of the MYB family of transcriptional regulators but it lacks the transcriptional activator domain, which suggests that it may act as a transcriptional repressor, repressing genes that promote non-hair cell identity. A model has been proposed in which the ratio of the levels of WER and CPC can specify epidermal cell identity (Lee and Schiefelbein, 1999). Cells with high WER:CPC levels develop as non-hair cells and those with lower ratios develop as root hair cells (Lee and Scheifelbein, 1999). A possible target for the WERuCPCmediated regulation is the GL2 gene that encodes a homeodomain, transcriptional regulator. GL2 mutates to a recessive, hairy phenotype, suggesting that GL2 is a transcriptional regulator required for the development of the non-hair cell (Di Cristina et al., 1996; Masucci et al., 1996). Possible roles for CPC and WER in the evolution of hair cell pattern The distribution of hair cell patterns among the angiosperms indicates that the striped pattern characteristic of the BrassicaceaeuCapparaceaeuResedaceae is derived from an ancestral state in which hairs could develop in any epidermal position relative to the underlying cortex. It is possible that changes in gene expression of key regulatory genes accompanied the evolution of the striped trait. At least two hypotheses (there are others) are proposed here to explain the evolution of pattern in the epidermis of the angiosperm root. (1) CPC and WER are expressed in every epidermal cell in the ancestral root. Root hairs develop in each location (over PCCW and ACCW) in the ancestral root and WER and CPC are not involved in the speci®cation of cellular identity in the ancestral species. It is proposed that the expression of these genes could have become restricted to particular cell types (i.e. in `stripes') at the same time as acquiring the ability to promote non-hair cell fate in cells over PCCW. (2) CPC and WER are already exclusively expressed in the epidermal cells located over the PCCW in the ancestral type. These genes then acquired the ability to transcriptionally activate genes that repress hair cell fate in cells in this location. The striped pattern of gene expression therefore already existed in the ancestral type and the cell fate mechanism co-opted the pre-existing pattern. Perspectives The recent deciphering of the molecular basis of the patterning of cell types in the root epidermis provides a mechanistic understanding of the development of pattern at the cellular level in plants. This information can now be used to examine the roles of key regulatory genes in the evolution of epidermal patterns in angiosperms. To meet this challenge, more detailed information is needed about the cellular patterns in the root epidermis in species from a number of key taxa. For example, if the epidermal cell patterns of some key families within the Capparales were known, it could more on®dently be stated how many times the striped pattern evolved in this group of plants. Having identi®ed important regulatory genes in Arabidopsis it is now important to identify orthologues in other species with different patterns of epidermal development. This will be instructive in understanding the role of these regulatory genes in the evolution and development of cell pattern in these other species. Similarly more detailed knowledge of the patterns of epidermal cells in the commelinoids (grasses, rushes, sedges etc.) will be instructive in terms of how many times the Vp pattern of cell division occurred. Understanding the molecular mechanisms underpinning the development of epidermis with asymmetric divisions is still some way off, but the analysis of root epidermal development in model monocot genetic systems will be instructive in this respect. This combined evolutionary and developmental analysis will offer insights into the molecular mechanism underpinning morphological change during evolution. Acknowledgements We are grateful to Jackie Nugent for comments on the manuscript and Ned Friedman for helpful comments and guidance. Root hair pattern in angiosperms We owe much to two very patient referees and Keith Skene for help in putting a comprehensible manuscript together. We are grateful to the BBSRC and the Gatsby Foundation for funding research in our laboratory. We are grateful to the Nottingham and Ohio Arabidopsis stock centres for seed stocks. References Berger F, Haseloff F, Schiefelbein J, Dolan L. 1998a. Positional information in root epidermis is de®ned during embryogenesis and acts in domains with strict boundaries. Current Biology 8, 421±430. Berger F, Hung C-Y, Dolan L, Schiefelbein J. 1998b. Control of cell division in the root epidermis of Arabidopsis thaliana. Developmental Biology 194, 235±245. È ber die Differenzierungsvorgange in der BuÈnning E. 1951. U Cruciferenwurzel. Planta 36, 126±153. Chase MW, Soltis DE, Olmstead RG, et al. 1993. Phylogenetics of seed plants: an analysis of nucleotide sequences from the plastid gene rbcL. Annals of the Missouri Botanical Garden XX, 528±580. Chase MW, Stevenson DW, Wilkin P, Rudall PJ. 1995. Monocot systematics: a combined approach. In: Rudall PJ, Cribb PJ, Cutler DF, Humphries CJ, eds. Monocotyledons: systematics and evolution, Vol. 2. Kew: Royal Botanical Garden, 685±730. Clowes L. 2000. Pattern in root meristem development in angiosperms. New Phytologist 146, 83±94. Cormack RGH. 1935. Investigations on the development of root hairs. New Phytologist 34, 30±54. 417 Cormack RGH. 1937. The development of root hairs by Elodea canadensis. New Phytologist 36, 19±25. Cormack RGH. 1947. A comparative study of developing epidermal cells in white mustard and tomato roots. American Journal of Botany 34, 310±314. Di Cristina M, Sessa G, Dolan L, Linstead P, Baima S, Ruberti I, Morelli G. 1996. The Arabidopsis ATHB10 (GLABRA2) is a HD-ZIP protein required for repression of ectopic root hair formation. The Plant Journal 10, 393±402. Dolan L, Janmaat K, Willemsen V, Linstead P, Poethig RS, Roberts K, Scheres B. 1993. Cellular organization of the Arabidopsis thaliana root. Development 119, 71±84. Lee MM, Schiefelbein J. 1999. WEREWOLF, a MYB-related protein in Arabidopsis, is a position-dependent regulator of epidermal cell patterning. Cell 99, 473±483. Leavitt RG. 1904. Trichomes of the root of vascular cryptograms and angiosperms. Proceedings of the Boston Society of Natural History 31, 273±313. Masucci JD, Rerie WG, Foreman DR, Zhang M, Galway ME, Marks MD, Schiefelbein JW. 1996. The homeobox gene GLABRA2 is required for position dependent cell differentiation in the root epidermis of Arabidopsis thaliana. Development 122, 1253±1260. Pinkerton ME. 1936. Secondary root hairs New Phytologist 98, 147±158. Rodman JE, Soltis PS, Soltis DE, Systma KJ, Karol KG. 1998. Parallel evolution of glucoseinolate biosynthesis inferred from congruent nuclear and plastid gene phylogenies. American Journal of Botany 85, 997±1006. Wada T, Tachibana T, Shimura Y, Okada K. 1997. Epidermal cell differentiation in Arabidopsis determined by a Myb homolog. CPC Science 277, 1113±1116.