Valve morphogenesis in the diatom genus Pleurosigma W. Smith
(Bacillariophyceae) – Nature’s alternative sandwich.
F.A.S. Sterrenburg*
Stationsweg 158
1852LN Heiloo
The Netherlands
Mary Ann Tiffany
Center for Inland Waters
San Diego State University
San Diego, CA 92-182-4614
USA
María Esther Meave del Castillo
Univ. Autónoma Metropolitana Iztapalapa
Lab. de Fitoplancton Marino
A.P. 55-535, México, D.F.
México
* Research Associate, National Natural History Museum “Naturalis”, Leiden, The Netherlands
Received October 12, 2003; received in revised form January 8, 2004.
Authors’ e-mail (√ = confirmed): √"Dr. Frithjof A.S. Sterrenburg" <fass@wxs.nl>,
"Dr. Mary Ann Tiffany":mgarrow@sunstroke.sdsu.edu,
mtiffany@sunstroke.sdsu.edu, mtiffany24@yahoo.com;, √"Dr. Maria Esther
Meave del Castillo" <mem@xanum.uam.mx>
1
Abstract:
The loculate (“chambered”) valve structure of centric diatoms like Triceratium favus Ehrenberg
has been mentioned time and again in the nanostructural literature. Here we draw attention to the
fundamentally different alternative sandwich model Nature developed in the genus Pleurosigma,
where it is non-loculate. This has so far been overlooked in nanostructural studies. We suggest
some mechanical aspects that would offer interesting avenues for experimental testing. The first
description of the natural fabrication process (“morphogenesis”) is presented. This begins with
the development of the raphe sternum, which then acts as a rigid backbone. The inner layer of the
sandwich-structured valve develops next, with relatively large ±round single internal foramina
not yet closed by a sieve-membrane, in offset arrangement. This serves as a substrate for rows of
stubby hollow pillars, also in offset arrangement. Then the outer layer of the sandwich develops
and two different patterns ("coarse-mesh" and "stellate bridges") have been observed. At first, the
external areolar foramina are relatively large and ±oval, gradually filling up until the tiny slits
characteristic of the genus remain. In Pleurosigma species with double internal areolar foramina,
small bridges grow from the opposite margins of the single foramina until they fuse. The sievemembranes then close the internal areolar foramina. The finished valve is a lightweight structure
expected to offer excellent strength with parsimonious expenditure of the raw material – silica.
2
Introduction
Bacillariophyceae (diatoms) are characterized by a siliceous exoskeleton (frustule) consisting of
two valves that fit together like a box and its lid. In the vast majority of species, this valve
consists of a single layer of silica, but in some genera – like Triceratium – there is a sandwich
structure which is called “loculate”, meaning that it consists of individual chambers like a
honeycomb. This loculate sandwich structure has been discussed and illustrated in the
nanostructural literature almost ad nauseam, but here we draw attention to a fundamentally
different (non-loculate) approach to a sandwich structure evolved by Nature, which has so far
been completely overlooked by nanotechnologists. We also publish the first description of its
natural fabrication process or morphogenesis and suggest some mechanical aspects that would
warrant further experimental exploration.
This paper is an offspin of two independent investigations: by Meave and Sterrenburg on
planktonic Pleurosigma species in the Mexican Pacific littoral, and by Tiffany on epiphytic
diatoms collected on the Pacific coast of California, USA. Puzzling morphological phenomena
seen in these studies were found to represent various stages in valve development and this
conclusion was briefly mentioned in Sterrenburg, Meave and Tiffany (20031).
Pleurosigma valve structure
The diatom genus Pleurosigma W. Smith possesses a sandwich-type valve ultrastructure
consisting of an inner and an outer layer separated by internal shoring. The inner layer (Fig. 1) is
perforated by round or oval and sometimes bisected (as in the figure) areolae equipped with a
sieve-membrane, the outer layer (Fig. 2) by extremely narrow longitudinal slits. The genus is
characterised by areolae arranged in one transverse and two obliquely intersecting rows; in Fig. 3
these three systems have been marked by lines. The genus Pleurosigma is also interesting from a
nanostructural point of view because its morphology is apparently under rigid mathematical
control: the intersection angle of the oblique stria systems (i.e. the ratio of pore spacing in the
3
transverse vs the oblique rows) varies by only a few degrees in a single species, whilst differences
of >50 degrees may exist between different species.
Diatom valve formation after mitosis
When diatoms multiply by mitosis (asexual reproductive cycle), the two extant valves are pushed
apart and each of the two daughter-cells forms an extra valve, inside the extant valves, to
complete the two new frustules. A comprehensive survey of the general aspects of this process,
which is unique to diatoms, was given in Pickett-Heaps, Schmid and Edgar (19902). In
Sterrenburg (19943), the sequence of events was illlustrated in the light-microscope (LM) for the
diatom Terpsinoë musica Ehrenberg, which thanks to its morphology clearly displays the various
stages of valve formation in the living cell. Findings from that investigation may serve as an
introduction to the mitotic multiplication process in diatoms here.
Figure 4 shows how after mitosis, the daughter nuclei are situated close to the newly formed cell
membranes, in the center of the parent cell. The gap representing the former cleavage furrow is
then temporarily widened by plasmolysis, creating an empty space between the retracted
daughter-protoplasts. Subsequently (Fig. 5) the two daughter-cell protoplasts again inflate and
touch and silica deposition for the two new valves begins in between the two daughter-nuclei,
which are now very close together. Then (Figs. 6 and 7) silica deposition spreads to the periphery.
This sequence can be called the “choreography” of valve formation and – as far as is known – is
the standard model in diatoms. The “mechanics” – the actual formative process of silica
deposition – is another matter and requires investigation in the electron-microscope (EM). No
observations on the sequence of valve formation during vegetative multiplication of Pleurosigma
species have been traced in the literature.
4
Materials and Methods
The samples with developing Pleurosigma valves we discuss here were collected from the Pacific
coasts of the USA and of Mexico.
The Californian sample, containing diatom species associated with Codium fragile and its
abundant epiphyte Ceramium sp., was collected at low tide from a tide pool at Bird Rock, La
Jolla on 22-3-2002. The macroalgae were treated with HNO3 and H2SO4. The developing
Pleurosigma valves in this sample could be identified as P. acus Mann by a direct comparison
with photomicrographs of the type of that species, kindly made available by Stuart R. Stidolph
and discussed in Stidolph (20034).
The Mexican samples are best represented by the material "San Blas", Nayarit, W. coast of
Mexico, 16-7-1999, plankton sample FpM 509. This is the type material of Pleurosigma
gracilitatis Sterrenburg, Meave and Tiffany, a slide of which has been deposited in the Hustedt
Arbeitsplatz, Bremerhaven, numbered BRM ZU 5/50. The material is also present as UAMID1483 in MEXU and the Sterrenburg collection as #661. This sample was oxidized with H2O2
and potassium dichromate. All slides were mounted in naphrax for LM examination. For SEM,
drops of the sample were dried on 12mm covers, mounted on stubs and sputtered with Au-Pd.
Observations
Fig. 8 presents an early stage of valve morphogenesis. The raphe sternum (marked “rs” ) is
already fully formed. As in other pennate species, it is the first structure to develop, constituting a
rigid "backbone" for the developing valve. The inner layer of the sandwich Pleurosigma valve is
the first to form and at this stage it possesses more or less round, large areolar foramina, marked
“af” in Fig. 9. It also serves as a substrate on which stubby hollow pillars are deposited, marked
“p” in Fig. 9.
To form the outer layer, the tops of these pillars are then joined by silica bridges in a pattern
differing for the two different species examined, as a coarse-mesh network (Fig. 10) in P. acus,
in a stellate arrangement (Fig. 11) for P. gracilitatis . Silica deposition then continues to fill in the
5
gaps, leaving a central opening that will constitute the external areolar foramen. At first, this is
fairly wide and more or less elliptic (Fig. 12), then silica continues to be deposited inward until
only the tiny external slits characteristic of Pleurosigma remain, as in Fig.2.
The internal areolae are divided by a narrow silica bridge in some species - including those
studied - but in these species also, they begin as single foramina, being divided only later on
(Fig. 13, developing dividing bars marked) and finally being closed by a sieve-membrane, Fig.
14.
It may be noted (Fig. 10) that at the edge of the valve, which in the final state consists of only a
single layer, silica deposition takes place in a very different pattern – apparently as transverse
bars, which finally fuse to form a smooth surface as in Fig. 2.
Discussion of some structural aspects
The end result is a sandwich structure in which the internal and external layers are shored by
more or less round or slightly oval pillars. We tentatively conclude that these pillars continue to
be hollow as they still are so at the last stage when their inside is observable (Fig. 10), but are not
yet able to definitively prove this for all cases, as their internal aspect is eventually occluded by
the developing external layer, as has occurred in Fig. 12. Some mature fractured valves we have
seen offered clear evidence that the pillars are not solid. For instance, Pleurosigma sumatricum
Peragallo, whose general aspect is shown in Fig. 18, has multiple cavities inside the pillars, Fig.
19. The inside of the finished valve may be compared to an underground car park: a wide-open
space, with pillars in a regular arrangement supporting the ceiling, Fig. 14.
It becomes clear that the traditional view of the Pleurosigma valve structure{e.g. Hustedt (192719665), Krammer & Lange-Bertalot (19866}, according to which it had a compartmentalized
(“loculate”) structure is incorrect. The small fragment of a mature valve in Fig. 17 offers clear
evidence that thanks to the columnar shoring, Pleurosigma spp. have an unobstructed valve
interior rather than a loculate honeycomb structure. There are no separate compartments, all
internal parts of the valve are in direct communication with each other. Round et al. (19907) do
6
state that the areolae are connected laterally and their use of the term “loculate” for the valve
structure is thus inappropriate if one considers the Latin root – “loculus” – for “cabinet with
compartments”.
The situation in Pleurosigma is, therefore, fundamentally different from the valve structure in
species from centric genera with a multi-layer valve like Thalassiosira eccentrica (Schmid et al.
19818), Coscinodiscus wailesii (Schmid and Volcani 19839) and Coscinodiscus granii (Tiffany
200310) where such dividing walls are formed on the basal layer during morphogenesis and a
honeycomb architecture results.
From an engineering point of view, the Pleurosigma structure appears to be a nice example of
parsimony. Experimental testing of the mechanical properties of diatom valves is still in an
embryonic stage and no comparative studies have been carried out as yet on the resistance to
torsion, shear etc. of the many types of diatom valve structures seen in Nature. From the outset,
the honeycomb loculate sandwich structure has caught the limelight in many nanostructural
publications but the non-loculate sandwich of Pleurosigma has so far been completely overlooked
by nanotechnologists. We here offer some tentative suggestions that in our view would warrant
further experimental investigation.
Two layers joined by hollow pillars resulting in a sandwich structure would appear to offer good
structural integrity (torsion, shearing) with a minimum of materials expenditure. The diatom
valve is to a certain extent able to cope with stress by flexing. The idea that silica could flex has
not always been taken into account by diatomists and has even been rejected by diatomist
reviewers when it was described to occur (S.R. Stidolph, pers. comm.). Our own
micromanipulation of Pleurosigma and Nitszchia valves – to mention two examples – for
mounting them on stubs, for instance, shows that the valves can flex by a surprising amount
without breaking. In view of the behaviour of glass fibres – now familiar from fibre-optics – this
should have long been evident.
The shortcoming of the Pleurosigma pillar-stiffened sandwich structure from an engineering point
of view might be that it is less resistant to fracture from bending than a sandwich with internal
honeycomb stiffening would be. This is clearly demonstrated by the typical "postage stamp"
fractures in overstressed specimens (Fig. 15). Load-bearing Pleurosigma-type structures would
7
require reinforcement, e.g. by spars and the raphe sternum can be considered as such a spar.
Another solution would be to add "speed bumps", thickenings joining adjacent pillars.When we
thought of this option we regarded ourselves as pretty clever engineers but in fact, Nature had
thought of this first, see the arrow in Fig. 16.
In addition, the genome of Pleurosigma species appears to contain “rules” that strictly govern the
angles under which silicate structures like the columns and areolae are being deposited. In the
genus, the intersection angle of the oblique stria systems is known to range from circa 35° to 90°,
within tight limits for an individual species. Could this be the result of a form of fractal coding?
Finally, a non-loculate sandwich structure is also observed in the genus Gyrosigma, with the
difference that the areolae are arranged in 2 perpendicular (instead of 3 obliquely intersecting)
systems. We have been able to examine the valve morphogenesis in Gyrosigma species also and
found it to follow exactly the same sequence. Detailed description will be published elsewhere,
here we may just point out that such complete agreement in fabrication of an otherwise unique
structure indicates a close phylogenetic relationship.
ACKNOWLEDGEMENTS
We would like to thank Steve Barlow for the generous use of the Electron Microscope Facility at
San Diego State University and Constance Gramlich for the collection of Codium fragile.
References:
1. F.A.S. Sterrenburg , María Esther Meave del Castillo, Mary Ann Tiffany. Studies on the genera
Gyrosigma and Pleurosigma (Bacillariophyceae). Pleurosigma species in the plankton from the
Pacific coast of Mexico. Cryptogamie-Algologie ((2003, in print).
2. J.
Pickett-Heaps, J., Anna-Maria Schmid and Leslie Edgar. The cell biology of diatom valve
formation. Progress in Phycological Research, 7, p. 1 – 168. Biopress, Bristol. (1990).
8
3. F. A. S. Sterrenburg. Terpsinoe musica Ehrenberg (Bacillariophyceae, Centrales) with
emphasis on protoplast and cell division. Netherlands Journal of aquatic ecology 28(1), 63-69
(1994).
4. S. R. Stidolph. Observations and remarks on the morphology and taxonomy of the diatom
genera Gyrosigma Hassall and Pleurosigma W. Smith. V. Pleurosigma types of A. Mann (1925):
a critical re-investigation. Micropaleontology 48, 3, 273-284. (2003).
5. F. Hustedt. Die Kieselalgen Deutschlands, Oesterreichs und der Schweiz. Rabenhorsts
Kryptogamenflora VII. Leipzig (1927-1966).
6. K. Krammer and H. Lange-Bertalot Bacillariophyceae, 1 Teil: Naviculaceae. In
Süsswasserflora von Mitteleuropa. p. 876, pl. 206, Stuttgart/New York (1986)
7. F. E. Round, R. M. Crawford and D. G. Mann. The Diatoms. Biology & Morphology of the
Genera. Cambridge Univ. Press. 1-747. (1990)
8. A.-M. M. Schmid, M. A. Borowitzka and B. E. Volcani, Morphogenesis and biochemistry of
diatom cell walls, in Cytomorphogenesis in Plants, edited O. Kiermayer, Cell Biol. Mon.,
Springer-Verlag, Vienna and New York 8, 63-97 (1981)
9. A-M. M. Schmid and B. E. Volcani, Wall morphogenesis in Coscinodiscus wailesii Gran and
Angst. 1. Valve morphology and development of its architecture. Journal of Phycology 19: 387402. (1983)
10. M. A. Tiffany, Diatom auxospore scales and early stages in diatom frustule morphogenesis:
their potential for use in nanotechnologyJournal of Nanoscience and Nanotechnology (2004)- this
volume
Legends for figures:
Figs 1, 2: Pleurosigma acus, SEM. The internal valve surface (Fig. 1) shows oval perforations
divided by a tiny silica bar and closed by a sieve-membrane, the external valve surface (Fig. 2)
shows very narrow slits. From this it is evident that the inner and outer surfaces represent two
entirely differently structured layers.
Fig. 3: Pleurosigma gracilitatis, LM, bar = 10 µm, valve view. The areolae (“dots”) are arranged
in 3 systems of intersecting lines.
9
Figs. 4-7. Terpsinoe musica, LM, girdle views illustrating the formation of new valves inside the
old ones when diatoms multiply by simple division. In Fig. 4 two daughter nuclei (small black
spheres in the center) have been formed, the two daughter-protoplasts are widely separated. In
Fig. 5 the daughter-protoplasts again touch and silica begins to be deposited (arrow) in between
the two daughter-nuclei. In Fig. 6 and 7 silica deposition (arrows) spreads outwards. Bar = 10 µm
Fig. 8- 19. SEM, stages in morphogenesis, all of Pleurosigma acus except Fig. 11, which shows
P. gracilitatis and Figs 18 and 19, which show P. sumatricum. All bars = 1 µm, except Fig. 18,
where it is 5 µm. All illustrations represent external valve views.
Fig. 8. Early stage: the raphe sternum (rs) is already fully developed and forms a rigid backbone.
The inner layer has formed and carries stubby pillars.
Fig. 9. Detail of early stage: inner layer shows oval areolar foramina (af), the pillars (p) arising in
between them are hollow
Fig. 10. Slightly later stage: the outer layer begins to develop in the form of a coarse mesh joining
the pillars, which are still hollow.
Fig. 11. In a different Pleurosigma species, the outer layer is not formed as a coarse mesh but as
fine bridges in a sort of stellate arrangement.
Fig. 12. Outer layer almost finished, the mesh now leaves only narrow slits free, which will
become even narrower as in Fig. 2.
Fig. 13 Inner layer, late stage. The originally (Fig. 2) oval foramina now begin to be bridged
(arrow) by narrow silica bars.
Fig. 14. Close-up of the finished structure in a broken valve. Internal openings now closed by
sieve-membranes.
Fig. 15. A typical “postage stamp” fracture in a developing valve.
Fig. 16. “Speed bumps” (arrow) between the pillars (p) ensure greater resistance to fracture.
Fig. 17. A fractured valve clearly shows that the interior of the sandwich has no walls dividing it
into compartments.
Fig. 18. Pleurosigma sumatricum, general aspect.
Fig. 19. Pleurosigma sumatricum, fragment of broken valve showing unobstructed interior space
and cavities in pillars.
10
11
12
13
14