J. Cell Sci. 63, 125-134 (1983) 125Printed in Great Britain © The Company of Biologists Limited 1983

ULTRASTRUCTURAL ASPECTS OF SPOROGENESIS INTHE APOGAMOUS FERN DRYOPTERIS BORRERI

E. SHEFFIELD*, S. LAIRD AND P. R. BELLDepartment of Botany and Microbiology, University College London, Gotuer Street,London WC1E 6BT, U.K.

SUMMARY

The events that accompany sporogenesis in the apogamous fern Dryopteris borreri parallel thoseseen in sexually reproducing ferns. Organelles dedifferentiate and redifferentiate, and form adiscrete band across the equator of dyads; nuclear vacuoles and lipid spherosomes appear duringprophase, and the major part of the ribosome population is removed and subsequently replacedduring meiosis. Similar events have been found to occur during sporogenesis in mosses, gymno-sperms and angiosperms, and are therefore characteristic of the meiotic transition from sporophyteto gametophyte, even in the absence of a transition from diplophase to haplophase.

The novel aspects of meiosis in D. borreri are largely those connected with the restitution eventthat precedes meiosis 1 and serves to maintain the sporophytic chromosome number throughout thelife cycle of this fern. Pre-meiotic cells are regularly found to be cleaved by annular wall ingrowths,which traverse the cytoplasm but not the nuclei. The significance of these ingrowths in relation totheories concerning apogamy and plant cell division are discussed.

INTRODUCTION

The ultrastructural changes that occur during sporogenesis in sexually reproducingferns have been described (e.g., see Sheffield & Bell, 1979; Marengo, 1979; Bell,1981) and resemble those seen during sporogenesis in higher plants (e.g., see Dickin-son & Heslop-Harrison, 1977). A regular sequence of nuclear events, organellededifferentiation and redifferentiation and ribosome elimination and replenishmentcharacterize the process in all the ferns, gymnosperms and angiosperms investigatedto date. These occurrences have been attributed to the change in phase fromsporophytic to gametophytic that takes place during meiosis (Sheffield & Bell, 1979).Meiosis in these plants is characterized by a reduction from the diplophase to thehaplophase chromosome number. No ultrastructural study has yet been made ofmeiosis in a life cycle in which the gametophytic chromosome number is unreduced.

The sporophytic ploidy level has been shown to be maintained throughout the lifecycle of apogamous ferns by means of a restitution event during sporogenesis and thesubsequent generation of a sporophyte from the gametophyte without oogenesis (e.g.,see Dopp, 1932; Manton, 1950). It is in such plants that one can expect to observefeatures associated with the phase change from sporophyte to gametophyte uncom-plicated by the change from diplophase to haplophase. The aim of the present inves-tigation was to examine meiosis in an apogamous species with the same chromosome

•Present address: Department of Botany, University of Manchester, Manchester M13 9PL,U.K.

126 E. Sheffield, S. Laird and P. R. Bell

number as related sexual species in the expectation that differences would becomeapparent from the normal sequence that could be related to the elimination of thehaplophase.

MATERIALS AND METHODS

Fertile fronds were taken from a cultivated plant of Dryopteris borreri, originally collected by MrsP. J. Newbould from St Germains, Cornwall and determined by her to be of the 'diploid' (2n = 82)strain (Manton, 1950). Individual son were excised and fixed in 3 % glutaraldehyde in 0-05 M-phosphate buffer (pH 6-9) at room temperature for 3 h. The material was then washed twice and leftovernight in ice-cold buffer. Post-fixation was in 2% aqueous osmium tetroxide for 2h at 0°C.Dehydration was in acetone and embedding in Durcupan or Epon.

Material for electron microscopy was first sectioned at 4 fun and appropriate stages photographedwith phase-contrast optics. These sections were remounted using the technique of Woodcock & Bell(1967), fine-sectioned and stained with uranyl acetate and lead citrate. They were examined in aHitachi HS9, Siemens Elmiskop 1 or Siemens 101 electron microscope.

Material for light microscopy was sectioned at 1-5 /Jm and stained either in 0-1 % CoomassieBrilliant Blue in 3: 1, ethanol/glacial acetic acid, or in a freshly made saturated solution of SudanBlack B in70%ethanol.

RESULTS

Pre-meiotic sporangia

Successive divisions regularly gave rise to sporangia containing eight archesporialcells characterized by richly populated cytoplasm of the type shown in Fig. 1. Therewas an abundance of ribosomes in these cells, and profiles of endoplasmic reticulumand dictyosomes with attached vesicles were numerous. The plastids were large, oftenexceeding 7 /im in maximum length, and contained starch, several osmiophilic globuliand internal lamellae. Profiles of endoplasmic reticulum were frequently visible ad-jacent to the plastids. The mitochondria were smaller than the plastids, never exceed-ing 3 /zm in maximum diameter, although profiles indicating that both these classesof organelle were undergoing division were found at this stage. The mitochondriacontained finger-like villi, and their stroma appeared less electron-opaque than thesurrounding cytoplasm. The cell walls of this and of preceding stages were highlyosmiophilic and generally formed a dark boundary around the cells, in both the lightand electron microscope.

Fig. 1. Cytoplasm of one of eight archesporial cells, which become spore mother cells.The organelles are well differentiated, ribosomes numerous, and the cell wall very osmio-philic. X 30 000. In all micrographs :m, mitochondrion;/), plastid;«, nucleus ;«i, cell wall.

Fig. 2. Portion of one of the spore mother cells from the inset, a slightly later stage ofdevelopment than that of Fig. 1. The dark lines traversing the cells of the inset can be seento consist of highly osmiophilic wall material. This material does not extend across thenucleus, although some discontinuities can be seen in the nuclear envelope. Fewerribosomes and profiles of endoplasmic reticulum can be seen and the plastids have dedif-ferentiated. No envelope can be distinguished at the boundary of the plastids. X16000.Inset: 4/Um section of the sporangium yielding Fig. 2. X220.

Sporogenesis in Dryopteris borreri 127

Figs 1-2

128 E. Sheffield, S. Laird and P. R. Bell

Pre-leptotene /prophase

The inset in Fig. 2 shows the phase-contrast appearance of the next stage ofdevelopment of the eight-celled sporangia, during which lengths of wall materialcould be seen extending across the cells. Fig. 2 shows the electron microscopic viewof this wall material, which traversed the cytoplasm but was never found to extendacross the nuclei. The growth of this partitioning coincided with the return of thenuclei to an interphase condition after an apparently normal prophase and metaphase.Some discontinuity of the nuclear envelope could be detected at this time, but en-velope dissolution and anaphase chromosome separation were never observed. Thecytoplasm of these cells contained fewer ribosomes than the preceding stage, but thenumber of vacuolar and vesicular profiles had increased considerably. The plastidsstill contained starch and a few globuli, but no internal lamellae or plastid envelopecould be resolved. The mitochondria remained unchanged, though appeared to havedarker stroma in relation to the more diffuse cytoplasm of these cells. The boundingand partitioning walls were very highly osmiophilic and contained numerous multi-layered regions.

Late prophase

Fig. 3 shows the stage of development that followed that of Fig. 2. The sporemother cells rounded off from one another, but remained connected by cytomicticchannels measuring up to 20 nm in width. The cytoplasm contained large numbersof vacuoles, endoplasmic reticulum and membranous profiles, but very fewribosomes. The nuclei were characterized by invaginations of the inner membrane ofthe envelope (nuclear vacuoles), which frequently occupied almost half the volumeof the nucleus. Fragments of electron-opaque material were visible in these vacuolesbut their chemical nature was not determined. The plastids were of widely rangingshapes and sizes, contained no starch, and still showed no signs of a bounding en-velope or internal lamellae. Electron-opaque inclusions, such as that illustrated by thebottom inset in Fig. 4, were found in the peripheral cytoplasm at this stage. Theseinclusions were similar in staining properties and appearance to aggregates ofribosomes and persisted throughout meiosis. A fine fibrillar layer was visible outsidethe plasmalemma (bottom inset, Fig. 4) that surrounded the spore mother cell.Another new class of inclusion appeared at this stage in the form of large, single

Fig. 3. Spore mother cell, prophase, showing extensive development of nuclear vacuoles(nv). Plastids differ widely in shapes and sizes, stroma are very electron-opaque. Arrowsindicate membrane-bound inclusions, cc, cytomictic channels; t, tapetum. X4000.

Fig. 4. Portion of dyad cytoplasm showing band of organelles forming the equatorialplate. Envelope profiles distinguishable at some parts of plastid boundaries. Membrane-bound inclusion (arrow) resembles those of spore mother cells. Note scarcity of freeribosomes. X40000. Top inset: 4ftfn section of spore mother cell in prophase showingphase-contrast appearance of membrane-bound inclusions (arrows). X780. Bottom inset:portion of spore mother cell peripheral cytoplasm showing aggregate of ribosome-likebodies. A fine fibrillar layer lies external to the plasmalemma. X45 000.

Sporogenesis in Dryopteris borreri 129

*<•'.

nv

VM

cc

• vlA

O

Figs 3-4

E. Sheffield, S. Laird and P. R. Bell

Figs 5-7

Sporogenesis in Dryopteris borreri 131

membrane-bound spherical bodies (arrows, Fig. 3; top inset, Fig. 4). The content ofthese inclusions was homogeneously electron-opaque and no sub-structure could bediscerned. They showed an affinity for Sudan Black B, but no reaction to CoomassieBlue when stained and examined in the light microscope.

Dyads

The cytoplasm of spore mother cells altered very little during the remainder ofmeiosis I, and Fig. 4 shows the appearance of dyads. These cells were traversed bya band of organelles running between the telophase nuclei; Fig. 4 illustrates thesimilarity between the plastids and mitochondria at this time. Some of the plastidswere found to contain starch and globuli, and most had regained an envelope profileand internal lamellae. Occasional profiles indicating organelle division were seen, andthe single membrane-bound inclusions of early meiosis I were unchanged in numberand appearance. Very few ribosomes could be detected, but some profiles of endoplas-mic reticulum were found to bear ribosomes.

Meiosis II

Figs 5 and 6 illustrate the appearance of newly formed spores. Fig. 5 shows the thicklayer of osmiophilic material that regularly lay outside the plasmalemma but internalto the fibrillar layer at this stage, some of which appeared to lie within the sporecytoplasm. Myelin figures could be discerned within this layer, but their fate couldnot be determined. The organelles had redifferentiated by this time, plastids showingstarch, globuli and lamellae and mitochondria, large shelf-like cristae (Fig. 6). Themembrane-bound inclusions remained unchanged in appearance and number.Endoplasmic reticulum profiles and ribosomes were slightly more frequent than in thepreceding stage.

Spores

Fig. 7 shows a spore, surrounded by degenerating tapetum, inside which is the finefibrillar layer seen in spore mother cells, which now lies around and between thespores. No sign of the myelin figures seen in the previous stage could be detectedaround the spores.

The cytoplasm still contained the membrane-bound inclusions seen previously andthe organellar components were slightly more differentiated but otherwise unchanged.

Fig. S. Boundary between young spore and degenerating tapetum (t). A region occupiedby highly osmiophilic myelin figures lies between the fibrillar layer and the spore plas-malemma, some profiles lie inside the cytoplasm. X22500.

Fig. 6. Cytoplasm showing redifferentiated organelles of young spore. Note distinctenvelope of plastid, cristae within the mitochondrion and membrane-bound inclusion.Ribosomes are still scarce. X70 000.

Fig. 7. One spore of tetrad showing no signs of osmiophilic material seen in Fig. 5.Organelles well differentiated, membrane-bound inclusions still present, also stackedendoplasmic reticulum on the far left and right of the spore. X7200.

132 E. Sheffield, S. Laird and P. R. Bell

The endoplasmic reticulum was found to be banked together in discrete regions,but the ground cytoplasm was otherwise rather empty, ribosomes still being in-frequent.

DISCUSSION

There are many similarities between sporogenesis in D. borreri and the sexuallyreproducing fern Pteridium (Sheffield & Bell, 1979). These include the developmentof cytomictic channels, a fall and then a rise in ribosome number, the aggregation oforganelles in a band between telophase nuclei at division I and a cycle of organellededifferentiation and redifferentiation; all of which have also been found to charac-terize sporogenesis in heterosporous ferns and higher plants (see Bell, 1981, forreview). The formation of nuclear vacuoles during prophase also resembles that seenin a wide range of plants and animals (Sheffield, Cawood, Bell & Dickinson, 1979),although the functional significance of these inclusions, and their form in vivo is stillunclear. The appearance of bodies resembling nucleoloids or aggregates of ribosomesin meiocytes of D. borreri also mirrors that seen in Pteridium at similar stages (Shef-field & Bell, 1979).

The major difference between the cytoplasmic events during sporogenesis in D.borreri and that of non-apogamous plants lies in the restitution stage of meiosis. Theincomplete wall formation that precedes meiosis in all the sporangia of the presentstudy cannot be reconciled with previous studies of sporogenesis in apogamous ferns.In her "generalized description of the features of meiosis in apogamous ferns", Man-ton (1950) identified four distinct types of sporangial development occurring on thesame plant. In the first, 16 spore mother cells result from four successive archesporialcleavages, and the 64 spores subsequently formed abort. In the second type, threearchesporial cleavages are normal, but the last division is imperfect, anaphase separ-ation does not occur, and interphase becomes re-established in the eight mother cells.The third type differs from the second in that "the cytoplasmic activities whichnormally result in cell cleavage may not be entirely suppressed, but may be presentin unco-ordinated forms.. .". "The nuclei become irregularly lobed, cell walls par-tially crossing the cell may be laid down, and sometimes cleavage into two unequalportions containing different-sized pieces of the restitution nucleus may result". Thistype of sporangium is suggested to be an 'imperfect' version of the type two sporangia.A fourth type of sporangium contains only four giant mother cells as a consequenceof two restitution events, and only 16 spores result. Exhaustive examination of thematerial used in the present study has so far failed to reveal sporangia of types one,two and four, yet there was no indication that sporangial development was in any wayimperfect, or that abortive spores were formed as a result of the events observed. Thismay be because the 'diploid' material examined behaved in a more uniform mannerthan the largely 'triploid' material of Manton. Further apogamous species will clearlyneed to be investigated at the ultrastructural level before the present study can beincorporated into a new generalized scheme of apogamous sporangial development.

The ingrowth of cell wall material observed during the restitution phase is of

Sporogenesis in Dryopteris borreri 133

considerable interest with regard to cell division in plants. Cytokinesis in somatic cellsof D. borreri occurs as a result of phragmoplast, followed by cell plate formation,regarded as the usual mode of division in vascular plants (Pickett-Heaps, 1969). Theannular furrowing seen during restitution in D. borreri mimics the mode of celldivision of prokaryotic organisms, and early stages of division oiSpirogyra (Fowke &Pickett-Heaps, 1969). Peripheral extension of cell walls has been observed in wheatcells in which karyokinesis is inhibited by caffeine (Pickett-Heaps, 1969) and exten-sions have been observed during normal meiosis in archegoniates. A small wall exten-sion can be seen in anaphase II meiocytes of Onoclea (Marengo, 1977, fig. 1) andsubstantial wall ingrowths characterize both divisions of moss meiosis (Brown &Lemmon, 1982a,b). It therefore seems that wall stubs are a normal feature of arch-egoniate meiotic divisions, but that the lack of a cell plate can result in their extensioninto deep ingrowths.

After preliminary examination of D. borreri sporangia it was suggested that themembrane-bound cytoplasmic inclusions seen in spore mother cells were of causalsignificance to the apogamous life cycle (Bell, 1979). Such bodies are certainly absentfrom Pteridium meiocytes but have now been found in spore mother cells of thesexually reproducing ferns Dryopteris filix-mas (Sheffield & Bell, unpublished ob-servations) and Onoclea (Marengo, 1979). Meiocytes and spores of mosses also con-tain similar bodies (Brown & Lemmon, 1982a), and they are now therefore thoughtto be unconnected with apogamy. Their appearance in D. borreri coincides with thedisappearance of the highly osmiophilic wall ingrowths, and it is possible that thelipidic component of these extensions contributes to these inclusions. The chemicalnature of these inclusions remains uncertain, but their affinity for osmium and SudanBlack B and lack of affinity for protein stains suggest that they are predominantlylipidic. Brown & Lemmon (1982a,b) and Marengo (1979) suggest that the inclusionsof Rhynchostegium, Amblystegium and Onoclea are lipidic, and the D. borreri bodiesare indistinguishable from these inclusions.

One unique feature of sporogenesis inZ). borreri remains, namely the highly osmio-philic cell walls. The exterior of the archesporial cells, meiocytes and young sporesare strikingly more osmiophilic than those of Pteridium or Onoclea, and are thus morereminiscent of those of mosses (Brown & Lemmon, 1982a). The myelin figuresexterior, and partially interior, to the young spores find no equivalent during the samestage in any other study of meiosis. Similar profiles characterize early prophase inspore mother cells of Pteridium (Sheffield & Bell, 1979) and Lycopodium (Pettit,1978) but their significance and subsequent fate during later stages of spore develop-ment in D. borreri remains a matter for conjecture.

In conclusion, the present study demonstrates that nuclear vacuole development,cycles of ribosome expunging and replenishment, and organellar dedifferentiationand redifferentiation do characterize the meiotic transition from sporophyte togametophyte, even in the absence of a transition from diplophase to haplophase.Whether they inevitably accompany the change in phase from sporophyte togametophyte in the absence of meiosis must await an ultrastructural study of apos-pory.

134 E. Sheffield, S. Laird and P. R. Bell

REFERENCES

BELL, P. R. (1979). The contribution of the ferns to an understanding of the life cycles of vascularplants. In Tlie Experimental Biology of Ferns (ed. A. F. Dyer), Experimental Botany, vol. 14,pp. 57—85. London: Academic Press.

BELL, P. R. (1981). Megasporogenesis in a heterosporous fern: features of the organelles in meioticcells and young megaspores. J. Cell Sci. 51, 109-119.

BROWN, R. C. & LEMMON, B. E. (1982a). Ultrastructure of sporogenesis in the moss, Amblys-tegium riparum. 1. Meiosis and cytokinesis. Am.J. Bot. 69, 1096-1107.

BROWN, R. C. & LEMMON, B. E. (19826). Ultrastructural aspects of moss meiosis: cytokinesis andorganelle apportionment in Rkynchostegitan serrulatum.J. Hattori Bot. Lab. 53, 41—50.

DICKINSON, H. G. & HESLOP-HARRISON, J. (1977). Ribosomes, membranes and organelles duringmeiosis in angiosperms. Phil. Trans. R. Soc. B. 277, 327-342.

DOPP, W. (1932). Die apogamie bei Aspidium remotum Al. Br. Planta 17, 86-152.FOWKE, L. C. & PICKETT-HEAPS, J. D. (1969). Cell division in Spirogyra. II. Cytokinesis. J.

Phycol. 5, 273-281.MANTON, I. (1950). Problems ofCytology and Evolution in the Pteridophyta. Cambridge University

Press.MARENGO, N. (1977). Ultrastructural features of the dividing meiocyte of Onoclea sensibiHs. Am.

J. Bot. 64, 600-601.MARENGO, N. (1979). The fine structure of the pre-meiotic stages of sporogenesis in Onoclea

sensibilis. Am. FernJ. 69, 87-91.PETTTT, J. M. (1978). Regression and elimination of cytoplasmic organelles during meiosis in

Lycopodium. Grana 17, 99-105.PICKETT-HEAPS, J. D. (1969). The evolution of the mitotic apparatus. An attempt at comparative

ultrastructural cytology in dividing plant cells. Cytobios 3, 257-280.SHEFFIELD, E. & BELL, P. R. (1979). Ultraatructural aspects of sporogenesis in a fern, Pteridium

aquilinum (L.) Kuhn. Ann. Bot. 44, 393-405.SHEFFIELD, E., CAWOOD, A. H., BELL, P. R. & DICKINSON, H. G. (1979). The development of

nuclear vacuoles during meiosis in plants. Planta 146, 597—601.WOODCOCK, C. L. F. &BELL, P. R. (1967). A method for mounting 4 fi resin sections routinely for

ultrathin sectioning. Jl R. microsc. Soc. 87, 485-487.

(Received 9 February 1983-Accepted 28 February 1983)