Sporoderm and tapetum ontogeny in Juniperus communis (Cupressaceae). Connective structures between tapetum and microspores

Review of Palaeobotany and Palynology 206 (2014) 23–44

Contents lists available at ScienceDirect

Review of Palaeobotany and Palynology

j ourna l homepage: www.e lsev ie r .com/ locate / revpa lbo

Research paper

Sporoderm and tapetum ontogeny in Juniperus communis(Cupressaceae). Connective structures between tapetumand microspores

Nina Gabarayeva a,⁎, Valentina Grigorjeva a, Svetlana Polevova b

a Komarov Botanical Institute, Popov st., 2, 197376, St. Petersburg, Russiab Moscow State University, Leninski Gory, 1, 119991, Moscow, Russia

⁎ Corresponding author. Tel.: +7 812 346 36 43.E-mail address: [email protected] (N. Gabarayeva).

http://dx.doi.org/10.1016/j.revpalbo.2014.03.0040034-6667/© 2014 Elsevier B.V. All rights reserved.

a b s t r a c t
a r t i c l e i n f o
Article history:Received 11 June 2013Received in revised form 18 December 2013Accepted 25 March 2014Available online 3 April 2014

Keywords:Sporoderm developmentTapetumSelf-assemblyConnective filament

The purpose of this work was to study in detail the successive stages of pollen wall development in Juniperuscommunis and intimate connection between the microspores and nutritive tapetum with TEM. Another goalwas to clear up the mechanisms underlying the developmental processes. The key stages observed are: theappearance of the initial outer glycocalyx layer as a row of spherical units at the early tetrad stage, followed bythe emergence of the underlying radially oriented string-like units, resulting in the formation of a reticulatelayer. After sporopollenin accumulation, spherical units turn into granules. String-like units start from themicro-spore plasma membrane, pierce the glycocalyx layer, pass between the primexine granules and through callosejacket and come out to the anther loculus, reaching the tapetal cells. Their substructure and size are similar toviscin threads, but not their function and chemical composition. All the structures, observed in sporoderm devel-opment, correspond to subsequentmesophases of amicellar system. The latter develops by self-assembly, undergenomic control, regulating increasing concentration of glycoprotein surfactants in the periplasmic space. A thicknetwork of filaments, connecting microspores and tapetum, supplies microspores with necessary nutrients andtestifies the opinion that, in the process of development, exine is a living, dynamic system, arranged fromplasmodesmata-like units.

© 2014 Elsevier B.V. All rights reserved.

1. Introduction

There are several objectives of this study. In a series of ontogeneticstudies, we have endeavored to confirm our recent hypothesis(Gabarayeva and Hemsley, 2006; Hemsley and Gabarayeva, 2007) onthe participation of self-assembly processes to spore/pollen wall devel-opment, working on remote groups of plants — ferns, gymnosperms(Cycadopsida, Gnetopsida), and basal and advanced angiosperms (seelist of references). We received a confirmation of our hypothesis, butPinopsida, in particular the Cupressaceae, were not yet studied. Besides,only a detailed ontogenetic study, when all the successive developmen-tal stages are traced, gives an opportunity to understand the structure ofall layers in themature pollen grainwall, layers which are oftenmaskedat maturity. Some previous studies have shown that details such as thesubstructure of the units, composing the sporoderm, allow a wideunderstanding of the exine as a living system, directly connected withnutritive tapetum via connective structures (Rowley and Morbelli,

2009), and one of our aims was to receive a confirmation of this pointof view.

Palynological and embryological bibliography of the genus Juniperusis not vast and restricted by LM and SEM data (Alexandrovski, 1971;Duhoux, 1972, 1975; Meyer and Bernard, 1973; Reguzova, 2004; Fujikiet al., 2005). The pollen grains of Juniperus communis were describedas monoporate, with granulate ornamentation and scattered Ubischbodies (Fujiki et al., 2005).

2. Material and methods

Male cones of Juniperus communis L. were collected in the Garden ofthe Komarov Botanical Institute, St. Petersburg, over 2 years in order toobtain all themain developmental stages for study. Stamens were fixedin 3% glutaraldehyde and 2.5% sucrose in 0.1 M cacodylate buffer, withthe addition of lanthanum nitrate (1%) for better preservation ofthe plasma membrane glycocalyx, at pH 7.4, 20 °C, 24 h. After post-fixation in 2% osmium tetroxide (20 °C, 4 h) and acetone dehydration,the samples were embedded in a mixture of Epon and Araldite. Ultra-thin sections were prepared using a LKB ultratome and diamond knife.Ultrathin sections were contrasted with 2% aqueous solution of uranylacetate and in 0.2% lead citrate. Sections were examined with a Hitachi

http://crossmark.crossref.org/dialog/?doi=10.1016/j.revpalbo.2014.03.004&domain=pdf

http://dx.doi.org/10.1016/j.revpalbo.2014.03.004

mailto:[email protected]

http://dx.doi.org/10.1016/j.revpalbo.2014.03.004

http://www.sciencedirect.com/science/journal/00346667

24 N. Gabarayeva et al. / Review of Palaeobotany and Palynology 206 (2014) 23–44

H-600 transmission electronmicroscope (TEM).Mature acetolysed pol-len grains for SEMweremounted on stubs, coated with gold–palladiumand examined with a Jeol JSM-6390 instrument.

3. Results

3.1. Meiosis

The typical organelle plate appears in anaphase meiocytes ofJuniperus communis in the course of the reduction division of meiosis I(Plate I, 1). The organelle plate includes plenty of mitochondria, largeplastids with starch grains and lipid globules. Two groups of chromo-somes, still very compact after metaphase-I, are located astride theorganelle plate. Later, at the telophase-I stage, nuclei are formed, andtheir chromatin is normally dispersed (Plate I, 2). Then, after pullingapart the organelles of the plate (band), a partial cleavage furrowingtakes place, resulting in the formation of two semi-cells, connected byan isthmus (Plate I, 3, 4, arrows). At the end of meiosis II, division bycentripetal furrowing is continued, resulting in the appearance ofpre-tetrads (Plate I, 5, arrowheads).

3.2. The tetrad period

After completion of meiosis the tetrads are covered with a joint mi-crosporemother cell envelope (MCE) and the callose jacket (Plate II, 1).The plasma membrane of the early tetrad microspores is slightlywavy and lacks a glycocalyx (Plate II, 1, 2 arrowheads). Many active

Plate I. Simultaneous meiosis of special type in Juniperus communis.

1. Anaphase of meiosis I. The equatorial organelle plate consists mainly of plastidscondensed, no nuclear envelopes. Thewall ofmeiocytes contains an electron-luScale bar = 1 μm.

2. Telophase of meiosis I. Nuclei with nuclear envelope are formed, whereas orga3. Telophase of meiosis I. Cytokinesis by partial infurrowing (arrows), resulting to4. Detail of the furrow, shown in 3 (arrows). Scale bar = 1 μm.5. A survey of a pre-tetrad: incomplete centripetal cytokinesis-II of meiosis (arrow

envelope, M = mitochondrion, MCE = microspore mother cell envelope, LG =

Plate II. Early tetrad stage in Juniperus communis. (see on page 4)

1. The plasmamembrane of the postmeioticmicrospores is slightlywavy and freeevery microspore. The tetrad is covered with the primary mother cell envelopsomes, Golgi vesicles, ER cisternae, lipid globules, mitochondria and plastids. S

2. A border of amicrospore and the tetradwall, consisting of callose layer and a comrather thick filaments which pierce through the primary envelope as such. The

3. A smallmagnification shows that filaments, startingwith themicrospore surfacanther loculus. Scale bar = 0.5 μm.

4. A high magnification fragments of 3 shows a thick net of filaments, spitting thestretched to the anther loculus (arrowheads). Some spherical orbicules are in tER=endoplasmic reticulum, F=filaments, LG= lipid globule,M=mitochondPS = periplasmic space, and Ta— tapetum.

Plate III. Young tetrad stage in Juniperus communis. First signs of the glycocalyx. (see on page 5

1. The appearance of the first depositions of the glycocalyx in the narrow periplastetrad surface (arrowheads). Many active dictyosomes in the peripheral micros

2. Details of the primary tetrad envelope, pieced by filaments (arrowheads). Scale3. A portion of a tapetal cell, adjacent to a tetrad. A thin glycocalyx layer (arrow

0.5 μm.Ca = callose envelope, D = dictyosome, M = mitochondrion, MC= m

Plate IV. Young tetrad stage. Glycocalyx in progress. (see on page 6)

1, 2. Osmiophilic droplets appear on the microspore plasma membrane (arrowhead0.5 μm.

3. Differentiation of the glycocalyx. A row of spherical units appears along the oumembrane (asterisk). Note many radially oriented filament-like units in the ca

4. Magnified portion, marked by an asterisk in 3. Note ring-like units (arrowhead5. Many active dictyosomes in themicrospore peripheral cytoplasm. Scale bar =

microspore cytoplasm, MCE = microspore mother cell envelope, PM = plasm

dictyosomes, Golgi vesicles, mitochondria and RER cisternae are seenin the microspore cytoplasm. A fibrillar character of the MCE is a com-mon feature for most species (Plate II, 1), but higher magnificationshows that fibrils not only cover the surface of the MCE, but also havecontinuations under it (Plate II, 2). Moreover, fibrils are seen undercallose, in the periplasmic space, where they are in contact with theplasma membrane (Plate II, 3). Higher magnification allows the obser-vation of a thick network of fibrils at the border of the MCE and callose,with some long fibrils stretched out to the anther loculus (Plate II, 4, ar-rowheads), and orbicules (which appear very early in the tetrad periodin this species) attached to fibrils (Plate II, 4, arrows).

Later on, the first signs of the glycocalyx appear in the narrow peri-plasmic space between the callose layer and the plasma membrane(Plate III, 1, arrows). Many active dictyosomes are observed in the pe-ripheral microspore cytoplasm. The primary tetrad wall (MCE) ispierced by fibrils (Plate III, 2, arrowheads). The tapetal cells secreteorbicules, developing (asterisks) and mature (arrowhead) orbicules.Developing (asterisks) and mature (arrowhead) orbicules are adjacentto their inner surface (Plate III, 3).

When the young tetrad stage progresses, newcomers are seen in theglycocalyx layer. These are first osmiophilic droplets (Plate IV, 1, 2,arrowheads). Then a row of spherical units appears alongside theouter border of the glycocalyx layer (Plate IV, 3, arrowheads). The struc-ture of the callose layer needs special attention: thewhole layer consistsof radially oriented filamentous-like units, some of them are pointed byarrows (Plate IV, 3). High magnification (Plate IV, 4, arrowheads, about×130,000) of the left part of the glycocalyx layer, shown in 3, reveals

(with large starch grains) andmitochondria. Anaphase sister chromosomes are still highlycentmatrix withmany finefibrils (mother cell envelope) and underlying callose envelope.

nelle plate persists. Scale bar = 2 μm.semicells, connected by an isthmus. Scale bar = 2 μm.

heads). Scale bar= 2 μm.Ach= anaphase chromosome, AL= anther loculus, Ca= calloselipid globule, OP = organelle plate, P = plastid, and TN = telophase nucleus.

of any depositions (arrowheads). A thick callose envelope surrounds thewhole tetrad ande which includes a fibrillar component. The microspore cytoplasm is full of active dictyo-cale bar = 0.5 μm.mon tetrad primary envelope. Highermagnification shows that thefibrillar component isplasma membrane still lacks the glycocalyx. Scale bar = 0.5 μm.

e, run through the periplasmic space, callose, and tetrad primary envelope and go out to the

primary mother cell envelope. Filaments are long and similar to strings of beads and arehe filamentous net (arrows). Scale bar = 0.5 μm.Ca = callose envelope, D = dictyosome,rion,MC=microspore cytoplasm,MCE=mother cell envelope, N=nucleus, P=plastid,

)

mic space between the plasmamembrane and the callose layer (arrows). Filaments on thepore cytoplasm. Scale bar = 0.5 μm.bar = 0.5 μm.s) is covered with developing (asterisks) and mature (arrowhead) orbicules. Scale bar =icrospore cytoplasm, MCE = mother cell envelope, P = plastid, and Ta — tapetum.

s). The glycocalyx layer in the periplasmic space becomes thicker (2, arrows). Scale bars=

ter border of the glycocalyx (arrowheads); clusters of osmiophilic substance at the plasmallose layer (arrows). Scale bar = 0.5 μm.s). Scale bar = 0.5 μm.0.5 μm.Ca= callose envelope, D= dictyosome, G= glycocalyx, LG= lipid globule, MC=a membrane, and V — vacuole.

25N. Gabarayeva et al. / Review of Palaeobotany and Palynology 206 (2014) 23–44

Plate II (caption on page 2).


Plate III (caption on page 2).



details of spherical units: some of them are rings rather than circles.Many active dictyosomes, their vesicles and lipid droplets persist inthe microspore peripheral cytoplasm (Plate IV, 5).

Plate IV (caption on pa

As a transition to the middle tetrad stage, the primexine starts todevelop. Clusters of roundish units appear in the periplasmic space atthe plasma membrane (Plate V, 1, asterisks), below the initial row of

ge 2).


spherical units (Plate V, 1, arrowheads). Radially oriented filamentous-like units in callose are easily discernible (Plate V, 1, arrows). Both tetra-hedral and planar tetrads are found in the same anther loculus (Plate V,2, 3). A layer of spherical units becomes prominent in the glycocalyx,establishing the initial primexine, and fibrillar filaments in the regionof MCE persist (Plate V, 4).

When themiddle tetrad stage progresses, a network of fibrils on thecallose surface becomes thicker (Plate VI, 1). It is clear from this imagethat radially oriented units, which constitute the callose layer (PlateVI, 1, black arrowheads), are large enough to appear different, dependingon the plane of section. They look barred or as spirals (Plate VI, 1,restricted by tiny counter arrowheads), as vials of a huge spiral or as astring (Plate VI, 1, double arrowheads) or as cylindrical units with centralwhite shanks (many other black arrowheads in Plate VI, 1). It is clearfrom this micrograph, that units, which pass through the callose layer,continue as filaments on its surface (Plate VI, 1, white arrowheads). Atthis stage radial units in the primexine emerge. They look like rodlets(Plate VI, 2, arrowheads), but at higher magnification they are similarto strings (Plate VI, 3, arrowheads). Most of the surface filaments arewavy, but some are straight for a short distance, and they can be seenas a long structure, stretched out to the anther loculus (Plate VI, 4,arrowheads). By the end of the middle tetrad stage the primexineconsists of an osmiophilic outer granular layer (Plate VI, 5,white arrow-heads) and an underlying layer which looks similar to a network (PlateVI, 5, asterisk). A fragment of the primordial endexine lamella appears atthe plasma membrane (Plate VI, 5, black arrowhead).

At the late tetrad stage, after SP accumulation on the primexine, theectexine is completed and consists of the outer granular layer andunderlying reticulate layer (Plate VII, 1, 3, distal side). In Plate VII, 3granules are shown by black arrowheads, and filaments are observedpassing between granules (Plate VII, 3,white arrowheads). The first end-exine lamella with a central white line is well developed (Plate VII, 1,arrow). Thick filaments on the microspore surface are rooted deeplyinto the callose layer (Plate VII, 1, arrowheads). From the proximalside, the ectexine has the same structure, and fragments of the secondendexine lamella emerge in the periplasmic space (Plate VII, 2, arrow-head). Coated pits of the microspore plasma membrane are observed(Plate VII, 2, white arrow), as well as some coated vesicles in thecytoplasm. The anther loculus is filled up with groups of orbicules(Plate VII, 4) and filaments.

At the end of the tetrad period, the disintegrating tetrads still pre-serve remnants of callose (Plate VIII, 1–3). The nucleus,with a nucleolus,occupies the central position in themicrospores (Plate VIII, 1).Manymi-tochondria undergo division (Plate VIII, 2). The sites of the local evagina-tions of the ectexine are especially demonstrative about the realsubstructural units of this layer. It becomes clear in such sites (PlateVIII, 3) that the reticulate layer of the primexine consists of radiallyoriented units, joined with each other by tangential connections.These radial units appear as strings (black thicker arrows). They are, infact, the same as neighboring units of the reticulate layer (black thinarrows), but stretched out at the evaginated sites. In the new space,which has appeared in result of callose degradation (asterisk), thesame string-like units pass through callose (arrowheads) and occuroutside of it, in the anther loculus (double arrowheads).

A survey of an anther loculus, including tapetal cells and micro-spores with callose remnants of the late tetrad stage, shows a networkof filaments, connecting the microspores with each other and with thetapetal cells (Plate IX). Filaments are wavy, so that it is possible to ob-serve their fragments only. Clusters of orbicules are seen here andthere (Plate IX, asterisks). Higher magnification shows the connectivefilaments in detail (Plate X). The connections between microsporesare evident going through callose remnants (Plate X, 1, arrowheads).Careful examination reveals (if a section has passed through the centreof a filament) that these structures are rooted into the ectexine; more-over, they are continuations of radially oriented string-like units of theectexine (Plate X, 2–5, arrowheads). The thickness (diameter) of

filaments is considerable — about 100 nm, and they have an electron-lucent core and electron-dense shell in the form of a spiral (Plate X, 5,arrowheads). The number of endexine lamellae gradually increases(Plate X, 2–4, arrows).

3.3. Post tetrad period

Young free microspores are still united with each other (Plate XI, 1)and with the tapetum (Plate XI, 2, 3, arrows) by numerous connectivefilaments.

At the middle free stage of microspores the number of endexine la-mellae has increased to 5–6 (Plate XII, 1–3). The typical central whitelines are seen in them (Plate XII, 3, arrowheads). Connective filamentspersist (Plate XII, 1–3), and their structure is string-like (Plate XII, 2,3). Clusters of very thin, hair-like structures, most of them are parallelto each other, are also seen in the anther loculus (Plate XII, 2 and 4).They look like typical nematic liquid crystals (NLC).

Somewhat later, as a result of massive sporopollenin accumulation,the central white lines of endexine lamellae become less pronounced(Plate XIII, 1). In the apertural site the ectexine and endexine areinterrupted (Plate XIII, 2). At the transition to the late free microsporestage sporopollenin of the exine becomes less osmiophilic (Plate XIII,3, 4), and connective filaments become thinner.

The sporoderm development is completed by the appearance of theintine (Plate XIV, 1–4). The late freemicrospores,with large nucleus andplastids with starch grains, are covered by adhered orbicules (Plate XIV,1). The intine layer, at first fibrillar, starts to develop in the periplasmicspace (Plate XIV, 2) and later acquires an alveolar structure (Plate XIV,3). Under the apertural site the intine layer is considerably thicker andforms a lens-like bulge (Plate XIV, 4). A mature pollen grain, with agranular surface, associated orbicules and a pore is shown in Plate XIV,5.

Plate XV shows other possible cases of connectivefilaments betweenmicrospores and tapetum, which we observed earlier in other species.We discuss these data together with new data on Juniperus communisin the next section.

4. Discussion

Simultaneous meiosis in Juniperus communis is of a special type. Anorganelle plate, typical for simultaneous meiosis, occurs, but the typeof cytokinesis is unusual: semi-furrows appear at telophase-I, and thefinal centripetal cleavage proceeds gradually at telophase-II. The sametype of cytokinesis was described in Magnolia (Brown and Lemmon,1992). Typical simultaneous meiosis with an organelle plate at the lateanaphase/early telophase was observed in some species from a numberof taxa, e.g. in ferns Psilotum nudum and Alsophila cetosa (Gabarayevaet al., 2011a,b).

Brief reconsideration of the main changes, observed in the periplas-mic space in the course of sporoderm development, in the callose layer,and in the microspore–tapetum interactions allows clarification of acomplex and harmonious system of correlations.

4.1. Ectexine framework: development of the glycocalyx

Both our earlier hypothesis (Gabarayeva and Hemsley, 2006;Hemsley and Gabarayeva, 2007) and the observations reported hereconverge to indicate that all events during exine development occur inthe frames of a self-assemblingmicellar system, located in the periplas-mic space, an arena for dynamic colloidal events.

The first structures, discernible in the early glycocalyx layer, arespherical units which occupy all the outer surface of the layer (PlateIV, 3). High magnification reveals many ring-like units through theglycocalyx layer, arranged in radial columns (Plate IV, 4). Sphericalunits correspond to the first micellar mesophase–spherical micelles,and ring-like units in column arrangement reflect evidently the process

Q2


of self-assembly of the second mesophase–cylindrical micelles orso-called strings, the latter are a kind of stretched or underdevelopedcylindrical micelles. The latter are discernible at later stages (Plate VI,2, 3, arrowheads), but in other places the inner part of the glycocalyxlooks like a reticulate layer (Plate VI, 1 and 5, asterisk; Plate VII, 1, 2).The explanation for this has come from images where the microsporesurface is slightly wavy (Plate VIII, 3): this image clarifies the substruc-ture of the inner part of the primexine. At the evaginated sites of themi-crospore surface, string-like units, at a variable extent of stretching, areevident (Plate VIII, 3, small and tiny black arrows), whereas in the leftpart of thismicrograph,where themicrospore surface is not evaginated,these units are not stretched, and the inner primexine layer looksreticulate. Cylindrical and string-like units-micelles correspond totufts, by Rowley – the universal units of exine (Rowley, 1990) which,as this author emphasized, can easily be stretched and compressed,changing their appearance.

Early SP accumulation on the outer layer of spherical micelles stabi-lizes this layer as granular (Plate VI, 5, white arrowheads).

4.2. Structure of the callose layer

Callose plays an important role in the establishment of the glycoca-lyx pattern (Gabarayeva and Hemsley, 2006; Blackmore et al., 2007;Gabarayeva et al., 2009). Being a swelling colloid (a gel: Ball, 1994),the callose envelope may behave actively in respect to the adjacentmacromolecules of glycoproteins of the glycocalyx, promoting astructure-forming process at the interface (Gabarayeva et al., 2009). Itwas also shown that callose is not an obstacle for passage of manysubstances and even rather large particles (Gabarayeva, 1992;Rodríguez-García andMajewska-Sawka, 1992; Gabarayeva and Rowley,1994).

Plate V. The transition to middle tetrad stage. The initiation of the primexine.

1. Clusters of spherical units (asterisks) in the inner part of the glycocalyx and a rothe callose layer (arrows). Scale bar = 0.5 μm.

2, 3. Different types of the tetrads: tetrahedral (2) and planar (3). Arrowheads poin4. A layer of spherical units becomes prominent in the glycocalyx, establishing an

primexine, M = mitochondrion, MC= microspore cytoplasm, MCE = micros

Plate VI. Middle tetrad stage. The appearance of radial units in the primexine. (see on page 10

1. Primexine becomes more prominent as sporopollenin accumulation increasescallose (arrowheads). Scale bar = 0.25 μm.

2, 3. The appearance of the radial units in the primexine (arrowheads). Scale bars =4. One of long filaments protruding from a tetrad wall to anther loculus (arrowh5. The appearance of a fragment of the first endexine lamella at the plasma mem

heads), with underlying reticulate pattern (asterisk). Scale bar = 0.25 μm.CaG = glycocalyx, PEx = primexine, M = mitochondrion, MC = microspore cyand TW — tetrad wall.

Plate VII. Late tetrad stage. The differentiation of the first endexine lamella. (see on page 11)

1, 3. Distal surface of amicrospore. Thickfilaments on themicrospore surface are rooarrowheads). The ectexine consists of outer granular layer and underlying retiScale bar = 0.5 μm.

2. Proximal surfaces of the tetrad microspores. Granular and reticulate layers of thesmall fragment of the second endexine lamellae appears in the periplasmic space (

4. A group of orbicules in the anther loculus. Scale bar = 0.5 μm.Ca = callose enMCE = microspore mother cell envelope, PM= plasma membrane, PS = per

Plate VIII. The end of the tetrad period. Disintegrating tetrads. (see on page 12)

1, 2. Fragments of the disintegrating tetrad. Note partly degrading callose layer. Sca3. Local evaginations of the ectexine show the real substructural units of this layer.

with each other by tangential connections. These radial units look as strings (b(thin arrows), but stretched out at the evaginated sites. In a new space, where caremnants (arrowheads) and come out of callose remnants to the anther loculuGV = Golgi vesicles, LG — lipid globules, M = mitochondrion, N = nucleus, N

It is known from organic polymer science that callose (β-1,3-glucane), getting to a water-based medium, swells and represents akind of amorphous homopolymer colloid. But our observations haveshown that in such biological systems as anther thecal liquid, callose isnot amorphous. Starting with the early tetrad stage, the structure ofthe callose layer in Juniperus communis looks highly ordered. The layerconsists of radially oriented or slightly bent rod-like units (Plates IV, 3,arrows; V, 1, arrows; VI, 1, arrowheads; VIII, 3, arrowheads). Such a struc-turing of the callose layer was observed earlier in other species(Gabarayeva and Grigorjeva, 2011 — Swida; Gabarayeva et al., 2011a,b— Symphytum), where procolumellae were rooted into callose. Butwhat is observed in development of Juniperus communis microsporesdeserves special consideration.

4.3. Connective filaments

“Hairy” mother cell envelopes (MCE), with a surface covered by fi-brils, are a common feature for many (if not all) species. This is alsothe case in Juniperus communis (Plates I, 1; II, 1). However, in this speciessuch structures play an outstanding role: they are connective filamentsbetween microspores and tapetum from the beginning to the end ofexine development. As early as the initial tetrad stage these filamentsare seen not only on theMCE, but also under it, stretching from the mi-crospore plasmamembrane through callose and the envelope ofmothercell (Plate II, 3, 4). These filaments are wavy; nevertheless, in someplaces it is possible to observe their long fragments, stretching out tothe anther loculus (Plates II, 4 and VI, 4, arrowheads). Starting with theappearance of the first glycocalyx accumulations, the connective fila-ments become thicker (Plates III, 2 and VI, 1), and the fact that theypass through the callose layer (Plate VII, 1) and further, into the antherloculus (Plate VIII, 3), becomes evident. The network of the connectivefilaments between the late tetrad microspores and tapetum is

w of spherical units at the outer border of the glycocalyx (arrowheads). Radial filaments in

t to orbicules (3). Scale bars = 1 μm.initial primexine. Scale bar= 0.5 μm.Ca= callose envelope, G= glycocalyx, IPEx= initialpore mother cell envelope, and PS = periplasmic space.

)

. Thick network of filaments on the tetrad surface. Note that filaments pass also through

0.25 μm.eads). Scale bar = 0.25 μm.brane (black arrowhead). The primexine appears as granular from outside (white arrow-= callose envelope, CS = cytosome (autolytic vacuole), D = dictyosome, F — filaments,toplasm, MCE = microspore mother cell envelope, P = plastid, PS = periplasmic space,

ted deeply into the callose layer (white arrowheads), and they pass betweengranules (blackculate layer. The first endexine lamella with central white line is well developed (arrow).

ectexine are well developed, and the first endexine lamella is pronounced (black arrow). Aarrowhead). Coated pit of the plasmamembrane is shownbywhite arrow. Scale bar=0.5 μm.velope, D = dictyosome, Ect = ectexine, F = filaments, GL = granular layer of ectexine,iplasmic space, and RL = reticulate layer of ectexine.

le bars = 1 μm.It is clear that the reticulate layer of the primexine consists of radially orientedunits, joinedlack thicker arrows). They are, in fact, the same as neighboring units of the reticulate layerllose has undergone degradation (asterisk), the same string-like units pass through calloses (double arrowheads). Scale bar = 1 μm.D = dictyosome, ER = endoplasmic reticulum,u = nucleolus, RCa = remnants of callose, and V = vacuole.


Plate VI (caption on page 8).


Plate VII (caption on page 8).


Plate VIII (caption on page 8).


Plate IX. Disintegrated tetrads with callose remnants. Wavy connective filaments between microspores and tapetum fill in the anther loculus.

1. A survey of the anther loculus with parietal tapetal cells, adjacent microspores and connective filaments between them. Groups of orbicules are shown by asterisks. Scalebar = 1 μm. F = connective filaments, CP = cup-like plastids, LG — lipid globules, N = nucleus, RCa = remnants of callose, and Ta = tapetum.


impressive (Plate IX, XI, XII, 1). Microspores from disintegrating tetradsare also connected with each other by filaments (Plate X). Free micro-spores still preserve them (Plates XII and XIII), but filaments disappearat the late free microspore stage, when intine is formed (Plate XIV). Im-ages in Plate X show unequivocally that connective filaments are

continuations of radially oriented string-like units of the ectexine. Thethickness (diameter) of connective filaments is considerable — about100 nm, and they have an electron-lucent core and electron-denseshell in the form of a spiral (Plate X, 5, arrowheads). Such long-livingand “first-hand” connections of microspores with tapetum testify to


the necessity of a close association between the nursing secretorytapetum and developing microspores.

Similar connective structures between tapetum and megasporeswere described for Selaginella species (Morbelli and Rowley, 1993),and in the pictures of megasporangium in Selaginella argentea(Rowley andMorbelli, 2009, Plates II and III) they are really impressive.The authors used the term “wick-like strands” for these connectivestructures, and their idea was that “wicks” were plasmodesmata-likestructures and appeared to be routes for transport of substances fromtapetum to developing megaspores. Thick connective filaments (their

Plate X. Details of the surface of the disintegrated tetrads.

1. Two adjacent microspores with remnants of callose. Connective filaments pass2. A border of a microspore. When sectioned centrally, filaments are thick (arr

membrane. Scale bar = 0.25 μm.3. Two endexine lamellae are formed (arrows), the outer one has the central whi4. Three endexine lamellae are formed (arrows), and the third is close to the plas5. Many active dictyosomes and coated vesicles are seen in the peripheral cytop

ectexine, F = filaments, GV = Golgi vesicles, LG — lipid globules, M = mitocremnants of callose, and V = vacuole.

Plate XI. Young free microspores and tapetum. (see on page 16)

1. Borders of two adjacent microspores of the late tetrad. Plenty of filaments cofilaments.

2, 3. Wavy connective filaments between microspores and tapetum (arrows).AL =

Plate XII. Middle free microspores. (see on page 17)

1. A survey of a border of tapetum and adjacent microspore. Note connective fila2. A border of a microspore and anther loculus contents. String-like connective fila

0.5 μm.3. The exine layers under higher magnification. Ectexine includes outer granular

evident here and there (arrowheads). Scale bar = 0.5 μm.4. A border of a microspore, of tapetal cells, and anther loculus contents. Typical n

that their source is tapetal cells. Scale bar = 1 μm.Ect = ectexine, End = endeN = nucleus, NLC = nematic liquid crystals, PM = plasma membrane, Ta = ta

Plate XIII.Middle free microspore stage in progress. (see on page 18)

1. A border of amicrospore. Ectexine and endexine are formed. The ectexine consilamellae, their central white lines become not well pronounced because of ma

2. An aperture site of a microspore. Note that exine is absent at the aperture site3, 4. Mature exine with associated star-like orbicules (arrows). All the elements of th

0.25 μm.Ap = aperture site, D = dictyosome, Ect = ectexine, End = endexineperiplasmic space, and V = vacuole.

Plate XIV. Late free microspore stage. The appearance of the intine. (see on page 19)

1. A survey of a microspore with associated orbicules. Large plastids with starch g2. A border of a microspore and peripheral cytoplasm. The intine is developing in3. A thick intine layer is formed. Scale bar = 1 μm.4. An aperture site of a pollen grain. Lens-like intine under the aperture. Scale ba5. A SEM image of a pollen grainwith associated orbicules and a pore (arrow). Sca

LG = lipid globule, M = mitochondrion, MC= microspore cytoplasm, Or – or

Plate XV. Connective filaments between microspores and tapetum, observed in other species.

1, 2. Microspores of Persea americana. 1. Tangential section through the surface of twinvaginations of the plasma membrane (black arrowheads). Some invaginationare seen in side view (white arrowheads). 2. The border of a free microspore wThin string-like filaments come out of the microspore surface and fill in the ant2010a, Annals of Botany 105: 939–955).

3. Young freemicrospore of Chamaedoreamicrospadix. The formation of endexineare long (black arrowheads). mc— microspore cytoplasm. Scale bar = 0.5 μm (

4. Young free microspore of Acer tataricum. Arc-like and straight units with dark-cture striae or connective filaments between microspores and tapetum. Scale ba

5. Early tetrad stage in Alsophila setosa. The border between tetraspore and tapetfrom: Gabarayeva et al., 2011a, Grana, 50 (4): 235–262).Ect= ectexine, G= glyand Ts = tetraspore.

substructure and size are similar to those of viscin threads), typical forrepresentatives of Onagraceae and Ericaceae (Skvarla et al., 1978;Hesse, 1983; Hesse, 2010; Hesse et al., 2000), were also demonstratedin cross sections of an anther in Betula pendula at cryofixation withSEM (El-Ghazaly, 2000; Rowley et al., 2003; Rowley and Morbelli,2009, Plate I). Rowley considered (personal communications) that suchconnective structures between tapetum and spores/microspores are acommon feature, and that in most species they were overlooked be-cause, being glycoproteins in composition, they can be revealed onlyby addition of ruthenium red, alcian blue or lanthanum nitrate during

between microspores (arrowheads). Scale bar = 0.25 μm.owheads). Endexine lamella is shown by an arrow, another one is formed at the plasma

te line. Thick filaments on the microspore surface (arrowheads). Scale bar = 0.25 μm.ma membrane. Connective filaments are long (arrowheads). Scale bar = 0.25 μm.lasm of a microspore. Scale bar = 0.25 μm.CV = coated vesicles, D = dictyosome, Ect —hondrion, MC = microspore cytoplasm, N = nucleus, PM = plasma membrane, RCa =

nnect both microspores with each other. Many orbicules (arrowheads) are seen between

anther loculus, MC= microspore cytoplasm, and Ta = tapetum.

ments between them (arrowheads). Scale bar = 1 μm.ments at the microspore surface and nematic liquid crystals at some distance. Scale bar=

and unner “reticulate” layer. The endexine consists of 5 lamellae, the central white line is

ematic liquid crystals, with their parallel orientation, are seen in the anther loculus. Notexine, F = filaments, M = mitochondrion, MC=microspore cytoplasm, Mi =microspore,petum, and V = vacuole.

sts of granular and underlying reticulate layers. The endexine includes 5–6 dark-contrastedssive sporopollenin accumulation. Scale bar = 0.25 μm.(arrowheads). Scale bar = 0.25 μm.e exine become less osmiophilic. Connective filaments have become thinner. Scale bars=, ER = endoplasmic reticulum, M = mitochondrion, MC = microspore cytoplasm, PS =

rains occupy a considerable part of the cytoplasm. Scale bar = 1 μm.the periplasmic space. Scale bar = 1 μm.

r = 1 μm.le bar=5 μm.Ap= aperture site, Ect= ectexine, End=endexine, Ex=exine, In= intine,bicule, P = plastid, PM= plasma membrane, and PS = periplasmic space.

(see on page 20)

o adjacentmicrospores of a former tetrad.Microspore surface processes are seen betweens, filled with osmiophilic substance, are seen in front view and appear circular, and othersith surface processes (arrows) and osmiophilic inclusions inside the surface invaginations.her loculus (arrowheads; see also 1). Scale bars = 1 μm (Fig. 6D, E from: Gabarayeva et al.,

lamellae, the first one is ready (white arrowhead). Outgrowths on the surface of the tectumFig. 8C from: Gabarayeva and Grigorjeva, 2010, Grana, 49: 91–114).ontrasted particles above columellae (arrows) are either a kind of predecessors for the fu-rs = 0.5 μm (Fig. 6c from: Gabarayeva et al., 2010b, Protoplasma 247, No. 1–2: 65–81).um. Long fibrils (arrows) are seen on the tetraspore surface. Scale bars = 0.3 μm (Fig. 9Ecocalyx, MC=microspore cytoplasm, PS=periplasmic space, T= tectum, Ta= tapetum,


Plate XI (caption on page 14).


Plate XII (caption on page 14).


Plate XIII (caption on page 14).


Plate XIV (caption on page 14).


Plate XV (caption on page 14).



fixation (Morbelli and Rowley, 1993). Our usual practice is to add lan-thanum during fixations (including the material of Juniperuscommunis), and in some other species we have found wick-like struc-tures (Plate XV), but in Juniperus communis they are especially evident:asmentioned above, the diameter of connectivefilaments is 100nm. If asection has passed through the centre of filaments, their structure andsize, with central transparent core and dense roundish “spots” on bothsides of the core (Plate X, 5, arrowheads), are similar to those shownfor Epilobium viscin threads (Rowley et al., 1983, Fig. 2; Rowley, 1986,Fig. 7, 1987, Fig. 11). However, this resemblance is purely morphologi-cal. The difference between viscin threads in Onagraceae and filamentsin Juniperus is apparently that the latter do not accumulate sporopollen-in. This explainswhy they are not preserved during preparation for SEM(Plate XIV, 5). However, the reason could also be that cryofixation isneeded to reveal these threads with SEM.

Pettitt (1976, 1979a,b) noted in his work on water fern spores thatthe exine was traversed by a system of interconnected channels, inopen communication with the sporophyte on one side and the gameto-phyte on the other. An interesting observation is that Morbelli andRowley (1993) have found wicks in Lycophyta to pass between exinecomponents, whereas comparable structures in angiosperms (viscinthreads) join with exine processes (Hesse, 1984; Rowley, 1987). Ouranalysis of Juniperus communis micrographs shows that in this speciesconnective filaments pass between primexine granules rather thanthey are a continuation of them (Plate VII, 1, 3, white arrowheads), sothe transport system from tapetum tomicrospores in Juniperus is closerto that of wicks in Lycophyta. But in the light of our micellar hypothesis(Gabarayeva and Hemsley, 2006; Hemsley and Gabarayeva, 2007), wesuggest that connective filaments in Juniperus and wicks in Lycophytaare long cylindrical micelles (=filaments) with the same function: atransport system for delivery of nutrients from tapetum to spores/microspores, whereas viscin threads in Onagraceae and Ericaceaespecies have primarily a pollen connecting function.

We have revised our previous ontogenetic data on a number ofspecies in search of such connective structures. The latter, identical tothose in Juniperus in diameter and structure, were observed in Perseaamericana microspores (Plate XV, 1, 2, arrows; Gabarayeva et al.,2010a,b, Fig. 6D, E), and similar filaments were found in Chamaedoreamicrospadix microspores (Plate XV, 3, black arrowheads; Gabarayevaand Grigorjeva, 2010, Fig. 8C), in Acer tataricum (Plate XV, 4, arrows;Gabarayeva et al., 2010a,b, Fig. 6c) and between tetraspores and tape-tum in the leptosporangiate fern Alsophila cetosa (Plate XV, 5, arrows;Gabarayeva et al., 2011a,b, Fig. 9E).

The main idea of Rowley and coauthors (Rowley et al., 1983, 2003;Rowley and Morbelli, 2009) was that exine is a living, dynamic system,arranged of plasmodesmata-like units (tufts), with the latter being con-nected by their processes with a nourishing tapetum (on one end) andare opened to the cytoplasm (on the other end). What we haveobserved in the course of ontogenetic studies, confirms this ideacompletely. The radially oriented glycocalyx units are rooted intothe cytoplasm and, from the other side, into callose in Trevesia(Gabarayeva et al., 2009), in Symphytum (Gabarayeva et al., 2011a,b),in Swida (Gabarayeva and Grigorjeva, 2011), in Magnolia (Gabarayevaand Grigorjeva, 2012), and in Passiflora (Gabarayeva et al., 2013a,b),and these connective units pass further into tapetal cells in Persea,Chamaedorea, Acer and Alsophila, and are especially prominent inJuniperus. These filaments are most probably long reversed cylindricalmicelles (=tufts) which, with their hydrophilic core and hydrophobicwrapper, are capable of delivering nutrients from the tapetum insidethe spore/microspore cytoplasm.

4.4. Endexine development

First signs of the initial endexine lamella appear at themiddle tetradstage as a shadow at the plasma membrane (Plate VI, 5, black arrow-head), and somewhat later well-developed primordial lamella, with

typical central white line, appears (Plate VII, 1, arrow). Further in devel-opment the number of lamellae increases (Plate X) up to 5–6 slightlyundulating layers, each with a central white line (Plate XII, 3; PlateXIII). The aperture site – a pore – lacks both ectexine and endexine(Plate XIII, 2). The endexine lamellae apparently develop, as in allother species with lamellate endexine we have studied, on the base oflaminate (“neat”) micelles, with their typical water-filled gap betweenbilayers, seen in TEM as a white line. The layers of this micelle systembefore sporopollenin accumulation are flexible, and this confirms theidea of Rowley (1987–1988b) that endexine lamellae are flexible andcan move. Besides, neat micelles are permeable to some substancesbut not others (Ball, 1994), so endexine lamellae can play a role of aselectivefilterwhich picks up the nutrients from the tapetum and trans-fers them through the plasma membrane to the microspore cytoplasm.

5. Conclusions

1. Simultaneous meiosis, with the organelle plate and a semifurrow attelophase-I, followed by the final centripetal cleavage at telophase-II, is characteristic for Juniperus communis.

2. The ectexine starts at the tetrad period as a rowof spherical units onthe glycocalyx framework. Spherical units correspond evidently tothe first micellar mesophase–spherical micelles.

3. The underlying layer, which appears later and looks like a reticulateone, consists in fact of radially oriented units — strings. The latterare a kind of stretched cylindrical micelles, the second micellarmesophase. This layer corresponds to the next mesophase — theso-called middle micelles.

4. At the middle and late tetrad stages spherical units of the outerprimexine layer accumulate sporopollenin, transforming intogranules.

5. Simultaneously primordial endexine lamella with typical centralwhite line appears in the periplasmic space. This white-lined lamellacorresponds to the next micellar mesophase — “neat”, or laminatemicelle, with its gap between bilayers.

6. In the disintegrated tetrads the number of endexine lamellaeincreases to 3, and up to the free microspore period it reaches upto 5–6.

7. Filaments with a diameter of 100 nm are observed as connectivestructures both betweenmicrospores andmicrospores and tapetum,starting with the early tetrad stage to the end of exine development.They start from the microspore plasma membrane, penetrate be-tween granules of ectexine, pass through callose and further throughthe anther loculus and join the tapetal cells. They are evidently verylong cylindrical micelles, stretched in some places up to strings.Their substructure and size are similar to those of viscin threads,but they differ from the latter in lacking sporopollenin and in theirfunction. These connective structures apparently deliver nutrientsfrom tapetum to microspores.

8. Reconsideration of our previous ontogenetic data on several remotespecies, obtained at a proper fixation, has shown the presence of sim-ilar connective structures between tapetum and spores/microspores.This confirms the idea of Rowley and coauthors that exine is a living,dynamic system, arranged of plasmodesmata-like units, the latterconnect the nourishing tapetum with spore/microspore cytoplasm.

9. After completion of the exine development, the intine appears as analveolar layer, with a lens-like bulge – an oncus – under a pore.

10. All the structures, occurring during exine development in Juniperuscommunis, appear most probably on the base of successive micellarmesophase. The same was observed in sporoderm ontogeny ofother species.

Acknowledgements

This work was supported by grant RFBR No. 11-04-00462 for NinaGabarayeva. Special thanks to Bruce Sampson for the language


improvement. I am thankful to my anonymous reviewers. We alsothank our engineer Peter Tzinman for assistance with the HitachiH-600 TEM.

References

Alexandrovski, U.S., 1971. The development of seed-buds and microsporogenesis inJuniperus species. Bot. Zhurn. (Leningrad) 56, 193–201 (In Russian).

Ball, P., 1994. Designing the Molecular World. Princeton University Press, Princeton.Blackmore, S., Wortley, A.H., Skvarla, J.J., Rowley, J.R., 2007. Pollen wall development in

flowering plants. New Phytol. 174, 483–498.Brown, R.C., Lemmon, B.E., 1992. Control of division plane in normal and griseofulvin-

treated microsporocytes of Magnolia. J. Cell Sci. 103, 1031–1038.Duhoux, E., 1972. Évolution structurale de la paroi du grain de pollen du Juniperus

communis L. (Cupressacées) cultivé in vitro, au cours de la phase d'hydratation. CRAcad. Sci. Paris 274, 2767–2770.

Duhoux, E., 1975. L'aperture dans l'exine et dans l'intine externe du pollen de Juniperuschinensis et de Juniperus communis. Pollen Spores 17, 191–201.

El-Ghazaly, G., 2000. New and improved methods in palynology. In: Nordenstam, B., El-Ghazaly, G., Kassas, M. (Eds.), Plant Systematics for the 21st Century. Wenner-Gren.Cent. Int. Symp. , 77, pp. 143–160.

Fujiki, T., Zhou, Z., Yasuda, Y., 2005. The pollen flora Yunnan, China. , vol. 1. Lustre Press/Roli Books p. 17.

Gabarayeva, N.I., 1992. Sporoderm development in Asimina triloba (Annonaceae). I. Thedevelopmental events before callose dissolution. Grana 31, 213–222.

Gabarayeva, N.I., Grigorjeva, V.V., 2010. Sporoderm ontogeny in Chamaedorea microspadix(Arecaceae). Self-assembly as the underlying cause of development. Grana 49,91–114.

Gabarayeva, N.I., Grigorjeva, V.V., 2011. Sporoderm development in Swida alba(Cornaceae), interpreted as a self-assembling colloidal system. Grana 50, 81–101.

Gabarayeva, N., Grigorjeva, V., 2012. Sporoderm development and substructure inMagnolia sieboldii and other Magnoliaceae: an interpretation. Grana 51, 119–147.

Gabarayeva, N.I., Hemsley, A.R., 2006. Merging concepts: the role of self-assembly in thedevelopment of pollen wall structure. Rev. Palaeobot. Palynol. 138, 121–139.

Gabarayeva, N.I., Rowley, J.R., 1994. Exine development in Nymphaea colorata(Nymphaeaceae). Nord. J. Bot. 14, 671–691.

Gabarayeva, N., Grigorjeva, V., Rowley, J.R., Hemsley, A.R., 2009. Sporoderm developmentin Trevesia burckii (Araliaceae). I. Tetrad period: further evidence for participating ofself assembly processes. Rev. Palaeobot. Palynol. 156, 211–232.

Gabarayeva, N.I., Grigorjeva, V.V., Rowley, J.R., 2010a. A new look at sporoderm ontogenyin Persea americana. Micelles and the hidden side of development. Ann. Bot. 105,939–955.

Gabarayeva, N.I., Grigorjeva, V.V., Rowley, J.R., 2010b. Sporoderm development in Acertataricum (Aceraceae). An interpretation. Protoplasma 247, 65–81.

Gabarayeva, N., Grigorjeva, V., Marquez, G., 2011a. Ultrastructure and developmentduring meiosis and the tetrad period of sporogenesis in the leptosporangiate fernAlsophila setosa (Cyatheaceae) compared with corresponding stages in Psilotumnudum (Psilotaceae). Grana 50, 235–262.

Gabarayeva, N., Grigorjeva, V., Polevova, S., 2011b. Exine and tapetum development inSymphytum officinale (Boraginaceae). Exine substructure and its interpretation.Plant Syst. Evol. 296, 101–120.

Gabarayeva, N., Grigorjeva, V., Kosenko, Y., 2013a. I. Primexine development in Passifloraracemosa Brot. Overlooked aspects of development. Plant Syst. Evol. 299, 1013–1035.

Gabarayeva, N., Grigorjeva, V., Kosenko, Y., 2013b. II. Exine development in Passifloraracemosa Brot.: post-tetrad period. Overlooked aspects of development. Plant Syst.Evol. 299, 1037–1055.

Hemsley, A.R., Gabarayeva, N.I., 2007. Exine development: the importance of lookingthrough a colloid chemistry “window”. Plant Syst. Evol. 263, 25–49.

Hesse, M., 1983. Dissimilar pollen tetrad development in Ericaceae and Onagraceaecauses family-specific viscin threads configuration. Plant Syst. Evol. 143, 163–165.

Hesse, M., 1984. An exine architecture model for viscin threads. Grana 23, 69–75.Hesse, M., 2010. Bonding single pollen grains together: how and why? In: von Byern, J.,

Grunwald, I. (Eds.), Biological Adhesive Systems. Springer Verlag, Wien, pp. 3–14.Hesse, M., Vogel, S., Halbritter, H., 2000. Thread-forming structures in angiosperms an-

thers: their diverse role in pollination ecology. In: Dafni, A., Hesse, M., Pacini, E.(Eds.), Pollen and Pollination. Springer-Verlag, Wien, pp. 281–292.

Meyer, N.R., Bernard, V.V., 1973. Electron microscopy investigation of pollen grains for-mation in Pinus silvestris L., Juniperus communis L., Larix sibirica Ledeb. Morphologyof Pollen and Spores of Extant Plants. Nauka, Leningrad, pp. 21–24 (In Russian).

Morbelli, M.A., Rowley, J.R., 1993. Megaspore development in Selaginella. I. “Wicks”, theirpresence, ultrastructure, and presumed function. Sex. Plant Reprod. 6, 98–107.

Pettitt, J.M., 1976. A route for the passage of substances through the developing pterido-phyte exine. Protoplasma 88, 117–131.

Pettitt, J.M., 1979a. Developmental mechanisms in heterospory: cytochemical demonstra-tion of spore-wall enzymes associated with β-lectins, polysaccharides and lipids inwater ferns. J. Cell Sci. 38, 61–82.

Pettitt, J.M., 1979b. Ultrastructure and cytochemistry of spore wall morphogenesis. In:Dyer, A.F. (Ed.), The Experimental Biology of Ferns. Academic Press, London, pp.213–252.

Reguzova, A.I., 2004. Morpho-physiological features of pollen grain formation andpollination in some species of Cupressaceae. Bot. Zhurn. (Leningrad) 87, 1111–1121(In Russian, with English summary).

Rodríguez-García, M.I., Majewska-Sawka, A., 1992. Is the special callose wall of microspo-rocytes an impermeable barrier? J. Exp. Bot. 43, 1659–1663.

Rowley, J.R., 1986. A model for plasmodesmata. Biology of Reproduction and Cell Motilityand Plants and Animals. University of Siena, Siena, Italy, pp. 175–180.

Rowley, J.R., 1987. Plasmodesmata-like processes of tapetal cells. Cellule 74, 229–241.Rowley, J.R., 1987–1988. Substructure within the endexine, an interpretation. J. Palynol.

23–24, 29–42.Rowley, J.R., 1990. The fundamental structure of the pollen exine. Plant Syst. Evol. 5,

13–29.Rowley, J.R., Morbelli, M.A., 2009. Connective structures between tapetal cells and spores

in Lycophyta and pollen grains in angiosperms — a review. Rev. Palaeobot. Palynol.156, 157–164.

Rowley, J.R., Dahl, A.O., Walles, B., Huynh, K.L., 1983. Viscin threads considered as connec-tive structures between pollen grains and tapetal cells. Fertilization and Embryogen-esis in Ovulated Plants. Veda, Bratislava (Czechoslovakia), pp. 89–92.

Rowley, J.R., Skvarla, J.J., El-Ghazaly, G., 2003. Transfer of material through themicrosporeexine — from the loculus into the cytoplasm. Can. J. Bot. 81, 1070–1082.

Skvarla, J.J., Raven, P.H., Chissoe, W.F., Sharp, M., 1978. An ultrastructural study of viscinthreads in Onagraceae pollen. Pollen Spores 20, 5–143.

http://refhub.elsevier.com/S0034-6667(14)00038-4/rf0005































































































Sporoderm and tapetum ontogeny in Juniperus communis (Cupressaceae). Connective structures between tapetum and microspores

Documents