Periplasmic cuticular calcium oxalate crystal deposition in Dracaena sanderiana


  • Svoboda V. Pennisi,

    1. University of Georgia, Horticulture Department, Coastal Plains Experiment Station, Tifton, GA 31793;
    Search for more papers by this author
  • Dennis B. McConnell,

    Corresponding author
    1. University of Florida, Environmental Horticulture Department, 1520 Fifield Hall, Gainesville, FL 32611–0670;
      Author for correspondence: Dennis B. McConnell Tel: +352 392 7932 Fax: +352 392 3870 Email:
    Search for more papers by this author
  • Laurie B. Gower,

    1. University of Florida, Materials Science and Engineering Department, 210 Rhines Hall, Gainesville, FL 32611–6400;
    Search for more papers by this author
  • Michael E. Kane,

    1. University of Florida, Environmental Horticulture Department, 1520 Fifield Hall, Gainesville, FL 32611–0670;
    Search for more papers by this author
  • T. Lucansky

    1. University of Florida, Botany Department, 3191 McCarty Hall, Gainesville, FL 32611–6400
    Search for more papers by this author

Author for correspondence: Dennis B. McConnell Tel: +352 392 7932 Fax: +352 392 3870 Email:


  • • The chemical nature and biological basis for crystal deposition in epidermal subcuticular areas are reported here for the angiosperm Dracaena sanderiana.
  • • Position, development and identification of crystals in adult leaves of D. sanderiana was carried out using X-ray diffraction, crystal morphology and scanning and transmission electron microscopy techniques.
  • • Numerous small (< 1–6 µm) calcium oxalate monohydrate crystals were found between the primary epidermal cell wall and the cuticle. Their formation was highly specific and predictable with respect to location and relative timing of development during leaf ontogeny. The crystals were perisplasmic as, at formation, the nascent epidermal cell wall was external to the crystals. Cuticular crystallization of calcium oxalate monohydrate in D. sanderiana occurred in crystal chambers situated between the plasma membrane and the primary cell wall. Crystal deposition did not occur in developing guard cells.
  • • The spatial pattern of calcium oxalate monohydrate within the epidermal cells, orientation of the crystallographic axes and the existence of crystal chambers outside the plasma membrane suggest biologically controlled crystal deposition in D. sanderiana.


Minerals are deposited by living organisms via a variety of processes collectively termed biomineralization (Lowenstam & Weiner, 1989). The formation of biogenic minerals may be ‘biologically induced’ or ‘biologically controlled’ (Mann, 1983). In the former process, minerals are formed via interactions between the organism and its proximal environment. Any control by the organism over the biomineralization process is minimal and typified by bulk extracellular and/or intracellular precipitation without specific interactions with organic matrices. By contrast, ‘biologically controlled’ precipitation implies implementation of organic matrices under rigid genetic control. An array of cellular activities directed towards structural, spatial, and chemical control over mineral deposition are characteristics of this process (Mann, 1983). An understanding of how organisms select, localize, and concentrate elements can be gained by investigating biologically controlled biomineralization. Such studies will yield information on how minerals are nucleated, spatially segregated, their internal microstructure and bulk shape determined, and how inorganic/organic interfaces are controlled (Birchall, 1989; Heuer et al., 1992).

Biogenic minerals have been well-documented within the plant kingdom (Franceschi & Horner, 1980). The most common phytocrystals are formed from the calcium oxalate (CO) hydrates, mainly the calcium oxalate monohydrate (COM), and calcium oxalate dihydrate (COD). Typically CO crystals appear intracellularly in specialized cells called crystal idioblasts. Extracellular CO crystal deposition is a characteristic feature of numerous gymnosperm species and ontogeny of extracellular deposits in coniferous gymnosperms indicates extracellular origin (Oladele, 1982; Fink, 1991a).

Compared with gymnosperms, extracellular CO crystals are not common in angiosperms. Berg (1994) demonstrated CO crystals embedded in outer cell walls on branchlets of Casurinaceae species. These deposits appeared soon after epidermal formation, suggesting that their crystal induction process differed from that within crystal idioblasts. Borchert (1984) reported large numbers of very small, irregular crystals in epidermal cell walls of honey locust (Gleditsia triacanthos). Extracellular CO crystal formation was reported in Nymphaea, which is a classical example of an angiosperm genus exhibiting crystals between primary and secondary cell walls in specialized astrosclereid cells (Arnott & Pautard, 1970; Franceschi & Horner, 1980). The crystallization process began within crystal chambers situated between the plasma membrane and primary cell wall (Kuo-Huang, 1992). The primary site of initial crystallization of extracellular crystals in Dracaena marginata and Sempervivum wulfenii is stated to be cytoplasmic with later translocation to the cell walls (Fink, 1991b).

Dracaena sanderiana possesses a variety of biogenic CO deposits in both intracellular compartments (leaf mesophyll cells) and apoplastic spaces of the epidermal cells beneath the cuticle (Vladimirova, 1996). The chemical nature and crystallographic properties of the CO deposits found in the foliar apoplastic space of D. sanderiana and the developmental aspects of crystal deposition are examined in this paper.

Materials and methods

Cuticular crystal extraction, scanning electron microscopy, and X-ray powder diffraction analysis

The following extraction procedures were used to separate cuticular epidermal crystals from intracellular deposits of Dracaena sanderiana hort Sander ex M.T. Mast. (Dracaenaceae). One gram of fresh mature leaf material was cut into 10 × 10 mm pieces and placed for 48 h in a 10 ml maceration solution containing cellulase (1.0% w/v), hemicellulase (1.0% w/v), and pectinase (0.1% w/v) (Protoplast Isolation Enzyme Solution I, Sigma). After maceration, adaxial and abaxial epidermal peels were obtained by gently pulling the epidermis from the underlying mesophyll. Peels were rinsed 3 times in deionized (di) water and dehydrated in an ascending ethanol series.

Isolated peels were ground to a fine paste in liquid nitrogen with a mortar and pestle. Part of the paste was used for X-ray diffraction analysis. A small quantity of the paste was placed on a glass slide and examined with a Phillips-Norelco X-ray diffractometer using CuKα radiation at 40 kV, 20 mA, and 10 min-10° scan speed for 2θ from 20° and 60°. Results were compared with the American Society for Testing Materials (ASTM) X-ray standards for COM (whewellite). ASTM data were obtained from the Joint Committee on Powder Diffraction Standards (JCPDS) – International Centre for Diffraction Data, 1996. The remaining paste was prepared for scanning electron microscopy (SEM) analysis by resuspending in 100% ethanol and centrifuging at 10 000 g. Crystals from the bottom of the centrifuge tube were collected with a pipette, placed on a circular glass coverslip, mounted on aluminium stubs with carbon conductive paste, air dried, sputter-coated with Au in an IB-2 ion coater (Eiko Engineering), and examined with a HITACHI S-4000 scanning electron microscope.

Demineralization procedure

Another group of isolated epidermal peels was used to observe crystal tubercles (spaces occupied by the cuticular crystals) after crystal removal. Peels were placed in Trump’s fixative (pH = 7.2) (McDowell & Trump, 1976) for 18 h at room temperature (20°C), rinsed in PBS (phosphate buffered saline, pH = 7.2) then deionized (d.i.) water, and placed for 3 h in a demineralizing solution (EDTA, sodium potassium tartrate, HCl, and sodium tartrate; S?P™ decalcifying Solution, Baxter, Scientific Products). After demineralization, epidermal peels were rinsed twice in di water, dehydrated in an ascending ethanol series, and critical point dried in CO2. Sputter-coating and SEM were as detailed above.

Freeze fracture of leaf tissue

To facilitate observation of cuticular crystals in situ, mature D. sanderiana leaves were placed in Trump’s fixative for 18 h at room temperature, rinsed in PBS, postfixed for 1 h at 4°C in 1% OsO4 (same buffer), rinsed in PBS and d.i. water, placed in 10% sucrose for 30 min, and cryofractured in liquid nitrogen. Fractured pieces were placed back in sucrose, dehydrated in an ascending ethanol series, and critical point dried in CO2.

Light and transmission electron microscopy

Epidermal peels obtained from fresh mature leaves of D. sanderiana, fresh shoot apices, and isolated crystals were observed with a Nikon Optiphot-Pol research microscope (Nippon Kogaku K.K. Tokyo, Japan) equipped with polarizing optics. Detailed measurements were made with an ocular micrometer. Photographs were taken with an automatic Nikon UFX-II camera attachment (Nippon Kogaku K.K. Tokyo, Japan). Shoot apices and mature leaf segments were also fixed and post-fixed as described for the freeze fracture procedure. After dehydration in an ascending ethanol series, samples were placed in an ascending acetone series, and embedded in Spurr’s low viscosity resin (Spurr, 1969). Thin (70–80 nm) sections were cut transversely to the main axis of the shoot apex and the leaf on a Reichert Supernova Ultramicrotome, collected on Formvar-coated copper grids (100 mesh), stained with ethanolic uranyl acetate and basic lead citrate, and examined with a HITACHI H-7000 transmission electron microscope (TEM).


General crystal morphology and X-Ray diffraction data

SEM micrographs of freeze-fractured epidermal peels revealed numerous crystalline deposits attached to the cuticle underside (Fig. 1a) with some crystals notably larger than others (Fig. 1b). The size of the deposits ranged from < 1 µm and up to 9 µm along the long crystal axis. Usually the predominant crystals were 5–6 µm long. Crystal polarity was observed (i.e. an exposed rounded crystal face and flat, smoother side faces (Fig. 1c–f)). The crystal sides facing the cuticle almost invariably had a smooth texture (Fig. 1d–f), while those facing the leaf interior had a rough texture and rounded outlines (Fig. 1b,e).

Figure 1 a–p.

Scanning electron micrographs of freeze-fractured Dracaena sanderiana cuticle. (a) Numerous small calcium oxalate monohydrate (COM) crystals embedded in the interior cuticle surface (arrow, cuticle). Bar, 10 µm. (b) Three large crystals on the interior cuticle surface. (c,d,f) The crystal can create a protuberance on the leaf surface (cuticle (black arrow)). Note smooth cuticle-exposed crystal face (white stars) and smooth side faces (black stars). (e–n) Isolated cuticular COM crystals. (e) Crystal displaced from original location with step-like layering on the leaf interior-exposed face (arrow). (j–k) Numbers in brackets are Miller indices for crystal faces. (g) A typical monoclinic COM crystal with expressed (1–01) and (010) planes. Parallel lines (white arrows) evident on (110) planes. (h–j) Variations on the same crystal form without the (010) plane. (k,l) Rounded (1–01) face, step-like and angular layers (white arrows). The (110) planes (black arrows) are less pronounced in these particular crystals. (m–n) Parallel lines (white arrows) on (110) crystal face (compare with (g)) and inner crystal envelope shaped as a parallelogram attached to the crystal (black arrow, point of attachment of the inner envelope part). (o,p) Demineralized cuticle. Crystal envelopes consist of an inner part, conforming to the crystal shape (black arrow, identical to the parallelogram in (n)), and an outer part, made of amorphous material (white arrow). Bars, 1 µm.

X-ray powder diffraction (Table 1) and crystal morphology (Fig. 1g–j) confirmed the identity of these deposits as COM, or whewellite. Monoclinic COM was characterized by a combination of planes, the most pronounced of which was (1–01) (the numbers are Miller indices identifying intersections of the crystallographic axes with the principal crystal faces) (Fig. 1g). Occasionally, the (010) face also developed (Fig. 1g,k). All isolated crystals had the typical (1–01) form or variations of it achieved by the development of additional planes (110) and (011) (Fig. 1h–l). Successive concentric layers and angular step-like layers associated with particular crystal faces, especially ones facing the leaf interior, were exhibited by some crystals (Fig. 1k,l). Multiple parallel lines traversing (110) planes were present in some crystals (Fig. 1g,m,n).

Table 1.  Comparison of American Society for Testing Materials data of calcium oxalate monohydrate with crystals extracted from the foliar cuticle/epidermis of Dracaena sanderiana
ASTM whewellitex CaC2O4·H2ODracaena sanderiana
D, ÅyI/I0zD, ÅI/I0
  1. xASTM data were obtained from Joint Committee on Powder Diffraction Standards (JCPDS) – International Centre for Diffraction Data, 1996. D, Åy is the wavelength spacings in Ångstroms. I/I0z is relative intensity of diffraction response compared to the primary peak.*The three major peaks are indicated by an asterisk (*) in each analysis.

5.79 30  –
3.78* 703.64 16.5
3.00  23.4 14.1
3.00  23.11 12.9
3.01 10  –
2.97* 452.97 53.9
2.92 10  –
2.84 10  –
2.00  42.53  3.8
2.49 18  –
2.36 30  –
2.35 12  –
2.21  62.21  7.9
2.08 14  –
1.98 101.98  9.8
1.95 10  –

Most crystals were isolated relatively unobstructed with foreign material (Fig. 1e–k), although some were enclosed in ‘envelopes’, readily observed after demineralization of the cuticle (Fig. 1o,p). The envelopes consisted of two parts: an outer amorphous to granular material, possibly cuticular waxes (Fig. 1o); and an inner thick, smooth layer adhering to the crystal with a different degree of tenacity, possibly compacted cuticular waxes (Fig. 1n,p).

Crystal development

Cuticular crystals appeared early in leaf ontogeny (Fig. 2) and were detectable in primordia approx. 1.5–2% of mature leaf size. During early developmental stages, the leaf primordium resembles an elongated cone (Fig. 2a–d). As a monocot species, D. sanderiana displays acropetal leaf maturation, and thus the oldest portion of the leaf is toward its distal end. As the primordium grew and elongated, cellular differentiation and maturation occurred (Fig. 2c,d). No leaf cuticular crystalline deposits were observed in young primordia (Fig. 2a,b) and cuticular crystals were not detectable until the primordium reached approx. 2500–3000 µm in length (Fig. 2c). Cuticular crystals were observed first in epidermal cells of the primordium tip (Fig. 2d).

Figure 2.

Light microscope micrographs of crystal development in Dracaena sanderiana shoot apex. (a–i) Whole mounts of shoot apices viewed in between crossed polars with partially (a,b) and fully crossed polars (c–e). (a,b) Dome-shaped meristem (black arrow in (a)) and first primordium (black arrow in (b)). The bright outline is a residual birefringence, not related to crystals (white arrow). Note the absence of any crystalline deposits at this developmental stage. Bar, 100 µm (c) Second leaf primordium. Acropetal maturation of cuticular epidermal crystals and raphide bundles, which form earlier (arrow). Bar, 150 µm (d) Crystal maturation zone of the second leaf primordium depicted in (c). Bar, 50 µm (e) Third leaf primordium with extensive cuticular crystal deposition. Bar, 200 µm (f–i) Crystal maturation zone depicted in (e). (f–h) Rod-like crystals are formed. (f,g) Partially crossed polars; (h,i) completely crossed polars. Crystal colours are related to the orientation of the crystal axis with respect to crossed polars, e.g. dark-blue crystals are extinct, bright-yellow crystals show maximum birefringence. (i) In mature epidermal cells most COM crystals are 4-and 5-sided. (j) The 6-sided shape of the larger crystals (black arrow) is typical of COM. Guard cell pairs are conspicuously crystal-free (white arrow). Bars, 10 µm.

Cuticular COM crystals appeared in a ‘crystal maturation zone’ of approx. 300–500 µm (Fig. 2e). Developing epidermal cells in this maturation zone were 7–10% of their mature length (Fig. 2f,g). Initially, two to five crystals per cell were visible (Fig. 2f). As cells continued to mature, more crystals formed per cell (Fig. 2g). At this early stage of their growth, all deposits appeared rod-like. Later crystals increased in size and visually appeared six-sided (Fig. 2h) with the shape typical of COM. Most mature crystals, however, visually appeared five-sided and four-sided based on their optical orientation (Fig. 2i). Guard cells, which differentiated at the same time, did not form cuticular crystals (Fig. 2j).

Ultrastructural elements in epidermal cells and their relationship to crystal deposition

The initial visible event in the deposition of cuticular COM crystals appeared to be the development of a plicate plasma membrane (Fig. 3a,b). The folds appeared to include parts of the cytoplasm. Highly pleiomorphic vesicles (approx. 0.1 µm in diameter) were closely associated with the plasma membrane. An electron-dense, rectangular shaped crystal chamber (approx. 0.5 µm in diameter) was present (Fig. 3b). This chamber developed in the region previously occupied by the cytoplasm, its outer boundary delineated by the immature cell wall. Crystals often protruded into the epidermal cell space in early growth stages, and the cytoplasm boundary was clearly delineated by the plasma membrane (Fig. 3c–f). These electron-dense crystal chambers were not readily discernible in late stages of crystal growth. The shapes of spaces occupied by the crystals in transverse sections varied but typically were rectangular (Fig. 3e,f). During initial stages of crystal development small electron-transparent vesicles (approx. 0.1 µm in diameter) were observed in the cytoplasm and outside of the plasma membrane in vicinity of the crystal (Fig. 3e). The vesicles appeared associated with long profiles of rough endoplasmic reticulum (RER). Prior to thickening of the outer tangential epidermal cell wall, crystals were approx. 2.5 µm in length and appeared to protrude into the cytoplasmic space (Fig. 3e,f). As the cell wall grew in thickness, the crystals were displaced away from the cytoplasm (Fig. 3g–j). The outer tangential epidermal cell wall in mature cells had concentric indentations near the crystals (Fig. 3i). The amorphous envelope differed from the fibrillar nature of the cellulose matrix of the cell wall (Fig. 3k,l).

Figure 3 a–l.

Transmission electron micrographs of developing crystals in immature Dracaena sanderiana epidermal cells. (a,b) Rectangular shaped electron-opaque crystal chamber and single and double membrane vesicles (verticle arrows) in proximity to the crystal chamber. Invaginations of the plasma membrane are evident (horizontal arrows). (c) In ultrathin sections crystals are not visible since they are not infiltrated by the resin and fall out during the sectioning process, however, the sites of their original location remain visible as white ‘holes’ with characteristic angular outlines. In early growth stages, crystals develop in close association with the plasma membrane (arrow); the nascent outer tangential cell wall is the outer boundary of the developing crystal. (d) Crystals can protrude deep within the cytoplasmic space (arrow, plasma membrane). (e) Vesicles appear in close association with RER (right arrow) and the plasma membrane (left arrow). (f) As crystals continue to grow, they enlarge into the periplasmic space. Crystal chamber membrane (arrow) is visible between the growing crystal and the cytoplasm. (g,h) Thickening cell wall (arrow) surrounds the epidermal cell. (i) In mature cells, periplasmic crystals are displaced from the cytoplasm and are closer to the cuticle. A marked indentation in the outer cell wall layers around the crystal is evident (arrow). (j) The crystal protuberance is present near the leaf surface. (k,l) Inner part of the crystal envelope (arrows) around the crystals. Bars, 1 µm. C, cuticle; CC, crystal chamber; CH, crystal hole; CW, cell wall; N, nucleus; RER, rough endoplasmic reticulum; VS, vesicle.


Crystal morphology and patterns of deposition

The present study is the first attempt to isolate and characterize cuticular crystals in a plant species, and also the first conclusive report to verify the hydration state of cuticular CO crystals. We expected COM since a comprehensive review of most studies has alluded to this CO form (Franceschi & Horner, 1980). The crystal morphology was similar to COM crystals grown synthetically (Sikes & Wierzbicki, 1996) and expression of typical faces (1–01), (010), (110), and (011) has been reported in intracellular plant crystals (Cody & Horner, 1984). The multiple parallel lines traversing (110) planes in some crystals (Fig. 1g,m,n) were especially intriguing. These lines could be cleavage planes (special planes in a crystal where the bonds between atoms of different layers are weaker), or intracrystalline inclusions of unknown nature. Intercalation of organic material on specific planes different from cleavage planes often occurs in biogenic minerals (Addadi & Weiner, 1989), and is the most likely explanation for the modified characteristics of these minerals. The polarity in surface texture of some crystals with respect to their cuticular orientation is interesting (Fig. 1b–f). The successive concentric layers and the rough texture associated with the crystal faces exposed to the leaf interior are indicative of crystalline matter precipitated within a membrane-bound sheath (Lowenstam & Weiner, 1989). The presence of crystal chambers also supports this interpretation. It is conceivable that a heterogeneity in the crystal chamber membrane from which the COM crystals precipitate, could be the causal factor in the observed polarity. This has been previously reported for biogenic minerals (Addadi & Weiner, 1989).

Initially-formed periplasmic crystals display morphological characteristics illustrated in Figs 4, 5. Clearly, cuticular COM shapes (Figs 2f–h, 5b), and the crystal shape depicted in Fig. 4(a), are very similar. Fig. 4(b) shows the plane of b and c axes is parallel to the plane of the crystal/epidermal cell interface, while the a axis is inclined. Such preferential orientation of the crystallographic axes has been reported for CaCO3 tablets in mollusc species, and could provide insights into the crystal nucleating sites (Lowenstam & Weiner, 1989). In gastropod nacre, aragonite crystals (CaCO3) have no preferential alignment of their b and a axes with respect to each other, but the c axes are aligned. This alignment means that adjacent crystals are randomly rotated about their c axes with respect to each other. Similarly, cuticular COM crystals in D. sanderiana displayed random rotation about their a axes with respect to their neighbours. The four-sided crystal shape is achieved by faster growth of (010) faces with respect to the other crystal faces (Fig. 5). Crystals that reached their final size, were characterized by the well expressed (1–01) faces and smaller (010) faces.

Figure 4.

Schematic illustrations of the orientation of the three crystal axes and principal crystal faces in developing periplasmic crystal from epidermis of Dracaena sanderiana (a); and orientation of the crystal axes of two growing calcium oxalate monohydrate (COM) crystals with respect to the epidermal cells and the leaf (b). Note that in both crystals the plane of b and c axes is parallel to the plane of the crystal–epidermal cell interface, while a axis is inclined. Arrow indicates direction of cell elongation. Compare with Fig. 2(g,h). Not drawn to scale.

Figure 5.

Schematic illustration showing a likely growth sequence of the periplasmic calcium oxalate monohydrate (COM) crystals in Dracaena sanderiana. (a) Principal crystal faces (101) and (010). (b) (010) faces grow faster than (1–01) faces. Growth along other planes is not shown, in reality crystals grow along all planes. The diagram illustrates relative growth of (010) planes with respect to other crystal planes. (c1) One (010) face has grown itself out of existence, while the other (010) face is still present, albeit smaller compared with (b). (c2) Both (010) crystal faces have disappeared. (d) Changes in crystal shape as seen in surface view in the epidermal cells of D. sanderiana. Compare with Fig. 2(g,h) and Fig. 2(i). Not drawn to scale.

Conflicting reports exist regarding the presence or absence of crystals associated with the apoplastic space above the guard cells. Alvin et al. (1982) reported that in Callitris endlicheri (Cupressaceae) ‘the guard cell cuticle was pitted by embedded CO crystals’. However, Borchert (1984) did not find CO crystals in the stomatal apparatus of Gleditsia triacanthos, a species, which deposits crystals abundantly in cuticular areas. Clearly, deposition of cuticular COM crystals was not part of the developmental sequence in D. sanderiana guard cells. This strongly suggests some controlling mechanism for CO deposition. For example, even when rhizospheric Ca2+ levels are elevated, the guard cells in D. sanderiana remain crystal-free (Pennisi, 1999). Similarly, as rhizospheric Ca2+ levels increase, specialized subsidiary cells in Commelina communis exhibit CO crystal deposition while the guard cells do not display this deposition (Ruiz & Mansfield, 1994).

Ultrastructural elements in epidermal cells and their relationship to crystal deposition

Initial events in deposition of cuticular COM crystals were development of invaginated plasma membrane and numerous pleiomorphic single and double membrane-bound vesicular bodies in the region facing the cuticle (Fig. 3a,b). Rectangularly shaped crystal chambers could have been derived by vesicle fusion. In all cases, crystal chambers were external to the plasma membrane. Crystal origin is most precisely described as periplasmic because the apoplast (i.e. the nascent epidermal cell wall, albeit very thin and immature at the time of crystal nucleation) is external to the developing COM crystals.

In coniferous gymnosperms, extracellular crystals were reported to originate in situ within the outermost cell wall layers of the mesophyll cells (Fink, 1991b). A different mode of extracellular crystal deposition was observed in D. marginata and S. wulfenii (Fink, 1991a). In the former species, crystals originated from ‘deeper layers within the cell walls’, and in the latter species, crystals were initially seen as ‘free-floating in the cytoplasm’ and then secondarily attaching themselves to the internal cell walls. The present study and previous ones (Kuo-Huang, 1992; Berg, 1994) suggest that periplasmic crystallization in angiosperm species occurs in crystal chambers, which may originate from the plasma membrane. In D. sanderiana, small vesicles appeared in close proximity to the cytoplasmic RER and the growing periplasmic crystals. There is a possibility that these vesicles were involved in the crystal growth, although the mode of action is unknown at this time. These vesicles may traverse the plasma membrane on their way to the apoplastic space and the developing crystals. It is conceivable that the vesicles contribute to crystal growth by delivering packaged ions g (i.e. Ca2+ and/or oxalate). This hypothesis is consistent with our finding that when D. sanderiana plants were deprived of mineral nutrients for extended time, the periplasmic crystals that formed were much reduced in size (Pennisi, 1999). Under conditions of limiting nutrient supply, the crystal maturation zone occurred later in leaf ontogeny. The thickening cell wall of epidermal cells separated the periplasmic crystals from the cytoplasm, thus presumably discontinuing the flow of Ca2+ and/or oxalate to the developing crystals.

Association of ER with extracellular crystals has been reported in secreting trichomes of chickpea (Cicer arietinum), where a ‘tubular–vesicular membrane network opened into the hole that contained a calcium oxalate crystal’ (Lazzaro & Thomson, 1989). As crystals were not the primary focus of Lazzaro & Thomson’s study, little additional information was given. Future work should focus on clarifying the role of cytoplasmic vesicles, the ER, and the plasma membrane in the deposition of periplasmic CO crystals.

In summary, this study presents evidence for the chemical nature and biological basis for CO deposits found in epidermal subcuticular areas of an angiosperm species. The periplasmic COM crystals were deposited in crystal chambers, closely associated with the plasma membrane. Crystal growth may have occurred via RER-derived vesicles. Biomineralization in D. sanderiana is, thus, implicated to be at least partially biologically controlled, a suggestion further supported by: the spatial pattern of crystal deposition; within the cells of the epidermis with crystal-free GC, the preferential orientation of crystallographic axes, and the existence of crystal chambers.


The authors thank Drs Karen Koch and Bart Schutzman for review of the manuscript and instructive criticism.