http://www.collembola.org/publicat/integum/discuss.htm Last updated on 2005.08.05 by Frans Janssens
Checklist of the Collembola: Some notes on the Ultrastructure of the Cuticula. Discussion

Luc De Bruyn, Department of Biology, University of Antwerp, Antwerp, B-2020, Belgium
Jean-Auguste Barra, Laboratoire de Zoologie, Université Louis Pasteur, Strasbourg, 67000, France
Frans Janssens, Department of Biology, University of Antwerp, Antwerp, B-2020, Belgium

This page is under construction.

Abstract

We propose that the formation of the surface structure of the laminate cuticula of Collembola and more in particular that the formation of the regular hexagonal pattern of the surface of the epicuticula is based on an epidermal by exocytosis regulated receptor mediated deposition of lipocuticulin micellae.

Discussion

Laminate Cuticula

The cuticula is formed as an inverse stack of polymer layers: the top layer is deposited first. Each layer is deposited by a receptor mediated exocytosis of in the Golgi apparatus synthesised lipoprotein micellae as described below. The lamination is caused by subsequent iterative synthesis and secretion cycles of the epidermis cells. Orientation of the polymerisation in the layers is random but consistent in each of the layers. This gives the cuticula its relative strength and rigidity.

Lipocuticulin Micella

The deposition of the cuticular layers is a dedicated specialisation of a much more general cell physiological process: the lipid transport mechanism. Lipid transport in intracellular and extracellular aquous fluids is made possible when fat droplets are surrounded by an amphipathic coating. The subspherical particles formed in such a way are called micellae. The basic structure of such a micella is a core of free fatty acids and glycerides that is coated with a monolayer of amphipathic phospholipids having their hydrophobe tails oriented towards the fatty core and their hydrophylic heads oriented to the aquous medium. In the phospholipid monolayer membrane of such micellae, proteins may be embedded. In this case the lipid micella is called a lipid-protein complex or lipoprotrein. Transmembrane proteins of lipoproteins are called apoproteins.

Fig.lp. Schematic model of hypothetical lipocuticulin micella
Janssens, F. 2000.
A model of a lipid-cuticulin complex or lipocuticulin micella (Fig.lp), in which cuticulin is the transmembrane apoprotein of the particle, shows the core of glycerides (Gl), the cuticulin apoprotein (Apo), and the membrane of phospholipids (Pl) (the phospholipid monolayer is partly cut away to show the glyceride core). Based on observations of Barra, 1977:209,210,Fig.4B of the initially deposited epicuticula in Tomocerus minor, we can deduce that the size of the micella is about 15 to 20 nm.

Fig.lpm. Schematic section through hypothetical lipocuticulin micella
Janssens, F. 2000.
Fig.lpm represents a transversal section through a scaled reconstruction of a lipocuticulin micella that models in a stylistic way the binding sites of the apoprotein cuticulin. The transmembrane cuticulin forms two binding sites at opposite poles of the micella. Note that the cuticulin is transmicellar from top to bottom pole. The two sites serve a different function. The convex bottom binding site has two functions. Its primary function is to bind to a receptor that is integrated in the apical cell membrane by exocytosis of Golgi transport vesicles. In this function it serves as a device to deposit lipocuticulinar monolayers. Its secondary function is to bind to the upper binding site of another micella that is stacked underneath it. In this function it serves as a device to arrange in space subsequently stacked micella. The concave top binding site serves as a device to center the micella in line with the micella that is previously deposited: the concave opening fits around the convex protrusion of the upper micella.

In analogy with the above described lipoprotein model, other proteins may take the place of the apoprotein of the lipoprotrein to form lipid-protein complexes such as lipoarthropodin, liposclerotin or liporesilin that participate in the formation of the exocuticular and endocuticular layers.

Deposition of Initial Lipocuticulinar Layer

Schmidt-Rhaesa & al. (1998) characterise the monophyletic taxon Ecdysozoa as follows: moulting under influence of ecdysteroid hormones, loss of locomotory cilia, trilayered cuticula and the formation of the epicuticula from the tips of the epidermal microvilli. During moulting in Macrobiotus hufelandi (Tardigrada), a new epicuticula is formed by secretion of numerous separate patches. The formation of a new epicuticula from the apices of the epidermal microvilli in a scalid of a larval Halicryptus spinulosus (Priapulida) is also characterised by a secretion in separate patches.

Fig.mv. Lipocuticulinar deposition at microvillar apices
Barra, J.-A. 1973 (unpublished).

Fig. mv. (Barra, 1973; 10000x) illustrates this in a transmission electron micrograph (TEM) of Pseudosinella theodoridesi: the first layer of the lipocuticulinar micellae form at a distance above the apices of the relatively short microvilli. This distance corresponds with the thickness of the glycolipid glycocalyx and is about as thick as the deposited layer itself. In the TEM, three microtubercles-in-formation of the epicuticular ornamentation can be recognised. The cuticular surface oblique ultra-thin section crosses two of the three microtubercles through their convex layer of micellae. Note the dark fading tips of the microtubercles; note the regular sphere packing pattern of the lipocuticulinar micellae in the dark fading areas.

We propose that the epicuticula is deposited by a mechanism of receptor mediated exocytosis of lipocuticulin micellae. Transmembrane glycoproteins with a cuticulin receptor domain bind the lipocuticulin micellae. Two pathways are possible to deposit the lipocuticulin particles at the apical epidermal surface:
1. the in the exuvial space released micellae bind to the cuticulin receptors of the glycoproteins of the epidermal plasma membrane.
2. the micellae bind to cuticulin receptors of glycoproteins of the membrane of the Golgi apparatus and these cuticulin-glycoprotein complexes are integrated into the plasma membrane by fusion of the membranes of the Golgi transport vesicles and the apical plasma membrane.
Both mechanisms might occur during the initial phase of the epicuticular deposition. But once the first layer is completed, it isolates the exuvial space from the underlying epidermal plasma membrane. Therefore, from that moment on, only the second mechanism can deposit further layers underneath the already deposited ones.

In this way the lipocuticulin micellae are deposited on top of the epidermal plasma membrane and the precursor of the initial epicuticular layer is formed. The lipocuticulinar particles form a luminal lining lying on top of the glycocalyx of the plasma membrane at the glycocalyx-exuvial lumen interface.

Fig.bl. Formation of a lipocuticulinar bilayer
Barra, J.-A. 1973 (unpublished).
The lipocuticulinar vesicles are arranged into a planar packing. At a critical highly ordered closest spheres packing, the phospholipid layer of the vesicles break at contact sites and the vesicles fuse together to form a lipocuticulinar bilayer. Fig. bl. (Barra, 1973; 10000x) illustrates the formation of such a relatively thick bilayer at the microvillar apices in a transmission electron micrograph of Pseudosinella s...: the lipocuticulinar bilayer is easely recognised; it is interrupted at the left, where probably a ridge of a crown is in formation; at the right, where a microtubercle is in formation. The lipocuticulinar bilayer is obviously thicker than the plasma membrane, about 3 times thicker, thus about 15 nm. This corresponds quite well with the estimated size of the lipocuticulinar micella described before. Observe the high density of ribosomes in the Golgi apparatus, indicating the protein synthesis activity required to build the lipocuticulin particles.

The following three figures try to illustrate the formation of the laminar structure of the epicuticula, based on the hypothetical lipocuticulinar micella model as described before.

Fig.lp1. Deposition of first monolayer of lipocuticulin micellae
Janssens, F. 2000.
The initial monolayer of lipocuticulinar micellae is deposited according to a spheres packing model. The micellae are bound to receptors that are integrated in the apical epidermal cell membrane. The receptors are delivered by exocytosis of Golgi transport vesicles. (mechanism to be illustrated).

Fig.lp2. Fusion of first monolayer of lipocuticulin micellae into a lipocuticulin bilayer
Janssens, F. 2000.
When more and more micellae get deposited, the packing model will approach more and more the closest spheres packing model, that corresponds with the maximum laminar volume packing of the micellae. At this critical spheres packing, the phospholid membranes of the lipocuticulin micellae break at contact sites and fuse together forming a continuous lipocuticulin bilayer. The transmicellar cuticulin plays an important role in the formation of the bilayer: it makes sure that the fusion of the micellae results in a laminar structure. Since it is anchored in the micellar phospholipid membrane, it prevents the formation of larger spherical micellae.

Fig.lp3. Deposition of subsequent monolayer of lipocuticulin micellae underneath lipocuticulin bilayer
Janssens, F. 2000.
Once the lipocuticulin bilayer is formed, the cuticulin receptor sites uncouple from their receptors. A new monolayer of lipocuticulin micellae is deposited underneath the bilayer. The micellae are centered and locked into position by aligning the concave top binding site of the new micellae with the convex bottom binding sites of the previous deposited bilayer. Note that the phospholipid membranes of the newly deposited micellae do not get in contact with the phospholipid membrane of the previously deposited bilayer. They are effectively isolated by the inbetween laying cuticulinar binding sites.

This deposition model eventually results in a hexagonal horizontal packing of lipocuticulinar micellae that fuse together into a lipocuticulinar bilayer. Due to the subsequent cubic vertical packing of lipocuticulinar micellae underneath the previously deposited lipocuticulinar bilayer, the subsequently deposited micellae fuse into a new bilayer that is isolated from the previous bilayer. In this way the inherently laminate structure of the epicuticula is shown.

The microvilli are positioned in a regular spaced closest sphere packing pattern. The space inbetween the microvilli is more or less equal to the space taken by a single microvillus. A remarkable correlation between the spatial pattern of the microvilli and the epicuticular crowns can be noted (see Fig.4C in Barra, 1977:210). A set of seven microvilli can be mapped to one crown: one microvillus corresponds with the center of the crown and six surrounding microvilli characterise the border of the circular basis of the crown. Note that the ridges and microtubercal pillars of a crown are aligned with the 'empty' space inbetween the microvilli.

Sabri & al. (2000) presented experimental evidence that cells use modulation of specific components of their glycocalyx to regulate their particle binding capacity. Their studies suggested that the increased adhesive potential is related to a decrease of O-glysosylation rather than N-glycosylation of surface glycoproteins.
We hypothise that the epidermal cells modulate the specific components of their glycocalyx in a spatial pattern in such a way that the modulation results in the tesselation of the epicuticular microtubercular ornamentation.

We propose that the convex elevations of the initial epicuticular layer is regulated by the formation of a thickness modulated glycocalyx. The modulation of the glycocalyx is constituted by glycophorin-like glycoproteins. Such a transmembrane glycoprotein has a convex folded hydrophobic amino acid chain in its extracellular domain. This hydrophobic fold is covered by hydrophylic branched oligosaccharide chains, forming a negatively charged mesh of carbohydrate extensions that form a dome, effectively shielding the hydrophobic extracellular protein fold. The 'glycophorins' are topographically placed inbetween three adjacent microvilli. The extracellular domain of the 'glycophorin' projects oligosaccharide chains into the exuvial space that span the apices of the three adjacent microvilli. The cytosolic domain of the 'glycophorin' in the cell cortex is associated with actin filaments that radiate into the cell. Two types of 'glycophorin' must occur, primarely characterised by their size. 'Glycophorin a' is the largest of the two. In Pseudosinella theodoridesi, its extracellular saccharide chains extend up to 130 nm from the membrane surface and its cytosolic actin filaments span the bases of the three adjacent microvilli. 'Glycophorin b' is smaller. Its saccharide chains extend up to 80 nm from the membrane surface and its associated actin filaments does not radiate as far as the adjacent microvilli.

We propose that the convex elevations of the initial lipocuticulinar monolayer are formed by glycocalyx constituted by 'glycophorin a'. The cytosolic domain of the 'glycophorin a' in the cell cortex radiates actin filaments into the cell that span an area corresponding to the three adjacent microvilli, in such a way that the microvilli are fully circumvented and that adjacent with actin filaments extended domains actually touch eachother. In this way, the 'glycophorin a' proteins and associated convex glycocalyx elevations are topographically distributed in a hexagonal pattern (see Fig.tobemade). This pattern corresponds with the hexagonal microtubercal pattern.
We conclude that the topographical distribution of the 'glycophorin a' molecules is the origin of the hexagonal pattern of the epicuticular microtubercles.

Epicuticular Hexagonal Tesselation

Fig.1. Model of hexagonal monolayer
Janssens, F. 2000.
The epicuticular texture is the result of assembling lipocuticulin micellae onto the glycocalyx at the apical surface of the epidermis cells. The micellae are deposited into a series of strata coplanar with the glycocalyx substrate. The layers of micellae assemble initially into small two dimensional islands due to membrane fusion as the result of exocytosis of glycoprotein-lipocuticulin complexes containing transport vesicles. Note that initially, the microvilli are absent. The islands and further subsequent layers of lipocuticulin micellae tend to grow laterally in a closest sphere packing formation. As a result of lateral growth of the multilayer isolated platelets are formed, as observed by Barra, 1977:210(Fig.4A) in Lepidocyrtus curvicollis. Each platelet is the central origin of a crown. The assembling micellae construct a well-ordered thin film of lipocuticulin micellae layer by layer, outside-in. The islands grow laterally by aggregation and coagulation of micellae. A scaled geometric closest sphere packing model is used to visualise the formation of a micellar monolayer of a crown (see Fig.1). Each lipocuticulinar micella with a diameter of 15 nm is modeled as a single sphere. It was decided to model a micella with a size of 15nm and not 20 nm, to take into account the size reduction due to the partial release of glycerides from the fatty core due to the coagulation. Assuming one centrally positioned micella, further micellae are packed in subsequent layers laterally in a single plane around the initial micella. In this particular simulation, 24 subsequent closest sphere packing layers make a hexagonal monolayer of lipocuticulin micellae with a diameter of 735 nm.

Fig.2. Model of three osculating adjacent monolayers
Janssens, F. 2000.
Eventually, growing adjacent platelets collide laterally. Note that two osculating monolayers do not meet in a hexagonal closest sphere packing but in a (more space consuming) cubic packing. The micellae of adjacent edges line up (see Fig.2). This cubic packing and lining up of micellae mark the borders of adjacent crowns. Due to the larger interstitial space of the cubic packing, the borders are more porous compared to the monolayers themselves. Due to the cubic packing, the interstitial space between micellae of stacked monolayers is concatenated tangentially and can form longitudinal pores or canals. The colliding platelet edges curl upwards into the exuvial lumen, forming the ridges of the crown (Barra, 1977:210,Fig.4B). Surface tension causes a ring-formed depression of the lipocuticulin floor at the edges near to the ridgal wall, as clearly can be seen in Barra, 1977:210,Fig.4B. Surface tensional forces also keep the base of the crown circular (Barra, 1977:210,Fig.4D).

When microvilli are formed, the initial density of epidermal microvilli is low, corresponding with one microvillus per eventually formed crown. The epidermal cells regulate the density and length of the microvilli at their apical surface. In this way the epidermal cells can control the size of the epicuticular crowns. The number of microvilli increases substantially during the formation of the cuticula.

The result is a periodic highly ordered epicuticular ultrastructure. The formation of this ultrastructure is mainly ruled by the receptor mediated binding of lipocuticulin micellae at the epidermal plasma membrane, the assembling of islands and multilayers of micellae and surface tension phenomena.

Longitudinal Ridgal Pore Canals

Fig.3. Model of formation of
ridgal pore canal
Janssens, F. 2000.
Based on the hypothetical ridge forming process described above, a geometric sphere packing model is used to visualise the formation of the longitudinal central canal in cross-section. Each lipocuticulin micella is modeled as a single sphere. In this scaled model it is assumed that two four layered lipocuticulin platelets grow laterally towards eachother. The colliding platelets curl upwards into the exuvial space. Due to the curvature of the micellar layers the intermicellar space will become larger for every subsequently curved layer. The computer generated image (Fig.3) simulates an orthogonal cross-section through a ridge with a thickness corresponding with the size of one lipocuticulin micella. It is clearly demonstrated that in the centre of the uprising ridge a canal is formed. In this model, with a micella size of 15 nm, the ridge is about 105 nm high, and 90 nm (at centre) to 165 nm (at saddle-edges) wide. The form of the cross-section computed by the sphere packing model confirms with the saddle-shaped cross-section such as observed by Massoud & Barra, 1980:255(Fig.2C) in Orchesella ariegica. The dimensioning of the reconstructed cross-section is also confirmed by the observations of Ghiradella & Radigan, 1974:302 in Pogonognathellus flavescens.

Formation of the uprising ridge at the edge of osculating layers requires a driving force that pushes the micellae towards eachother. The force that drives the micellae to eachother is provided by the fusion of the membrane of the transport vesicle with the apical plasma membrane. During the fusion, the spherical vescular membrane is converted into a plane circular plasma membrane. This circular membrane piece is actually increasing the surface of the apical membrane. The membrane fusion creates a centrifugal drift of the micellae.

Note: King, Pugh, Fordy, Love & Wheeler, 1990 describe ridges with two parallel longitudinal canals in Anurida maritima (Hopkin, 1997:55(Fig.4.9.a)). This might be due to an artefact introduced during preparation. The ultra-thin sections used in transmission electron microscopy of biological material are typically 30 to 200 nm thick. In case of a non-orthogonal cross-section through a ridge, due to the thickness of the ultra-thin section, the projected image of the canal opening at the front of the section and that at the back of the section do not superimpose and might be interpreted as two parallel canals. To be verified.
On the other hand, the parallel canals that can be seen in images made by Massoud & Barra, 1980:253 of Folsomia candida can be explained by taking into account the relative distance of the cross-section to the centre of the pillar that is connected with the ridge(s). If the section is made through two to the pillar connected ridges and close enough to the centre of the pillar in which case the pillar column is also sectioned, the section will show two parallel canals, such as in Massoud & Barra, 1980:253(Fig.1D). These parallel canals are in reality the canals of two different ridges that are both connected to thesame pillar. In Massoud & Barra, 1980:253(Fig.1E), the leftmost section is made at a larger distance from the centre of a pillar, but still through the epicuticular crown, showing two parallel canals with a relatively much larger distance. This again demonstrates that the so-called parallel canals are in fact the canals of two different ridges.

Three dimensional reconstruction of the epicuticular ornamentation

The reconstruction is based on a lipoprotein micellae packing model as discussed in more detail before. The floor of a crown is based on a hexagonal closest spheres packing, while the ridges are based on a cubic closest spheres packing.

Fig.3D. Three dimensional reconstruction of the hexagonal tesselation
Janssens, F. 2000.

The reconstruction is done by packing ten thousands of metaspheres. Each metasphere represents a lipoprotein micella. A metasphere simulates a sphere having a kind of gravity field. Approaching metaspheres will attract eachother and this will produce a deformation of the metasphere's surface (try to picture this: while two metaspheres are approaching eachother their shape will change from the original sphere (when they are still far apart) to an egg-like shape (due to the gravity field), to droplet-like shapes interconnected by a jelly-like thread (when close enough to make contact), to a larger sphere (when the 2 centers coincide)). Packed metaspheres model effectively packed lipoprotein micellae in a coagulated fashion.
This metasphere model seems even useful to simulate some of the what seems to be artefacts on Helen Ghiradelli's close-up of the epicuticula: on the crown floors, one can recognise small spherical bumps randomly distributed across the floor. In the 3D reconstruction one can see a similar artefact on the floor: in the raised center part of the floor one can observe a radial linear elevation. This is caused by the imperfection to occupy all space evenly of the hexagonal closest spheres packing in the circular envelope: it will not always be possible to fill the circular enveloped area in hexagonal closest spheres packing with equally sized metaspheres. Some metaspheres will coincide to produce larger metaspheres. Compare this with coagulated lipoprotein micellae that combine into a larger micella. Due to the fact that an algoritm is used to fill the circular floor part starting always from the same angle, all the larger metasheres get lined up to produce this linear elevation. By modifying the algoritm to start at random the bumps should be reproduced more randomly as in Ghiradelli's SEM.

Transcuticular Pore Canals

The cuticle is traversed by numerous vertical lines (Wigglesworth, 1965:28):
"In surface view, these appear as minute dots, often arranged in polygonal fields separated by clear boundaries, corresponding to the limits of the epidermal cells. On focussing the microscope up and down they appear to rotate, indicating that they run a spiral course. These vertical lines were called 'pore canals' by Leydig (1855), and were regarded by him as filamentous processes of cytoplasm from the epidermal cells, around which the non-living substance of the cuticle was secreted. That interpretation is probably correct."

Fig.4. Oblique tangential section through the procuticula of Tomocerus minor
Barra, J.-A. 1973.
In TEM's of the procuticula, made during the pro-exuvial period, numerous canals containing several filaments can be observed; the canals are arranged in a hexagonal pattern (Fig.4 after Barra, 1973:Figure 16A; x 21000). The spatial pattern of the canals is synchronous with that of the epicuticular tesselation. The canals contain several filaments. The filaments originate from a mesh-like network of filaments at the apical surface of the epidermal cell cortex (Fig.4a). Locke, 1961, 1964 considers these filaments as 'wax-canal filaments' that transport wax to the epicuticular surface (Barra, 1973:58). Filshie, 1970 considers the filaments not made from wax but from a precursor of wax (Barra, 1973:63). Barra, 1973:63 calls the filaments 'filaments tubulaires osmiophiles'. Taken into account that the canals are arranged in a hexagonal pattern that coincides with the epicuticular ornamentation of which the hexagonal tesselation originates from the hexagonal pattern of the epidermal microvilli, we conclude that the canals are the remains of the microvillar epidermal cell membrane protrusions and that the filaments of the pro-exuvial phase are the actin microfilaments of the microvillar cytoskeleton, while the filaments of the post-exuvial phase are microvillar microtubuli.

Fig.4a. Oblique tangential section through the procuticula of Tomocerus minor
Barra, J.-A. 1973.

In the tangential section in Fig.4a (after Barra, 1973:Figure 23; x 15000), the terminal web at the bases of the microvilli at the apical surface of the epidermal cells is clearly demonstrated.

The microvilli of the epidermal cells continue to extend after the epicuticula has been deposited. Epicuticular pillars have been formed at the apices of the microvilli. The epicuticula is porous enough to let pass the relatively small molecules from the breakdown of the old cuticula in the exuvial space. In this way, the microvilli can resorb those products and recycle them to synthesise the lipoproteins necessary to form the new procuticula. The procuticular lipoproteins are secreted in the intermicrovillar space inbetween the bases of the microvilli. To summarise: resorption of breakdown material from the old cuticula occurs at the microvillar shaft, while secretion of the material for the deposition of the new cuticula occurs at the intermicrovillar base.

Fig.5. Transvers section through the 'pore canals' of the cuticula of Tomocerus minor, end of pro-exuvial period
Barra, J.-A. 1973.
Fig.6. Transvers section through the 'pore canals' of the cuticula of Tomocerus minor, begin of post-exuvial period
Barra, J.-A. 1973.

During the pro-exuvial period, the microvilli have a relatively large diameter; in Pseudosinella subduodecima, up to 300 nm (Barra, 1973:58). The actin microfilaments are not approaching the cell membrane and are quite loosely bundled (Fig.5 after Barra, 1973:Figure 16B; x 45000). The spontaneous depolymerisation and repolymerisation of the spiral actin microfilaments creates a retrograde flux from the apex of the microvillus towards its base. The spatial configuration of actin filaments supports the free transportation of the from the exuvial space resorbed material to the epidermal cell body.

After the ecdysis, the resorption function of the microvilli becomes obsolete. In Pseudosinella subduodecima, the diameter of the microvilli is correspondingly reduced to 100 nm (Barra, 1973:58). The filaments are now positioned more close to the cell membrane (Fig.6 after Barra, 1973:Figure 16C; x 45000). This might be functional: the filaments may influence the direction of the deposited chitin fibers? During the post-exuvial period, the main function of the filaments is maintaining the shape of the microvillus. Therefore, it is possible that the filaments are not actin microfilaments but hollow microtubules.
According to Barra, 1973:63:
"Dans les canaux cuticulaires les filaments sont enfermés dans un gaine de 200 á 230 Å, le terme de tubule conviendrait mieux; les tubules sont vidés de leur contenu dense")
Cytoskeletal microtubules have a diameter of about 25 nm. To be investigated in more detail.

The spiral aspect of the 'pore canals', as described by Wigglesworth, indicates that the loci of the microvilli are rotationally displaced while the cuticula layers are formed. This 'cork-screw' shape of the microvilli might serve as an anchoring system: it anchors the cuticula actively against the epidermis and/or vice versa. To be investigated.

Exuvial glands

Up to date, it is generally accepted that during the post-exuvial period, wax is transported via transcuticular pore canals to the even finer wax canals and finally to the epicuticular surface. Just after ecdysis, it is observed that the newly exposed epicuticula is waxed already. We postulate that just before ecdysis, the epicuticular wax deposition is probably controled by the exuvial glands (Fig.e1 after Barra, 1973:Fig.17C; x 17000). Those exuvial glands were first described by Philiptschenko, 1906 in Tomocerus vulgaris (cf. Barra, 1973:64).

Fig.e1. Lumen of exuvial gland
of Tomocerus minor
Barra, 1973.
Barra, 1973 confirmed Philiptschenko's findings in Tomocerus minor. Fig.e1 is a transvers cross-section through the apical lumen of a unicellular lateral cephalic exuvial gland of 6 á 7 micron. The 'pre-exuvial' wax cannot be transported to the procuticular surface through the pore canals since the epidermal microvilli still fill the pore canals. Observe the actin filaments of the microvilli in the cross-sections of the procuticular pores in Fig.e1 (nCu = new cuticula, Ex = exuvial space, Gl = lumen of gland).
The exuvial gland primary function is to secrete the fluid that fills the exuvial space that has been created between old and new cuticula at the moment of apolysis. Just before ecdysis, wax is deposited onto the epicuticular lipoprotein micellae and the wax layer next to the micellae forms an oriented monolayer based upon the hydrophilicity of the phospholipid micellar coating.

Note: reverse pinocytosis of the exuvial fluid can be observed at the bases of the microvilli of the lumen of the exuvial gland. Fig.e1 shows three phases of the reverse pinocytosis. See the arrows in the left lower corner, from left to right:
- vesicle approaching lumenal membrane,
- membrane of vesicle begins to join with lumenal cell membrane, communicating the vescular lumen with the glandular lumen
- vesicle membrane completely joined with cell membrane, vescular contents expulsed into glandular lumen

The relative high density of microvilli of the glandular lumen indicates that the gland has a resorption functionality. We hypothise that the exuvial glands start resorbing the exuvial fluid just prior to the moment of ecdysis and continue to do so during ecdysis.
Fig.e2. Paracrystal in vesicle of exuvial gland
of Tomocerus minor
Barra, 1973.

At the end of the pre-exuvial period, in the unicellular exocrine exuvial glands one can observe vesicles containing paracrystaline lattices (Fig.e2 after Barra, 1973:Fig.17B; x 90000). The lattice is a cubic packing of 10 nm spheres. This two-dimensional paracrystal is formed as a subsquare plate. A side of the plate is about 40-50 spheres long. Probably, the spheres are lipid micellae or low density lipoprotein micellae (having a high lipid to protein proportion), a precursor of the wax deposition compound.

Fig.e3. Free paracrystals in lumen of exuvial gland
of Tomocerus minor
Barra, 1973.
A merocrine secretion of the vesicles inbetween the bases of the microvilli of the gland's lumen releases the paracrystals into the apical lumen of the exuvial gland. The paracrystals are stored inbetween the microvilli. (Fig.e3 after Barra, 1973:Fig.17D; x 90000). Since the paracrystals are only found inside the lumen of the gland and not in the exuvial space, we can assume that the paracrystals are 'dissolved' into isolated micellae before they are communicated to the exuvial space. This process results in a colloidal dispersion (emulsion) of the lipid and/or lipoprotein micellae. The micellae re-assemble in a multilayer on top of the lipocuticulin micellae. After ecdysis, during the drying process of the epicuticula, the micellae coagulate into a multilaminar lipid layer, resulting into a wax layer deposition, such as described by Edney, 1977:?.

During ecdysis, the apical microvilli of the epidermal cells begin to retract. This results in the creation of helical transprocuticular pores. The microvillar retraction creates a vacuum in the pores. This vacuum will suck the liquid phase of the epicuticular wax into the interstitial spaces (=wax canals?) of the micellae of the outer epicuticula, anchoring the wax layer effectively to the epicuticular surface.

During the post-ecdysial period, by shortening or elongating the microvilli in the cuticular pores, the epidermal cells can infuence the structure of the epicuticular wax impregnation. Shortening the microvilli will suck the wax deeper into the procuticular matrix, making the 'waxed' cuticular layer thicker, and therefore making the cuticula less permeable to water. Elongating the microvilli will create a pressure into the pores that will expulse wax from the interstitial outer epicuticular spaces, reducing the thickness of the 'waxed' layer and making the cuticula more permeable to water. In this way, the epidermal cells can control the water permeability of the cuticula dynamically in response to fluctuating environmental circumstances.

To conclude: just before moulting, a multilayer of low density lipoprotein micellae is deposited onto the cuticulinar epicuticula by the exuvial glands via the exuvial space. After ecdysis, a wax layer is formed by the drying process of the lipoprotein micellae. Once the procuticula is exposed, wax is transported to the epicuticular surface through cuticular wax channels (cf. Edney, 1977). Tbi.

Summary of exuvial gland functionality (put in a dynamic context):
- secretion of exuvial fluid (from apolysis to ecdysis)
- secretion of epicuticular wax (prior to ecdysis)
- resorption of exuvial fluid (just prior to ecdysis)

Questions to be answered: what happens with the exuvial gland after ecdysis? What happens with the exuvial gland opening in the cuticula after ecdysis?
Dr Johan Mertens, who is doing a study on pheromone communications systems in Collembola by means of droplets appearing on the tergites of Hypogastrura as a reaction to danger (pers. comm.). Possibly this is a pheromone communication to warn other individuals. Maybe the exuvial glands play a role in this: they might produce the liquid wax dispersion containing in lipids soluble pheromones. This would add a fourth functioniality to the exuvial glands:
- secretion of wax droplets containing pheromones (after ecdysis)

Wax secretion is associated with oenocytes and fat body cells.

To be completed.

Miscellaneous

Lemburg (1998) investigated the ultrastructure of the cuticula of Priapulus caudatus and Halicryptus spinulosus (Priapulida) and found that chitinase dissolves the unsclerotized cuticular layers almost completely. The epicuticle and the sclerotized cuticula are not affected by chitinase. Pronase dissolves all exocuticular layers, but not evenly. The sclerotized regions of the exocuticle are not affected as well as the complete epicuticle and the endocuticle.

To be completed.