| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
Structure and Development |
2Institute of Botany, Academia Sinica, Nankang, Taipei, Taiwan 11529 Republic of China; 3Department of Botany-Microbiology, Ohio Wesleyan University, Delaware, Ohio 43015 USA
Received for publication September 17, 2002. Accepted for publication December 19, 2002.
 |
ABSTRACT |
|---|
 TOP ABSTRACT INTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONLITERATURE CITED |
|---|
Key Words: Annonaceae • anther development • pollen • polyad • septate anther • sporogenesis • tapetum
|
INTRODUCTION |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONLITERATURE CITED |
|---|
and in the Ternstroemiaceae (Lin, 1998
). Both types of anthers occur within the Annonaceae, a magnoliid family of 130 genera and 2500 species. Septate anther morphology in Annonaceae has been reported by a long series of authors: Moore (1895)
illustrated septa in the anthers of Cardiopetalum calophyllum Schlechtendal; Lecomte (1896)
described anther septa in a species of Monodora; Herms (1907)
described and illustrated strips of sterile tissue in the developing anthers of Asimina triloba (L.) Dunal; and Samuelsson (1914)
reported similar structures in the anthers of Annona. The recorded number of genera with septate anthers has continued to increase with closer study: Fries (1959)
reported septate anthers in eight genera, Walker (1971a)
in nine, Van Heusden (1992)
in 11, and Johnson and Murray (1995)
in 17. Steinecke (1993)
further distinguished the anther septa of Annonaceae into "dauerhafte" and "vergangliche" types, representing persistent and ephemeral septa, respectively.
Conflict exists in the literature, however, as to whether or not these septa exist for certain genera. Periasamy and Swamy (1959)
described septum formation in the anthers of Cananga odorata (Lam.) Hook. f. & Thomson, but septa had not been observed in this genus by Oes (1914)
. The anthers of Disepalum (as the segregate genus Enicosanthellum) were reported as septate by Ban (1975)
, but both Walker (1971a)
and Johnson (1989)
concluded that Disepalum anthers were aseptate. Locke (1936)
described the anthers of Goniothalamus as aseptate, but they were reported by Fries (1959)
and Walker (1971a)
to be septate.
In studies of anther development in three genera of Annonaceae, Periasamy and colleagues (Periasamy and Swamy, 1959
; Periasamy and Kandasamy, 1981
; Periasamy and Thangavel, 1988
) suggest that complexity of anther development in Annonaceae, and the technical difficulty of examining it, may explain such inconsistencies. Three different pathways of septate anther development were described. In Cananga odorata (Periasamy and Swamy, 1959
), the tapetal septa are developed occasionally as a result of sterilization of already differentiated sporogenous cells. In other genera differentiation of anther septa appears to occur earlier. Periasamy and Kandasamy (1981)
reported that in Annona squamosa L. the archesporial cells formed a continuous mass in each anther lobe, but some of them divided periclinally earlier and their inner daughter cells became sterile and formed tapetal septa, whereas those archesporial cells that divided later would give rise to fertile sporogenous cells. In Xylopia nigricans Hook. f. & Thomson, the archesporium, i.e., mass of archesporial cells, was segmented in each anther lobe and the sterile tissue between the archesporia developed into the parenchymatous septa (Periasamy and Thangavel, 1988
). These reports also suggest that the pattern of septum formation in Annonaceae is not uniform in this family but varies among the genera. Endress and Stumpf (1990)
, in contrast, had proposed that all parenchymatous septa in angiosperms have a primary origin (present from the beginning of sporogenous tissue differentiation), while tapetal septa have a secondary origin (arising only secondarily in an initially homogeneous sporogenous tissue).
Our initial interest in these septate anthers was in establishing their reliability as a character for phylogenetic analyses. In some genera, e.g., Xylopia, the septa are readily visible in dehisced anthers, while in others the septa appear to be present in undehisced anthers yet are not visible after dehiscence (Murray, 1993
; Johnson and Murray, 1995
). It appeared at the outset that one of the problems with inconsistencies had to do with different techniques of observation or that different developmental stages were possibly being observed. We undertook a comparative developmental study of anthers of Annonaceae to determine the presence and nature of anther septa in the Annonaceae. We wanted to examine anther development as broadly as possible to understand how processes such as wall development, sporogenesis, formation of the tapetum, and development of pollen units might be related. In particular, we hoped to examine the relationship between the presence of anther septa and the formation of pollen polyads that had been previously reported by several authors (Fries, 1959
; Walker, 1971a
).
|
MATERIALS AND METHODS |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONLITERATURE CITED |
|---|
|
|
RESULTS |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONLITERATURE CITED |
|---|
|
|
|
|
|
From this point, developmental patterns varied considerably among the species examined. Variations involved the synchrony of periclinal divisions of archesporial cells, the fate of sporogenous initials, the number of primary sporogenous cells produced per anther lobe, the degree of mitotic activity of sporogenous cells, the formation of septa, the type of tapetum, the type of pollen unit formed, and the number of pollen units formed per anther lobe. At the male stage of anthesis, anthers dehisced longitudinally and the septa were always degenerated or shrunken so that pollen grains were dispersed from two thecae in the usual angiosperm fashion.
In the following description of the processes of sporogenesis and pollen development, we present the species according to whether they have aseptate anthers, tapetal type (T-type) septa, or parenchymatous type (P-type) septa, and then summarize the characteristics within each species.
Species with aseptate anthers: Artabotrys hexapetalus and Cananga odorata
Artabotrys hexapetalus (Figs. 9–13)
The archesporial cells in the same sporangium differentiated and proceeded to periclinal division synchronously (Fig. 9). After periclinal division, the outer daughter cells (the primary parietal cells) soon underwent further periclinal divisions (Fig. 10), whereas the inner daughter cells (the sporogenous initials) developed into a homogeneous mass of two or three columns of sporogenous cells (Fig. 11). Prior to the pollen mother cell (PMC) stage, the innermost wall layer that directly enclosed the sporogenous mass differentiated as tapetum, and the two (occasionally three) middle layers were already somewhat compressed (Fig. 11). All of the sporogenous cells differentiated into PMCs; each sporangium contained, depending on the size of the anthers, 50–100 spherical PMCs 20–25 µm in diameter. Meiosis type, determined by the distribution of the four nuclei, the orientation of the spindle fibers, and the form of the meiotic tetrads (Sastri, 1957
), was successive (Fig. 12). After meiosis, the callose wall soon disintegrated and the microspores were released. The tapetal cells were enlarged and binucleate before meiosis took place. The tapetal cells remained intact and were gradually compressed during pollen development; the tapetum was therefore of the glandular type (= secretory type of Johri et al., 1992
). At anthesis, pollen was shed in monads, the long axis of which was 45 µm, and the exine was thinner than 1 µm. Tannin cells and small starch grains were present in the connective (Fig. 13). The connective of this species was massive, and the endothecium was continuous between the two medial sporangia (Fig. 13).
|
Species with anthers having tapetal-type (T-type) septa: Annona glabra, Annona montana, Asimina triloba, Fissistigma oldhamii, Monodora minor, and Rollinia mucosa
Six studied species exhibited regular formation of one- or two-cell-layered, T-type septa in their anthers, which made the anther polysporangiate (Figs. 19, 22, 29, 34, 39). Thicker parenchymatous septa developed occasionally (Figs. 21, 22, 34). In each sporangium only one primary sporogenous cell developed, and the primary sporogenous cell directly functioned as the pollen mother cell (Figs. 19, 22, 29, 39); the tapetum was of the plasmodial type (= amoeboid type of Johri et al., 1992
) in Asimina triloba (Fig. 24), Fissistigma oldhamii (Fig. 31), and probably Monodora minor as well (Fig. 35) and of the glandular type in Annona glabra (Fig. 20), A. montana, and Rollinia mucosa (Fig. 40). The tapetum, including the septa, was digested completely during pollen development.
|
|
|
Before meiosis, tapetal cells were enlarged, ca. 25 µm thick, highly vacuolated (Fig. 19), and binucleate. The PMCs were large, spherical, and ca. 70 µm in diameter. Meiosis proceeded successively (Fig. 19). After callose digestion, the four microspores of a tetrad remained attached and enlarged drastically. The tapetal cells lost their cell walls, but protoplast fusion was not observed; the tapetal cells disintegrated gradually during pollen development (Fig. 20). Only one pollen tetrad was produced in each sporangial chamber, so that 60–70 pollen tetrads were released from a single anther. At anthesis, most pollen tetrads were 150 µm long, and the exine was ca. 4 µm thick.
Annona montana
The information is identical to that for A. glabra, except that there were usually 18–20 sporogenous cells in each anther lobe. Thus, 70–80 pollen tetrads were produced by each stamen. The pollen tetrads of A. montana were 150–170 µm long.
Asimina triloba (Figs. 22–25)
The earliest stage observed was of meiotic tetrads. Anthers were wide but short, with each anther lobe containing ca. 20 sporangial chambers and thus 20 PMCs in more or less two longitudinal rows. Septa formed both transversely and longitudinally (Fig. 23). The meiotic tetrads were mostly tetragonal. After callose digestion, the four microspores of each meiotic tetrad stayed together; the tapetal cells then lost the cell wall gradually and diffused into the locule (Fig. 24). The mature pollen was shed in tetrads, each tetrad having a long axis of 75–85 µm; the pollen exine was 2.5–3 µm thick (Fig. 25). At the time of anther dehiscence, starch grains were found in the anther locules, deriving from the decomposing middle layers (Fig. 25).
Fissistigma oldhamii (Figs. 26–33)
Several columns of archesporial cells developed in each anther lobe. The cells then underwent periclinal division asynchronously. As in Annona glabra, those sporogenous initials produced earlier became primary sporogenous cells (Figs. 26, 27), and those formed later became septal initials. There were 40–50 primary sporogenous cells/sporangial chambers per anther lobe. The septa were one cell layer thick, running horizontally, vertically, or obliquely among the sporogenous cells (Fig. 28). The PMCs were 25–30 µm in diameter and meiosis was successive (Figs. 29, 30). Prior to meiosis, the tapetal cells were binucleate, relatively small, and had dense cytoplasm. After meiosis, they gradually lost their cell walls, but the protoplasts remained intact. After callose digestion, the microspores separated to form monad pollen units, and the tapetal protoplasts fused to form a periplasmodium (Figs. 31, 32). The pollen monads were 30–37 µm long and the exine was 1.5–2 µm thick (Fig. 33).
Monodora minor (Figs. 34–36)
The earliest stage observed was immediately after callose digestion. The four microspores remained together as a tetrad, which was tightly enclosed by the tapetum, and microspores had not yet divided mitotically (Fig. 34). The anthers were wide but short. In each anther lobe there were ca. 20 sporangial chambers in more or less two columns (Fig. 35). Rarely, there was only a single column. Two-cell-layered septa formed both transversely and longitudinally (Fig. 35). The tapetal cells were intact at first, but then the cell boundaries were no longer discernible: the protoplasts seemed to fuse together but did not diffuse all over the locule (Fig. 35). Microspores at this stage had a strongly wavy distal exine. This distal exine later stretched up when the tapetum degenerated and the pollen fully expanded. The mature pollen was shed in tetrads, each tetrad having a long axis of 75–85 µm; the pollen exine was 3–4 µm thick (Fig. 36). At anther dehiscence, crystal druses were found mixed with the pollen tetrads in the pollen thecae (Fig. 36). Secondary wall thickenings occurred in both epidermal and endothecial cells.
Rollinia mucosa (Figs. 37–40)
Several columns of archesporial cells developed in each anther lobe, each cell then undergoing periclinal division asynchronously (Fig. 37). The earlier produced sporogenous initials resulting from those earlier periclinal divisions became primary sporogenous cells, and those formed later became septal initials. There were 25–40 primary sporogenous cells/sporangial chambers per anther lobe, distributed in mostly two columns (Fig. 38). The septa were only one cell layer thick and formed transversely, longitudinally, and obliquely among the sporogenous cells. The PMCs were 25–30 µm in diameter, and meiosis was successive. The tapetal cells were relatively small, vacuolated, and with small crystals before meiosis (Fig. 39). The microspores were released into the anther locule after callose digestion, but the tapetal cells retained their cell walls and formed a tapetal framework; the tapetal cells gradually degenerated without cytoplasmic fusion (Fig. 40). The pollen monads were 35–40 µm long, with an exine thinner than 1 µm.
Species with anthers having parenchymatous-type (P-type) septa: Cymbopetalum brasiliense, Froesiodendron amazonicum, Goniothalamus amuyon, Hornschuchia polyantha, Porcelia magnifructa, Xylopia arenaria, and Xylopia parviflora
In these seven species, the sporangial chambers were arranged in a single column in each anther lobe, and only transverse parenchymatous septa formed (Figs. 41, 52, 53, 55, 57, 62, 66). The septa were lined on both surfaces by a single layer of tapetal cells (Figs. 41, 53, 56, 59, 63); however, in the two Xylopia species, the tapetum degenerated before meiosis (Fig. 67). After meiosis, all of the microspores in the same sporangial chamber stayed together as one tetrad or polyad (Figs. 46, 54, 58, 59, 64, 68), therefore each sporangial chamber produced only a single pollen unit. After the tapetum degenerated, the parenchymatous portions of the septa maintained the compartments during pollen development (Figs. 46, 54, 64, 68); the septa were eventually compressed and broken down (Figs. 48, 60, 69). At anthesis, anthers dehisced longitudinally, forming two locules. Pollen tetrads or polyads formed two columns in each locule; usually there was a gap present between two adjacent pollen units (Figs. 48, 60, 69), which was the space left after the breakdown of the parenchymatous portion of the septum. Polyads were usually large, with the long axis 100–250 µm (Figs. 48, 54, 61, 64). The exine developed only on the distal face of each individual grain and was discontinuous between adjacent grains, i.e., acalymmate (Figs. 48, 54, 60, 64). The exine was up to 17 µm in thickness in some species. Pollen of the two Xylopia species was exceptional for the small size of the pollen tetrads (50–60 µm long) and lack of exine on the grains.
Cymbopetalum brasiliense (Figs. 41–48)
A column of 12–14 sporangial chambers developed in each anther lobe. The septa were 5–7 cell layers thick (Fig. 41). Each sporangial chamber contained two sporogenous cells; in some of the preparations, cytoplasmic connections were visible between the contact faces of the two cells (Fig. 41), suggesting that they originated from the same primary sporogenous cell. Prior to meiosis, the tapetal cells were vacuolated, binucleate, retaining the cell wall, and about 25 µm thick (Fig. 41). The PMCs were hemispherical to nearly spherical and were 45–50 µm in diameter (Fig. 41). During meiosis, the two PMCs in the same sporangial chamber underwent meiosis synchronously. Each PMC had an individual callose envelope; after meiosis, the two callose envelopes in the same sporangial chamber, each enclosing a meiotic tetrad, were in contact but not fused, with a distinct boundary between them (Fig. 43). Meiotic division was not clear-cut but left cytoplasmic connections between adjacent daughter cells (Figs. 43, 44). After the callose walls were digested, the eight microspores in a sporangial chamber formed an octad. The tapetal protoplasts formed a periplasmodium (Fig. 45) and soon became digested. At maturity, pollen octads were 200–260 µm long, with an exine 15 µm thick (Figs. 47, 48). During pollen development, the two middle layers of the anther wall, as well as the parenchymatous portion of the septa, accumulated numerous starch grains. When these cells broke down, the grains were released into the anther locule and mixed with the pollen grains (Fig. 48).
Froesiodendron amazonicum (Figs. 4, 49–54)
Archesporial development and periclinal division of archesporial cells were asynchronous in each anther lobe (Figs. 49–51). About 10–13 primary sporogenous cells developed in each lobe, which were differentiated from earlier produced sporogenous initials. Each primary sporogenous initial soon divided to form two secondary sporogenous cells oriented in various directions (Figs. 49–51); thus each sporangial chamber had two sporogenous cells, recognizable by their larger size and paired organization. The septa were 5–7 cell layers thick (Fig. 4). Before meiosis, the PMCs were in pairs; each pair was more or less spherical, and individual PMCs were therefore hemispherical (Figs. 4, 52, 53). The diameter of the adherent pair of PMCs was ca. 40 µm. The cell layer enclosing the pair of sporogenous cells differentiated as tapetum; the tapetal cells were well expanded, ca. 20 µm thick, highly vacuolated, and binucleate (Fig. 4). After meiosis, the two meiotic tetrads in each sporangial chamber joined to form one octad (Fig. 54). The behavior and structure of tapetum were not traced. Pollen octads eventually reached 200 µm in length, with the exine as thick as 15–17 µm. Large, prismatic crystals were present in the mature anther locules (Fig. 54).
Goniothalamus amuyon (Figs. 55–58)
The youngest anthers available of this species showed enlarged sporogenous cells with an undifferentiated tapetum (Fig. 55). There were usually 12–13 sporangial chambers in each anther lobe, but some longer anthers had up to 16 per lobe. There was only one sporogenous cell in each sporangial chamber and the septa were 4–5 cell layers thick (Fig. 55). The primary sporogenous cells functioned directly as PMCs. Before meiosis, each PMC was 30–35 µm in diameter, and the tapetal cells were binucleate and ca. 15 µm thick. The meiotic tetrads were mostly tetrahedral. The tapetal cells remained intact for some time after meiosis (Figs. 55, 56) and then lost the wall and fused into a periplasmodium (Fig. 57). The pollen grains were shed in tetrads, each tetrad having a diameter of ca. 100 µm; the exine was thin, ca. 1 µm thick.
Hornschuchia polyantha (Figs. 59–61)
A single column of 13–18 sporangial chambers developed in each anther lobe. The septa were 5–7 cell layers thick. In each sporangial chamber there were two sporogenous cells. The PMC stage was not observed. Meiosis proceeded synchronously in the two PMCs in the same sporangial chamber, but asynchronously in different sporangial chambers of the same anther lobe. Each meiotic tetrad had an individual callose envelope; the adjacent callose envelopes were in close contact but not fused. After the callose walls were digested, the two tetrads in each sporangial chamber formed an octad. Then the tapetal cells lost their cell walls and their protoplasts seemed to fuse together (Fig. 59), but the periplasmodium quickly disintegrated. Occasionally, thinner, 3–4-cell-layered septa formed, which broke down after callose digestion and microspore expansion; in such cases two octads fused together to form a 16-grained polyad (Fig. 60). The pollen octads were 100–120 µm long and the exine was 6–7 µm thick (Fig. 61). During pollen development, the two middle layers of the anther wall and the parenchymatous portion of the septa accumulated numerous starch grains. When these cells broke down, the grains were released into the anther locules and surrounded the pollen grains (Fig. 60). Starch grains in mature anthers were up to 10 µm in diameter.
Porcelia magnifructa (Figs. 62–64)
A column of 12–14 sporangial chambers developed in each anther lobe, the chambers delimited by septa 5–7 cell layers thick. Six sporogenous cells were produced in each chamber and were irregularly organized (Fig. 62). Prior to meiosis, the tapetal cells were vacuolated, binucleate, and about 15 µm thick. The PMCs were polyhedral and ca. 25 µm wide (Fig. 63). During meiosis, the six PMCs in the same chamber underwent meiosis synchronously, but different chambers in the same anther lobe varied in the timing of meiosis. Every PMC had its own callose envelope and, after meiosis, each meiotic tetrad was enclosed within its own callose wall. Adjacent callose envelopes were in contact but not fused, with a distinct boundary between them. After the callose walls were digested, one pollen polyad with 24 grains formed in each sporangial compartment (Fig. 64). The polyads soon reached 140–160 µm long and did not expand much further. Meanwhile, tapetal protoplasts seemed to fuse to form a periplasmodium enclosing the young polyad, but became digested shortly. Mature polyads were 140–160 µm long, with the exine 5–6 µm thick. Numerous starch grains and some large prismatic crystals were released from the middle layers and the septa to the pollen sac (Fig. 64).
Xylopia arenaria (Figs. 7, 65–67, 69)
The youngest anthers available showed asynchronous periclinal division in archesporial cells (Fig. 65) and primary sporogenous cells differentiated from the earlier produced sporogenous initials. There was a single column of 15–16 primary sporogenous cells developed in each anther lobe, separated by septa 4–5 cell layers thick (Fig. 66). The primary sporogenous cells functioned directly as PMCs. The spherical PMC was ca. 25 µm in diameter (Fig. 67). The endothecium of the anther wall enlarged quickly during sporogenesis, and the secondary wall thickenings started to develop even before meiosis. The one or two middle layers remained until just after meiosis, but the tapetal cells did not enlarge or divide; instead, they began to degenerate just prior to meiosis (Fig. 67) and had nearly disappeared by the end of meiosis. Meiosis was mostly successive and produced tetragonal tetrads. The callose wall that formed around the tetrads was not digested after meiosis, and the microspores remained attached in tetrads. Pollen internal structure developed normally: the cytoplasm was dense, the vegetative nucleus was located at the center, and the crescent-shaped generative cell was next to it. Pollen wall formation, in contrast, was unusual: the callose wall was degraded eventually, and a basically intinous and exineless pollen wall was formed (Figs. 7, 69). At anthesis, the pollen tetrads were 50–60 µm wide. Druses secreted from the connective were mixed with the pollen grains in the anther locules.
Xylopia parviflora (Fig. 68)
Archesporial cells proceeding to asynchronous periclinal division were also observed. Each anther lobe contained a single column of 11–14 sporangenous cells separated by septa 4–5 cell layers thick. The information on the development from sporogenesis to pollen maturation is identical to that described for X. arenaria except that at anthesis, starch grains but not druses were mixed with pollen tetrads in the locule.
|
DISCUSSION |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONLITERATURE CITED |
|---|
), when archesporial cells in one anther lobe undergo synchronous periclinal division and the resulting sporogenous initials all differentiate into a homogeneous mass of primary sporogenous cells, the anther is aseptate (Figs. 9, 11). In contrast, when periclinal division of archesporial cells is asynchronous, as reported here for Annona glabra (Fig. 5), A. montana, Fissistigma oldhamii (Figs. 26, 27), Froesiodendron amazonicum, Rollinia mucosa (Fig. 37), Xylopia arenaria (Fig. 65), and X. parviflora, some of the medial sporogenous initials enlarge and become sporogenous cells while the remaining sporogenous initials do not enlarge and give rise to sterile septa. These septa eventually compartmentalize pollen mother cells within discrete sporangial chambers.
The nature of the septa (T-type vs. P-type) is determined by the number of layers of sterile cells formed between adjacent sporogenous cells: when there are only one or two layers of sterile cells, a T-type septum results and when there are three or more layers of sterile cells, a P-type septum is found. The correlation between thickness of the sterile cell layer and septum type conforms to the findings of Tobe and Raven (1986)
for Onagraceae and Endress and Stumpf (1990)
for a number of other angiosperm families. As far as we can determine, the two types of septa in Annonaceae are both of primary origin, i.e., present from the beginning of sporogenous tissue differentiation and homologous in ontogeny. The exception to this generalization is the occasional septa of Cananga odorata, which seem to be secondary in origin.
Our results agree with those of Buss et al. (1969)
and Lersten (1971)
, who attribute septal formation in Fabaceae to "potentially sporogenous cells" (= sporogenous initials). Our conclusions conflict, however, with those of Periasamy and colleagues, who claimed that the septal initials in Annona squamosa are derived from earlier produced sporogenous initials (Periasamy and Kandasamy, 1981
), while the multisporangia of Xylopia nigricans result from the formation of multi-archesporangia (Periasamy and Thangavel, 1988
).
Steinecke's (1993)
classification of the septa as enduring and ephemeral types should be viewed in the context of anther development and function. Neither P-type nor T-type septa are functional beyond pollen maturation. Tapetal cells disintegrate completely before or during the middle stages of pollen development, thus the T-type septum completely disappears at that time and is not traceable in mature anthers (Figs. 23, 33, 36). The P-type septum, although it persists much longer because of its more resistant parenchymatous portion (Figs. 46, 54, 58, 68), eventually becomes shrunken and compressed during late pollen development and inevitably breaks apart during anther dehiscence (Figs. 48, 60, 61, 69), although vestiges may be visible in open thecae (Fig. 69; see also Johnson and Murray, 1995
, fig. 2b). The ephemeral nature of both T-type and P-type septa helps to explain why the determination of anther septation can be contradictory when different methods are employed and why the presence of septa, especially the T-type septa, has been underreported.
In septate anthers, i.e., polysporangiate anthers, each sporangial chamber confined by septa, regardless of septum type, contains only a single primary sporogenous cell, as far as we have observed. In the two species with septate anthers and monad pollen (F. oldhamii, R. mucosa), the primary sporogenous cell functions as the pollen mother cell directly, and eventually each sporangial chamber produces four pollen monads only. In the species with septate anthers and compound pollen, the primary sporogenous cell either becomes a PMC directly, or divides to form two, four, or eight PMCs, which proceed to meiosis synchronously; after digestion of the callose wall, all of the 4, 8, 16, or 24 microspores cohere together, thus each sporangial chamber produces only one pollen polyad. This latter developmental pattern, is also found in some mimosoid Fabaceae, such as Acacia, Parkia, and Mimosa, which also have septate anthers and compound pollen (Kenrick and Knox, 1979
; Johri et al., 1992
).
Tapetum
Tapetum type of Annonaceae species has been another area of confusion in the literature. Davis (1966)
reported the tapetum of Annonaceae to be glandular (secretory). Parulekar (1970)
reported the secretory type of tapetum for Artabotrys, Cananga, and Miliusa and the plasmodial type for Annona, Goniothalamus, and Uvaria. Periasamy and Kandasamy (1981)
claimed that in Annona squamosa the peripheral tapetum was secretory and collapsed early and that the septal tapetum then contributed to formation of a plasmodial type tapetum. Steinecke (1993)
documented the presence of a secretory tapetum in species of 16 genera of Annonaceae but did not report a plasmodial-type tapetum for any Annonaceae. In her review table (table 25 in Steinecke, 1993
), the tapetum type for Annona, Artabotrys, Cananga, and Goniothalamus is either secretory or plasmodial in different reports. Our experience in this investigation is that tapetum type is difficult to distinguish in Annonaceae with septate anthers without extremely careful study: after callose digestion, the tapetal cells are heavily pressed by the growing microspores/pollen. Meanwhile, the primary walls of the tapetal cells are already decomposing, and it is usually difficult to see the presence or absence of the cell boundary plasmalemma, which determines whether or not cytoplasmic fusion has taken place. In addition, it is also difficult to judge if a periplasmodium-like structure results from degeneration of tapetal protoplasts or from the fusion of active tapetal protoplasts. In our opinion, tapetum type should be determined based on the latter; however, the critical stages could not always be found in our study. It is noteworthy that in the two species of Xylopia in which the tapetum starts to degenerate before meiosis (Fig. 67), the pollen does not develop a typical exine wall (Figs. 68, 69).
Anther development has long been modeled in plants as involving formation of three "germ layers" that give rise to specific cell lineages: the L1 layer gives rise to the epidermis, the L2 layer produces the endothecium, middle layers, outer tapetum, and archesporial cells, and the L3 layer gives rise to the connective and inner tapetum (Satina and Blakeslee, 1941
; Goldberg et al., 1993
). It has also been well accepted that the tapetum is normally of dual origin, from the parietal layer and from the connective, in the angiosperms (Bhandari, 1984
). Nevertheless, our present study suggests that tapetum differentiation may instead be induced by chemical signaling from neighboring sporogenous cells and that ontogenetic origin has little or no significance for tapetum formation. This idea is based on the fact that the tapetum differentiates, without exception, from the single cell layer or rarely two layers immediately enclosing the sporogenous cells and that tapetum differentiation takes place when its associated sporogenous cells are already well-developed (Biddle, 1979
; comparing Fig. 28 with Fig. 29 and Fig. 38 with Fig. 39 of this study). Specialized structures such as the septate anthers of Annonaceae (Steinecke, 1993
; this study) and Fabaceae (Buss et al., 1969
) are especially good examples for showing that the tapetum may be constructed from several different tissue layers, as long as the cell is in close contact with a sporogenous cell. In other words, tapetum differentiation is position-dependent rather than cell autonomous. In the septate anthers of Annonaceae, when the septum is only one or two cell layers thick, the septum can only be of the tapetal type because all septal cells have direct contact with sporogenous cells.
Function of anther septa
Lersten (1971)
proposed that septum formation in angiosperms was an intermediate stage of anther morphology arising under evolutionary pressure for size reduction of anthers and of their sporogenous tissue. This hypothesis does not apply well to Annonaceae, because those genera with septate anthers do not have smaller anthers than those with aseptate anthers; on the contrary, in some members of tribe Bocageeae, e.g., Hornschuchia leptandra D. M. Johnson and H. bryotrophe Nees, the septate anthers are among the largest in the family (Johnson and Murray, 1995
). Nevertheless, the number of sporogenous cells per anther is indeed much reduced in the septate group. For instance, the number of primary sporogenous cells per anther is less than 100 in the majority of septate members in this study. It should be noted as well that the number of anthers per flower in Annonaceae can range from six to several hundred and has no positive or negative correlation with the presence of anther septation in the samples here studied.
A second hypothesis supported by a number of other authors (e.g., Dnyansagar, 1954
; Pacini et al., 1985
; Tobe and Raven, 1986
) is that anther septa provide better contact between the tapetum and developing pollen grains. The tapetum plays a major role in meiocyte/spore nutrition, production of exine precursors, production and release of callase, and production of sporophytic proteins, enzymes, and pollenkitt (Pacini and Franchi, 1993
; Parkinson and Pacini, 1995
). Annonaceae in which septa occur generally produce large, compound, and presumably energetically expensive, pollen units. The strong correlation between anther septation and compound pollen in the Annonaceae has been previously noted (Fries, 1959
; Walker, 1971a
; Guinet and Le Thomas, 1990
). In addition, the compound pollen in the Annonaceae is often enormous, forming the largest nonpollinium pollen units in the angiosperms (Walker, 1971b
).
According to the pollen monograph of Annonaceae (Walker, 1971a
) and our present data, 11 genera (Annona, Asimina, Cananga, Cardiopetalum, Cymbopetalum, Duckeanthus, Froesiodendron, Fusaea, Goniothalamus, Hornschuchia, and Porcelia) have individual pollen units over 100 µm in length (= the "Very Large" pollen category of Erdtman, 1966
). All of these genera have pollen shed in tetrads or polyads. Among these species with large polyads are also found the thickest exines known in flowering plants, e.g., exine 17 µm thick in Froesiodendron amazonicum and 15 µm thick in Cymbopetalum brasiliense (this study). Except for Cananga (with partially and secondarily septate anthers) and Duckeanthus (anther type unknown), all of these genera have septate anthers (septa in Cardiopetalum and Fusaea from authors' unpublished data). Without exception, Annonaceae with septate anthers and compound pollen units seem to have only one pollen unit per sporangial chamber. These correlations suggest that anther septation in Annonaceae maximizes the surface coverage of pollen units by tapetal cells, which provides efficient nutrient supply during pollen development that might be critical for achieving large size and thick exine. The anthers of Cananga might represent an intermediate condition: Cananga has only one row of ca. 14–18 PMCs in each sporangial chamber, but the T/P surface ratio is still very high because each developing PMC/tetrad has direct contact with the tapetum to acquire highly efficient nutrient support, and the pollen tetrads eventually reach ca. 100 µm in length.
Constraints of anther architecture may also partially explain the relationship between an increase in pollen unit size and the decrease in number of pollen mother cells per anther lobe. If pollen units of an anther were to increase in diameter from 70 to 140 µm, for example, the volume of the anther would have to increase nearly eightfold to accommodate an equal number of pollen grains, which might require a major change in overall floral structure. In many species of Annonaceae, pollen enlargement was achieved by forming compound pollen and decreasing the number of pollen grains in an individual anther, but without significant change in anther size or shape. The latter could have led to the reduction of sporogenous tissue and the transformation of potential sporogenous cells into sterile septa. Formation of septa seems consequential to the trends of enlarging pollen unit and reducing sporogenous tissue. The T-type septum provides highly efficient physiological support, and the P-type septum is even more functional: here the tapetum provides the physiological support and the parenchymatous portion the physical support, which maintains well each developing tetrad/polyad in its specific compartment.
However, there is also a phylogenetic component to the presence of anther septa. Rollinia, always considered a specialized genus within the Annona group (Fries, 1959
; Maas et al., 1992
), has small pollen grains in monads, but T-type septa in the anthers. This finding does not support our hypothesis, but the fact that tapetal septa seem to be universal in Annona (Samuelsson, 1914
; Periasamy and Kandasamy, 1981
; Steinecke, 1993
; this study) suggests that septa in Rollinia were present in the common ancestor of the two genera and that there has simply been no selection against septa as smaller single grains evolved in this genus. Wider sampling is in progress to provide data for phylogenetic analysis of septate anther distribution in the family.
Our results clearly show that it will not be sufficient to classify Annonaceae as having aseptate anthers simply because no traces of anther septa can be detected at the time of anther dehiscence, but that microscopic study of development is required. The aseptate state has now been demonstrated by careful developmental study in species of Artabotrys, Cananga (this study), Miliusa (Periasamy and Swamy, 1959
), Anaxagorea, Cleistopholis, Cremastosperma, Duguetia, Friesodielsia, Guatteria, Orophea, Pseuduvaria, and Unonopsis (Steinecke, 1993
), although it certainly occurs more widely. Similarly, septa, particularly T-type septa, are probably underreported in the literature because of their tendency to disappear before pollen is mature and must be documented using the same techniques.
|
FOOTNOTES |
|---|
4 Author for reprint requests (dmjohnso{at}owu.edu
)
|
LITERATURE CITED |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONLITERATURE CITED |
|---|
Bhandari N. N. 1984 The microsporangium. In B. M. Johri [ed.], Embryology of angiosperms. Springer-Verlag, New York, New York, USA
Biddle J. A. 1979 Anther and pollen development in garden pea and cultivated lentil. Canadian Journal of Botany 57: 1883-1900[ISI]
Buss P. A., Jr. D. Galen N. R. Lersten 1969 Pollen and tapetum development in Desmodium glutinosum and D. illinoense (Papilionoideae; Leguminosae). American Journal of Botany 56: 1203-1208[CrossRef][ISI]
Davis G. L. 1966 Systematic embryology of the angiosperms. John Wiley, New York, New York, USA
Dnyansagar V. R. 1954 Embryological studies in the Leguminosae. VI, Inflorescence, sporogenesis, and gametophytes of Dichrostachys cinerea W. and A. and Parkia biglandulosa W. and A. Lloydia 17: 263-274
Endress P. K. S. Stumpf 1990 Non-tetrasporangiate stamens in the angiosperms: structure, systematic distribution and evolutionary aspects. Botanische Jahrbücher für Systematik, Pflanzengeschichte und Pflanzengeographie 112: 193-240
Erdtman G. 1966 Pollen morphology and plant taxonomy. Angiosperms. Hafner, New York, New York, USA
Fries R. E. 1959 Annonaceae. In H. Melchior [ed.], Die natürlichen Pflanzenfamilien, ed. 2, 17aII, 1–170. Duncker and Humblot, Berlin, Germany
Goldberg R. B. T. P. Beals P. M. Sanders 1993 Anther development: basic principles and practical applications. Plant Cell 5: 1217-1229
Guinet P. A. Le Thomas 1990 Current knowledge and interpretation of the pollen characters in Annonaceae and Leguminosae, subfamily Mimosoideae. Review of Palaeobotany and Palynology 64: 109-127[CrossRef][ISI]
Herms W. B. 1907 Contributions to the life history of Asimina triloba. Ohio Naturalist 8: 211-217
Johnson D. M. 1989 Revision of Disepalum (Annonaceae). Brittonia 41: 356-378[CrossRef][ISI]
Johnson D. M. N. A. Murray 1995 Synopsis of the tribe Bocageeae (Annonaceae), with revisons of Cardiopetalum, Froesiodendron, Trigynaea, Bocagea, and Hornschuchia. Brittonia 47: 248-319[CrossRef][ISI]
Johri B. M. K. B. Ambegaokar P. S. Srivastava 1992 Comparative embryology of angiosperms. Springer-Verlag, New York, New York, USA
Kenrick J. R. B. Knox 1979 Pollen development and cytochemistry in some Australian species of Acacia. Australian Journal of Botany 27: 413-427[CrossRef]
Lecomte H. 1896 Sur la formation du pollen chez les Anonacées. Bulletin du muséum d'histoire naturelle (Paris) 2: 152-153
Lersten N. R. 1971 A review of septate microsporangia in vascular plants. Iowa State College Journal of Science 45: 487-497
Lin L. K. 1998 Theaceae (2) Ternstroemioideae. In Flora of Republicae Popularis Sinicae, vol. 50, 70–176. Science Press, Beijing, People's Republic of China (in Chinese)
Locke J. F. 1936 Microsporogenesis and cytokinesis in Asimina triloba. Botanical Gazette (Crawfordsville) 98: 159-168[CrossRef]
Maas P. J. M. et al 1992 Rollinia. Flora Neotropica Monographs 57: 1-189
Moore S. Le M. 1895 The phanerogamic botany of the Matto Grosso expedition, 1891–1892. Transactions of the Linnaean Society of London (Botany) 4: 296-305
Murray N. A. 1993 Revision of Cymbopetalum and Porcelia (Annonaceae). Systematic Botany Monographs 40: 1-121
Oes A. 1914 Beiträge zur Entwicklungsgeschichte der Annonaceen. Verhandlungen der Naturforschenden Gesellschaft in Basel 25: 168-178
Pacini E. G. G. Franchi 1993 Role of the tapetum in pollen and spore dispersal. Plant Systematics and Evolution, Supplement 7: 1-11
Pacini E. G. G. Franchi M. Hesse 1985 The tapetum: its form, function, and possible phylogeny in Embryophyta. Plant Systematics and Evolution 149: 155-185[CrossRef][ISI]
Parkinson B. M. E. Pacini 1995 A comparison of tapetal structure and function in pteridophytes and angiosperms. Plant Systematics and Evolution 198: 55-88[CrossRef][ISI]
Parulekar N. K. 1970 Annonaceae. Bulletin of Indian National Academy of Sciences 41: 38-41
Periasamy K. M. K. Kandasamy 1981 Development of the anther of Annona squamosa L. Annals of Botany 48: 885-893
Periasamy K. B. G. Swamy 1959 Studies in the Annonaceae I: microsporogenesis in Cananga odorata and Miliusa wightiana. Phytomorphology 9: 251-263
Periasamy K. S. Thangavel 1988 Anther development in Xylopia nigricans. Proceedings of the Indian Academy of Science (Plant Science) 98: 251-255
Samuelsson G. 1914 Über die Pollenentwicklung von Anona und Aristolochia und ihre systematische Bedeutung. Svensk botanisk tidskrift utgifven af svenska botaniska föreningen 8: 181-189
Sastri R. L. N. 1957 On the division of pollen mother cells in
