| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
Anatomy and Morphology |
Department of Ecology and Evolutionary Biology, University of Tennessee, Knoxville, Tennessee 37996 USA
Received for publication 14 June 2007. Accepted for publication 20 January 2008.
ABSTRACT
The vast majority of flowering plant seeds contain a triploid endosperm formed by fertilization of a monosporic, Polygonum-type female gametophyte. However, evolutionary transitions to six other genetic constructs of endosperm are widespread, and six of seven known patterns are found in the order Piperales. Within Piperaceae, Manekia has not been described, and we report its female gametophyte to be tetrasporic and 16-nucleate at maturity. Manekia ontogeny is generally characterized by early establishment of a bipolar or weakly bipolar body plan and a binucleate central cell at maturity (Drusa-type pattern); however, ca. 16% of early stages had distinctly tetrapolar organization, and ca. 21% of mature specimens had a tetranucleate central cell (Penaea-type pattern, not previously reported in Piperaceae). An evolutionary developmental analysis indicates heterochrony, heterotopy, novelties, and sequence deletions have each played roles in modulating variation within Piperales. Our data suggest the common ancestor of Piperaceae was tetrasporic and retained a plesiomorphic bipolar body plan, producing a "functionally bisporic" form of triploid endosperm derived from the lineal descendants of two megaspores and a sperm. Developmental modifications of this tetrasporic, bipolar ontogeny can account for the origin of all three other known "true" tetrasporic endosperm genetic constructs, formed from derivatives of all four megaspores and a sperm. These derived endosperms in turn have higher ploidy, higher potential heterozygosity, and reduced genetic conflicts.
Key Words: Costa Rica • endosperm • evolution of development • heterochrony • Manekia • Piperaceae • Piperales • tetraspory
There has been considerable recent interest in the origin and early evolution of putative angiosperm-defining features, motivated in part by improved resolution of the base of the angiosperm phylogenetic tree. Surprisingly, endosperm, the unique bisexual, embryo-nourishing tissue of angiosperm seeds, and a primary food source of humans and many other organisms, has turned out to be exceptionally variable among early-diverging flowering plant lineages (Floyd and Friedman, 2000
; Williams and Friedman, 2002; Friedman and Williams, 2003; Friedman, 2006; Rudall, 2006; Tobe et al., 2007). Endosperm genetic constitution is monosporic diploid in Nymphaeales and Austrobaileyales; monosporic triploid in Amborella, eumagnoliids, and basal lineages of monocots and eudicots; and polysporic pentaploid or higher in at least five other derived taxa within these lineages (Williams and Friedman, 2004). In fact, all seven known genetic constructs of endosperm can be found within these so-called basal angiosperm lineages.
Over the years, a large body of theory has accumulated to explain the evolution of endosperm genetic constitution (reviewed in Friedman et al., 2008). Such theories can be evaluated by evolutionary developmental studies because endosperm evolution is tightly bound to the constraints of female gametophyte ontogeny. The central cell of the female gametophyte is the precursor cell to the endosperm and before it becomes fertilized it can contain from one to 14 nuclei (Maheshwari, 1950). The number and meiotic descent of central cell nuclei directly affects endosperm ploidy, heterozygosity, maternal-to-paternal genome ratio, and the potential for genetic conflict.
The family Piperaceae has historically been a focal point for hypotheses about female gametophyte evolution, extending back to just after the discovery of double fertilization. Campbell (1902) suggested that the tetrasporic, 16-nucleate female gametophytes of Piperaceae represented the ancestral angiosperm "type" because they were intermediate between the relatively large female gametophytes of all gymnosperms and the highly reduced Polygonum-type, eight-nucleate female gametophytes thought to have been present in early flowering plants. This view was supported over the years by other embryologists (Ernst, 1908; Pearson and Thomson, 1917; Schürhoff, 1926; Fagerlind, 1944; Gvaladze and Akhalkatsi, 1990), and by the fact that Piperaceae has long been thought to have diverged from an ancient angiosperm node (Bessey, 1915; Burger, 1977; Donoghue and Doyle, 1989; Zimmer et al., 1989; Hamby and Zimmer, 1992; Tucker et al., 1993).
Alternatively, Johnson (1900, 1902) proposed that the 16-nucleate female gametophytes of Piperaceae were developmental descendants of an ancestral monosporic, eight-nucleate (Polygonum-type) female gametophyte (see also Maheshwari, 1937). The view of Johnson is now prevalent because of the nested phylogenetic position of Piperaceae within extant flowering plant clades that are characterized by Polygonum-type female gametophytes (Donoghue and Scheiner, 1992; Williams and Friedman, 2004; Soltis et al., 2007, Wanke et al., 2007a).
Tetraspory has originated at least 46 times within angiosperms, based on a conservative estimate, and it has also been seen as an irreversible evolutionary step (Palser, 1975). Piperaceae are known for the diversity of their tetrasporic, 16-nucleate female gametophytes (in Piper 16 chromosome sets are partitioned into eight nuclei; see Williams and Friedman, 2004). Within the family, there are at least four known ontogenetic patterns and these produce all four known genetic constructs of tetrasporic endosperms in flowering plants. Monospory and/or bispory are predominant in other families within Piperales. Given such variation, a comparative analysis of female gametophyte ontogenies in Piperales can pinpoint the evolutionary developmental changes involved in the origin of tetrasporic endosperms and in their subsequent evolution.
To date, embryological work in Piperaceae has been confined to three genera. All are tetrasporic, initiating female gametophyte ontogeny from four free megaspores (the haploid products of meiosis), which then undergo two mitoses to ultimately produce 16 haploid chromosome sets (partitioned into either eight or 16 nuclei during development). In Piper, the Fritillaria type has been reported (eight-nucleate with pentaploid endosperm) (Maheshwari and Gangulee, 1942; Swamy, 1945; Yoshida, 1960; Kanta, 1962, Nikiticheva, 1981, Prakash et al., 1994); in Peperomia, the Peperomia type (16-nucleate with 8-15 ploid endosperm) (Campbell, 1901, 1902; Brown, 1908; Johnson, 1914; Swamy, 1944; Nikiticheva et al., 1981; Plyushch, 1982; Smirnov and Grakhantseva, 1988), and in Zippelia, the Drusa type (16-nucleate with triploid or higher endosperm) (Lei et al., 2002). The status of Manekia and Verhuellia remain completely unknown.
Phylogenenetic placement of genera in Piperaceae has received recent and ongoing attention (Jaramillo and Manos, 2001; Jaramillo et al., 2004; Wanke et al., 2007a, b), and these studies provide strong support for Manekia (= Sarcorharchis, Arias et al., 2006) as sister to Zippelia, which is in turn sister to a clade comprised of Piper and Peperomia. Recently, Verhuellia, a genus allied with Peperomia was found to be sister to the rest of Piperaceae (Wanke et al., 2007b). Piperaceae itself has long been placed as sister to Saururaceae (Tucker et al., 1993), and this clade is in turn nested within Piperales in the eumagnoliid clade (Wanke et al., 2007a). Verhuellia, Manekia, and Zippelia are exceedingly species poor, whereas Piper and Peperomia are among the most speciose basal angiosperm lineages, each with >1500 species (Wanke et al., 2007b).
In this study, we report primarily on the embryology of Manekia naranjoana (C. DC.) Callejas. Little is known about the reproductive biology of Manekia because it flowers high in the canopy of lowland and montane tropical rain forests. Endress and Igersheim (1998) have reported on gynoecium structure and Burger (1971) and Steyermark (1971) on floral morphology. Our main goals were (1) to describe female gametophyte development of M. naranjoana and to infer its endosperm ploidy and genetic constitution and (2) to place Manekia female gametophyte ontogeny in a comparative context to reconstruct historical sequences of developmental modifications underlying the great diversity of endosperms in Piperaceae.
MATERIALS AND METHODS
Study species
The genus Manekia comprises ca. four species of vines (Arias et al., 2006) and is widely distributed in the neotropics, from southern Nicaragua to northern Peru and from the Lesser Antilles to the Atlantic forest of southern Brazil (Trelease and Yuncker, 1950; Arias et al., 2006). Two morphological features distinguish it from Piper, Peperomia, Zippelia, and Verhuellia: perianthless flowers embedded within the inflorescence rachis and shoots with both terminal and axillary inflorescences (Figs. 1, 2) (Miquel, 1843; Trelease, 1927). Manekia has a combination of features of early and late successional species and is found in both mature and secondary rain forests (T. Arias, personal observations). The primary study species, M. naranjoana, is distributed in Central America from southern Nicaragua to southern Panama. Some material of M. sydowii (Trel.) Arias, Callejas and Bornstein, collected in the Ventanas Paramo of Colombia, was also viewed. An extensive survey of over 600 herbarium records from Central American collections from 24 herbaria indicates that M. naranjoana can flower in any month of the year (T. Arias, personal observations).
|
Collections
For the study of pollen tube growth and female gametophyte development in M. naranjoana, floral morphology was described in the field, and over 200 inflorescence samples at different developmental stages were collected. Subsequently, over 1000 flowers were viewed microscopically. Several inflorescences were self-pollinated to test for self-incompatibility, which is known in Piper (de Figueiredo and Sazima, 2000).
Inflorescences were either fixed for 24 h in 3:1 95% ethanol (etOH):acetic acid (HAc) and stored in 75% ethanol or fixed in FAA (50 ml 95% etOH:5 ml glacial HAc:10 ml 40% formaldehyde:35 ml dH2O) and stored in 75% ethanol (Williams and Friedman, 2004). Reproductive material was dehydrated through an ethanol series and was infiltrated and embedded in glycol methacrylate (JB-4 embedding kit; Polysciences, Warrington, Pennsylvania, USA). Serial sections (5 µm thick) were cut and stained with aniline blue (flowers fixed in 3:1 etOH:HAc), or 0.1% toluidine blue in dH20 (w/v). Structural features of the pollen, female gametophyte, and embryo were observed using fluorescence and light microscopy. Stigmatic receptivity was measured using Peroxtesmo KO peroxidase test paper (Sigma #90606, St. Louis, Missouri, USA). Images were processed with a Zeiss axiocam digital photo system (Oberkochen, Germany) and Adobe (San Jose, California, USA) Photoshop, version 7.0. Composite images were derived from multiple focal planes in a few instances; however, all other manipulations were applied to the whole image (unless noted in figure legends). "Camera lucida" drawings were made by tracing structural features of the female gametophyte over images in Photoshop. Vouchers were deposited in the National Herbarium of Costa Rica (CR), University of Costa Rica Herbarium (USJ), and the University of Antioquia Herbarium (HUA) (see Appendix).
Character analyses
Character evolution was studied using the most recent molecular and morphological phylogenetic trees for eumagnoliids (Qiu et al., 2005), and Piperales and Piperaceae (Nickrent et al., 2002; Jaramillo et al., 2004; Wanke et al., 2007a, Wanke et al., 2007b). The families belonging to Piperales are as circumscribed in APG II (2003), although monophyly of Aristolochiaceae is uncertain (see Wanke et al., 2007a). Placement of genera within families in Piperales was based on Wanke et al. (2007b). Ancestral states for discrete characters (as determined from this study and previously published works) were determined using simple parsimony, as implemented in the program MacClade, version 4.03 (Maddison and Maddison, 2001). All characters were treated as unpolarized and unordered (all transitions among states are equally probable).
RESULTS
Flower, carpel, and ovary development
The Manekia naranjoana flower is bisexual, subtended by a single bract that becomes totally immersed in and fused with the adjacent parenchymatic tissue of the rachis during development (Figs. 1–3). As a consequence, there is no distinction between the external ovary wall and the rachis (Figs. 3, 4). The bract is hypopeltate, persistent, with marginal filamentous muticellular hairs and abaxial oil cells. The ovary is unilocular with a single, basal, orthotropous ovule (Figs. 3–6). It is formed by postgenital fusion (Endress and Igersheim, 2000) of four (3–5) carpel bases, as judged by the occurrence of a single vascular strand leading to each of 3–5 free stigmatic lobes and is thus syncarpous (Figs. 3, 4). When stigmas are receptive, the apical portions of the carpels remain unfused (Fig. 4), and the stigmatic lobes are decurrent. The stigma has unicellular papillae and ethereal oil cells immersed in the substigmatic tissue (Fig. 4). A relatively undifferentiated transmitting tract is present in the region of postgenital fusion (Figs. 4, 21).
|
|
Ovule development
The single, basal ovule of M. naranjoana originates at the center of the floral apex in between the separate, early-developing carpels (Fig. 3). It is orthotropous, crassinucellar, and bitegmic (Figs. 4–6). The outer integument initiates growth after the inner, growing to enclose the full length of the ovule but not participating in the formation of the micropyle (Figs. 5, 6). The micropyle is in contact with the wall of the ovarian cavity at maturity. The inner integument is three cell layers thick, and the inner two cell layers are densely cytoplasmic. The outer integument is also three cell layers thick, but its cells are large and highly vacuolate (Figs. 5, 6). A hypostase is formed at the base of the ovule during late stages of female gametophyte ontogeny (Fig. 6).
Numerous degenerate or unformed ovules were seen alongside normal ovules at all stages of development from ovule initiation to fruit formation. Fewer than 10 totally or well-developed seeds per inflorescence were seen in all specimens of Manekia, including herbarium material.
Female gametophyte development
Megasporogenesis
The single archesporial cell can be first recognized by its hypodermal position and larger size relative to other nucellar cells. The archesporial cell eventually gives rise to the megasporocyte through mitosis by a series of periclinal divisions (Figs. 7–9). As a result a four-layered (3–5) parietal tissue is formed (Figs. 10–11). As the megasporocyte enters meiosis I, the large nucleus is variable in its position, and the cell is ca. 35 µm long.
|
During meiosis II, the spindle of the micropylar nucleus in the dyad is parallel to the vertical axis of the ovule, while the spindle of the chalazal nucleus is more or less perpendicular to the vertical axis of the ovule (Fig. 14). Following meiosis II, four free megaspore nuclei are formed within an ovoid coenocyte, which is still about the same size as that of the mature megasporocyte (ca. 35 µm long; see Fig. 15). Cytokinesis was never observed at either stage of meiosis.
As summarized in Fig. 16, at the onset of megagametogenesis, the arrangement of megaspores varies from linear, to bipolar with a 1+3 arrangement of one micropylar nucleus and three chalazal nuclei (the most common pattern, Fig. 17), to tetrapolar (11 of 67 coenocytes). Coenocytes with bipolar and tetrapolar organization represent late stages of postmeiotic ontogeny because a large vacuole is present and the four nuclei are entering prophase (e.g., Fig. 17), whereas four-nucleate coenocytes just after meiosis (as judged by their smaller interphase nuclei and many vacuoles) were not yet polarized (Fig. 15).
|
In some cases, the mature female gametophyte has a tetrapolar organization. In these, four domains of cellularization are present, and four free nuclei are found in the central cell, each apparently a lineal descendant of a different megaspore (Fig. 19). In one case, four free nuclei can be seen fusing within the central cell to form a tetraploid central cell nucleus (Fig. 20). Of 35 mature female gametophytes, 19 had binucleate central cells, five had tetranucleate central cells, and 11 were undetermined. Irrespective of mature organization, cellularized female gametophytes always possessed three cells forming an egg apparatus at the micropylar pole (Fig. 21).
These descriptions correspond to a continuum of ontogenetic variation between two previously defined extremes, the Drusa type (bipolar organization, binucleate central cell; Hakansson, 1923) and the Penaea type (tetrapolar organization, tetranucleate central cell; Stephens, 1909).
Pollen tube growth and fertilization
Pollen grains on the stigmatic surface are generally clumped in clusters within a mucilaginous substance. Pollen did not germinate on self-pollinated stigmas, and the papillae were covered with callose depositions by the time they were collected (Fig. 4). In naturally pollinated flowers, pollen germinates on the stigma and pollen tubes grow between stigmatic papillae to penetrate solid ground tissue and then elongate intercellularly toward the region of postgenital fusion of the carpels (above the ovarian cavity; Fig. 22). We did not see pollen tubes growing within the narrow open space formed by the upper unfused portions of the carpels. Upon entering the region of postgenital fusion, pollen tubes grow toward the ovarian cavity and directly into the micropyle and the nucellus.
|
Integration of floral and female gametophyte ontogenies
The following sequence of reproductive events in M. naranjoana was determined based on the observation of 40 inflorescences that initiated development simultaneously on a single vine in Alberto Brenes Biological station from 25 May to 28 June 2006 (Fig. 26).
|
Day 5
The inflorescences are light yellow and 5–7 cm long. The two laterally inserted stamens are evident at the surface of the rachis. The pistil is immature and covered by the bract. Carpels had elongated but are not yet fused. The inner integument has formed a well-defined micropyle, and the megasporocyte is present.
Day 12
Inflorescences are yellow to light brownish-yellow and 10–14 cm long. The lateral and apical stamens are mature, the basal stamen has formed but is still immature, and the apical stamen is just initiating. Stigmatic lobes are formed, and stigmas test positively for receptivity. Stages of megagametogenesis are seen in this material (Fig. 26). The coenocyte is ca. 35–40 µm wide x 50–60 µm long just before cellularization.
Day 18
Inflorescences are light brownish-yellow with dark brown dots and 14–22 cm long. The lateral stamens have abscised, and the basal stamen is either mature or has abscised. The apical stamen is present but immature. The stigmatic surface is only weakly receptive, or it is necrotic with callose deposits on its upper surface. Female gametophytes are cellularized, and early postfertilization stages are present.
Day 25
Inflorescences are brownish green and 22–25 cm long. All stamens have abscised, and fruits are being initiated. The apical stamen is mature.
Day 35
Infructescences are brownish green and >25 cm long. Few fruits have formed.
DISCUSSION
Megagametogenesis and megasporogenesis of Manekia and other Piperaceae
In Manekia a single, hypodermal archesporial cell ultimately gives rise to a single megasporocyte 3–5 cell layers deep within the nucellus. The female gametophyte is tetrasporic and 16-nucleate at maturity. Megasporogenesis takes place over roughly five to eight days and megagametogenesis about 10 days (Fig. 26). A similar interval of ca. 18 days is known in some cereals (Cass et al., 1985). Fertilization is porogamous, and the pollen tube enters a degenerate synergid. Endosperm development is free-nuclear, as it is in Zippelia and most other Piperaceae (Lei et al., 2002).
The mature female gametophyte of Manekia always had an organized egg apparatus comprised of two synergid cells and an egg cell. In the majority of mature female gametophytes, the central cell was binucleate and the remaining nuclei were largely confined to the chalazal domain and were variable in number (up to 11 nuclei and nine cells; Fig. 16). The two polar nuclei, one contributed from each pole, were seen in various stages of fusion, including fertilization, within the central to chalazal region. These are Drusa-type female gametophytes because of their bipolar or weakly bipolar organization and binucleate central cells (Hakansson, 1923; Maheshwari, 1950). However, 16% of four-nucleate coenocytes were characterized by a distinctly tetrapolar organization, and a similar percentage (21%) of mature female gametophytes had a tetrapolar arrangement of four quartets of three cells and one free nucleus, and hence a tetranucleate central cell. These are definitive characteristics of the Penaea-type female gametophyte (Stephens, 1909; Tobe and Raven, 1984), which has not yet been reported in Piperaceae. The Drusa-type has been reported in Zippelia (Lei et al., 2002). The Drusa and Penaea ontogenetic patterns are both quite rare in angiosperms (Palser, 1975) and only occur together in one other family, the Apiaceae (Johri et al., 1992).
Two critical features characterize Drusa female gametophytes: the 1+3 organization of the coenocyte after megasporogenesis and the binucleate central cell (whose functional offspring is a triploid endosperm derived from two megaspore progenitors and a sperm). Zippelia shares both these features, but apparently not always a triploid endosperm. Lei et al. (2002, p. 54) found that "a large primary endosperm nucleus with many nucleoli is formed by the fusion of the fertilized polar cells or nuclei along with the 11 antipodal cells" [our italics]. Thus, Zippelia and Manekia share similar landmarks of megagametogenesis (1+3 coenocyte and binucleate central cell), but ontogenetic variation in each produces different functional products. InZippelia, endosperms are typically triploid and "functionally bisporic" but can range up to 14-ploid (Lei et al. 2002), whereas in Manekia endosperms fall into two discreet classes: triploid, functionally bisporic or pentaploid, "true tetrasporic" endosperms.
The two main features of the Penaea-type are early tetrapolar organization and a tetranucleate central cell. Within Piperaceae, the 16-nucleate Peperomia-type also has early tetrapolar organization. In Peperomia, however, fewer peripheral cells are formed, so more than four nuclei are contributed to the central cell. In Manekia, we never saw more than four nuclei fusing within the central cell. Early tetrapolar organization is also a characteristic of eight-nucleate Adoxa-type female gametophytes and has been found as a developmental variant among the predominantly Drusa-type pattern in Ulmus (Hjelmqvist and Grazi, 1965). Within Piperaceae, the tetrapolar Adoxa-type was reported in Piper by Johnson (1902, 1910; see also Swamy, 1944), but others (Maheshwari and Gangulee, 1942; Swamy, 1944, Swamy, 1945) found the bipolar Fritillaria-type (1+3 early organization) for the same taxa. We agree with Maheshwari (1937; 1946a, b) that the figures in the earlier works are consistent with the subsequently recognized Fritillaria-type.
The origin of tetraspory in Piperaceae
Bisporic and tetrasporic development of the female gametophyte are derived, homoplastic characters in flowering plants, but little is known about their functional significance. Such female gametophytes produce genetically heterogeneous endosperms, and because of their importance in transferring nutrients to the embryo, a large body of theory has developed to explain their evolution. Comparative studies can be used to reconstruct historical sequences of developmental modifications to see if they are consistent with patterns predicted by theory.
Bisporic and tetrasporic female gametophytes have been reported in some Piperales; thus, we briefly review these here. There is a single report of bispory in Piperaceae, in an undetermined species of Piper (Fagerlind, 1939; see also Maheshwari, 1946a). If confirmed, bispory would likely be a derived state within the genus, given the many reports of tetrasporic, Fritillaria-type female gametophytes in Piper and Macropiper (see Lei et al., 2002).
Within Hydnoraceae, a bisporic Allium-type female gametophyte was reported in Prosopanche bertoniensis (Chodat, 1916) and a tetrasporic Adoxa-type in Hydnora (Dastur, 1921). Neither includes complete developmental sequences, and P. Maheshwari (1946a) concluded that Hydnora could represent either monospory or bispory, whereas S. C. Maheshwari (1955, p. 82) stated that "both of these accounts are probably incorrect" and should be reinvestigated (see also Fagerlind, 1938).
Within Saururaceae, bisporic development was reported in Anemopsis californica (Johnson, 1900; Quibell, 1941) and in Saururus (Nikiticheva, 1981); however only Quibell (1941) provided figures. Other studies of these same two taxa included clear depictions of cell walls between all megaspores, thus indicating monospory (Saururus loureiri, Yoshida, 1961; Anemopsis californica, Raju, 1961). All other reports within Saururaceae also show monospory (Murty, 1960; Raju, 1961, Nikiticheva, 1981).
All studies of Lactoridaceae and Aristolochiaceae depict monosporic, seven-celled/eight-nucleate female gametophytes (Tobe et al., 1993; González and Rudall, 2003; Madrid and Friedman, In press), and all three other eumagnoliid orders have been polarized as having this same ancestral state (Williams and Friedman, 2004). Thus, an evolutionary transition from a monosporic (seven-celled/eight-nucleate) female gametophyte to a tetrasporic female gametophyte can be inferred to have occurred along the lineage leading to Piperaceae, either before or after the divergence of Verhuellia from the rest of the family (Verhuellia has not been studied yet). It is possible that bispory was an intermediate condition in this historical sequence, given the many (albeit unsubstantiated) reports of bispory in the Piperales. Toward this end, knowledge of Verhuellia and more certain phylogenetic placement of Hydnoraceae are needed.
There are now at least five female gametophytic variants known within Piperaceae: the 16-nucleate Drusa-type with a diploid central cell in Manekia (this study) or diploid and higher in Zippelia (Lei et al., 2002), the 16-nucleate Penaea-type with a tetraploid central cell in Manekia (this study), the Fritillaria-type in Piper with a seven-celled/eight-nucleate female gametophyte with a tetraploid central cell (Maheshwari and Gangulee, 1942; Swamy, 1945; Yoshida, 1960; Kanta, 1962; Nikiticheva et al., 1981; Prakash et al., 1994), and two different functional variants of the 16-nucleate Peperomia-type in Peperomia generally with either eight- or 14-nucleate central cells (Swamy, 1944; Nikiticheva et al., 1981; Plyushch, 1982; Smirnov and Grakhantseva, 1988; older studies reviewed in Maheshwari, 1937).
Typological classifications of female gametophytes tempt one to infer ancestral character states on the basis of idealized whole organism descriptions (i.e., the mature female gametophyte type). However, angiosperm female gametophytes, though exceptionally small, are still complex individuals that evolve through small modifications of ontogeny (Herr, 1967; Friedman, 2006) as well as by large-scale duplications or deletions of developmental modules (Williams and Friedman, 2002; Friedman and Williams, 2003). Modifications of the modular construction of female gametophytes are especially apparent in the origin of tetrasporic female gametophytes, in which specific ontogenetic changes in early development appear to underlie a great lability of late ontogenetic and mature structure within the family. As with larger organisms, one still needs to understand ancestral states of individual traits to reconstruct the ancestral organism (Williams and Friedman, 2004).
Evolutionary developmental consequences of the origin of tetraspory in Piperaceae
Figure 27 highlights four shared derived features of development found in Piperaceae: tetraspory, polarization of the coenocyte at the four-nucleate stage just after meiosis II, formation of 16 haploid chromosome sets that are partitioned into either eight nuclei (Piper) or 16 nuclei (Zippelia, Manekia, Peperomia) at maturity, and endosperm derived from more than one megaspore derivative. The last three features are each causally related to the origin of tetraspory and point to evolutionary developmental transitions that must have occurred at some point along the lineage leading to Piperaceae (with the caveat that Verhuellia is not yet known).
|
). In the strict sense, however, predisplacement results in an extension of developmental time because offset timing is conserved, and thus tetrasporic female gametophytes in Piperaceae would be expected to retain the plesiomorphic pattern of three mitoses (as in Polygonum-type) and possess 32 nuclei at maturity. Because Piperaceae female gametophytes undergo two mitoses, not three, there has also been an early offset of ontogeny (hypomorphosis; Reilly et al., 1997).
The failure to form cell walls during megasporogenesis explains the earlier onset of ontogeny but not why there has been an apparent deletion of the terminal stage, the "third mitosis" (Fig. 27). Broad surveys of all angiosperms indicate that cellularization of the female gametophyte always occurs upon placement of four nuclei within the micropylar domain (Battaglia, 1989). In all tetrasporic female gametophytes, free-nuclear ontogeny initiates at meiosis I, but the onset of polarity occurs after meiosis II, when the four-nucleate coenocyte becomes organized such that a single nucleus is placed in the micropylar domain. Free-nuclear development within the micropylar domain ends when four nuclei are formed after the second mitosis, at which time cellularization occurs resulting in the apparent sequence deletion of the third mitosis (sensu Hufford, 1996). Thus, the origin of the 16-nucleate female gametophyte from an eight-nucleate ancestor is a correlated outcome of the shift from monosporic to tetrasporic ontogeny and the conservation of the four-nucleate micropylar developmental module (Fig. 27).
The occurrence of genetically heterogeneous endosperm is another consequence of the origin of tetraspory because polar nuclei are contributed to the central cell from the lineal descendants of either two or four megaspores, instead of just one. Variation in central cell genetics and in mature female gametophyte structure results from spatial variation in polarization of the coenocyte, a form of heterotopy. As long recognized (Porsch, 1907; Coccuci, 1973; Favre-Duchartre, 1984), the angiosperm female gametophyte body plan consists of one, two, or four domains of ontogeny, each of which has been described as a developmental module in which a single nucleus is placed within a cytoplasmic domain of the coenocyte and undergoes two mitoses to produce four nuclei, which are usually partitioned into three cells leaving one free nucleus within the former coenocyte, or central cell (Friedman and Williams, 2003). The number of modules initiated depends upon how many nuclei migrate to unfilled domains of the coenocyte early in megagametogenesis.
Module initiation of bipolar Polygonum-type gametophytes occurs as a result of directed nuclear migration after mitosis I and has been shown to be a genetically based trait in Zea (Vollbrecht and Hake, 1995). In bipolar Drusa- and Fritillaria-type coenocytes, a similar migration occurs after meiosis II, in which one of two nuclei from the micropylar meiotic dyad becomes positioned in the chalazal region. In the Fritillaria-type of Lilium, nuclear repositioning after meiosis II is developmentally similar to that in Zea; it is caused by a longitudinal orientation and attachment of microtubules to distant portions of the coenocyte (Flint and Johansen, 1958).
Zippelia, Manekia, and Piper share a bipolar body plan, established at the end of meiosis II. Given that the monosporic and/or bisporic female gametophytes of all outgroups of Piperaceae also have a bipolar body plan, bipolar coenocyte organization was conserved during the evolutionary transition to Piperaceae, though its establishment was shifted to an earlier ontogenetic stage (from just after mitosis I in a Polygonum-type ancestor to just after meiosis II in the immediate ancestor of Piperaceae; Fig. 27). Bipolar, tetrasporic female gametophytes initiate development from a "1+3" body plan, but in Piper there is an additional novel step of fusion of the three chalazal nuclei (Maheshwari and Gangulee, 1942; Swamy, 1945; Maheshwari, 1955; Yoshida, 1960). It is reasonable to infer that the positioning of three nuclei close to each other in the chalazal domain must have evolved before the fusion of these nuclei could originate (and in fact free-nuclear ontogeny could not occur at all if nuclear fusion was the ground state).
As shown in Fig. 27, tetrapolar organization is a derived state, found in Peperomia and in Penaea-type variants within Manekia. Tetrapolar body plans are established either by the breakdown of bipolar organization or as a more organized process mediated by the origin of a novel nuclear positioning event within a plesiomorphic bipolar gametophyte.
Endosperm genetics and developmental plasticity in Piperaceae
Endosperm genetic constitution is said to have a large impact on the fitness of its associated embryo (sporophyte) and also on the fitness of the maternal sporophyte (Brink and Cooper, 1947; Stebbins, 1976; Haig and Westoby, 1989 Queller, 1989). In Piperaceae, the endosperm varies in ploidy level (3n–15n) and megaspore constitution (2–4 megaspores; Fig. 27), resulting in at least four distinct genetic patterns. Theory predicts the genetic constitution of endosperm will evolve toward higher ploidy, heterozygosity, and maternal to paternal genome ratios, and toward reduced exposure to genetic conflicts (the magnitude of each is described in Friedman et al., 2008).
We have reconstructed a historical sequence of female gametophyte evolution in which bipolar body plans and bisporic (diploid) central cells were plesiomorphic character states, from which new combinations of body plan organization and central cell genetics evolved. Our scenario of female gametophyte evolution (see Fig. 27) produces changes in endosperm genetics that are consistent with theoretical predictions. Monosporic, triploid endosperm (Polygonum type) initially gave rise to "functionally bisporic," triploid endosperm (Drusa-type endosperm derived from only two of four megaspores). Only one developmental change is needed to produce two other types of tetrasporic endosperms: (1) tetrasporic, pentaploid endosperm in Piper can originate via novel fusion of the three chalazal nuclei after Drusa-type 1 + 3 bipolar organization is established (Fritillaria-type); (2) tetrasporic, pentaploid endosperm of Manekia can originate via a shift from bipolar to tetrapolar organization (Penaea-type). The tetrasporic, 9–15-ploid endosperms of Peperomia originate in at least two steps, first by a transition from bipolar to tetrapolar early organization (or by the loss of strong bipolar organization), and second by the loss of partitioning of three or six lateral/antipodal nuclei into cells, resulting in more nuclei being contributed to the central cell. Tetrapolar Penaea-type organization stands out as an intermediate step in the origin of Peperomia-type female gametophytes.
True bisporic female gametophyte development may have been the developmental precursor to either functionally bisporic or to true tetrasporic endosperms (because its formation involves the incomplete, rather than complete, loss of cell walls during megasporogenesis followed by establishment of a plesiomorphic, bipolar body plan). Thus, the many unsubstantiated reports of bisporic development in Piperales are of interest. Bisporic and functionally bisporic (Drusa-type) endosperms are both formed from a central cell containing only two of the four megaspore derivatives, drawn from either the same meiotic dyad or from separate dyads (see Friedman et al., 2008). Exclusion of some megaspore derivatives during formation of the central cell means that endosperms on a single plant are genetically heterogeneous, each one containing anywhere from one half to all of the maternal genome (Bulmer, 1986; Haig, 1986). These endosperms are thus exposed to variable levels of genetic conflict.
Not surprisingly, functionally bisporic Drusa-type female gametophytes are quite variable in their ontogenies. In Limnanthaceae (Palser, 1975), the number of megaspores participating in the endosperm is reduced during early ontogeny by the phenomenon of "strike" (see Haig, 1990). On the other hand, in Zippelia the number of megaspores participating in endosperm can be increased via inclusion of antipodals, and similarly, in Manekia early tetrapolar organization achieves the same end. The Penaea-type, Fritillaria-type, and Peperomia-type ontogenies represent derived patterns that guarantee stable transmission of the entire maternal genome to endosperm, thus diluting the effect of non-imprinted selfish paternal genes (Friedman et al., 2008 and refs therein). It is worth noting that within Piperaceae, evolutionary transitions to the Fritillaria-type and the Peperomia-type are associated with two of the most speciose genera among basal angiosperms, Piper and Peperomia.
On the basis of likely pathways of developmental change and unstable transmission of the maternal genome, we suggest that bisporic and functionally bisporic female gametophytes are intermediate, transitional stages in endosperm evolution. If so, bisporic or functionally bisporic endosperms should be widespread but rare. This prediction is borne out, since Drusa-type endosperms are known in at least 11 unrelated families and are rare in each (Palser, 1975; Johri et al., 1992; this study). The rarity, the breadth of occurrence in phylogeny, the relationships among ontogenies, and the nature of selective forces acting on functionally bisporic endosperms all suggest that functionally bisporic endosperms represent a transitional stage in the origin of tetrasporic endosperms from monosporic or true bisporic endosperm ancestors. There is no empirical data on the relative fitness of endosperms, and the relationship between endosperm genetics and fitness in these and other endosperms is well worth exploring.
Systematic implications of floral and embryological characteristics of Piperaceae
On the basis of morphological characters, Manekia was largely considered to be a part of Piper (de Candolle, 1923) or to be separate but closely allied to Piper (Trelease, 1927; Jaramillo and Manos, 2001; Lei et al., 2002). Recent molecular phylogenetic analyses (Jaramillo and Manos, 2001; Jaramillo et al., 2004; Wanke et al., 2007a, Wanke et al., 2007b) and the developmental evidence found in this study suggest Manekia is more closely related to Zippelia. Both Manekia and Zippelia share a predominantly Drusa-type pattern of ontogeny despite considerable variation. Both have early bipolar organization and a two-nucleate, diploid central cell at maturity, and these features do not occur together in Peperomia or Piper female gametophytes. Zippelia and Manekia are also distinguished from Piper by a micropyle formed by the inner integument and by the stamen initiation sequence (Jaramillo et al., 2004). In both Manekia and in some species of Piper with four stamens, the two laterally inserted stamens mature first, but in Manekia the basal stamen matures next, and in Piper the apical stamen matures next (Tucker, 1982; Jaramillo et al., 2004).
Conclusions
Embryological traits support both the removal of Manekia from Piper and its sister relationship to Zippelia— Manekia has a tetrasporic, 16-nucleate female gametophyte quite different from that of Piper and shares the rare Drusa-type female gametophyte ontogeny with Zippelia. In Manekia, variation in early spatial patterning within the coenocyte establishes two distinctly different body plans: (1) a bipolar or weakly bipolar pattern that produces a binucleate central cell at maturity (Drusa type) or (2) a tetrapolar pattern that produces a tetranucleate central cell (Penaea type). The Penaea type has not been reported before in Piperaceae.
An evolutionary developmental analysis suggests that tetraspory originated in a female gametophyte with bipolar organization in the common ancestor of Piperaceae. It produced a triploid endosperm derived from one sperm cell and two polar nuclei (lineal descendants of two of the four megaspores). Such functionally bisporic, triploid endosperm is on average of higher heterozygosity and less subject to kin conflict than is the monosporic, triploid endosperm of a Polygonum-type ancestor (Friedman et al., 2008). Functionally bisporic, triploid endosperm may be an intermediate step in the origin of other endosperms produced by tetrasporic female gametophytes. All other female gametophytes within Piperaceae can be seen as phylogenetically and developmentally derived (directly or indirectly) from a functionally bisporic endosperm: the Penaea-type variants within M. naranjoana, the Fritillaria-type within Piper, and the Peperomia-type (and its variants) within Peperomia all ensure inclusion of the entire diploid maternal genome into endosperm (all four megaspores are represented). These endosperms have an even higher average heterozygosity, higher ploidy (5n –15n), and a lower potential for kin conflict than do the functionally bisporic Drusa-type and the monosporic and bisporic endosperms of other Piperales (Friedman et al., 2008). It will be important to see what insights Verhuellia can provide.
Appendix. Voucher specimens of Manekia used in this study. CR = National Herbarium of Costa Rica; USJ = University of Costa Rica Herbarium; HUA = the University of Antioquia Herbarium.
|
FOOTNOTES
1 The authors thank B. Hilje for help in the field in Costa Rica and T. Feild, P. Rudall, E. Schilling, M. Taylor, and one anonymous reviewer for helpful comments on the manuscript. They are grateful for financial support from the McClure Fund (International Center, University of Tennessee), the Department of Ecology and Evolutionary Biology, and IdeaWild (Fort Collins, CO) to T.A. and from the National Science Foundation (DEB 0640792) to J.H.W. 
2 Author for correspondence (e-mail: tavd4{at}mizzou.edu); present address: 109 Tucker Hall, Division of Biological Sciences, University of Missouri, Columbia, MO 65211 USA
LITERATURE CITED
Alberch, P., S. J. Gould, G. F. Oster, AND D. B. Wake. 1979. Size and shape in ontogeny and phylogeny. Paleobiology 5: 296–317.[Abstract]
APG II [Angiosperm Phylogeny Group II]. 2003. An update of the angiosperm phylogeny group classification for the orders and families of flowering plants: APG II. Botanical Journal of the Linnean Society 141: 399–436.[CrossRef][ISI]
Arias, T., R. Callejas, AND A. Bornstein. 2006. New combinations in Manekia, an earlier name for Sarcorhachis (Piperaceae). Novon 16: 205–208.[CrossRef][ISI]
Battaglia, E. 1989. The evolution of the female gametophyte of angiosperms: An interpretative key. (Embryological questions: 14). Annali di Botanica (Roma) 47: 7–144.
Bessey, C. E. 1915. The phylogenetic taxonomy of flowering plants. Annals of the Missouri Botanical Garden 2: 109–164.[CrossRef]
Brink, R. A., AND D. C. Cooper. 1947. The endosperm in seed development. Botanical Review 13: 423–477.[CrossRef]
Brown, W. H. 1908. The nature of the embryo sac of Peperomia. Botanical Gazette (Chicago, Ill.) 46: 445–460.[CrossRef]
Burger, W. C. 1971. Family 41. Piperaceae. In W. C. Burger [ed.], Flora costaricensis. Fieldiana Botany 35: 5–218.
Burger, W. C. 1977. Piperales and the monocots: Alternative hypotheses for the origin of monocotyledonous flowers. Botanical Review 43: 345–393.[CrossRef]
Bulmer, M. G. 1986. Genetic models of endosperm evolution in higher plants. In S. Karlin, and E. Nevo [eds.], Evolutionary processes and theory, 743–763. Academic Press, Orlando, Florida, USA.
Campbell, D. H. 1901. The embryo-sac of Peperomia. Annals of Botany 15: 103–117.
Campbell, D. H. 1902. Recent investigations upon the embryo sac of angiosperms. American Naturalist 36: 777–786.[CrossRef][ISI]
Cass, D. D., D. J. Peteya, AND B. L. Robertson. 1985. Megagametophyte development in Hordeum vulgare. 1. Early megagametogenesis and the nature of cell wall formation. Canadian Journal of Botany 63: 2164–2171.
Chodat, R. 1916. La végétation du Paraguay. III. Hydnoracées. Bulletin de la Société Botanique de Genêve 8: 186–201.
Coccuci, A. E. 1973. Some suggestions on the evolution of gametophytes of higher plants. Phytomorphology 23: 109–124.[ISI]
Dastur, R. H. 1921. Notes on the development of the ovule, embryo sac and embryo of Hydnora africana Thunb. Transactions of the Royal Society of South Africa 10: 27–31.
DeFigueiredo, R. A., AND M. Sazima. 2000. Pollination biology of Piperaceae species in southeastern Brazil. Annals of Botany 85: 455–460.
de Candolle, C. 1923. Piperacearum clavis analytica. Candoella 1: 65–415.
Donoghue, M. J., AND J. A. Doyle. 1989. Phylogenetic analysis of angiosperms and the relationships of Hamamelidae. In P. R. Crane, and S. Blackmore [eds.], Evolution, systematics, and fossil history of the Hamamelidae, vol. 1. Introduction and "lower" Hamamelidae, 17–45. Clarendon Press, Oxford, UK.
Donoghue, M. J., AND S. M. Scheiner. 1992. The evolution of endosperm: A phylogenetic account. In R. Wyatt [ed.], Ecology and evolution of plant reproduction, 356–389. Chapman and Hall, New York, New York, USA.
Endress, P. K., AND A. Igersheim. 1998. Gynoecium diversity and systematics of the paleoherbs. Botanical Journal of the Linnean Society 127: 289–370.[CrossRef][ISI]
Endress, P. K., AND A. Igersheim. 2000. Gynoecium structure and evolution in basal angiosperms. International Journal of Plant Sciences 161 ( Supplement_6): S211–S223.[CrossRef][ISI]
Ernst, A. 1908. Ergebnisse neuerer Untersuchungen über den Embryosack der Angiospermen. Verhandlungen der Schweizerischen Naturforschenden Gesellschaft 91: 230–263.
Fagerlind, F. 1938. Wo kommen Tetrasporische durch drei Teilungsschritte vollentwickelte Embryosäcke unter den Angiospermen vor? Botaniska Notiser 461–498.
Fagerlind, F. 1939. Kritische und revidierende Untersuchungen über das Vorkommen des Adoxa (Lilium)-Typus. Acta Horti Bergiani 13: 1–49.
Fagerlind, F. 1944. Der tetrasporische Angiospermen-Embyrosäcke und dessen Bedeutung für das Verhältnis der Entwicklungsmechanik und Phylogenie des Embryosacks. Arkiv för Botanik 31A: 1–71.
Favre-Ducharte, M. 1984. Homologies and phylogeny. In B. M. Johri [ed.], Embryology of angiosperms, 697–734. Springer-Verlag, Berlin, Germany.
Flint, F. F., AND D. A. Johansen. 1958. Nucleocytoplasmic relationships in the Fritillaria type of megagametogenesis. American Journal of Botany 45: 464–473.[CrossRef][ISI]
Floyd, S. K., AND W. E. Friedman. 2000. Evolution of endosperm developmental patterns among basal flowering plants. International Journal of Plant Sciences 161: S57–S81.[CrossRef][ISI]
Friedman, W. E. 2006. Embryological evidence for developmental lability during early angiosperm evolution. Nature 441: 337–340.[CrossRef][Medline]
Friedman, W. E., E. N. Madrid, AND J. H. Williams. 2008. Origin of the fittest and survival of the fittest: Relating female gametophyte development to endosperm genetics. International Journal of Plant Sciences 169: 79–92.[CrossRef][ISI]
Friedman, W. E., AND J. H. Williams. 2003. Modularity of the angiosperm female gametophyte and its bearing on the early evolution of endosperm in flowering plants. Evolution 57: 216–230.[ISI][Medline]
González, F., AND P. J. Rudall. 2003. Structure and development of the ovule and seed in Aristolochiaceae, with particular reference to Saruma. Plant Systematics and Evolution 241: 223–244.[CrossRef][ISI]
Gvaladze, G. E., AND M. S. Akhalkatsi. 1990. Is the Polygonum-type embryo sac primitive? Phytomorphology 40: 331–337.
Haig, D. 1986. Conflicts among megaspores. Journal of Theoretical Biology 123: 471–480.[CrossRef][ISI]
Haig, D. 1990. New perspectives on the angiosperm female gametophyte. Botanical Review 56: 236–274.[CrossRef]
Haig, D., AND M. Westoby. 1989. Parent-specific gene expression and the triploid endosperm. American Naturalist 134: 147–155.[CrossRef][ISI]
Håkansson, Å. 1923. Studien über die Entwicklungsgeschischte der Umbelliferen. Acta Universitatis Lundensis 18(7): 1–120.
Hamby, R. K., AND E. A. Zimmer. 1992. Ribosomal RNA as a phylogenetic tool in plant systematics. In P. Soltis, D. Soltis, and J. J. Doyle [eds.], Molecular systematics of plants, 51–90. University of Illinois Press, Urbana, Illinois, USA.
Herr, J. M. Jr. 1967. On the nature of variation. Phytomorphology 17: 200–207.[ISI]
Hjelmqvist, H., AND F. Grazi. 1965. Studies on variation in embryo sac development. Botaniska Notiser 118: 329–360.
Hufford, L. 1996. Ontogenetic evolution, clade diversification, and homoplasy. In M. J. Sanderson, and L. Hufford [eds.], Homoplasy: The recurrence of similarity in evolution, 271–301. Academic Press, San Diego, California, USA.
Igersheim, A., AND P. K. Endress. 1998. Gynoecium diversity and systematics of the paleoherbs. Botanical Journal of the Linnean Society 127: 289–370.[CrossRef][ISI]
Jaramillo, M. A., AND P. S. Manos. 2001. Phylogeny and patterns of floral diversity in the genus Piper (Piperaceae). American Journal of Botany 88: 706–716.
Jaramillo, M. A., P. S. Manos, AND E. A. Zimmer. 2004. Phylogenetic relationships of the perianthless Piperales: Reconstructing the evolution of floral development. International Journal of Plant Sciences 165: 403–416.[CrossRef][ISI]
Johnson, D. S. 1900. On the endosperm and embryo of Peperomia pellucida. Botanical Gazette (Chicago, Ill.) 30: 1–11.[CrossRef]
Johnson, D. S. 1902. On the development of certain Piperaceae. Botanical Gazette (Chicago, Ill.) 34: 321–340.[CrossRef]
Johnson, D. S. 1910. Studies in the development of the Piperaceae. I. The suppression and extension of sporogenous tissue in the flower of Piper betel L. var. monoicum C. DC. Journal of Experimental Zoology 9: 715–749.[CrossRef][ISI]
Johnson, D. S. 1914. Studies of the development of the Piperaceae. II. The structure and seed development of Peperomia hispidula. American Journal of Botany 1: 323–339.[CrossRef][ISI]
Johri, B. M., K. B. Ambegaokar, AND P. S. Srivastava. 1992. Comparative embryology of angiosperms. Springer-Verlag, Berlin, Germany.
Kanta, K. 1962. Morphology and embryology of Piper nigrum L. Phytomorphology 12: 207–211.[Medline]
Lei, L.-G., Z. Y. Wu, AND H. X. Liang. 2002. Embryology of Zippelia begoniaefolia (Piperaceae) and its systematics. Botanical Journal of the Linnean Society 140: 49–64.[CrossRef][ISI]
Maddison, W. P., AND D. R. Maddison. 2001. MacClade 4: Analysis of phylogeny and character evolution, version 4.03. Sinauer, Sunderland, Massachusetts, USA.
Madrid, E. N., AND W. E. Friedman . In press. Female gametophyte development in Aristolochia labiata Willd. (Aristolochiaceae). Botanical Journal of the Linnean Society<
