| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
Invited Special Papers |
2Duke University, Department of Biology, Box 90338, Durham, North Carolina 27708 USA; 3Department of Plant Biology, Southern Illinois University, Carbondale, Illinois 62901-6509 USA
Received for publication January 5, 2004. Accepted for publication June 15, 2004.
 |
ABSTRACT |
|---|
 TOP ABSTRACT INTRODUCTIONANTHOCEROTOPHYTAMARCHANTIOPHYTABRYOPHYTACONSPECTUSLITERATURE CITED |
|---|
Key Words: Anthocerophyta • Bryophyta • bryophyte phylogeny • hornworts • liverworts • Marchantiophyta • mosses • tree of life
|
INTRODUCTION |
|---|
|
TOPABSTRACTINTRODUCTIONANTHOCEROTOPHYTAMARCHANTIOPHYTABRYOPHYTACONSPECTUSLITERATURE CITED |
|---|
). Within mosses and liverworts, leafy gametophytes are the norm, rivaling the leafy sporophytic growth forms of some tracheophytes, especially lycophytes (Renzaglia et al., 2000
). However, because they depend on water for sexual reproduction, the gametophytes of bryophytes are small relative to most vascular plant sporophytes. Sexual reproduction in bryophytes involves release of motile male gametes into the environment and requires successful navigation of these naked cells from the male to the female sex organs via an external water source.
Sporophytes of bryophytes are without exception monosporangiate and matrotrophic throughout their life span (Graham and Wilcox, 2000
). Ephemeral and dependent on the gametophyte for nutrition and protection, they never exhibit the modular, indeterminate growth form of the gametophyte generation. In their greatest structural complexity, bryophyte sporophytes consist of a nutritive foot, elongating pedicel or seta, and a single terminal sporangium or capsule. Formative divisions in the embryo produce all precursor components of the sporophyte; i.e., distinct embryonic regions are determined to develop into the three organographic zones of the mature sporophyte: sporangium, seta, and foot. In contrast, an apical meristem initial develops in the embryo of tracheophytes and is subsequently responsible for continuous production of repeated shoot and root modules in these plants (Bierhorst, 1971
; Kato and Imaichi, 1997
). Capsules of bryophytes are structurally elaborate and, in some instances, exhibit complicated mechanisms for spore production and dispersal. Basal sporophyte elongation with nonsynchronized spore production in hornworts, elaters in liverworts, and peristomes of mosses provide examples of this complexity.
General treatments of bryophyte morphology can be found in Leitgeb (1874–1881)
, Campbell (1895)
, Goebel (1905)
, Smith (1955)
, Parihar (1965), Watson (1971)
, Puri (1973)
, Richardson (1981)
, Schofield (1985)
, and Crum (2001)
. The Manual of Bryology, edited by Verdoorn (1932)
, contains authoritative treatments of selected bryology topics that summarized the state of our knowledge at that time, and the New Manual of Bryology, edited by Schuster (1984)
, provided expanded updates more than fifty years later. Both manuals are still useful. Other edited volumes on various aspects of bryophyte biology especially relevant to the tree of life include Clarke and Duckett (1979)
, Smith (1982)
, and Shaw and Goffinet (2000)
.
The crucial position of bryophytes in embryophyte evolution
An unambiguous conclusion from the multitude of contemporary phylogenetic investigations of streptophytes is that bryophytes are the first green plants to successfully radiate into terrestrial niches. These small, inconspicuous plants have existed for several hundreds of millions of years and have played a prominent role in shaping atmospheric and edaphic change and the subsequent evolution of all forms of plant life on land. Explorations of life history phenomena in bryophytes and a solid understanding of interrelationships among them are necessary to reconstruct the early evolution of embryophytes.
The concept that the embryo/sporophyte evolved in land plants through intercalation of mitotic divisions between fertilization and meiosis is widely accepted (Graham, 1993
; Graham and Wilcox, 2000
). Based on this axiom, land plant evolution proceeded in the direction of progressively more elaborate sporophytes. Although generally true, unconditional acceptance of this trend leads to conclusions that ignore processes of reduction and parallel/convergent evolution, phenomena that have occurred repeatedly during bryophyte diversification (Schuster, 1992
; Niklas, 1997
; Boisselier-Dubayle et al., 2002
). A defining characteristic of embryophytes is the meiotic production of spores in tetrads and sporopollenin-impregnated spore walls. Because of their resistance to degradation, fossil spores have provided valuable clues to the initial stages of land colonization (Taylor, 1995
; Wellman and Gray, 2000
; Wellman et al., 2003
). The earliest confirmed land plant fossils are spores, speculated to be from an ancient liverwort dating to the middle Ordovician, some 475 million years ago (mya) (Wellman et al., 2003
).
Gametophytes of bryophytes also provide critical clues about land plant evolution. Thalloid and filamentous growth forms are shared with pteridophytes, but the completely subterranean and nonphotosynthetic life histories found in many lycophytes and some ferns show no homology in bryophytes (Bierhorst, 1971
). The achlorophyllous gametophyte of the liverwort Cryptothallus is a recent acquisition within a strictly photosynthetic lineage (Renzaglia, 1982
). Unlike pteridophytes, bryophyte gametophytes frequently show organ development (leaf, stem, and rhizome) and extensive tissue differentiation, including conducting and supportive tissues (Hé bant, 1977
; Ligrone et al., 2000
). Production of multicellular gametangia was an innovation in embryophytes that was a necessary precursor to embryo development (Graham and Wilcox, 2000
). Among land plants, only mosses and liverworts produce superficial gametangia, which are variously protected by elaborate appendages, including leaves. Hornworts sequester vulnerable organs in internal compartments (Renzaglia et al., 2000
; Renzaglia and Vaughn, 2000
).
A lack of intermediate forms in both life history phases and the potential to interpret morphological transitions in opposite directions have obscured relationships among bryophytes and pteridophytes. Understanding morphological evolution requires unambiguous establishment of phylogenetic relationships among and within bryophyte lineages. Over the past decade, great strides have been made toward reaching this goal; however, fundamental questions remain.
In addition to elucidating early patterns of morphological diversification in embryophytes, bryophytes are crucial to understanding plant genome evolution. Approximately 66% of genes identified from expressed sequence tag analyses of gene expression in gametophytes of Physcomitrella patens have homologues in the Arabidopsis genome, consistent with the hypothesis that genes expressed in the diploid plant body of angiosperms were expressed in the gametophytes of early land plants and were recruited for sporophytic morphogenesis later in plant phylogeny (Nishiyama et al., 2003
). Phylogenetic and functional analyses of genes expressed in Physcomitrella gametophytes have clarified the phylogenetic history of several important gene families, including MIKC-type MADS-box genes (Krogan and Ashton, 2000
; Henschel et al., 2002
; Hohe et al., 2002
) and homeobox genes (Champagne and Ashton, 2001
). Phylogenetic analyses of the KNOX (homeobox) gene family across the land plant tree of life have provided insights into the history of gene duplication and functional divergence during embryophyte history (Champagne and Ashton, 2001
). Because KNOX genes are involved in expression of meristematic activity in vascular plant sporophytes, functional analyses of KNOX genes in mosses, liverworts, and hornworts are central to understanding evolution of plant development in embryophytes. Comparable studies of genes involved in flower development are underway, and, in the context of phylogenetic analyses of bryophytes, the early evolution of these genes is now a tractable problem for investigation (Himi et al., 2001
).
Relationships among the three lineages
Relationships among the three lineages of bryophytes remain one of the major unresolved questions in plant evolutionary biology (Goffinet, 2000
). Virtually every conceivable hypothesis has been put forth in regards to primary branching patterns at the base of embryophytes. Most commonly, bryophytes are viewed as a grade of three monophyletic lineages, with an uncertain branching order (Mishler et al., 1994
; Qiu et al., 1998
). Controversy often focuses on which bryophyte group is sister to all other embryophytes, with two hypotheses most frequently supported: liverworts as sister to other embryophytes vs. hornworts as the sister group (Mishler et al., 1994
; Hedderson et al., 1996
, 1998
; Malek et al., 1996
; Garbary and Renzaglia, 1998
; Qiu et al., 1998
; Beckert et al., 1999
; Duff and Nickrent, 1999
; Nishiyama and Kato, 1999
; Soltis et al., 1999
; Nickrent et al., 2000
; Renzaglia et al., 2000
; Stech et al., 2003
). A moss-plus-liverwort clade has been recovered in several of these analyses (Hedderson et al., 1996
, 1998
; Nishiyama and Kato, 1999
; Nickrent et al., 2000
; Renzaglia et al., 2000
). Recently, it was postulated that hornworts, not mosses, are the closest living relative of tracheophytes. This speculation finds support in sequence data as well as in structural genomic features (Samigullin et al., 2002
; Kelch et al., in press
). In contrast, recent analyses of amino acid sequences based on entire plastid genomes provided support for a monophyletic bryophyte assemblage; however, these results must be viewed with caution because of severe limitations in taxon sampling (Nishiyama et al., in press
).
The focus of this review is to present the current state of knowledge on phylogenetic relationships within, not among, hornworts, liverworts, and mosses. Emphasis is placed on synthesizing results of recent molecular investigations that have revolutionized interpretations of genetic and morphological diversification within each of these groups. Intriguing new perspectives on character evolution have emerged from these studies.
|
ANTHOCEROTOPHYTA |
|---|
|
TOPABSTRACTINTRODUCTIONANTHOCEROTOPHYTAMARCHANTIOPHYTABRYOPHYTACONSPECTUSLITERATURE CITED |
|---|
; Leitgeb, 1879
; Campbell, 1895
, 1917
, 1924
; Goebel, 1905
; Lang, 1907
; Bower, 1935
). In no other branch of the green tree of life does extension of each sporophyte involve continuous, presumably indeterminate, basipetal growth of a single elongated sporangium. All stages of spore development, from undifferentiated cells through pre-meiotic/meiotic spore mother cells to sequentially more mature spores, can be found in a single hornwort sporangium. A constant production of spores therefore ensures dispersal throughout the growing season for as long as the gametophyte persists. This mode of sporophyte development has no counterpart in other plant groups, thus obscuring the phylogenetic position of hornworts among green plants.
Hornworts have remained relatively unexplored at all levels of phylogenetic inquiry (Renzaglia and Vaughn, 2000
; Stech et al., 2003
; Duff et al., in press
). The perception that hornworts are invariable, elusive, and difficult to identify has contributed to the paucity of systematic studies within the group. This small, homogeneous assemblage contains 100–150 poorly delineated species (Schuster, 1992
). Concepts of interrelationships among hornworts based on morphology, and the resulting classification schemes, show virtually no consensus at the generic, familial, and ordinal levels (Mishler and Churchill, 1984
; Hasegawa, 1988
; Hässel de Menéndez, 1988
; Schuster, 1992
; Hyvönen and Piippo, 1993
; Renzaglia and Vaughn, 2000
). Twelve genera of hornworts have been named, Anthoceros, Dendroceros, Folioceros, Notothylas, Megaceros, Phaeoceros, Aspiromitus, Hattorioceros, Leiosporoceros, Nothoceros, Mesoceros, and Sphaerosporoceros, of which only the first six are widely recognized.
Even with the advent of molecular systematics and a renewed interest in early land plant phylogeny, hornwort sampling has been sparse, with one to three taxa included in most analyses (Katoh et al., 1983
; van de Peer et al., 1990
; Mishler et al., 1994
; Bopp and Capesius, 1996
, 1998
; Hedderson et al., 1996
, 1998
; Malek et al., 1996
; Qiu et al., 1998
; Beckert et al., 1999
; Duff and Nickrent, 1999
; Nishiyama and Kato, 1999
; Soltis et al., 1999
; Nickrent et al., 2000
). Among the dozens of papers on bryophyte phylogeny over the past ten years, there is only one comprehensive molecular analysis of within-hornwort relationships, based on rbcL gene sequences from 20 hornworts (Duff et al., in press
). A second study utilizing the plastid trnL intron sampled nine hornworts but focused on the position of the group among land plants (Stech et al., 2003
). Results of these analyses are congruent and reveal novel but intuitive relationships. The rbcL analysis provided much greater resolution of hornwort interrelations because of more extensive sampling, including additional species of the five genera included in the trnL study and representatives of three other genera (Folioceros, Leiosporoceros, and Nothoceros). The discussion that follows will focus on taxonomic inferences and morphological character evolution that emerge from scrutiny of the consensus phylogenetic pattern supported by these pioneering studies (Fig. 1).
|
). Growth form (Fig. 2), chloroplast structure and number (Fig. 3), antheridial number and jacket cell organization, Nostoc colony organization, and presence of mucilage canals and thallus outgrowths are taxonomically useful gametophytic characters. Taxonomically informative features of the sporophyte include degree of development of histogenic regions (Fig. 4), spore and pseudoelater architecture and ultrastructure (Fig. 5), and the presence of columella and stomata (Fig. 6).
|
|
|
|
|
, 1988)
segregated it into an autonomous family and order, whereas Hasegawa (1988)
and Schuster (1992)
recognized it as a subgenus of Phaeoceros. In contrast, Hyvönen and Piippo (1993)
supported a sister relationship between Leiosporoceros and Folioceros, based primarily on morphologically similar pseudoleaters (Fig. 5A). More detailed examination of Leiosporoceros dussii reveals morphological and molecular features heretofore undescribed in any hornwort. Moreover, Leiosporoceros gene sequences have extremely low levels of RNA editing (J. Duff, University of Akron, unpublished data) and thus differ from other hornworts that have been shown to have extensive editing (Yoshinga et al., 1996
; Duff et al., in press
). The gametophyte resembles that of Phaeoceros and Megaceros in that it is fleshy and lacks internal mucilage canals. Number of antheridia per cavity is greater than 20, a feature shared only with Anthoceros and Folioceros (Cargill et al., in press
). However, unlike other hornworts in which Nostoc is in discrete spherical colonies within the ventral thallus (Fig. 2B), those in Leiosporoceros occur in branching strands that run longitudinally and are sequestered in the thallus midregion. Ventral mucilage clefts that enable Nostoc to enter and establish colonies in other taxa (Fig. 6B) are lacking in Leiosporoceros. Chloroplasts of Leiosporoceros are also readily differentiated from those in other hornworts. Starch is neatly aggregated around the periphery of the organelle, with a central elaboration of photosynthetic membranes; the chloroplasts have no pyrenoids (Fig. 3A). The sporophyte of Leiosporoceros is elongated and robust, and its anatomy departs significantly from that of other hornworts (Fig. 4A). The suture is highly differentiated and visible as a deep longitudinal groove. The assimilative and sporogenous regions are massive when compared with other hornworts. Several layers of small spore tetrads are surrounded by mucilage and interspersed with groups of large, elongated pseudoelaters (Fig. 5A). Stomata are abundant and appear similar to those in more derived taxa (Fig. 6A). Clearly, the shared traits between Leiosporoceros and other hornworts provide insight about plesiomorphies within the group. For example, stomata and large numbers of antheridia are best interpreted as ancestral hornwort traits. On the other hand, unique morphological traits that characterize Leiosporoceros are presumed autapomorphies and likely reflect the deep evolutionary separation of this genus from other hornworts. Further molecular and morphological studies are required to evaluate these hypotheses.
After Leiosporoceros, Anthoceros plus Folioceros form a clade sister to other hornworts (Fig. 1). Taxonomic treatments have generally recognized a sister relationship between Anthoceros (including Folioceros) and Phaeoceros, placing them in the same family or subfamily. Thus genetic divergence between Anthoceros plus Folioceros and the remaining taxa appears problematic at first glance. Similarities between Anthoceros, Folioceros, and Phaeoceros include rosette-like habits (Fig. 2A), large solitary chloroplasts with well-developed pyrenoids (Fig. 3B), and comparable sporophyte anatomy (Fig. 4B). However, clearly defined features distinguish Anthoceros and Folioceros from other hornworts; these include dorsal lamellae, schizogenous mucilage cavities, and antheridia in large groups of up to 50 per cavity, as compared to 1–4 (–6) antheridia per cavity in other hornworts. Darkly pigmented spores with well-defined trilete marks also serve to differentiate these taxa from other hornworts (Fig. 5B). Molecular evidence that Anthoceros plus Folioceros form a clade sister to all other hornworts (except Leiosporoceros) has reinforced the taxonomic value of diagnostic morphological features that are restricted to these two genera.
Close affinity between Folioceros and Anthoceros is well supported by rbcL data and reflected in most current classifications (Hasegawa, 1988
, 1994
; Schuster, 1992
; Hyvönen and Piippo, 1993
). In contrast, Hässel de Menéndez (1988)
segregated Folioceros into a monotypic family and order based on spore ornamentation and pseudoelaters. Folioceros has thick-walled, reddish brown, highly elongated pseudoelaters, whereas Anthoceros has short, thin-walled multicellular pseudoelaters similar to those of Phaeoceros (Fig. 5D). Additional differences are found in the placenta and chloroplasts of these two taxa (Vaughn et al., 1992
; Vaughn and Hasegawa, 1993
). With only a single species included in the rbcL sequence analysis, it is not possible to evaluate monophyly of Folioceros.
The remaining hornworts form a monophyletic group that includes two well-supported assemblages: Phaeoceros laevis sensu lato (represented in Fig. 1 by P. carolinianus) plus Notothylas and Megaceros plus Dendroceros. A close affinity between Phaeoceros and Notothylas was suggested by Hässel de Menéndez (1988)
, who placed these two genera in the family Notothyladaceae. Both genera have chloroplasts with prominent pyrenoids, spores with an equatorial girdle (Fig. 5C), and 2–4 (–6) antheridia per chamber. However, because of the distinctive sporophyte of Notothylas (Fig. 2D), most systematists have segregated this genus into a monotypic subfamily, family, or order (Singh, 2002
). Notothylas is the only hornwort taxon in which growth of the sporophyte is abbreviated, spore production appears synchronized, stomata are absent, and the columella is normally absent to poorly developed, a combination of characters that indicate affinities with liverworts. Consequently, it has been suggested that the Notothylas sporophyte is plesiomorphic, representing a structural "link" with other bryophytes. Under this interpretation, hornwort radiation involved an elaboration of sporophytes in more derived taxa (Campbell, 1895
; Mishler and Churchill, 1984
; Graham, 1993
; Hyvönen and Piippo, 1993
; Hasegawa, 1994
). An alternative hypothesis, supported by molecular data, is that sporophytes in Notothylas are not representative of the ancestral condition in hornworts but are highly reduced and specialized (Lang, 1907
; Bartlett, 1928
; Proskauer, 1960
; Renzaglia, 1978
; Schofield, 1985
; Schuster, 1992
). Features such as the existence of a relictual and largely nonfunctional suture in some species support the derived nature of the Notothylas sporophyte. If parallel reduction in sporophye complexity occurred among hornwort genera, Notothylas may be polyphyletic (Lange, 1907
; Proskauer, 1960
). An evaluation of this hypothesis requires increased taxon sampling across the hornworts.
Diversity within Phaeoceros is particularly evident in spore morphology (Schuster, 1992
). Phaeoceros laevis s. l., includes species with spiny papillate spores, whereas ornamentation in the remaining species varies from vermiculate to blunt, wartlike projections. As described later, the three representatives of Phaeoceros with vermiculate spores included in molecular analyses are more closely related to Megaceros than to P. laevis s. l.
A close relationship between Megaceros and Dendroceros is evident in morphological characters such as spiraled pseudoelaters (Fig. 5E, F), absence of stomata, and solitary antheridia. The only epiphytic hornwort, Dendroceros, has a thickened, central midrib with perforated wings (Fig. 2B, C); large, central pyrenoids in each plastid; and multicellular spores (Fig. 5E) (Hasegawa, 1980
; Renzaglia and Vaughn, 2000
). Diagnostic features of Megaceros include unicellular green spores with distal mammilla (Fig. 5F), the absence of a pyrenoid, and multiple plastids per cell (Hasegawa, 1983
; Valentine et al., 1986
; Vaughn et al., 1992
). However, as discussed next, the demarcation between Megaceros and Dendroceros is not always well defined, especially with regard to growth form (Proskauer, 1953
; Hässel de Menéndez, 1962
).
A clade containing two species of Dendroceros is sister to a monophyletic assemblage that includes species previously placed in Megaceros, Phaeoceros, Nothoceros, and Dendroceros (Fig. 1). This taxonomically heterogeneous group in turn consists of two clades: the first includes two Old World species of Megaceros, the Austral-Asian M. flagellaris and M. denticulatus (Hasegawa, 1983
; Glenny, 1998
), and the second is an assemblage of species from four generic segregates. Three Phaeoceros species, P. coriaceus (Steph.) Campbell, P. hirticalyx Steph., and Phaeoceros chiloensis (Steph.) Hässel de Menéndez, are included within the second Megaceros clade (Stech et al., 2003
; Duff et al., in press
). All these species have spores with markedly different architecture from those of other Phaeoceros (Fig. 5C). Moreover, these species have monoandrous androecia and multiple chloroplasts that lack pyrenoids, two diagnostic characters of Megaceros (Duff et al., in press
). However, lack of spiraled pseudoelaters, existence of stomata, and yellow, not green, spores are defining features of Phaeoceros. Bartlett (1928)
and Proskauer (1951)
noted that morphological boundaries between Megaceros and Phaeoceros are blurred and that similarities in growth form and chloroplast structure suggest a close relationship between the two. Molecular analyses have indicated that Phaeoceros with vermiculate spores and mamillae on the distal faces are more closely related to Megaceros than to Phaeoceros with papillate spores (P. laevis s. l.) (Fig. 5B).
The crown group of Megaceros consists of M. aenigmaticus, the only North American representative of the genus; M. vincentianus, the only species from the Neotropics; Dendroceros canaliculatus (= M. canaliculatus), and Nothoceros giganteus (= M. giganteus). The last two species have thickened midribs and wings, which accounts for previous, and apparently inappropriate, placements in Dendroceros or the newly delineated Nothoceros. The existence of unicellular, mamillate spores, and plastids devoid of pyrenoids clearly place these species in Megaceros. Thus, scrutiny of the morphology of these seemingly disparate hornwort species reveals features that solidify their inclusion in the Megaceros clade. The well-developed costa and monostromatic wings in these taxa were likely a result of parallel evolution with Dendroceros.
Inferences about morphological evolution in hornworts from molecular analyses
One intriguing feature of hornworts is the large, solitary chloroplast with a prominent pyrenoid, which is shared with green algae but has no parallel in any other embryophyte group. Within hornworts, pyrenoids appear to have been lost multiple times. Similar pyrenoid losses (and gains) have been described in several algal lineages (Hoham et al., 2002
; Nozaki et al., 2002
). In hornworts, chloroplast compartmentalization characterizes several taxa, including Leiosporoceros and certain species of Phaeoceros, Anthoceros, and Megaceros (Burr, 1970
; Valentine et al., 1986
; Vaughn et al., 1992
; Duff et al., in press
). This arrangement is consistent with a carbon-concentrating mechanism typical of organisms with pyrenoids, including other hornworts (Smith and Griffiths, 1996
, 2000
; Hansen et al., 2002
). It has been speculated that the "pyrenoid-like" area evident in certain hornworts represents a transitional state from presence to complete absence of the pyrenoid (Burr, 1970
). An evolutionary inference supported by this interpretation, in addition to the phylogenetic topology presented in Fig. 1, is that a solitary plastid with a pyrenoid is plesiomorphic in hornworts. In Leiosporoceros, the plastid is solitary but without a pyrenoid, the remnant of which is a compartmentalized organelle with peripherally aggregated starch and centralized grana and plastoglobuli (Fig. 3A). Independent losses of the pyrenoid with or without organellar compartmentalization occurred at least once each in Megaceros, Notothylas, and Anthoceros (Vaughn et al., 1992
; Singh, 2002
; Duff et al., in press
). In Phaeoceros hirticalyx, P. coriaceus, and P. chilioensis, species that are probably better placed in Megaceros, loss of the pyrenoid may be interpreted as preceding the evolution of spiraled pseudoelaters and stomatal loss.
As structures that facilitate gas exchange, stomata are important innovations in the diversification of land plants. Their presence in hornworts has been viewed either as a synapomorphy with mosses and tracheophytes or as a homoplastic acquisition within hornworts (Mishler and Churchill, 1984
; Kenrick and Crane, 1997
; Renzaglia et al., 2000
). The presence of stomata in Leiosporoceros, Anthoceros, and Folioceros supports the contention that these structures are plesiomorphic in hornworts and may be homologous to those in mosses and/or tracheophytes. A clear case of homoplasy is the loss of stomata in at least three, possibly four, hornwort lineages: Notothylas, Dendroceros, and Megaceros. Stomatal loss may have accompanied modifications in sporophyte development, e.g., maturation of the sporophyte within the protective gametophytic involucre where gas exchange is limited (Notothylas, Fig. 2D and Dendroceros, Fig. 2C). Stomatal loss in Megaceros is associated with occurrence of these species in periodically inundated habitats. The existence of P. coriaceus, P. hirticalyx, and P. chiloensis, three terrestrial species with stomata, supports this hypothesis. The topology presented in Fig. 1 necessitates at least two losses of stomata in the Megaceros clade.
The interpretation set forth by Proskauer (1951)
and Schuster (1992)
that mucilage clefts on the ventral side of the gametophyte in hornworts (Fig. 6B) are homologous to sporophytic stomata (Fig. 6A) is not supported by molecular analysis. Absence of mucilage clefts in Leiosporoceros and the specialized function of these structures in all other hornworts indicate that gametophytic "stomata" evolved after hornwort diversification simply as an entryway for the cyanobacterium, Nostoc.
It is reasonable to hypothesize that the habit of extant members of Anthoceros, Leiosporoceros, and Phaeoceros represent the ancestral condition in hornworts and may be related to their common occurrence on exposed soil. Morphological diversity in other taxa likely results from radiation into and consequent adaptations to specialized habitats; e.g., Dendroceros is an epiphyte, and Megaceros is restricted to tropical or temperate sites where it often occurs submerged in streams. Diversification of Dendroceros may be correlated with the evolution of angiosperms, which provided abundant new bark and leaf habitats (Ahonen et al., 2003
). Notothylas is an ephemeral hornwort that grows as a pioneer on soil. Unlike other genera in which spores are wind dispersed, Notothylas spores are dispersed by water or facultatively by insects or other animals, thus eliminating the "need" for vertical elongation of the sporophyte.
|
MARCHANTIOPHYTA |
|---|
|
TOPABSTRACTINTRODUCTIONANTHOCEROTOPHYTAMARCHANTIOPHYTABRYOPHYTACONSPECTUSLITERATURE CITED |
|---|
; Schuster, 1984
; Crandall-Stotler and Stotler, 2000
). Within this monophyletic assemblage are several morphologically isolated elements that represent products of deep divergences (Garbary and Renzaglia, 1998
; Renzaglia et al., 2000
). Morphological heterogeneity in the group is particularly evident in growth form of the gametophyte, which shows the greatest range of variability among bryophytes. Since the starting point of liverwort nomenclature (Linnaeus, 1753
) and the beginning of their systematic treatment (Endlicher, 1841
), hepatics have been organized into three groups based on growth form: (1) complex thalloids, (2) simple thalloids, and (3) leafy liverworts. Conflicting concepts of diversification have led to opposing views on the directionality of change within liverworts; that is, whether thalloid or leafy forms are viewed as ancestral (see literature review in Crandall-Stotler and Stotler, 2000
and Davis, in press
[Figs. 7, 8]). Morphological studies supported the concept that simple thalloid liverworts are more closely related to leafy types than to complex thalloids. Classification schemes reflect this interpretation with hepatics typically divided into two groups: marchantioid or complex thalloid liverworts (Marchantiopsida, Marchantiidae) and jungermannioid liverworts, including the leafy (Jungermanniopsida, Jungermanniidae) and simple thalloid taxa (Jungermanniopsida, Metzgeriidae). Complex thalloid types usually have air chambers with dorsal pores and differentiated internal tissues (Fig. 9A). Less commonly, the thallus resembles the simple thalloid type in the lack of internal or epidermal differentiation (e.g., Sphaerocarpos, Monoclea, and Dumortiera). Gametophytes of leafy liverworts range from radially symmetrical with three rows of morphologically similar leaves (isophyllous) to dorsiventral with two rows of lateral leaves and an additional row of reduced (to absent) ventral underleaves or amphigastria (anisophyllous; Fig. 9C). Simple thalloid (metzgerialean) organizations show less variability, from fleshy undifferentiated thalli to those with prominent midribs and monostromatic wings (Fig. 9B). Leaflike lobes or lobules in some taxa blur the distinction between leafy and simple thalloid forms. Internal differentiation of water-conducting tissue is restricted to Haplomitrium and certain simple thalloid taxa, whereas conducting parenchyma is widespread among both complex and simple thalloid forms, but not leafy taxa (Hébant, 1977
; Kobiyama and Crandall-Stotler, 1999
; Ligrone et al., 2000
).
|
|
|
), oil bodies provide valuable taxonomic information because their size, shape, number, and color are taxon specific (Crandall-Stotler and Stotler, 2000
). Unlike hornworts, but comparable to mosses, is the production of a variety of organized external appendages, most of which function in protecting fragile tissues. For example, various mucilage papillae, hairs, scales, bracts, cups, or flask-shaped structures protect the meristem, gemmae, and other vegetative organs. Especially vulnerable are the superficial sex organs that often occur in clusters protected by flaps of tissue, leaf lobes, young leaves, or modified branches (Fig. 9B).
The uniformity and uniqueness of liverwort sporophytes provide compelling evidence for monophylly of hepatics. Unlike mosses and hornworts, sporophytes of liverworts reach maturity within the confines of protective gametophytic tissue that develops from the shoot/thallus (= perigynium or coelocaule) and/or archegonium (= calyptra). Additional gametophytic structures such as perianths, pseudoperianths, bracts, scales, and involucral flaps may further surround the sporophyte and associated protective tissue. In such a milieu, photosynthesis is limited, and the sporophyte derives nourishment from the gametophyte through a placenta. The seta is pale to hyaline, and the capsule is devoid of stomata. The majority of liverwort sporophytes are differentiated into foot, seta, and capsule; in the occasional marchantioid taxon (e.g., Riccia, Corsinia), the seta and/or foot is vestigial or absent. At completion of meiosis and spore development, cells of the seta typically undergo rapid elongation through water imbibition and thus elevate the capsule away from the substrate. Sterile, elongated elaters have hygroscopic, spiraled, inner-wall thickenings, that are strategically interspersed among spores to facilitate their separation and dispersal (Fig. 9E). Capsule dehiscence normally entails a patterned separation into four longitudinal valves, but variations range from two valves through irregular fragments or plates to cleistocarpous capsules.
The backbone of liverwort phylogenetic relationships
Crandall-Stotler and Stotler (2000)
used morphological characters in a cladistic analysis of liverworts. Their analyses included 34 taxa and 61 characters, and they resolved two main lineages: complex thalloids (Marchantiopsida) and simple thalloids plus leafies (Jungermanniopsida: Metzgeriidae, Jungermanniidae, respectively). However, their sampling was not extensive enough to address phylogenetic issues within any of the major clades. There are a few taxa for which placement relative to the three large groups is ambiguous on the basis of morphological, ultrastructural, and chemical features. These include Treubia and Apotreubia (Treubiales), Monoclea (Monocleales), Sphaerocarpos, Geothallus and Riella (Sphaerocarpales), Blasia and Cavicularia (Blasiales), and Haplomitrium (Haplomitriales [Calobryales]). Early molecular analyses of the liverworts were limited to single genomic regions with limited taxon sampling (e.g., Lewis et al., 1997
; Bopp and Capesius, 1998
; Beckert et al., 1999
; Stech and Frey, 2001
) but recent multigene analyses with increased sampling have begun to clarify phylogenetic relations among (and within) the major groups of liverworts. Phylogenetic relationships within Marchantiopsida (complex thalloids) from DNA sequence data were analyzed by Bischler (1998)
, Wheeler (2000)
, and Boisselier-Dubayle et al. (2002)
. Forrest and Crandall-Stotler (in press)
focused on Metzgeriidae (simple thalloids), whereas He-Nygrén et al. (in press)
sampled a wide diversity of liverwort taxa. Davis (in press)
provided the most extensive analysis of relationships among leafy liverwort genera available to date.
Davis (in press)
reconstructed "backbone" relationships among liverworts based on a combined data set including two nuclear, three mitochondrial, and eight loci sequenced from 20 liverworts and three outgroup mosses (Fig. 7). The data were analyzed using maximum parsimony, maximum likelihood, and Bayesian inference, and most of the results were robust to these alternative methods. The liverworts are resolved as monophyletic, as are class Marchantiopsida (complex thalloids) and Jungermanniidae (leafies). Metzgeriidae are resolved as a grade paraphyletic to Jungermanniidae, in agreement with earlier studies. Although Forrest and Crandall-Stotler (in press)
sampled different species, results of their analysis of five plastid loci are congruent with those of Davis (in press)
.
Although Haplomitrium has generally been regarded as an early-diverging lineage within the liverworts (Smith, 1955
; Schuster, 1984
; Renzaglia et al., 1994
), the precise placement of this genus remains problematic. The gametophyte of Haplomitrium is erect and radially symmetrical and therefore reminiscent of both jungermannialean liverworts and mosses. Prior to the discovery of antheridia and sporophytes in Takakia (Smith and Davison, 1993
; Renzaglia et al., 1997
), Haplomitrium was considered closely related to Takakia because of gametophytic similarities (Schuster, 1972
, 1984
). More recent molecular and morphological data have come together to solidify the placement of Takakia among mosses (see later). Divergent opinions have been expressed with regard to the relationship of Haplomitrium to other hepatics. A conclusion from Bartholomew-Began's (1990
, 1991)
extensive morphogenetic reevaluation of Haplomitrium was that the genus is a member of the simple thalloid lineage. In their analysis of land plant relationships based on rbcL sequences, Lewis et al. (1997)
noted that the precise position of the genus depended on the data set analyzed (1st and 2nd vs. 3rd positions, all positions, "ts/tv" weighting); Haplomitrium fell out sister to all other embryophytes, sister to all other liverworts, or nested within the liverworts and sister to the leafy taxa. Nuclear 18S rDNA sequences resolved Haplomitrium (without bootstrap support) as sister to the class Jungermanniopsida (i.e., leafies plus simple thalloids; Hedderson et al., 1996
).
Recent multigene analyses have focused on two hypotheses: Haplomitrium is either sister to Jungermanniopsida or sister to all other liverworts. In contrast to almost all other nodes on her tree, Davis (in press)
reported that the placement of Haplomitrium varied among analyses. Under parsimony, likelihood, and Bayesian methods, Haplomitrium is resolved with strong support as sister to Jungermanniopsida (simple thalloids plus leafies), and this inclusive clade is in turn sister to Marchantiopsida (complex thalloids; Fig. 7). However, the most complex heterogeneous Bayesian substitution model, with 21 partitions, yielded Haplomitirum as the sister group to all other liverworts. Forrest and Crandall-Stotler (in press)
and Qiu (2003) reported that Haplomitrium plus Treubia form a clade sister to all other hepatics. However, the sister-group relationship was unsupported. When Treubia was excluded from the analysis by Forrest and Crandall-Stotler (in press)
, the position of Haplomitrium was unresolved. Thus, the affinities of Haplomitrium are not yet satisfactorily resolved; Davis (in press)
felt that the weight of the current evidence supports a position for the genus as sister to the class Jungermanniopsida, whereas Qiu (2003) and Forrest and Crandall-Stotler (in press)
favor a position as sister to all other hepatics.
Although unexpected, the affinity between Treubia and Haplomitrium finds support in morphology. Both are "leafy" taxa with gametangia situated in leaf axils or lobules. Treubia is decisively more dorsiventral, with an oblique to transverse leaf insertion (succubous) and small dorsal lobules (Renzaglia, 1982
), whereas some species of Haplomitrium tend toward anisophylly and succubous insertion (Bartholomew-Begin, 1991
). In both genera, a tetrahedral apical cell is responsible for shoot growth. Perhaps the most compelling evidence for a close relationship between these two genera icomes from the peculiar yet similar sperm cells that they produce. Cladistic analyses based on spermatogenesis consistently recovered a Treubia plus Haplomitrium clade that is sister to the remaining liverworts (Garbary et al., 1993
; Renzaglia and Garbary, 2001
). Stech et al. (2000)
elevated Treubia to class Treubiopsida based on trnL intron sequence divergences between it and other liverworts.
Systematics and phylogeny of the Marchantiopsida (complex thalloid liverworts)
Unlike other hepatic groups, the complex thalloid liverworts include relatively drought-resistant species. Many morphological features of Marchantiopsida indicate xeromorphic adaptations (Schuster, 1992
; Wheeler, 2000
). In addition to air chambers in the dorsal part of the thallus (Fig. 9A), marchantioid liverworts are characterized by two types of rhizoids (smooth and pegged), archegonial involucres, unlobed spore mother cells, four primary androgones in the antheridium, six rows of neck cells in the archegonium, idioblastic oil body cells, ventral thallus scales, unistratose capsule walls, and a simple locomotory apparatus in the small biflagellated sperm cell (Schuster, 1966
, 1992
; Renzaglia et al., 2000
; Renzaglia and Garbary, 2001
). Of these, only features of the sperm appear to be universal in all species.
Although Marchantiopsida are resolved as monophyletic, traditional relationships among taxa generally are not supported by molecular data. The classical morphological separation of this liverwort class into three orders; i.e., Monocleales, Sphaerocarpales, and Marchantiales, is challenged by nucleotide sequence data (Wheeler, 2000
; Boisselier-Dubayle et al., 2002
). Incongruence between morphological and molecular patterns may be attributed to parallel changes in multiple lineages (Boisselier-Dubayle et al., 2002
).
The multigene analyses of Davis (in press
, Fig. 8) and Forrest and Crandall-Stotler (in press)
provided strong support for the placement of Blasia as a member of the complex thalloids, a result that conflicts with the traditional placement of this liverwort within the simple thalloids (Renzaglia, 1982
). Sperm cell features, persistent ventral scales, a small wedge-shaped apical cell, and a Monoclea-like female involucre provide morphological evidence for the inclusion of Blasia in the complex thalloid lineage (Renzaglia and Duckett, 1987
; Pass and Renzaglia, 1995
; Renzaglia and Garbary, 2001
). Previous molecular analyses based on one or two gene sequences do not agree in the placement of Blasia. Stech and Frey (2001)
resolved Blasia as sister to Jungermanniopsida (simple thalloids plus leafies) and described the new class, Blasiopsida. Their study was based solely on trnL intron sequences (ca. 500 bp), and the relationship was without bootstrap support. Wheeler (2000)
found that Blasia grouped with the simple thalloids (Metzgeriidae) based on 26S nrDNA (also without bootstrap support), and He-Nygrén et al. (in press)
resolved Blasia as sister to the remaining liverworts.
After Blasia, Sphaerocarpos is the next divergent taxon (Fig. 8). A position for Sphaerocarpales (Sphaerocarpos, Riella, Geothallus) among complex thalloids is generally supported by morphology (Smith, 1955
; Bishler, 1998
; Crandall-Stotler and Stotler, 2000
; Boisselier-Dubayle et al., 2002
). However, with additional taxon sampling, the sister relationship between Sphaerocarpales and the remaining Marchantiopsida is called into question. Based on LSU rDNA sequences, Wheeler (2000)
and Boisselier-Dubayle et al. (2002)
reported that Sphaerocarpales were placed within Marchantiaceae. Similarly, Sphaerocarpos nested between Neohodgsonia and Marchantia in the five-gene analysis of Forrest and Crandall-Stotler (in press)
. The implication from these results is that the relatively simple morphology of both generations in Spharocarpales may not be plesiomorphic but rather the product of extreme simplification in ephemeral or aquatic habitats.
Air chambers are found in the crown group taxa (Marchantia, Preissia, Targionia, Riccia) (Wheeler, 2000
). One lineage, Monoclea plus Dumortiera, has secondarily reverted to a morphologically simple thallus devoid of chambers, perhaps adaptations to the semi-aquatic habit of these plants (Wheeler, 2000
). The production of archegoniophores (carpocephala) that elevate sporophytes above the gametophyte also evolved within the crown Marchantiopsida group. Independent losses of these structures occurred in riccioid taxa (Riccia, Ricciocarpos, Oxymitra) and Monoclea (Wheeler, 2000
). Reduction in sporophyte complexity is likewise a derived feature of riccioid liverworts (Renzaglia et al., 2000
; Boisselier-Dubayle et al., 2002
). With a jungermannioid-like sporophyte elevated on a fragile and highly elongated seta, Monoclea seems inappropriately placed within this crown group. Additional characters of the genus, including a free nuclear embryo and monoplastidic meiosis in some species (Schofield, 1985
; Renzaglia et al., 1994
), support a more traditional placement of Monoclea in Marchantiopsida. However, congruence among the multigene analyses provided support for Monoclea close to Dumortiera.
Systematics and phylogeny of Metzgeriidae (simple thalloid liverworts)
Clearly not a monophyletic group, Mezgeriidae traditionally include some 30 highly diverse genera of "leafy" and thalloid forms. Although four apical cell types are found in the group (Fig. 10), a unifying feature of apical growth in these plants is development of wings and leaves from a central wedge cell (single initial) that forms in the newly produced apical derivative (Fig. 11A) (Renzaglia, 1982
). Simple thalloid genera are distinguished from leafy liverworts (Jungermanniidae) in that they are anacrogynous: archegonia are produced along the mid-thallus of either the main, lateral, or ventral shoots. Consequently, the apical cell is not transformed into permanent tissue after archegonial development, and sporophytes do not terminate the shoot as in acrogynous Jungermanniidae. Additional features that unify the simple thalloid taxa, but are also found in leafy liverworts, are the development of antheridia from two primary androgones, oil bodies in all cells, lobed sporocytes, smooth rhizoids, and five rows of neck cells per archegonium.
|
|
95). Parsimony bootstrap values 
