close

Вход

Забыли?

вход по аккаунту

?

Biomineralization of Unicellular Organisms An Unusual Membrane Biochemistry for the Production of Inorganic Nano- and Microstructures.

код для вставкиСкачать
Reviews
E. B‰uerlein
Biomineralization
Biomineralization of Unicellular Organisms: An
Unusual Membrane Biochemistry for the Production of
Inorganic Nano- and Microstructures
Edmund B‰uerlein*
Keywords:
biomineralization ¥ magnetosomes ¥
unicellular organisms ¥ vesicles
Dedicated to Richard B. Frankel
Angewandte
Chemie
614
¹ 2003 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
1433-7851/03/4206-0614 $ 20.00+.50/0
Angew. Chem. Int. Ed. 2003, 42, No. 6
Angewandte
Chemie
Biomineralization of Unicellular Organisms
With evolution, Nature has ingeniously succeeded in giving rise to an
impressive variety of inorganic structures. Every organism that synthesizes biogenic minerals does so in a form that is unique to that
species. This biomineralization is apparently biologically controlled. It
is thus expected that both the synthesis and the form of every specific
biogenic mineral is genetically determined and controlled. An investigation of the mechanism of biomineralization has only become
possible with the development of modern methods in molecular biology. Unicellular organisms such as magnetic bacteria, calcareous algae, and diatoms, all of which are amongst the simplest forms of life,
are particularly suited to be investigated by these methods. Crystals and
composites of proteins and amorphous inorganic polymers are formed
as complex structures within these organisms; these structures are not
known in conventional inorganic chemistry.
1. Introduction
There is much to indicate that bacteria or archaea
(formerly called archaebacteria), which can be assigned to
the prokaryotes, appeared in the late Archean and Proterozoic eons as the first life forms on Earth.[1] Microfossils[2]
suggest that they existed prior to about 3.6 Ga (Ga ¼ giga
years, 109 years). Even today both can survive or live under
such extreme conditions that rarely occur in the inhabited
world. Such extremophiles[3] thrive very well at high or low
temperatures (thermo- or psychrophiles), under highly acidic
or alkaline conditions (acido- or alkalophiles), high salt
concentrations (halophiles), and high levels of g radiation
(radiation-resistant microorganisms).
Even though fossils of microbial life forms of carbon-like
material exist, the oldest fossils are found in petrified forms
and correspond to the morphologies of bacteria, magnetite
crystals can be regarded as the most reliable direct evidence
of early life (Figure 1 A). This applies only if they display the
five most important characteristics of biologically controlled
magnetite,[4] as found in today©s magnetic bacteria.[5] These
properties are:
1) crystal dimensions of a single magnetic domain and
restricted anisotropic width/length ratios
2) chemical purity
3) crystallographic perfection
4) unusual crystal morphologies
5) crystallographic direction of extension of magnetic crystals (described in detail in reference [4b, c]).
The ability to form linear chains is an important additional
property of biogenic magnetite crystals, which in living
magnetobacteria are surrounded by a special phospholipid
membrane and are thus called magnetosomes.[6] Biogenic
magnetite crystals were also found in ™magnetofossils∫ in
terrestrial sediments up to two billion years old[7a±c] and
possibly in meteorites from Mars (Figure 1 B).[7d] After their
isolation from a culture of Magnetospirillum gryphiswaldense,
magnetosomes were arranged in parallel for the electron
microscope with a bar magnet to obtain information on their
purity and magnetosomal membrane.[7ef] As long chains were
Angew. Chem. Int. Ed. 2003, 42, 614 ± 641
From the Contents
1. Introduction
615
2. Biomineralization on Bacterial
Surfaces
616
3. Formation of Nanocrystal
Forms from Fe3O4 in Magnetic
Bacteria
618
4. The Formation of Complex
Crystal Morphologies of CaCO3
in Unicellular Algae
624
5. The Formation of
Nanostructured Cell Walls of
Amorphous Polysilicic Acid in
Diatoms
629
6. Summary and Outlook
636
found in the presence of the bar magnet instead of the original
short chains (Figure 1 C), this experiment could be a model
for the formation of magnetite crystal chains on Mars,
provided short chains were released after the death of the
bacterial cells.
During the period between two billion years ago to the
Precambrian Era about 550 million years ago, the terrestrial
™magnetofossils∫ were the only witnesses to biomineralization, until evidence was found at the end of this eon for a
™matrix-mediated∫ formation of a calcium mineral by an
invertebrate, Cloudina.[8] Several hundred million years
previously the main animal phyla appear to have already
divided.[9, 10] In the early Cambrian Period (525±510 million
years ago) a tremendous increase in biomineralization
occurred, called the Cambrian Explosion. In the first 10 million years, a plethora of biomineralizing animal species was
formed almost exponentially.[4a 11±13] Most of the more important skeletal materials were then produced. Thereafter only
corals, a few algae, and the vertebrates in marine habitats
formed new skeletons.[13]
This knowledge prompted Kirschvink and Hagedorn to
formulate a bold and comprehensive hypothesis:[14] The
complex mineral structures of animals and plants as they
occur today could have evolved from an already existing,
simpler system. Thus gene patterns that developed for a
special function would have been changed through gene
duplication, mutation, and adaptation in another biological
system for a new role, a new mineral structure.[15] The
palaeontological findings suggested that the formation of
[*] Prof. em. Dr. E. B‰uerlein
Abteilung Membranbiochemie, Max-Planck-Institut f¸r Biochemie
Am Klopferspitz 18 A, 82152 Martinsried (Germany)
Fax: (þ 49) 89-8578-3777
E-mail: e_baeuerlein@yahoo.de
¹ 2003 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
1433-7851/03/4206-0615 $ 20.00+.50/0
615
Reviews
E. B‰uerlein
not only the ability to form magnetite crystals is clearly
present in many higher organisms. The system based on
intracytoplasmic vesicles through which biominerals are
formed could also have been transferred from the magnetotactic bacteria to the unicellular eukaryotic organisms. If this
hypothesis were to be verified, the explanation of the
mechanisms that lead to the formation of complex structures,
including bones and teeth, would be simpler.
2. Biomineralization on Bacterial Surfaces
2.1. Immobilization of Ions
Figure 1. Fossil chains of magnetite crystals of magnetotactic bacteria
(magnetofossils): A) terrestrial magnetofossils from marine sediments.
They originate from cores from the South Atlantic (Deep Sea Drilling
Project (Leg 73), Angola Basin); their age is about 50 million years,
length of the bar represents 100 nm. (Unpublished electron microscope image (TEM), with kind permission of N. Petersen and M. Hanzlik, LMU Munich). B) Magnetofossils in the Mars meteorite ALH
84 001. Electron microscope image (SEM-BSE) of a long chain of magnetite crystals (arrow) which were taken through the intact surface of a
freshly fractured specimen. The magnetite crystals are located on the
rim of carbonate globules in ALH 84001. The length of the bar represents 200 nm. (Section from Figure 4 A of reference [7d] with kind permission of I. Friedman and the National Academy of Science, USA).
c) Long magnetosome chains. Electron microscope image of isolated
magnetosomes from Magnetospirillum gryphiswaldense which during arrangement for electron microscopy had formed long chains in the
weak magnetic field. (From reference [7d] with kind permission of D.
Sch¸ler.) SEM ¼ scanning electron microscopy, BSE ¼ back-scattered
electron mode.
magnetite crystals of magnetic bacteria could be the preliminary stage of biogenic mineral formation in eukaryotes. This
assumption finds support in that magnetite crystals have also
been found in unicellular eukaryotic organisms such as
Euglena algae (Euglenophyta)[16] and dinoflagellates[4b] as
well as in higher species such as salmon,[17] trout,[18] carrier
pidgeons,[19] army ants,[19] and also in the human brain.[20] But
Edmund B‰uerlein was born in 1932 in
Hˆchst (Germany). He studied chemistry in
Saarbr¸cken, Munich, and Frankfurt (Germany) and he completed his PhD with Prof.
Th. Wieland on biologically relevant hydroquinones. He then moved to the MaxPlanck-Institut f¸r medizinische Forschung,
Heidelberg (Germany), as a research group
leader. He completed his Habilitation at the
Fakult‰t f¸r Chemie, Universit‰t Heidelberg
in 1974, where he was appointed Professor
in 1980. In 1984 he moved to the MaxPlanck-Institut f¸r Biochemie in Munich, Abteilung Membranbiochemie, where he was research group leader until the
end of his active service in 1997.
616
Bacterial cells have a volume of 1.5±2.5 mm3,[21] and
consequently a high surface/volume ratio. To give an idea of
what this means, a cube with an edge of 1 cm can be divided
into 1012 cubes with an edge of 1 mm. Thus the whole surface
of these 1012 small cubes is 10 000 times larger than that of the
large cube.[22] The large surface area of a bacterial cell is
suitable primarily for the uptake of nutrients and the
elimination of waste materials, as both processes are totally
dependent on the diffusion gradients of the respective
substances.
Moreover, the different structures of the cell walls of the
two groups of bacteria, which were originally named Gram
positive or Gram-negative depending on their color behavior
with Gram coloration, provide a large number of possibilities
to bind cations and also anions. Thus the cell wall of Grampositive bacteria is made up of a highly crosslinked polymer, a
25-nm-thick peptidoglycan layer, also known as murein
sacculus, which is rich in carboxylate groups. Secondary
polymers, such as teichoic acid, chains of 8±50 glycerol or
ribitol molecules that are esterified with each other through
phosphate bridges, or teichuronic acid, which is essentially
based on uronic acid, extend the arsenal of negatively charged
compounds (phosphate and further carboxylate groups). In
contrast, the cell wall of Gram-negative bacteria are made up
of a thin peptidoglycan layer of 4 nm which is protected from
the environment by the lipid/protein bilayer of its outer
membrane. Therefore the externally extending lipopolysaccharides (LPS), which are anchored to the outer membrane
with the lipophilic ends, are probably the preferred centers of
cation binding as a result of their many phosphate and
frequently also carboxylate groups.
Additional layers of the outer surface of both the Grampositive and the Gram-negative bacteria are formed from acid
mucopolysaccharides: as capsules if they are bound to the cell
walls,[23] or as a slime if they can move freely between the
bacteria.[23] Sheaths (hollow cylinders of filament-forming
bacteria) are made up of a heteropolysaccharide. This
heteropolysaccharide contains gluconic acid cations through
which the sheaths bind. Such an immobilization can also occur
actively through enzymes that are located in the sheaths: for
example, they oxidize MnII to MnIV, which precipitates as
MnO2.[23]
Many bacteria and archaea form paracrystalline surface
layers with a two-dimensional protein arrangement, the
S layer (surface layer).[24] The polar amino acid functions
Angew. Chem. Int. Ed. 2003, 42, 614 ± 641
Angewandte
Chemie
Biomineralization of Unicellular Organisms
appear to be buried mainly within the S-layer proteins.
However, such functions in the channels that penetrate
through the S-layer network can explain the observed
biomineralization.[23] Remarkably, the S-layer proteins are
the first glycoproteins to be found in prokaryotes, initially in
the archaeon Halobacterium salinariun (previously H. halobium).[25]
From this overview of all important bacterial surfaces, a
negative surface charge ensues from the COOH and PO(OH)
groups at the neutral pH of the environment. Thus in the
simplest case the positively charged metal ions are bound by
electrostatic interactions. However, because of the complex
structure of the surfaces there is little concrete evidence of
actual metal complexes. A three-pK model for the cell walls
of bacilli obtained from acid±base titration suggests not only
the participation of carboxy and phosphoryl groups, but that
participation of hydroxy groups may be involved in complex
formation.[23, 26] Stability constants of the metal carboxylates
have been determined for CdII, CuII, PbII, and AlIII.[26]
Although the surfaces of many bacteria are negatively
charged, anions such as carbonate and silicate can be bound at
neutral to acidic pH values. Diaminopimelic acid, which is
present in the peptidoglycan of Gram-positive bacteria such
as Bacillus subtilis, and N-acetylfucosamine, which is present
in the lipopolysaccharide (LPS) of Gram-negative bacteria
such as Pseudomonas aeroginosa, provide a limited number of
positively charged functional groups after protonation.
Since the binding of silicate is high relative to the number
of positive groups, it is assumed that multivalent metal ions
such as FeIII exercise a bridging function in which they first
bind to negatively charge sites and then immobilize the
silicate ion. If no multivalent metal ions are present in
solution it is highly probable that the positive groups bind the
silicate ions.[27]
2.2. Nucleation and Crystal Growth of Biominerals
Living bacteria that are partially or totally covered in
minerals are found in many, principally aqueous, environments. Microfossils[2] consist mainly of a special mineral layer
around a volume that resembles that of bacteria. Extensive
investigations into their formation on living bacteria have
been carried out to corroborate the biogenic origin of these
fossils.[2] Bacteria are able not only to bind ions, but are also
able to produce minerals on their highly reactive surface. To
produce the necessary local supersaturation, they can decrease the amount of free energy required for precipitation.
This is possible when they change the pH value or redox
potential on their surface by means of their metabolism. The
ion-binding sites are assumed to be the centers of this
growth.[28] In this case, poorly ordered to crystalline phases of
iron or manganese oxides can be precipitated several orders
of magnitude more rapidly by enzymatic redox reactions than
by chemical redox reactions from a solution without bacteria.[23] Metal ions such as FeIII, alone or together with AlIII,
form not only bridges between the silicates and carboxy or
phosphate groups of the surface, but are also critical for the
formation of amorphous to poorly ordered, fine-grained
Angew. Chem. Int. Ed. 2003, 42, 614 ± 641
polysilicates.[23, 27] An unusual example of nucleation and
crystal growth, in this case of carbonate minerals, has been
found on the S layer of photosynthetic bacteria.[29] In this case,
the S layer consists of copies of a 104-kDa protein in an
hexagonal configuration. S layers are to date the best-known
organic matrices, although not in but on the surface of an
organism. In weak light, gypsum (CaSO4) is formed on these
layers; however, during photosynthesis, calcite (CaCO3) is
formed through an increase in the pH value in the immediate
neighborhood.
2.3. Biomolecules as Matrices
The stability of the extended two-dimensional lattice of
S layers allows the isolation and use of these biomolecules as
matrices for the preparation of molecular-templated nanostructures.[24e, 30, 31] Thus metallic platinum could be deposited
on an isolated S layer by reduction of K2PtCl4 in such a way
that it forms highly ordered and periodically arranged metal
clusters. Well-separated and almost spherical particles (dark)
are arranged along the crystalline protein matrix, and the
tetragonal structure of the S layer is reproduced (Figure 2).[31a]
Figure 2. Platinum nanocluster, highly ordered on a crystalline bacterial
S layer (surface layer). A) TEM image of platinum clusters produced
on isolated S layers of Sporosarcina urea by reduction of a platinum
salt. B) Distribution of platinum clusters (black) with an average diameter of 1.9 0.6 nm on the two-dimensional protein matrix, visualized
by image processing. The unit cell of the cluster lattice has a size of
13.2 î 13.2 nm2. (From reference [31a] with kind permission of W.
Pompe and H. Engelhardt, and Springer Verlag, Heidelberg.)
2.4. Metal or Mineral Formations inside Bacteria–A Transition
from the Outside to the Inside?
All previously described minerals are formed on the
surface of bacteria (Section 2.1).[2a] In the following section,
the biomineralization of magnetite (Fe3O4) crystals in special
phospholipid vesicles of the cell interior (cytoplasm) is
discussed. The question arises as to whether on a pathway
from the exterior to the interior, a primitive stage of
biomineralization occurs in the cytoplasm, that is, without a
membrane. Indeed, gold ions in the form of AuIII or AuCl4
can actually penetrate into the cytoplasm of endospores, that
is, metabolically resting bacteria, and form numerous metallic
gold nanoparticles with a diameter of 5±20 nm diameter in a
redox reaction (Figure 3).[2a] Although AuIII ions are strong
617
Reviews
E. B‰uerlein
ferrihydrite (2-line ferrihydrite[35]) these bacteria not only
produce magnetite (Fe3O4) outside the bacteria. When after a
short time FeII is detectable in the reaction mixture, they form
about 60 iron oxide particles of 30±50-nm diameter in their
cytoplasm. These are very probably surrounded by a lipid
bilayer membrane and give rise to reflections in selected-area
electron-diffraction (SAED) experiments, which may be
attributed to magnetite or maghaemotite.[35] In contrast to
the previously described experiment,[34] these minerals were
formed at iron concentrations similar to those found in soils
and sediments.[35a]
3. Formation of Nanocrystal Forms from Fe3O4 in
Magnetic Bacteria
3.1. Occurrence of Magnetic Bacteria
Figure 3. Formation of gold nanoparticles in the cytoplasm of the endospores of Clostridium botulinum. If these endospores are suspended
in a solution of AuCl3 for 10 min at room temperature gold particles
with a diameter of 5±20 nm are formed. In a thin section without negative contrasting the gold particles are found by electron microscopy almost exclusively in the cytoplasm. The length of the bar represents
500 nm. (From reference [2a] with kind permission of T. J. Beveridge.)
oxidizing agents, in contrast to HgII ions, their conversion into
metal in the cytoplasm[32] resembles the detoxification and
resistance behavior of bacteria towards HgII ions. However,
metallic gold is deposited, whereas elemental mercury, which
in this case actually first arises through the action of cytosolic
mercury reductase, diffuses out of the bacteria as vapor.[33]
Intracellular particlelike iron sulfide deposits, which were
irregularly distributed and consist of amorphous material,
have been found in a few sulfate-reducing bacteria. They are
probably not surrounded by a membrane since after lysis of
the cells they could not be isolated by density-gradient
centrifugation.[34] Moreover, spherical particles have been
found in a few purple bacteria in which anaerobic photosynthesis occurs–that is, they originate from the earliest stages
of photosynthesis development–when cultured in media with
relatively high concentrations of iron. These particles, which
were surrounded by a kind of membrane and formed a chain
within the cell, were probably responsible for the magnetic
reaction of the bacteria. Such purple bacteria could thus be an
earlier stage of biomineralization in vesicles.[34]
A bacterium has recently been described as a ™missing
link∫ between bacterial biomineralization on the surface on
the one hand and in the cytoplasm or intracytoplasmic
vesicles on the other, as it probably allows both at the same
time.[35a] This bacterium, Shewanella putrifaciens CN 32, a
Gram-negative and facultatively anaerobic bacterium that
can reduce FeIII as electron acceptor to FeII in its anaerobic
energy generation (dissimilatory iron reduction, see also
Section 3.4). Under the strict anaerobic conditions of an H2/
Ar atmosphere and only in the presence of the FeIII oxide
618
Magnetic bacteria, also known as magnetotactic bacteria,
are widespread in aqueous habitats and prefer microaerophilic regions (oxic-anoxic transition zone, OATZ).[36, 37] Such
characteristic habitats exist mainly at the water±sediment
interface of fresh-water pools, lakes, and rivers, where most
magnetic bacteria have also been found. The Chiemsee
(Bavaria) is an excellent example.[7f] Numerous magnetotactic bacteria have also been found in the OATZ in
transitions from brackish water to sea water, which are
characterized by special layers (for example, vertically
arranged chemical layers that exhibit oxidizing conditions
above, but have a reducing environment below).[38] Although
unusual, these magnetotactic bacteria are found a few meters
deep in a water column, which depending on the season,
moves in a vertical direction. Even in the salt water of the
South Atlantic, magnetotactic bacteria have been discovered
in surface sediments up to a depth of 3000 m at the
continental edges of both the eastern part (Angola) and the
western part (Brazil).[39]
The magnetotactic bacteria are found in all significant
morphotypes, as cocci, spirilli, vibrios, rod bacteria, and also
as multicellular bacteria[40] in the different habitats. As is the
case with many other sediment bacteria, they are usually
difficult to culture. For this reason, since the isolation of the
first magnetotactic bacteria in 1979, today called Magnetospirillum magnetotacticum MS-1[41] (previously Aquaspirillum
magnetotacticum[42]) only ten pure cultures have been described.[7f, 43] It was, however, still possible to describe an ecology
and an initial phylogenetic pedigree of the magnetotactic
bacteria, thanks to an elegant method that combines the
comparative sequence analysis of ribosomal ribonucleic acid
(rRNA)[44] with in situ hybridization of fluorescent oligonucleotide probes (FISH) for bacteria that could previously not
be cultured (Figure 4).[45]
3.2. Species-Specific Magnetite Crystals
Electron microscopic studies have shown that magnetite
crystals with specific sizes and morphologies are found in the
numerous forms of bacteria (morphotypes); both size and
Angew. Chem. Int. Ed. 2003, 42, 614 ± 641
Biomineralization of Unicellular Organisms
Angewandte
Chemie
Figure 4. 16S rRNA-based tree reconstruction showing the phylogeny of magnetotactic bacteria (shadowed region). The tree was prepared with
the ARB software package and is based on a ™parsimony∫ analysis. It was corrected in accordance with the results of ™maximum likelihood∫–and
™neighbor joining∫ analysis, when unstable branches were indicated by multifurcations. The tree topology is based on almost full-length 16S rRNA
sequences. The partial sequences of the strains CS103, MC-1, MV-1, MMP, and RS-1 have been added with the parsimony algorithm without allowing changes to the overall tree topology. (From reference [45] with kind permission of R. Amann.)
morphology seem to be specific to the respective species of
bacteria. This assumption was confirmed by the isolation and
culturing of pure specimens: their ™mature∫ crystals correspond in size (maximum dimension) and morphology.[5b] In
contrast to the above essentially undistorted ™mature∫
crystals, a number of deviations were conspicuous when the
isolates were obtained from natural habitats;[46] these deviations are probably brought about by changes in the bacterial
environment, for example changes in iron concentration,
oxygen content, and temperature. As a simplification, the
crystal morphologies found in magnetotactic bacteria can be
classified into: 1) cubooctahedral, 2) pseudohexagonal, and
3) bullet-shaped (Figure 5). It should be emphasized here that
all the magnetite crystals found have the cubic face-centered
crystal lattice of magnetite, but species-specific crystal
morphologies.[4b, c] Idealized crystal morphologies of these
biogenic crystals are shown in Figure 6. In addition, the
morphologies found for greigite (Fe3S4), the magnetic sulfur
analogue of magnetite, is also shown.[4b, c]
The unexpected and unusual features of these biogenic
magnetite crystals is not only a narrow size distribution, but
above all, a diameter range of 40±120 nm, which thus allocates
them the highest magnetic moment. This diameter range
corresponds to magnetite crystals with a single magnetic
domain.[47] Since magnetite crystals from inorganic synthesis
offer neither a similarly narrow size distribution (and thus
also no such uniform magnetic properties) nor such a variety
Angew. Chem. Int. Ed. 2003, 42, 614 ± 641
of crystal morphologies, the formation of these biogenic
magnetite and greigite crystals is probably under strict
biological control. Research into this control has not yet
been reported.
Figure 5. Crystal morphologies and intracellular organization of magnetosomes from magnetotactic bacteria: A) cubooctahedral; b) bulletshaped; c±d) pseudohexagonal. The magnetosomes are arranged in
one (C) or more (D) chains. The lengths of the bars represents
100 nm.
619
Reviews
E. B‰uerlein
Figure 6. Idealized crystal morphologies of magnetite (Fe3O4) and greigite (Fe3S4)
derived from high-resolution TEM investigations on magnetosomes from magnetotactic bacteria. Top row: variations of pseudohexagonal prisms; bottom row
(from left to right): cubooctahedra, which occurs with magnetite and greigite in
magnetosomes; elongated cube with angular truncated corners, which was found
only in magnetosomes with greigite; elongated cubooctahedra present in a few
magnetosomes with magnetite. (Modified from Devouard et al.[4c] with kind permission of R. B. Frankel).
3.3. The Magnetosome: a Phospolipid Vesicle for the Production
of Species-Specific Magnetite Crystals.
3.3.1. The Magnetosome Membrane
In their pioneering work, Blakemore, Frankel, and coworkers were not only successful in the isolation[36a] and
cultivation[42] of Magnetospirillum magnetotacticum, but also
in the identification of the magnetite crystals.[5b] With the
assistance of transmission electron microscopy (TEM) they
were also able to show that almost all these crystals are
surrounded by an electron-dense layer, probably a lipid
bilayer. The term magnetosome was therefore introduced for
such enveloped, magnetic, inorganic crystals.[6] An initial
analysis of the lipids of small quantities of isolated magnetosomes revealed the presence of neutral lipids, free fatty acids,
glycolipids, sulfolipids, and phospholipids.[48] A pure culture
was obtained with the phylogentically related Magnetospirillum gryphiswaldense (Figure 7 A, B) which, unlike the
highly oxygen-sensitive M. magnetotacticum, was essentially
oxygen tolerant.[41, 49] A magnetotactic bacterium that provided the highest yields of highly purified magnetosomes
could be developed. A new technique in which magnetic
columns are used allowed the isolation of the magnetosomes
in a highly pure form.[50] As a result, the lipid composition of
the magnetosome membrane, as well as that of the outer and
cytoplasmic membrane, was determined quantitatively (Table 1).[7f, 51] Contrary to what was previously thought,[48] the
lipid profile of the magnetosome membrane differed from
that of the other two membranes. The amount of phosphatidylglycerol was about three times larger than in the external
and cytoplasmic membrane, whereas the amount of phosphatidylcholine was eight times larger. In contrast, ornithinamide
620
Figure 7. TEM image of A) Magnetospirillum gryphiswaldense with a
chain of cubooctahedral magnetite crystals and one flagellum at each
pole (the length of the bar represents 500 nm) and b) its isolated and
purified magnetosomes enveloped by a membrane (the length of the
bar represents 20 nm).
lipid and another, previously unknown lipid (X-NH2), which
also contained no phosphate group, could only be found in the
outer and cytoplasmic membranes, but not in the magnetosome membrane. Both lipids were valuable markers in the
separation of the membranes. These results suggest at least
that the magnetosome membrane follows its own synthesis
pathway, and could also be responsible for the maximum size
of the crystals as a deformation of the phospholipid vesicle by
crystals that were too large has not been previously observed.
Table 1: Lipid composition of the membranes of M. gryphiswaldense
Lipid
Outer
Cytoplasmic Magnetosome
membrane membrane
membrane
[mol %]
[mol %]
[mol %]
Phosphatidyl ethanolamine
Phosphatidyl glycerol
Ornithineamide lipid
Phosphatidyl choline
XNH2
59.9 5.1
13.9 0.9
18.6 5.8
±
7.6 0.7
70.7 0.5
12.6 1.9
4.7 0.4
1.1
7.2 1.9
52.8 5.5
38.3 5.7
±
8.9 0.5
±
Angew. Chem. Int. Ed. 2003, 42, 614 ± 641
Biomineralization of Unicellular Organisms
Also, varying the compositions of the phospholipids of the
membrane and thus its surface could also play a role in the
respective crystal morphology of magnetite crystals.
Functionally even more important than phospholipids are
the proteins of the magnetosome membrane; it is presumed
that the latter are involved in the mechanism of magnetite
crystal formation. They should be associated in a biologically
controlled biomineralization,[4a, 52] whose first principle is the
intracellular spatial compartmentalization with the lipid
bilayer of the magnetosome. Iron transport proteins that
produce local saturation in the magnetosomes, proteins that
catalyze crystal nucleation, and redox proteins that produce a
FeIII/FeII stoichiometry of 2:1 for magnetite (Fe3O4) are
expected to control this mechanism.
Consequently, interest in the chemistry of proteins and in
molecular biology has been directed at the proteins of the
magnetosome membrane;[48, 53] these proteins are probably of
pivotal importance for the simplest mechanism of a controlled
biomineralization. Because of the electrostatic properties of
their membrane, isolated magnetosomes form stable suspensions. Detergents are able to dissolve the magnetosome
membrane, upon which the magnetite crystals immediately
agglomerated.[7f] By means of sodium dodecyl sulfate polyacrylamide gel electrophoresis (SDS-PAGE), nine specific
protein bands were initially found in these extracts from the
magnetosomes of M. gryphiswaldense,[54] whereas in those of
M. magnetotacticum either two[48] or three[53a] protein bands
were found for the respective magnetosome membrane. In
the most recent investigations on M. gryphiswaldense, the
number of specific bands has risen to thirteen.[55]
3.3.2. Cloning and Sequence Analysis of the Genes of Magnetosome Membrane Proteins
The first step in the molecular biology of magnetosome
membranes, carried out by Y. Fukumori and co-workers
seven years ago with SDS-polyacrylamide gel on the Nterminal amino acid sequence of a 22-kDa protein (MAM22)
from M. magnetotacticum, led to the cloning and sequence
analysis of the corresponding gene.[53a] Likewise, the amino
acid sequence of a 24-kDa protein (MamA) from M. gryphiswaldense was determined, and was found to have a 91 %
similarity to the protein from M. magnetotacticum.[54] Although very different protein compositions were found in the
magnetosome membranes of a number of Magnetospirillum
strains by SDS-PAGE, Western blots (identification of the
protein by a specific antibody) showed that this 24-kDa
protein occurred in all membranes.[53b] It showed significant
homology with the tetratricopeptide repeat (TPR) family of
thus far at least 25 proteins.[56] The primary structure of the
TPR motif is a degenerate tetratrico (34) amino acid repeat
that is arranged mostly in tandems of up to nine single motifs.
This motif appears in many organisms, from bacteria to
humans. TPR proteins can exercise numerous biological
functions, which range from transcription control to protein
folding, and include protein transport and cell division. The
tetratricopeptide motif is repeated up to six times and appears
to play an important role in protein±protein interactions.
Angew. Chem. Int. Ed. 2003, 42, 614 ± 641
Angewandte
Chemie
Unlike typical membrane proteins, MamA does not contain a
hydrophobic membrane-spanning segment, and it can thus be
assumed that it is bound electrostatically to the magnetosome
membrane.[54] The homologous MAM22 protein has been
suggested to act as a receptor as well as playing a role in the
interaction with the cytoplasmic proteins.[57]
The recent investigations on M. gryphiswaldense[55] mentioned in Section 3.3.1 have led to the cloning and sequencing
of four further genes of important proteins of the magnetosome membrane. These genes (mamB, mamC, mamD,
mamE) and their derived proteins have thus far not been
associated with the formation of magnetite crystals in any
other magnetic bacterium. The mamB gene codes for a
protein whose band in SDS-PAGE corresponds to a molecular mass of 33.3 kDa.[55] The significant sequence homology
with proteins of the ubiquitous cation-diffusion facilitation
(CDF) family[58] makes MamB a candidate for iron transport,
as this protein family is not only responsible for the transport
of heavy metals, but also for resistance to certain heavy
metals. Such a resistance can lead to their export
(for example, the evaporation of mercury whose ions are
initially converted into elemental mercury by a mercury
reductase in the bacterial cytoplasm[33]) or equally to deposition of heavy metals in the vesicles.[59] The mamC gene
encodes a 15.5-kDa protein, the main protein in the magnetosome membrane. The mamD gene encodes a protein that
corresponds to a protein of 21.9 kDa.[55] No information on
proteins homologous to MamC and MamD has been found in
data banks.
During this work the genome of the two magnetic bacteria
Magnetospirillum magnetotacticum and Magnetococcus MC-1
(together with those of 13 nonmagnetic bacteria) were
published almost fully sequenced (that is, more than 95 %)
within four weeks by the Joint Genome Institute (JGI) of the
U.S. Department of Energy and immediately made accessible
in the public domain.[60] Thus, the homologous genes with
significant similarity to mamA and mamB from M. gryphiswaldense could be identified in the genome sequences of the
two magnetic bacteria.[55] The Mam proteins from M. gryphiswaldense and M. magnetotacticum have almost identical
sequences (91±97 % similarity), whereas the similarity between the amino acid sequences of Magnetospirillum species
and Magnetococci-MC1 lies at 46±67 %. By means of a
further comparative analysis it was possible to assign the
fourth new gene, mamE, to the N-terminal sequence of the
36.3-kDa protein from M. gryphiswaldense.[55]
3.3.3. The mamAB Gene Cluster in M. Gryphiswaldense,
M. Magnetotacticum, and Magnetococcus MC-1
Not only are genes homologous to mamA and mamB
from M. gryphiswaldense found in the two genomes of
M. magnetotacticum and Magnetococcus MC-1, but additionally the neighboring genes, open reader frames (ORFs), are
also arranged collinearly in all three bacterial strains (Figure 8). Furthermore, the corresponding ORFs belong predominantly to the same known protein families, or have
nothing in common with known proteins.[55] Clearly, the
621
Reviews
E. B‰uerlein
Figure 8. Molecular organization of the mamaAB gene cluster in M. gryphiswaldense, M. magnetotacticum, and Magnetococcus MC-1. The arrows indicate the direction of gene transcription. The different marking of the graphically depicted arrows ORFs (open reader frames), which belong to
the respective homologous families of genes. These families occur in every mamAB cluster of the three magnetotactic bacteria investigated. The
dashed lines connect equivalent genes, that is, the most similar homologues. (From reference[55a] with kind permission of D. Sch¸ler and the
American Society of Microbiology.)
following protein families in the mamAB cluster may be
assigned to these three, possibly also to all known, magnetic
bacterial strains (Figure 8):
1) TPR ProteinsThe mamA genes in all three strains show
similarities to the genes of the TPR proteins (Section 3.3.2).
2) CDF ProteinsNot only are the proteins of the mamB
genes in all three strains (Section 3.3.2) homologous to the
CDF proteins, but additional CDF homologues have also
been found in the mamAB clusters of M. magnetotacticum
and Magnetococcus MC-1 (Figure 8)
3) HtrA-like Serine ProteasesIn addition to the mamE gene
of M. magnetotacticum, additional genes whose proteins
are similar to these serine proteases[61] have been found in
the mamAB clusters of all three strains: ORF 2 in
M. gryphiswaldense, ORF 7 in M. magnetotacticum, and
ORF 2 in Magnetococcus MC-1 (Figure 8).
4) LemA-like ProteinsIn each of the three bacterial strains,
an ORF has been found, with a similar sequence to that of
the lemA-like genes, between the mamA and mamB
genes. The functions of this protein family[62] are currently
unknown.
In parallel, two further classes of genes occur in the
mamAB clusters of all three magnetic bacteria: the ORFs 8
and 9 in M. gryphiswaldense, the ORFs 13 and 14 in M. magnetotacticum, and the ORFs 8 and 9 in Magnetococcus MC-1
(Figure 8).[55] As no significant sequence similarities have
been found for the derived proteins of prokaryotes or
eukaryotes, they could represent important functions of
magnetite biomineralization. In addition to these details, it
can be concluded from a glance at Figure 8 that the current
622
genes probably have an operon-like organization and are
connected functionally with each other. Thus they could play
a specific role in magnetite biomineralization.
As is now known (D. Sch¸ler, personal communication)
the genes mamC[55] and mamD[55] are found in a second gene
cluster.
3.4. The Mechanism of Magnetite Crystal Formation in
Magnetosomes
Before the detailed mechanism of magnetite crystal
formation is uncovered by switching off of individual genes
(knock-out method) the results of classical microbiology on
magnetite biomineralization will be addressed once more at
this point.
The unusually high iron uptake of the magnetotactic
bacteria under microaerobic conditions and 10±20 mm FeSO4
leads to a maximum in both magnetism and cell growth. The
iron content ranges from about 3 % in most bacteria to 10 %
of dry weight in Magnetobacterium bavaricum.[37, 45a] Iron ions
are not taken up as FeII but as FeIII, which is formed from FeII
by rapid oxidation in the medium, although the maximum
solubility of FeIII in water at biological pH is only about
1018m. This is possible without precipitation because this
enormous FeIII uptake take place in a medium in which a
stationary culture of M. gryphiswaldense was previously
cultured under moderate iron deficiency and which greatly
stimulates FeIII uptake.[63] However, there is no indication that
the complexing agent is one of the known siderophores[64]
(FeIII-specific complexing agent, MW. ca. 1000 Da, stability
Angew. Chem. Int. Ed. 2003, 42, 614 ± 641
Angewandte
Chemie
Biomineralization of Unicellular Organisms
constant Kf ¼ 1030±1050) as synthesized by many, primarily
aerobic, nonmagnetic bacteria under the stress of low iron
concentrations. This was also observed with the equally airtolerant Magnetospirilium sp. AMB-1.[65] The FeIII uptake of
M. gryphiswaldense follows Michaelis±Menton kinetics, with
the constants KM ¼ 3 mm FeIII and Vmax. ¼ 0.86 nmol
FeIII min1 mg1 (dry weight). These data correspond to a
transport system of low affinity and high rate.[66]
Also of considerable importance for the mechanism is
whether and when FeIII is reduced to FeII during transport
across the three membranes (outer, cytoplasmic, and magnetosome membranes), that is, whether in the final instance FeIII
or FeII ions arrive at the ™empty∫ magnetosome vesicles (here,
empty means without magnetite). A ferrireductase such as
that isolated from the external membrane of a dissimilatory
iron-reducing bacterium[67a, b] and identified in Candida albicans[67c] has hitherto not been found in magnetotactic cells.
However, a soluble and cytoplasmic ferrireductase, which is
bound loosely to the cytoplasmic membrane and is inhibited
by ZnII ions, has been isolated from M. magnetotacticum. In
that way the average magnetosome count falls and the
number of nonmagnetic cells increase.[57, 68] If FeII ions were
transported into the empty magnetosome vesicles, two
protons would be formed by the oxidation of two FeII ions
to give two FeIII ions [Eq. (1)], since Fe3O4 contains stoichiometrically two FeIII and one FeII.
2 Fe2þ þ 2 H2 O ! 2 ½FeðOHÞ2þ þ 2 Hþ þ 2 e
bacterial cell growth and the oxygen concentration in the
medium.
When radioactive 55FeCl3 is added to a culture of nonmagnetic M. gryphiswaldense when the oxygen concentration
for magnetite biomineralization is reached,[50] the extensive
iron uptake, clearly closely coupled with the parallel increase
in magnetism, begins immediately (Figure 9 A). As a concentration of 1 mm iron in the medium is sufficient for cell growth,
this result supports the assumption that most of the added
iron (final concentration 30 mm) is transported into the
vesicles, the empty magnetosomes (Figure 9 B). No clusters
of stored iron are found in the cytoplasm.
With magnetically induced light scattering, magnetism is
detected just 5±10 min after the addition of iron. If after a
total of 30 min cells are removed and killed, particles are
ð1Þ
Formally this would be associated with an acidification of
the intravesicular medium. In analogy to the coccolith vesicles
of coccolithophores, in which calcification (calcite formation)
also leads to the formation of protons (Section 4.4.2), here too
the protons would have to be transported from the magnetosomes to produce a defined, alkaline microenvironment for
magnetite crystal formation.
Y. Fukumori and co-workers were the first to discover FeII
oxidation activity in a culture of M. magnetotacticum, which
uses nitrate as electron acceptor. They were able to attribute
the high FeII nitrite oxidoreductase activity to a nitrite
reductase, a cytochrome cd1, in the periplasma.[57, 68] As FeIII
is clearly taken up by bacteria (Section 5.4) it could be
reduced to FeII by this enzyme and transported into the empty
vesicle. However, if this enzyme oxidizes FeII to FeIII the
divalent ions would have to be (re)transported into the
periplasma.
The formation of magnetite crystals and thus magnetism
has provided the unique opportunity to develop a simple
spectroscopic method by which to measure the time-resolved
formation of a mineral, magnetite, in living magnetic bacteria.[69] In a differential light scattering induced by magnetism, a light beam is passed through a suspension of cells of
M. gryphiswaldense subjected to a strong homogeneous
magnetic field both parallel (Emax) and vertical to (Emin) to
the light beam. The ratio of the two scattering intensities
(Cmag ¼ Emax/Emin) correlates well with the average number of
magnetic particles of different cell populations.[69] This
sensitive and rapid method makes it possible to investigate
the magnetism in parallel with the iron uptake as well as with
Angew. Chem. Int. Ed. 2003, 42, 614 ± 641
Figure 9. Close coupling of increased iron transport and magnetism during
the growth of M. gryphiswaldense: A) iron was added as 55FeCl3 after 14.5 h (arrow) and adjusted to an initial concentration of 30 mm. Under microaerobic
conditions, iron uptake (Feintra) takes place almost parallel to the formation of
magnetism (M): & cell density (OD), * cellular magnetism (M), and ^ intracellular iron content (Feintra corresponds to nmol Fe per mg dry weight). (Modified from reference [50]). B) Representation of empty vesicles. Thin section of
M. magnetotacticum after iron deficiency over several generations. Three almost magnetite-free magnetosomes (empty vesicles) clearly show a bilayer
membrane (arrow). The length of the bar represents 20 nm. (From reference [2a] with kind permission of T. J. Beveridge.)
623
Reviews
Figure 10. TEM images of the formation of magnetite crystals in
M. gryphiswaldense: A) Magnetite particles 30 min after induction of
Fe3O4 biomineralization by the addition of 30 mm FeCl3 (see Figure 9 A.
These early particles, which already have a chainlike alignment, have a
diameter of 5±20 nm and correspond mostly to the order of magnitude
of superparamagnetic crystals.[47] B) Chain of magnetosomes of a cell
that has grown for several hours in the presence of 30 mm FeCl3. The
™mature∫ crystals, which are mostly situated in the middle of the
chain, are cubooctahedral and have a maximum diameter of 42±
45 nm.
found in the transmission electron microscope image which,
even though spatially separated, begin to indicate a chain
(Figure 10 A).[50] The initial particles are 5±20 nm in size and
consequently exhibit superparamagnetism,[47] that is, they do
not have a permanent magnetic dipole moment. This rapid,
local supersaturation and crystal formation presumes the
existence of empty vesicles. The chain of essentially ™mature∫
crystals, which in comparison to inorganic magnetite are of
surprising regularity and have a diameter of 42±45 nm
(Figure 10 B), is the result of this biomineralization. The
elemental assignment of iron and oxygen was carried out by
electron microscopy, which operates under low-dose conditions and uses the automatic three-window method of energy
spectroscopy imaging (ESI).[69b, 70]
3.5. Magnetotaxis
Magnetotactic bacteria are so named because they
typically align themselves 80±90 % like a magnetic needle in
the Earth©s magnetic field of 50 mT (Tesla).[71] This alignment
is possible because they have a permanent magnetic dipole,
which is large enough within the Earth©s magnetic field to
overcome the thermal forces that would otherwise produce a
statistical disorder. Such a dipole is made up of the abovementioned magnetosomes, intracellular phospholipid vesicles
in which single-magnetic-domain magnetite crystals organize
themselves along a chain through magnetic interactions. This
chain is probably bound to the inside of the cytoplasmic
membrane with the help of each individual magnetosome and
a protein of the TPR family[55] (Section 3.3.2). In this way the
bacterial cell as a whole acquires a magnetic dipole moment,
which corresponds approximately to the sum of the individual
magnetic moments of the magnetosomes[72] and is orientated
parallel to their axis of rotation. Aligned passively in the
Earth©s magnetic field, the bacterial cells then swim along the
magnetic field lines. They are driven by their flagella motors
and steered by an aerotaxis, that is, by a movement that
follows an oxygen concentration gradient, possibly to more
favorable microaerobic habitats. Discussion can therefore
624
E. B‰uerlein
only be of a magnetically supported aerotaxis,[73] not of
magnetotaxis.
However, because magnetic bacteria already require
microaerobic conditions to synthesis their magnetite crystals,
as was demonstrated experimentally with M. gryphiswaldense
(Figure 9 A), the above-mentioned aim of a magnetically
supported aerotaxis also appears questionable.
However, one new development should be mentioned
briefly here. There are at least two theories to describe the
mechanism of a biological compass in animals: a) the
magnetic±mechanical interaction of the Earth©s magnetic
field with minute magnetic crystals (< 100 nm) in tissues,[74]
and b) the effect of changes in magnetic fields on biochemical
reactions.[75] Although both theories are similarly well-corroborated, the former theory is currently preferred because it
has been possible to identify magnetite in the tissues of
different animals. For example, superparamagnetic magnetite
particles (grain size < 20 nm) have been found in carrier
pigeons and soldier ants,[19] and with grain sizes that are
typical of magnetic bacteria (30±120 nm) in, for example,
salmon[17] and trout[18] (Section 1). For the latter, a mechanism
with a certain analogy to the formation of the magnetosome
chain has been proposed, that is, an orientation of the chain in
the magnetic field. The following theory has been developed
for the accumulation of superparamagnetic magnetite particles: In an external magnetic field, a membrane vesicle of a
few mm diameter which contains a number of superparamagnetic particles is extended in parallel to the magnetic field axis
and contracted across it. As this effect can be reinforced by a
change in osmotic pressure, these membrane vesicles, also
known as ferrovesicles, have been called magnetic osmometers.[76] There are probably at least two types of magnetoreception in animals. It would be of interest to know whether
vesicle-dependent magnetoreception is the older of the two,
and whether it developed from the magnetosomes of
magnetic bacteria.
4. The Formation of Complex Crystal Morphologies
of CaCO3 in Unicellular Algae
4.1. The Occurrence of Coccolithophores (Calcareous Algae)
The best-known species of marine algae, mainly unicellular, which form a large part of the phylum Haptophyta
(Prymnesiophyta)[77] are the coccolithophores (also called
coccolithophorides). They are covered by a shield, the coccolithosphere, a frequently spherically-shaped aggregation of
shells, platelets, or rods which are called coccoliths and are
made of unusual calcite crystals. Together with the foraminifera,
which do not contain any chlorophyll, the calcareous algae form
a main part of marine plankton and are the most important
carbonate source for the deep-sea sediments of the oceans.
Fossil coccolithophores from the late Triassic Period and
the Palaeozoic Era are thus far known in only a few examples
and in a small number of species.[78] However, they are found
abundantly in a larger variety of species in the widespread
deposits of light calcareous rocks of the early Jurassic Period
(190±170 million years ago) and above all in the earlier
Angew. Chem. Int. Ed. 2003, 42, 614 ± 641
Biomineralization of Unicellular Organisms
Cretaceous Period (95±63 million years ago). The end of the
Cretaceous Period and the transition to the Tertiary Period
was characterized by the extinction of about two thirds of the
50 genera of the coccolithophores of that time. New species
developed during the Tertiary Period to reach a new highpoint during the Eocene Epoch (50 million years ago).[78, 79]
On the basis of coccolith ultrastructure, attempts have been
made to draft a type of phylogeny for the coccolithophores.[79c]
4.2. Holo- and Heterococcoliths
A differentiation is made between holo- and heterococcoliths with respect to their calcite crystal morphology.
Holococcoliths are composed merely of simple calcite elements of rhombohedral or prismatic crystal forms. They are
generally formed as extracellular structures.[80]
In analogy to the magnetite crystals in the magnetosomes,
the calcite crystals of the heterococcoliths are likewise
species-specific, and their complicated and complex morphologies are not found in inorganic chemistry.[81] Moreover,
the property of the heterococcoliths, rare for eukaryotes, to
also form their calcite structure intracellularly and in membrane vesicles suggests an evolutionary relationship to
bacterial biomineralization.
Although the two unicellular algae Emiliana huxleyi and
Pleurochrysis carterae have been intensively investigated to a
similar extent, only the latter is described here, as coccolith
vesicles have hitherto only been isolated from Pleurochrysis
carterae (Figure 11).[82] Moreover, it has been possible to find
Figure 11. SEM image of a Pleurochrysis cell showing the mineralized
scales of the coccosphere (the calcareous shield). These scales, called
coccoliths, consist of an oval, organic base plate (x) and two discs of
CaCO3 crystals on their rim (arrow points). The length of the bar represents 1.0 mm. (From reference [89] with kind permission of M. E.
Marsh and Springer Verlag, Vienna).
Angew. Chem. Int. Ed. 2003, 42, 614 ± 641
Angewandte
Chemie
or produce mutants of this calcareous alga with which the
respective function of (now) three acid polysaccharides in the
formation of the complex crystal morphologies have been
explained (see also Section 4.5.1).[83]
4.3. Heterococcolith Structures of the Unicellular Alga,
Pluerochrysis carterae
On the surface of Pleurochrysis, the heterococcoliths
consist of an oval organic base plate and, attached to it at the
rim, a ring of calcite crystals. These crystals form two parallel
discs, which extend radially from the rim of the coccolith
(Figure 12 A, B).[81] The ring itself is formed by the interlocking of individual crystals with alternate radial (R) or
vertical (V) alignment (Figure 13). This R and V orientation
corresponds to the alignment of the crystallographic c axis of
the crystals to the plane of the coccolith. With the exception
of the distal shield element, the disc element that is more
distant from the base plate and unusually appears as a rather
curved crystal surface (Figure 12 B), all other surfaces are
sheetlike elements (Figure 13).[81]
Figure 12. A) TEM images of an isolated, mature coccolith from Pleurochrysis whose plane is tilted about 308; B) thin section which shows a
cross-sectional view of a mature coccolith of Pleurochrysis in its vesicle
prior to secretion into the coccosphere. The V and R crystal units are
located on the rim of the base plate (b). The distal shield (d), the proximal shield (p), the inner tube (i), and the outer tube (o) as elements
of the mineral ring are indicated. The length of the bar represents
100 nm. (From reference [81b] with kind permission of M. E. Marsh
and Springer Verlag, Vienna)
625
Reviews
E. B‰uerlein
which in well-founded analogy to the c subunit of the
ATP synthase was named ™proteolipid∫.[84b] It was
established by cloning and sequencing that its ORF
has a derived molecular weight of 16.2 kDa. Its apparent molecular weight was, however, 24 kDa. Immunoblotting and immunofluorescence microscopy confirmed that this 24-kDa protein corresponds to the
16.2 kDa protein and was apparently modified posttranslationally.[86]
4.4.2. Proton, Hydrogen Carbonate, and Calcium Transport
in the Coccolith Vesicles
Figure 13. Schematic representation of the V and R crystal units and their precisely fitted
and mutually interlocking structure on the rim of the coccolith of Pleurochrysis viewed
from the interior of the coccolith. For simplicity the crystal elements are represented as
if they consist of thin plates. The crystal surfaces, which correspond to the distal and
proximal shield, and the outer and inner tubes are identified. (From reference [81b] with
kind permission of M. E. Marsh and Springer Verlag, Vienna.)
4.4. Biomineralization in Coccolith Vesicles of Pleurochrysis
Carterae
4.4.1. Membrane Proteins of the Coccolith Vesicles
The exceptional difficulties in isolating coccolith vesicles
have hitherto been overcome only with Pleurochrysis carterae.[82] A successful combination of a) dissolution of the
cytoskeleton, which connects the organelles of the coccolithophores, with b) an elaborate sucrose gradient centrifugation, which is possible thanks to the somewhat higher density
of the coccolith vesicles, c) the possible differentiation of the
coccolith vesicles from the outer coccoliths, and d) the
selective dissolution of these free, outer coccoliths has led
to the first isolation of purified coccolith vesicles (Figure 14 A, B).
Similar to the magnetosome membrane from the magnetic bacterium M. gryphiswaldense,[55] a whole range of
polypeptides (in this case as many as 20) have been found
by SDS-PAGE during the analysis of the purified coccolith
vesicles. This large number is explained primarily by the now
confirmed presence of a V(vacuole)-ATPase, whose total
complex in many species contains between 12 and 13
subunits.[84]
Of the polypeptides, whose apparent molecular weights
lie close to the most important subunits of the V1(A±E) and
V0(a, c) part, two have been identified: 1) The subunit B,
which reacted specifically with the 2E7-monoclonal antibody
raised against the corresponding subunit B from oats.[85] The
isolated and purified vesicles were made visible on the basis of
the same antibody (Figure 14 A).[82] 2) The ™16-kDa subunit∫,
626
The formation of calcium carbonate leads to the
release of many protons in the coccolith vesicles.
Depending upon whether hydrogen carbonate
[Eq. (2)] or carbon dioxide [Eq. 3] is transported into
the vesicle, this amounts to one or two protons per
molecule. Both the coccolithophore cell and the magnetic bacterial cell must produce a precise pH within a
defined microenvironment of their vesicles which leads
to crystal nucleation, crystal growth, and the mature end
product, of which the crystal morphology and size are
specific to the coccolith and magnetite crystal, respectively (Section 3.4).
HCO3 þ Ca2þ ! CaCO3 þ Hþ
ð2Þ
CO2 þ H2 O þ Ca2þ ! CaCO3 þ 2 Hþ
ð3Þ
Extensive ATPase activities have been found in the
coccolithophore cell,[87] , for example, a CaII-dependent ATPase of the P type on the plasma membrane and a V-ATPase
in the coccolith vesicles. As the coccolith vesicles stem from
the trans-Golgi network, (Figure 15) V-ATPases are also
found in the latter.[84b] An ATP-dependent proton transport
that is inhibited by nitrate has been found in purified isolated
Figure 14. Immunofluorescence microscopy of coccolith vesicles. Isolated and aldehyde-fixed vesicles were exposed to 2E7 monoclonal and
secondary antibodies, which had been labeled with FITC fluorochromium. Vesicles visualized by epifluorescence (A) and by Normaski optics
(B). The length of the bar represents 1 mm. (From reference [82a] with
kind permission of E. L. Gonzalez.)
Angew. Chem. Int. Ed. 2003, 42, 614 ± 641
Biomineralization of Unicellular Organisms
Angewandte
Chemie
established: in Pleurochrysis, calcium ions form 25 -nm
particles in which the ions complex with the polysaccharides
PS1 and PS2.[89] A large number of these particles, the
coccolithosomes,[89] are formed in the middle Golgi cisterns
and are probably infiltrated to and into the outermost transGolgi cisterns in small vesicles of Golgi membranes. They are
present in these coccolith vesicles both during calcite
nucleation and during the whole crystal growth (Figure 15).[81]
4.4.3. Formation of a Heterococcolith in a Coccolith Vesicle
Figure 15. Schematic representation of coccolith formation in the Golgi
apparatus of Pleurochrysis. A coccolith vesicle is shown before (1), during (2), and after (3) mineral deposition. Ca±PS1/PS2 complexes are
formed in medial Golgi cisternae. These complexes are present as single 20-nm particles (p) before and during mineralization (1 and 2). After the mineralization ceases (3 and coccosphere) the crystal surfaces
are surrounded by an amorphous polyanion coat (cc). (The coccolith
base plate (hatched marks), coccoliths (coc), unmineralized scales (s),
chloroplasts (chl), endoplasmic reticulum (er), nucleus (n), plasma
membrane (pm).) (From reference [97] with kind permission of M. E.
Marsh and Springer Verlag, Vienna.)
coccolith vesicles. In this way the function of the previously
identified V-ATPases could also be established.[86] Furthermore, with an antibody against its subunit c of the V0 part the
V-ATPase could be localized on the coccolith vesicles.
However, against expectation, the protons were pumped into
and not out of the vesicles. This orientation of the V-ATPase
corresponds to that found in secretory vesicles in all
eukaryote cells, which likewise arise out of the trans-Golgi
network.[84b] It is thus still an open question as to how the
protons are transported out of the vesicles. An antiporter,
which for example transports two protons out and one
calcium ion in, has so far not been found.
Little is known about the molecular mechanism of the
transport of HCO3 or CO2, generally known as dissolved
inorganic carbon (DIC). However, the release of protons
during calcite formation appears to be an important requirement the for the carbon-concentrating mechanism (CCM) in
chloroplasts, which in spite of the low concentration of
dissolved DIC in the ocean allows a higher photosynthesis
rate than would be expected at this lower DIC concentration.[88]
Even today the mechanism of CaII transport into the
coccolith vesicles is also essentially unknown. The hope that
one of the two CaII-dependent ATPases–the P-type or Vtype, which represents the widespread ATPase activity in
Pleurochrysis–could be connected with CaII transport is still
unfulfilled, both for the plasma membrane and the coccolith
vesicles.[82] However, the assumption that in place of diffusion
by the cytosol, calcium ions are also transported, immobilized
in vesicles, into the coccolith vesicles is becoming more firmly
Angew. Chem. Int. Ed. 2003, 42, 614 ± 641
Before crystal formation begins, many coccolithosomecontaining vesicles attach themselves to the outer rim of the
basal plate (Figure 16 A). This rim is covered by a small ring
of a special organic material called the coccolith ribbon. The
first crystallites form in accumulations of coccolithosomes
(Figure 16 B); in each case, a crystallite corner is in contact
with the coccolith ribbon (Figure 16 C) and each developing
crystal is firmly attached. A ring of 24 small crystals of
parallelepipeds, which exhibit alternating radial (R) or
vertical (V) orientation relative to their crystallographic
c axis, is formed from the crystallites (Figure 17 A). A single
fold of the coccolith ring that would run parallel to the rim of
the basal plate can explain why crystals of similar structure
form on both sides of the fold with different orientation.[81b]
The maturing of crystals to their complex structures begins at
the V unit (Figure 17 B) with the extension of a (1014) surface,
which forms the platelike surface of the outer tube elements
(Figure 17 C), and ends with the structure of parallel double
discs, which is characteristic of mature coccoliths. The detailed crystallographic analysis[81b] is not discussed here.
Figure 16. Thin sections showing successive early stages of mineralization in Pleurochrysis with a grazing view of the coccolith rim: A) Base
plate with dense, polyanion-rich particles (coccolithosomes) associated
with their rim (arrow head). B) Base plate with a ring of small, rectangular crystals between the polyanion-rich particles (arrow point). Sections without negative contrasting. C) Cross-section showing a small
crystal (arrow) above the base plate. It is clearly attached to a small
band of organic material, the coccolith ribbon, which is located on the
distal rim of the base plate. The lengths of both bars represent
100 nm. (From reference [81a, b] with kind permission of M. E. Marsh
and Springer Verlag, Vienna.)
627
Reviews
E. B‰uerlein
work on unicellular eukaryotic organisms has gained in
importance. In particular, the pioneering investigations of
Marsh et al. on Pleurochrysis carterae[81a, 96] have for the first
time allowed the localization of crystal nucleation in a
unicellular organism, in this case calcite, and an unusual
crystal growth. The isolation of acidic polysaccharides,[90a] the
production of associated antibodies,[89, 96] and a mutant
analysis[83, 96] of these polysaccharides, together with morphological studies, have also contributed.
4.5.1. The Acidic Polysaccharides PS1, PS2, and PS3
Figure 17. a) Electron microscope image of an isolated protococcolith
from Pleurochrysis. It consists of a ring of 24 small crystals in the shape
of rectangular parallelepideds. The crystals have alternating V and R
unit orientation. B) An isolated, still immature coccolith at a later development stage. As observed in projection, the R elements have the
structure of double parallelograms, which will develop into the inner
tube and proximal shield elements. C) Cross-section of a coccolith at a
similar stage showing an immature V element on the rim of the coccolith base plate. The lengths of all bars represent 100 nm. (From reference [81a, b] with kind permission of M. E. Marsh and Springer Verlag,
Vienna.)
The process of mineralization is apparently stopped by the
pronounced swelling[89] of the coccolith vesicle (Figure 15),
probably when the amount of ion product of the calcium
carbonate is so greatly decreased by the huge influx of fluid
that no further precipitation is possible. Pleurochrysis crystal
growth also ends with dissociation of the polyanion-rich
particles, the coccolithosomes. In addition, the residual
calcium ions and numerous polyanions are released; the
latter then form a protective layer on the crystal surface.[90]
Subsequently the ™mature∫ coccolith is transported to the
upper surface of the algae, the coccolithosphere, by exocytosis.
4.5. Mechanism of Formation of Complex Crystals in the
Coccolith Vesicles of Pleurochrysis carterae
For some time it has been assumed that highly acidic
macromolecules play an important role in the biomineralization of eukaryotes. Of these, those that can bind a large
number of calcium ions and occur in high concentrations at
the centers of mineralization of a number of tissues have
aroused special interest.[91, 92] For example, the phosphophoryn proteins, which are associated with the dentin of vertebrates, are rich in aspartic acid and phosphoserine.[93] In the
presence of calcium ions, they form 25-nm particles, as does
one of the acidic polysaccharides.[94] The considerable difficulties associated with the elucidation of the biomineralization of multicellular organisms because of the growing
number of sequenced proteins and because of their different
possible activities[95] can be judged from the mother-of-pearl
of crustacean shell, which has been extensively investigated
because of its hardness and fracture resistance. Therefore
628
The three acidic polysaccharides PS1, PS2, and PS3, have
been isolated from the coccoliths of Pleurochrysis and
purified.[90a] In polyacrylamide gel without SDS, PS1 and
PS3 migrate as relatively narrow bands that appear to
represent a narrow molecular weight range. The fact that, in
contrast, PS2 forms a ladder of many discrete bands over the
whole gel suggests a broad variation in the degree of
polymerization. PS2 is the most abundant coccolith polyanion
in Pleurochrysis. The mass ratio of PS2/PS1 is 77:22. The
smaller but very important fraction[96] of PS3 lies close to 2 %.
PS2 is made up essentially of the repeating motif [!4)dglucuronate(b1!2)meso-tartrate(3!1)glyoxylate(1-]n (formula of the calcium salt of PS2 reproduced from references [89, 90a] with kind permission of M. E. Marsh), a motif
that has not previously been found in any other polysaccharides. With its four carboxylate groups or four negative charges
per motif, PS2 is the most acidic polyanion that has thus far
been described for mineral storage.[90, 97] The primary structures of PS1 and PS3 are still unknown. PS1 is a polyuronide
that consists essentially of glucuronic acid and galacturonic
acid (1:3 ratio).[90] PS3 is a galacturonomannan with a
significant number of sulfate ester groups.[81a]
The fate of PS1 and PS2 was followed during the whole
mineralization process by TEM with well-characterized antibodies and immunogold labeling of thin sections; the analysis
began with crystal nucleation and continued to the end of
crystal growth.[89] The many detailed electron microscope
images cannot be shown here for reasons of space. Their
results, which are also discussed together with the following
mutation experiments, will be summarized briefly here (see
Figure 15 for a schematic representation).
The polysaccharides PS1 and PS2 are synthesized in the
central cisterns of the Golgi apparatus. There, together with
calcium ions, they form numerous individual 25-nm particles
called coccolithosomes. Vesicles that contain a series of these
particles are derived from the edges of the central cisterns.
These PS1/PS2±Ca particles are then transported in these
Angew. Chem. Int. Ed. 2003, 42, 614 ± 641
Angewandte
Chemie
Biomineralization of Unicellular Organisms
small vesicles to the coccolith vesicles, with which they fuse[81a]
The coccolith vesicles contain an oval basal plate on the rim of
which the PS1/PS2±Ca particles now attach themselves. The
mineralization process then begins there during which these
particles and thus also the polyanions PS1 and PS2 are
present, from crystal nucleation to shortly before the end of
crystal growth. In the end phase of coccolith maturing, the
crystal receive a polysaccharide coating of PS1 and PS2, which
is not detectable on them during the growth phase.
Until recently no well-characterized antibodies for the
localization of PS3 by immunogold labeling could be produced. However, it was possible to show by pulse-chase
experiments with radioactive carbonate, sulfate, and calcium
in kinetic experiments[97] that the cellular half-life of PS3 is
more than three times that of PS1 and PS2. It is concluded
from this that PS3 is synthesized and secreted in a different
way to that of PS1 and PS2.[81a, 97]
4.5.2. Mutants from Pleurochrysis that do Not Synthesize These
Acidic Polysaccharides
More than 100 spontaneous and chemically induced
mutants have been isolated from Pleurochrysis. The respective polysaccharides were isolated from the cells after pulse
labeling with 14C carbonate and 35S sulfate and characterized
by gel autoradiography.
The coccoliths of a ps1 mutant, which had been induced
chemically
with
N-methyl-N’-nitro-N-nitrosoguanidine
(MNNG) and no longer expresses PS1, are clearly similar in
every respect to the wildtype coccoliths.[81] That means that
PS1 probably does not play a role in the formation of these
coccoliths, which possess the morphology of the wildtype.
The hypothesis that PS2 is a functional intermediate of
calcite biomineralization in Pleurochrysis was supported by
three independent mutants.[83] Two of these ps2 mutants
formed spontaneously, a third was induced chemically with
MNNG. All three are phenotypically identical and do not
express PS2. The mineral content of each of these mutants is
less than 5 % of the content of the wildtype, which also means
that most of their coccoliths are not mineralized. It is
extremely important for the functional analysis of the ps2
mutants, however, that they deposit the small amount calcium
carbonate not only on the rim of the basal plate as calcite, like
the wildtype, but also that they form the V and R structures of
the wildtype crystal with the more mature crystals. Thus PS2
has no influence on the crystal morphologies of this specific
crystal form. Moreover, the merely reduced calcite mineralization on the coccolith rim in the absence of PS2 indicates
that PS2 is not absolutely necessary for crystal nucleation, and
that the PS1/PS2±Ca particles are clearly not the only source
of calcium for coccolith formation.
How does PS2 contribute to the formation of unusual,
anvillike calcite crystal? On a small ribbon of organic
material (Figure 16 C) a large number of PS1/PS2±Ca particles are bound to the outer rim of the basal plate in the
wildtype (Figure 16 A). The result is a local concentration of
6 m polyanion-bound calcium ions[90b] before crystal formation
begins.[89] In contrast to the ps2 mutants, the PS1/PS2±Ca
particles in the wildtype clearly accelerate the rate of calcite
Angew. Chem. Int. Ed. 2003, 42, 614 ± 641
crystal nucleation. It is still not known whether the PS1/PS2±
Ca complex exercises its action through its structure, that is, as
part of a postulated crystal nucleation complex, or through its
capacity as a buffer for calcium ions.
Mutant analysis, the ideal pathway for an understanding
of the mechanisms of complex biological systems, has, as
previously described, come close to an explanation of crystal
nucleation. The particular potential of this method was plain
to see in the first described ps3 mutants,[81a] which were
induced by MNNG and were now termed ps31:[96] when PS3
was no longer expressed, the complex anvillike calcite crystals
were also not formed; instead, only simply orientated crystals
that resemble the rectangular parallelepipeds of the protococcoliths of the wildtype were found (Figure 17 A). Amongst
500 new mutants of Pleurochrysis, a second mutant (named
ps32) has been recently found that also exhibits the protococcolith phenotype and does not express PS3. In addition, it
was also now possible to produce monospecific anti-PS3
antibodies. In this way it was possible to demonstrate that
PS3, unlike PS1 and PS2, is localized between the membrane
of the coccolith vesicle and the crystal surface and is directly
associated with the growth and the shaping of the calcite
crystals to their complex anvillike morphologies (M. Marsh,
personal communication).[96]
5. The Formation of Nanostructured Cell Walls of
Amorphous Polysilicic Acid in Diatoms
5.1. Occurrence of Diatoms
Diatoms are unicellular eukaryotic organisms, algae of the
class Bacillariophyceae. The number of living species known
today is about 100 000.[98] They are found in sea water,
brackish water, and in fresh water. These diatoms make up
the dominant part of phytoplankton, which together with
other marine organisms such as silicoflagellates, radiolaria,
and sponges, converts annually the exceptional amount of
6.7 gigatons (giga ¼ 109) of silicon biogenically during the
formation of their silica skeletons.[99]
According to recent age estimates, the oldest fossil
diatoms could not have existed earlier than 266±238 Ma
(million years) ago.[100] Based on the discovery of the hitherto
earliest fossil diatoms (dated to the early Jurassic Period,
185 Ma ago),[101] the rRNA clock developed for this purpose
shows that the diatoms must have originated between
266±238 Ma and 185 Ma ago. In spite of this, it is still an
open question whether or not these unicellular diatoms could
have originated during the Cambrian Explosion of biomineralization[4a, 11, 12, 14] (Section 1) between about 525±510 Ma ago,
that is, much earlier than the Permo-Triassic boundary of
around 250 Ma ago. According to this timescale, about 96 %
of all marine species, 70 % of all terrestrial vertebrates, and
many main groups of terrestrial plants were, amongst other
factors, destroyed by the most violent period of continental
volcanic activity in the Earth©s history.[100, 102] Thus, unlike the
coccolithophores,[78] no fossil diatoms have yet been found
which survived the Permian±Triassic Periods mass extinction
to allow a new age estimation (Section 4.1).
629
Reviews
5.2. The Structure of Diatoms
Diatoms can be readily recognized by their characteristic
ornamented cell walls, which are made of amorphous polysilicic acid, opal. Such a silica cell is also known as the frustule.
Since the morphologies are species specific, a genetic control
of this biogenic structure formation is very likely.
The architecture of the diatoms (class Bacillariophyceae)
allows this class of algae to be subdivided into two orders.[98]
Thus the silica shells of Pennales are extended and bilaterally
symmetrical (lancet-shaped or elliptical). Cylindrotheca fusiformis is shown here as an example which for about the last
ten years has become established as a model organism for
molecular biological investigations of biomineralization processes (Figure 18 A).[103, 104] The silica shells of Centrales on the
other hand are often radially symmetrical, but with many
deviations. An almost perfect example of central diatoms is
Cosconodiscus granii (Figure 18 B).
Common to all diatom cell walls (frustule) is that they
consist of the hypotheca, the box, and the epitheca, the lid
(Figure 19). The epitheca for its part consists of two parts, the
Figure 18. Morphologies of diatom cell walls: a) TEM image of Cylindrotheca fusiformis. The length of the bar represents 2.5 mm. b) SEM
image of Coscinodiscus granii. The length of the bar represents 50 mm.
(From reference [109] with kind permission of M. Sumper.)
Figure 19. Structure of the diatom cell wall (from reference [109] with
kind permission of M. Sumper.)
630
E. B‰uerlein
flat upper part, called the epivalve, the upper box and a ringshaped side wall, which consists one or more girdle bands.
This bipartition applies also to the hypotheca, in this case the
hypovalve and the lower girdle bands with the difference that
the girdle bands of the lid overlap those of the box.[98] In this
way the protoplast is completely enclosed and protected. This
concept is strengthened in that organic layers probably join
the silica shells (valves) and the girdle bands exactly. In
addition, all silicified components are protected by a casing of
organic material, probably for protection against desilification.[105]
As was recently shown by atomic force microscopy
(AFM) on living, hydrated Pinnularia viridis (Nitzsch),[106]
their cell walls (frustules) are coated with a thick mucilaginous material that is only interrupted in the vicinity of the
raphe fissure (Figure 20 A).[107] (The raphes are probably
responsible for the rapid, gliding motion of diatoms; maximum of about 20 mm s1).[108]
Figure 20. Images of Pinnularia sp. obtained by field emission scanning
electron microscopy (FESEM): A) Cross-section of fractured valve with
a view of the surface that faces the concealed cytoplasm. B, C) The enclosed chambers (stars) result from a cleavage along the AA’ axis,
which is labeled in (A). D) High enlargement of the region framed in
(A) shows tiny pores organized in precise arrays. The lengths of the
bars represent: A) 5 mm, B±D) 500 nm. (From reference [105c] with
kind permission of R. Wetherbee)
Since the cells are constantly in contact with their
environment, the shells (valves) and to a lesser extent the
girdle bands are equipped with numerous holes such as pores,
slits, etc. which allow the necessary exchange reactions. At the
same time these holes form the fascinating ornamentation of
the diatoms (Figure 20 B±D).
Angew. Chem. Int. Ed. 2003, 42, 614 ± 641
Biomineralization of Unicellular Organisms
5.3. Biomineralization of the Cell Wall in Silica Deposition
Vesicles (SDV)
Extensive electron microscope studies on a variety of
different diatoms of both orders, the Pennales and the
Centrales, have established the different morphological stages
of the biomineralization process,[105a] which is summarized
schematically (Figure 21).[109] Similar to the coccolithophores
(calcareous algae), whose biomineralization takes place in
specialized vesicles, the coccolith vesicles (Figure 15); special
vesicles, called silica deposition vesicles (SDV), were also
found in diatoms.[105] It has been concluded from investigations on other unicellular eukaryotic organisms (protists) that
the SDV is the common organelle for their silica (SiO2¥n H2O)
biomineralization process.[110] The ornamented silica shell
(frustule) is formed in the SDV, whose membrane, called the
silicalemma, consists of a typical lipid bilayer. The origins for
the species-specific ornamentation of the silica shells is thus
sought in the SDV, their components, and their structure.
5.3.1. Silicic Acid Transport and Cell Wall Synthesis
Angewandte
Chemie
In the cytoplasm the orthosilicic acid must find the path to
its deposition for the construction of the hypotheca in the
SDV under clearly stringent structural conditions. These
conditions differ from those in magnetic bacteria and
coccolithophores as the new silica wall is first formed after
cytokinesis (division of the cytoplasm) and before cell
division (Figure 21). The daughter cells thus remain in close
proximity in the region in which the two silica deposition
vesicles (SDVs) form the new hypovalves (Figure 22), a
region that is thus shielded from the extracellular space. For
that reason the orthosilicic acid can only be transported
through the cytoplasmic membrane when the latter is still
connected with the surroundings. Since many diatoms form
their silica walls within one to two hours,[113] large amounts of
orthosilicic acid must be transported through the cytoplasm
without being polymerized.
The orthosilicic acid uptake of the diatom cells follows
Michaelis±Menten kinetics with the kinetic parameters KM ¼
0.2±7.7 mm and Vmax ¼ 1.2±950 fmol Si cell1 h1.[114] The transport of Si(OH)4 in marine diatoms is coupled with sodium
ions in the form of a sodium/orthosilicic acid symporter.[103, 111]
Transport proteins of the cytoplasmic membrane for the
formation of the silica walls in diatoms, its silicic acid
transporters (SITs), have already been characterized and
sequenced.[103, 111, 112] Thus the first proteins that specifically
bind silicon, whose predominant form in sea water is as
undissociated orthosilicic acid [Si(OH)]4, and transport it into
the cells have been described. This transport is coupled with
and controlled by the incorporation of orthosilicic acid into
the cell wall,[112] the frustule, a relationship also found
between the transport of FeIII ions and the formation of
magnetite crystals in magnetic bacteria (Figure 9 A, B).
Figure 21. Diatom cell cycle: 1) cytokinesis (division of the cytoplasm)
and formation of an SDV for a valve (silica shell) in each daughter protoplast; 2) and 3) expansion of the SDV and formation of a new hypovalve within each SDV; 4) exocytosis of the contents (hypovalves of
each SDV); 5) separation of daughter cells; 6) formation of the first
girdle band SDV; 7) consecutive formation and secretion of girdle
bands; 8) DNA replication. (From reference [109] with kind permission
of M. Sumper.)
Angew. Chem. Int. Ed. 2003, 42, 614 ± 641
Figure 22. Features of silicic acid uptake, its intracellular transport,
and its deposition in a dividing cell. The scheme shows these three activities after separation of the cytoplasm (cytokinesis) into two daughter cells, which are still located in the silica shell, the epi- and hypotheka, of the mother cell during the synthesis of their silicic acid walls.
The outer black lines represent the silica wall of the mother cell. The
plasma membrane of each of the two daughter cells forms the respective large rectangle with shaded borders. Other components are labeled. Black dots represent silicic acid. The lower daughter cell illustrates three main components of transport into the cell and polysilicic
acid formation: the silicic acid transporter, the soluble silicon pool,
and SDVs. The upper daughter cell illustrates three possible means of
intracellular transport into the SDVs: direct transport by intracellularly
localized silicic acid transporting proteins, ionophore-mediated transport by silicate ionophoretic activities, and transport by silicon transport vesicles (STVs). (From reference [112a] with kind permission of
M. Hildebrand.)
631
Reviews
If the orthosilicic uptake is investigated in diatom cells that
have been cultured for an extended period under substrate
deficiency and their intracellular stores have then been
essentially emptied, after addition to the medium Si(OH)4 is
taken up at maximum rate with replenishment of the stores.
This massive increase in orthosilicic acid transport occurs in a
manner similar to that when DNA synthesis starts.[115] On this
basis cDNA libraries were set up from C. fusiformis, and five
cDNAs with identical common sequences were identified and
sequenced (SIT 1±5).[112] As no significant homologies for the
derived proteins were found in either prokaryotes or eukaryotes, SIT 1±5 represent a new class of transporters,[112a]
although they have a signature sequence (AX3LX3GR) for
sodium symporters at amino acids 216±225 (Figure 23).[116]
The transmembrane domain is highly conserved with 87±99 %
amino acid identity, as is seen from a comparison of the five
SITs,[112a] and is probably the region in which the undissociated orthosilicic acid is transported through the membrane.
As the C-terminal domains in other transporters probably
control their activity,[117] it is assumed that the five SITs
display respectively different activities in different regions of
the cytoplasmic membrane (Figure 23.[111b] In the intracellular
stores, which depending on the conditions are filled by these
orthosilicic acid transporters, up to 50 % of the total silicon is
present as soluble orthosilicic acid.[118] Not only do the
concentrations vary from 19±340 mm soluble orthosilicic
acid,[112] but they are maintained far above the saturation
limit of the respective solubility for orthosilicic acid.[119]
Two mechanisms have been proposed for this observation:First, on the basis of investigations in which it was
possible to precipitate 80 % of the soluble orthosilicic acid
with trichloroacetic acid, it was suggested that organic
orthosilicic acid binding components inhibited its polymer-
E. B‰uerlein
ization.[120] The capacity of these components can also
determine the extent of the silicic acid stores. It was possible
to extract the silicic acid ionophores form Nitzschia alba with
organic solvents, but they could not be characterized. These
ionophore activities were induced sixfold in cells that had
been cultivated under orthosilicic acid deficiency. This
indication of their direct relationship with orthosilicic acid
metabolism has led to the prediction that they serve to
maintain high intracellular concentrations of unpolymerized
orthosilicic acid.[121] Thus catechol and its derivatives, for
example, catecholamines such as epinephrine, bind and
dissolve orthosilicic acid rapidly. This class of compounds is
synthesized by bacteria in the form of the catechol siderophores; during iron deficiency, these compounds capture
extracellular FeIII under aerobic conditions or dissolve FeIII
from its oxides. In these catechol siderophores, 2,3-dihydroxybenzoic acid groups are bound to 1,4-diaminobutane (putrescine) or N-(3-aminopropyl)-1,4-diaminobutane (spermidine)–two amines that play an important role in silaffins,
which were isolated from the cell walls of diatoms and will be
described later (Section 5.3.3).[122] However, during the biomineralization of magnetite by magnetic bacteria an ionophore is secreted that holds the available FeIII in solution, is
not a siderophore, and stimulates iron uptake.[63, 65] The
chemical nature of this compound is also unknown (Section 3.4).
Second, as an alternative mechanism it has been suggested
that intracellular transport of orthosilicic acid takes place
through small vesicles in which it is stored, either solubilized
or undissolved.[123±125] These vesicles have been named silicon
transport vesicles (STVs) by Schmid and Schulz.[123] They
were found during several electron microscope studies in the
region of active silica deposition vesicles (SDVs).[123, 125, 126]
These STVs correspond to the highly characterized, small
transport vesicles of coccolithophores (Section 4.5.1). The
size of the STVs corresponds to that of the polysilicic acid
particles, which attach themselves to the growing silica shells.
This indicates that these particles are used as building blocks.
Their size is probably determined by polyamine-modified
polypeptides, called silaffins, or by their polyamines alone
(Section 5.3.3). Such polysilicic acid particles were already
found by AFM in Pinnularia viridis with remarkably low
variation in their size: 44.8 0.7 nm in the valves and 40.3 0.8 nm in the girdle bands.[106] Whether orthosilicic acid is
contained in STVs has not yet been investigated by X-ray
microanalysis.[127]
5.3.2. Silica Deposition Vesicles
Figure 23. Topological model of the silicic acid transporters (SITs) from
C. fusiformis, based on SIT4, in a lipid bilayer membrane. Bottom ¼ intracellular. IL ¼ intracellular loops, EL ¼ extracellular loops,
INS ¼ intracellular amino segment, ICS ¼ intracellular carboxy segment,
CC ¼ coiled coil, (þ) and () orders the charged amino acids,
C ¼ cystein residue. (From reference [112a] with kind permission of M.
Hildebrandt.)
632
It was recognized quite early that morphogenesis of
structures containing silicic acid takes place in silica deposition vesicles not only in diatoms but also in other unicellular
eukaryotic organisms.[105a, 110, 128] In spite of this their formation
in diatoms is still a matter of debate, mainly because of the
plethora of different vesicles. However, there are numerous
indications that the SDVs originate in the Golgi apparatus,
both in the central diatoms[123, 126, 129] and in the pennate
diatoms.[130] The analogy to the coccolith vesicles (Section 4)
supports these findings (Figure 15).
Angew. Chem. Int. Ed. 2003, 42, 614 ± 641
Biomineralization of Unicellular Organisms
A methodologically important step was made with the
introduction of the cationic, lipophilic fluorescent dye rhodamin 123 (R123), which depending on the electrochemical
potential accumulates in the SDVs and there indicates an
acidic pH value.[131] R123 is also incorporated into actively
polymerizing polysilicic acid and this process can be monitored quantitatively.[105c, 112a, 131]
The extraordinarily extensive and until now mainly
morphological studies by electron microscopy suggest that
orthosilicic acid is directed to the site of morphogenesis, the
SDVs, either by ionophore supported diffusion[121] (Figure 22)
or by transport vesicles (STVs)[123, 126] as particles of a certain
size (Section 5.2.2). However, it has thus far not been possible
to isolate the SDVs to investigate orthosilicic acid transport
through their membranes, the silicalemma, because of the
difficulty in removing the silica shell.
5.3.3. Proteins of the Diatom Cell Wall: Frustulins, Pleuralins,
Silaffins
If the cell walls of the diatoms are compared with those of
the coccolithphores, whose complex structures are made up of
either amorphous silica (opal) or crystalline calcium carbonate, it is immediately clear when searching for analogies that
many more organic components are found during work-up of
the diatom cell walls. This difference is attributed to the fact
that unlike the coccoliths, diatom cell walls possess pores and
slits from an ornamentation process, also called micromorphogenesis.[105c, 121] Whereas the three polysaccharides PS1,
PS2, and PS3 of the coccolithophore cells can be isolated most
simply by treatment with trichloroacetic acid, or even with
EDTA,[90a] the diatom cell wall must be treated with
anhydrous hydrogen fluoride (HF) to release the very firmly
bound and incorporated components.[104, 105b, 132, 133] Three
groups of proteins has been isolated from the cell wall of
C. fusiformis and characterized: frustulins,[104, 105b, 132] pleuralins,[133] and silaffins.[134]
Frustulins are found everywhere on the cell wall and
clearly form a protective protein coating deep into the
internal space (Section 5.3.1; see also reference [106]). The
protein coating can be completely destroyed with EDTA, a
complexing agent for calcium ions.[104, 105b, 132] The frustulins
resemble the polysaccharides PS1 and PS2 in both properties;
the latter also cover the calcite crystals of the coccoliths as a
protective coat and can be released with EDTA (Section 4.5.1). Furthermore, the frustulins in C. fusiformis are a
family of four, essentially glycosylated proteins, which are
found together with apparent molecular weights of 75±
200 kDa.[104, 105b, 132] The frustulins are conspicuous for,
amongst other properties, five to six repeated domains that
are rich in acids and cystein.
Pleuralins, the second family of proteins of diatom cell
walls from C. fusiformis, could only be obtained by dissolving
the silica cell wall with hydrogen fluoride, which is why they
were originally called HEPs (HF-extractable proteins).[133]
Three pleuralins, called pleuralin 1 (200 kDA), pleuralin 2
(180 kDa), and pleuralin 3 (150 kDa), have been isolated and
sequenced (Figure 24 A). Like the frustulins, these proteins
Angew. Chem. Int. Ed. 2003, 42, 614 ± 641
Angewandte
Chemie
Figure 24. A) Schematic primary structures of the three pleuralins.
b) Immunogold labeling of a thin section of C. fusiformis. The TEM image shows the overlapping region of the girdle bands (framed). The
black arrow indicates the plasma membrane and points towards the
extracellular space. The rows of oval structural elements outside the
protoplasts represent the girdle bands, from the left of the hypotheka
and from the right of the epitheka. An anti-pleuralin 1 antibody was
used for indirect immunogold labeling. The length of the bar represents 200 nm. (Adapted from reference [109, 133b] with kind permission of M. Sumper and N. Krˆger.)
contain repeated special domains, in this case the PSC
domains (proline-serine-cystein domains).
The important discovery of the (hitherto) only signal
sequence of proteins of biomineralization (the pleuralins),
which is important for an entry into the endoplasmic
reticulum, points to a relationship or analogy between
biomineralization in eukaryotes and the secretory transport
pathways of proteins, exocytosis. This analogy is extensively
supported by the biomineralization of the coccoliths of
Pleurochrysis carterae, in which almost all processes previously described take place in the Golgi system (Figure 15).
However, no proteins were detected; without exception, acid
polysaccharides have hitherto been found.
With immunogold labeling, all pleuralins have been
identified at a specific site of the diatom cell wall that
corresponds to their cell-cycle-dependent development. With
this method it was first found that the protein composition of
the two theka, the epitheka and the hypotheka, are clearly
different. During cell division the pleuralins bind to the newly
formed girdles of the hypotheka, whereas previously they
were still found on those of the epitheka (Figure 24 B). This
process runs parallel to the simultaneous conversion of the
former hypotheka into the epitheka of the daughter cell[133]
(Figure 21). Clearly, this development-controlled cooperation
of the pleuralins with the diatom cell wall is involved in the
hypotheka±epitheka differentiation.
The silaffins make up the third family of proteins of the
diatom cell wall in C. fusiformis. Like the pleuralins, they
could also only be found after dissolution with anhydrous
hydrogen fluoride. They are three polypeptides of low
molecular weight: silaffin 1A (4 kDA), silaffin 1B (8 kDa),
and silaffin 2 (17 kDa). With the help of the N-terminal amino
acid sequence of silaffine 1B the corresponding sil1 gene was
633
Reviews
E. B‰uerlein
polyamines, which consist of 5±10 methylated propylamine
groups, and by e-N,N-dimethylation of the other two lysine
residues. In the silaffin 1A1 isoform, one of these dimethyllysines is replaced by a d-hydroxy-e-N,N,N-trimethyllysine
(Figure 25 B).[134, 136]
The unique property of the silaffins to form polysilicic
acid in vitro suggests that they probably play a pivotal role in
the construction of the silica cell wall. Whereas in the absence
of silaffins a solution of orthosilicic acid at an end concentration of 1m (freshly prepared from tetramethyl orthosilicate
in 1 mm hydrochloric acid) is stable for at least a few hours,
each silaffin as well as the mixture of silaffins can precipitate
polysilicic acid from this orthosilicic acid solution within
seconds. Although the amount of polysilicic acid is essentially
proportional to the amount of individual silaffins or their
mixture (Figure 26 A), spherical particles of very different
diameters are formed. Thus silaffin 1A induces the formation
of particles with diameters of 500±700 nm (Figure 26 B),
whereas the natural mixture of silaffin 1A, silaffin B, and
silaffin 2 produces much smaller particles (diameter < 50 nm).
This relationship between different silaffins and particle size,
which can also be of considerable interest to materials
science, has led to even more detailed investigations. Recently, the complete chemical structures of silaffin 1A1 and
silaffin 1A2 were elucidated: they are made up of 15 and
18 amino acid residues, respectively (Figure 25 B).[134] In the
modifications of their respective four lysine units, they differ
only in the d-hydroxy-e-N,N,N-trimethyl residue; the quaternary ammonium ion has hitherto been found only in
silaffin 1A1.[134, 136] However, this difference has practically
Figure 25. Formation of the silaffin 1A isoforms: A) Schematic structure of the precursor of all silaffins, the polypeptide sil1p. The black
pentagons show the repeating R1±R7 elements from which all silaffins
are formed. The amino acid sequences of the elements R2 and R3±R7
are shown in parenthesis. Silaffin B is formed from element R1. The
white bar represents the signal sequence, the gray ellipse the highly
acidic prosequence.[134b] B) Chemical structure of silaffin 1A1 and silaffin 1A2. The chemical structures of the side chains are only illustrated
for the modified lysine residues. (From reference [134b] with kind permission of M. Sumper and N. Krˆger and the American Society for Biochemistry and Molecular Biology.)
cloned. As is the case with the frustulins[104, 105b, 132] and the
pleuralins (Figure 24 A), the protein sil1p coded there shows a
series of repeating sequences (Figure 25 A), which are termed
R1±R7. Unlike the frustulins and the pleuralins, sil1p is not
used in the cell in this repetitive structure, but is cleave
proteolytically. The R1±R7 units are released as single
peptides, R1 corresponds to the precursor of sillafin 1B
whereas R2±R7 correspond to the precursors of the isoforms
of silaffin 1A (Figure 25 A).[134]
The active peptides, the silaffins, are formed in the cell by
a unique posttranslational modification of four lysine residues
of their precursors, which are present as two pairs or as one
pair and two individual residues (Figure 25 A). This modification is an alkylation of each of two lysine units by linear
634
Figure 26. Silaffin-induced silica precipitation: A) Correlation between
the silaffin concentrations (mg protein in 0.1 m silica solution) and the
amount of polysilicic acid (nmol) that is precipitated from the silicic
acid solution. Dotted line: silaffin mixture. Solid line: pure silaffin 1A.
B, C) SEM image of silica precipitated (B) by silaffin 1A and (C) by the
natural silaffin mixture. The diameter of the particles is 500±700 nm
(B) and < 50 nm (C). The length of the bar represents 1 mm. Modified
from reference [109] with kind permission of M. Sumper and N. Krˆger.)
Angew. Chem. Int. Ed. 2003, 42, 614 ± 641
Biomineralization of Unicellular Organisms
no influence on the amount of polysilicic acid precipitated
over the pH range of 5±8, nor on the diameter of the spherical
particles produced.[134]
Moreover, since the unmodified peptide pR5 cannot
induce polysilicic acid precipitation in an acid solution with a
pH value lower than 6,[134] but polysilicic acid deposition in
vitro takes place in the acid plasma of the SDVs, the
polyamine side chains clearly play a crucial role in the
biological system.
5.3.4. Species-Specific Polyamines, Structure, and In Vitro Production of Polysilicic Acid Nanoparticles
Until now, almost all investigations have been carried out
on C. fusformis, the model organism for diatom research.
Since the formation of valves with their pores, slits, and other
openings in the SDVs (Figures 19, 20 A±C) leads to speciesspecific structures, it is probable that each species has
different silaffins or silaffin-like molecules.
To test this hypothesis, organic cell-wall components of six
different diatoms from a broad range of species were
isolated.[135a] In a special SDS-PAGE, which also has the
ability to separate compounds of low molecular weight,[137]
the assumption has already been confirmed that, in comparison to C. fusiformis, each of the five other species has a
respectively specific silaffin composition.[135a] In addition,
compounds were also discovered whose molecular weights
are smaller than 3.5 kDA and, with the exception of
C. fusiformis, appear to be either one of or the main
components of all other diatoms. After purification and mass
spectrometric analysis, a different number of polyamines
were found for each type, in each case with different
molecular weights. All compounds are linear polyamines
composed of N-methylaminopropyl or aminopropyl units,
that is, they exhibit a different degree of methylation. As in
the polyamine synthesis of bacteria and eukaryotes,[135b]
ornithine (2,5-diaminopentanoic acid) and its decarboxylation product putrescine (1,4-diaminobutane) form the basis of
the alkylation with aminopropyl residues. These two residues
as well as the amino groups of the terminal aminopropyl
residue can be methylated differently. A simplified polyamine
structure is shown schematically. The aminopropyl groups and
ornithine are not added here, and putescine is shown within a
frame. (Reproduced with kind permission of M. Sumper and
N. Krˆger.)
The assumption that in addition to the silaffins these
polyamines are also major components of biological polysilicic acid formation can be also confirmed by in vitro
experiments. The largest number of polyamines has been
isolated from Nitzschia angularis.[135a] Their molecular weights
vary between 600±1250 Da. They were investigated with
respect to both the molecular weight and to their ability to
Angew. Chem. Int. Ed. 2003, 42, 614 ± 641
Angewandte
Chemie
precipitate polysilicic acid at different pH values. Thus
spherical polysilicic acid particles of mainly 0.8±1 mm were
produced with polyamines of 1000±1250 Da at pH 5 (Figure 27 A). With polyamines of 600-700 Da, spherical polysilicic acid particles of only 100±200 nm were found (Figure 27 B). A close relationship between the morphology of
polysilicic acid particles and pH emerges when the pH value
of the orthosilicic acid is increased in the presence of the
natural polyamine mixture from N. angularis. The diameter
falls from 700 nm to 50 nm with increasing pH value
(Figure 27 C±F). An additional potential for structure formation is presented by three silaffins from N. angularis, as can be
deduced from the formation of blocklike structures In
vitro,[135] which have also been found in vivo as areolae
(honeycomb) walls.[138] When the silaffins are combined with
the polyamines from N. angularis, these blocklike structures
vary in that a flat side with oblate polysilicic acid particles is
formed.[135]
Recently it was possible to isolate a SP41 protein (SP41 ¼
scale protein 41 kDa) from the silica-containing platelets and
spines of Mallomonas slendens, a photosynthetic Chrysophycee (gold±brown alga), which is closely related to the diatoms.
With the help of polyclonal anti-SP41 antibodies it was
possible to show for the first time that during the biogenesis of
platelets and spines SP41 is deposited together with polysilicic
acid in special SDVs.[139] Thus a whole arsenal of compounds is
available to investigate the relationship between polysilicic
acid polymerization, particle size, and structure formation,
and ornamentation.
Figure 27. Silica precipitates induced with polyamines from Nitzschia
angularis. Polyamines of molecular weight MW 1000±1250 Da (A) and
600±750 Da (B) were used. The natural mixture of polyamines
(MW ¼ 600±1250 Da) was used to investigate the pH dependence of
the polyamine-induced silica precipitation: (C) pH 5.4, (D) pH 6.3, (E)
pH 7.2, (F) pH 8.3. The polyamine concentration in each solution was
0.85 mg ml1. The length of the bars represents 1 mm in (A) and (B),
500 nm in (C), (D), (E), and (F). (From reference [135a] with kind permission of M. Sumper and N. Krˆger and the National Academy of
Science USA.)
635
Reviews
Based on the structural investigations,[105a, 138] the isolation
of organic compounds from the silica cell walls,[109, 135, 139] and
the chemical syntheses for the preparation of inorganic
materials from polysilicic acid with defined porosity–with
the use of organic substrates with defined shape[140±144c]–a
series of model concepts were proposed.[105a, 145] These revolve
around either preformed structures of the mould-prepattern
hypothesis[145e, f] or the phenomena of phase separation.[146]
Recently Sumper[146a] outlined a simple phase-separation
model for the formation of honeycomblike structures (areolae) frequently found in diatoms and also formed by chemical
synthesis.[144b, c] His model describes repeated phase separation in the SDVs through which emulsions of micro- and
subsequently nanodroplets or micelles are produced as
centers of polysilicic acid deposition. Volkmer et al. used an
in situ video microscope to monitor time-resolved phase
separation during the formation of radial silica shells. [146b]
Lipophilic, detergent-stabilized drops with limited amounts of
a mixed titanium silicon oxide developed the ability to form
and shape an inorganic matrix in a dynamic dissipation
process, that is, in a spontaneous formation of increasingly
smaller drops. However, thus far only spinelike shapes have
been found (honeycomb shapes were not observed). The
choice of mixture of anionic and cationic surfactants is critical
for the success of such investigations. Moreover, van Blaaderen et al were able to show that colloidal crystals, threedimensional periodic structures from small, suspended colloidal particles, can also develop into particles whose sizes are
dependent on the periodic distance of the hollow sections in a
polymer layer.[147a] This process was named colloidal epitaxis.[147b]
The importance and utility of silicate compounds in
biological systems has been discussed for quite some time.[148a]
The question has been raised as to whether the biomineralization of silicic acid offers any ecological advantages. Several
proposals have been put forward about the function of the
silicic acid cell wall of the diatoms, for example, as a UV filter
or as a form of protection against zooplankton (eukaryotic
unicellular organisms that do not contain chlorophyll).
However, none of these hypotheses have been verified
experimentally. On the contrary, Milligan and Morel have
shown that the polysilicic acid plays an active role as a pH
buffer in the cell walls of diatoms.[148b] Diatoms have
extracellular carboanhydrase on their surfaces that catalyzes
the reaction between HCO3 and CO2. The high catalytic rate
of this enzyme allows it to act as a pH buffer by either
providing or rapidly binding protons. The buffer action of
polysilicic acid could be verified in studies on cells of
Thalassiosira weissflogii or in the presence of isolated
polysilicic cell walls with isolated soluble carboanhydrase
(bovine): the polysilicic acid accelerates the reaction of the
enzyme. It has been postulated that the increased release of
CO2 by the extracellular carboanhydrase is part of the carbonconcentrating (CCM). This results in the increase of the
concentration of CO2 in the vicinity of the most important
carboxylase, the ribulose-1,5-diphosphate carboxylase oxygenase (RubisCO), and thus in an increase in the rate of
photosynthesis, even though the concentration of CO2 in sea
water is only 1/100 that required for the full activity of the
636
E. B‰uerlein
enzyme. In this way, the global cycle of silica (essentially the
biomineralization of polysilicic acid) can be coupled to that of
carbon (that is, photosynthesis) in diatoms. This could be
proven directly in the diatomaceous Phaeodactylum tricornutum, which can be isolated as both silicified and nonsilicified
morphotypes. The silicified morphotype preferentially fixes
carbon.
6. Summary and Outlook
A great deal of new information on biomineralization is
currently being obtained rapidly by the methods of molecular
biology. What for a long time appeared to be extremely
difficult to achieve (an understanding of the mechanism of
biomineralization at least in stages) has already been
achieved by Marsh et al. with the unicellular coccolithophores
(calcareous algae). A type of multicomponent complex was
revealed in which the fixation of a first crystallite edge is
associated with an accumulation of 25-nm-large PS1/PS2±Ca
particles which release the polyanion-bound calcium ions in
high concentrations. Here the first crystals are produced as
small rectangles. The acid polysaccharide PS3, which is
associated both with the coccolith vesicle membrane and
with the above crystals, then induces the growth of the
respective protocrystal to the unusual anvil-shaped crystals of
Pleurochrysis (Figure 12).[96]
Unfortunately as yet no coccolithophore genome has been
sequenced. In contrast, two genome sequences have been
known for the magnetic bacteria since April 2001. This has
contributed to the observation that according to the extensive
work of Sch¸ler and co-workers on proteins of the magnetosome membrane, not only are these proteins found in two
gene clusters, but in addition a series of protein families has
now been found there in a similar arrangement. (Figure 8).
Thus, it is now possible to investigate the mechanism of
magnetite crystal formation in the magnetosome vesicles by
switching off individual genes. Furthermore, the accompanying empty vesicles could now be isolated and investigated.
The magnetosome vesicles have also led to a comprehensive and discussion-stimulating hypothesis of biomineralization by Kirschvink and co-workers in which they propose
evolution of the genes of bacterial magnetite biomineralization through doubling, mutation, adaptation, and variation up
to bone formation.[14] However, the work of Sch¸ler has
demonstrated that until now all genes of proteins from the
magnetosome membrane occur solely in magnetic bacteria.
Moreover, no homologues has been found in other organisms.[55]
The analogy of coccolith formation to the transport of
proteins through the endoplasmic reticulum (ER) and Golgi
apparatus to exocytosis has in this article led to the working
hypothesis that the magnetosome vesicles, as perhaps the first
intracytoplasmic vesicles in bacteria, could be the prokaryotic
precursors of the endomembrane system of the ER and the
Golgi apparatus. The formation of the magnetosome membrane is unknown, endocytosis stands against de novo synthesis. The absence of the endomembranes in early eukaryotes, the diplomonads, is conspicuous.[149] Endosymbiotic
Angew. Chem. Int. Ed. 2003, 42, 614 ± 641
Angewandte
Chemie
Biomineralization of Unicellular Organisms
events of such early protists with a magnetic bacterium could
have led to the formation of the endomembranes.
The previously described parts of biomineralization
machineries are clearly too complicated when isolated in full
to be used in vitro for production. As already shown by
polyamines and silaffins (Section 5), the use of parts is of
considerable interest for the chemist, but only methodological
simplification and variation of these compounds can open up
new areas. Probably the first example was shown by Stone and
co-workers[150] who used the unmodified polycationic precursor peptide pR5 of silaffin 1A1 that does not precipitate
polysilicic acid at physiological acidic pH, but does so at
neutral pH. This peptide was incorporated in a mixture of a
hydrophilic and a lipophilic acrylate by photopolymerization:
a hologram with a periodicity of 1.33 m was imprinted into the
polymer by means of a holographic process (Figure 28 A, C).
The peptide is clearly fixed in the H2O-enriched depression as
by slow addition of 0.1m [Si(OH)4] a two-dimensional
arrangement of polysilicic acid particles with a mean diameter
of 452 nm ( 81 nm) formed in the periodicity of the hologram (Figure 28 B, D).
In a fascinating analogy to studies on the surfaces of
semiconductors and metals[151] Stone and co-workers[152]
introduced a ™combinatorial phage display library∫ to biomimetic material science in order to discover peptides that
precipitate polysilicic acid. Thus a very large number of
peptides were expressed from twelve amino acids as fusion
proteins with a phage coat protein. The peptide with the
highest yield of spherical particles contains a serine residue
and five histidine units. As with the silaffins, the amount of
silicic acid was proportional to the amount of peptide phage
particles (Figure 25),[152] so that such polysilicic acid particles
are a mixture of organic and inorganic compounds.
A different and novel method to produce spherical
polysilicic acids particles is the previously described experiment of Stone and co-workers.[150] Only parts of the pR5
peptide, which probably catalyze the precipitation of the
polysilicic acid, extend out of the hydrophilic part of the
polymer. The resulting particles contain only small amounts
of organic substances and are not a composite. The comparison of both methods may contribute to a clarification of the
question of what determines the diameter of the particles and
how a preparation of more narrow diameter variations may
be obtained.
Addendum
Important progress in biosilica morphogenesis was published recently by Sumper and co-workers.[146c] Ammonium
fluoride was used instead of anhydrous hydrogen fluoride to
extract silaffins (Section 5.3.3.) by dissolving diatom biosilica
under mild conditions. The apparently native silaffins were
obtained, all of whose serine residues were found to be
phosphorylated. In silaffin molecules, therefore, not only are
many positive charges accumulated by the modification with
polyamines, but also many negative charges by the abovementioned phosphoryl groups. These zwitterionic molecules
are able to self-assemble, probably forming a template that
promotes silica precipitation in large units.
Figure 28. Two-dimensional array of ordered silica nanospheres
formed within a hologram (lattice) that had previously been produced
by a holographic process in an exceptionally clear and colorless polymer: A) SEM images of the control hologram after being treated with
dissolved silicic acid; B) the polysilicic acid nanostructure formed by
reaction of polysilicic acid with the peptide-modified hologram;
C) AFM images of the control polymer before and d) the hybrid structure after treatment with silicic acid. (Modified from reference [150]
with kind permission of M. O. Stone and Nature.)
Angew. Chem. Int. Ed. 2003, 42, 614 ± 641
I thank Dr. Dirk Sch¸ler (Bremen, Germany), Professor
Mary E. Marsh (Houston, TX), as well as Dr. Susan Glasauer
and Professor T. J. Beveridge (Guelph, Canada) who shortly
before the conclusion of this article provided me with preprints
of their highly relevant publications. I have received critical
help from my colleagues Professor Lars-Oliver Essen (Marburg, Germany), Professor Nicolai Petersen (Munich, Germany), Dr. Dirk Sch¸ler (Bremen, Germany), and Dr. Michael
Winkelhofer (Southampton, UK), who kindly reviewed the
manuscript. I thank wholeheartedly Professor Dieter Oesterhelt, who in my active retirement gave me the opportunity as a
guest in his department to be able to dedicate myself to this
article and more. This article would not have been completed
without the inestimable support of my wife Cornelia and my
son Felix, who as computer specialists made almost everything
possible, including the presentation of the many figures.
Received: February 18, 2002 [A519]
637
Reviews
[1] M. R. Walter in Earths Earliest Biosphere: Its Origin and
Evolution (Ed.: J. W. Schopf), Princeton University Press,
Princeton 1983, pp. 187 ± 213.
[2] a) T. J. Beveridge, in Metal Ions and Bacteria (Eds.: T. J.
Beveridge, R. J. Doyle), Wiley, New York 1989, p. 1 ± 29;
b) F. G. Ferris, W. Shotyk, W. S. Fyfe in Metal Ions and Bacteria
(Eds.: T. J. Beveridge, R. J. Doyle), Wiley, New York 1989,
p. 415 ± 442.
[3] a) K. Hausmann, B. P. Kremer, Extremophile, Mikroorganismen in ausgefallenen Lebensr‰umen, VCH, Weinheim, 1995;
b) K. O. Stetter, FEMS Microbiol. Rev. 1990, 75, 117 ± 124.
[4] a) H. A. Lowenstam, Science 1981, 211, 1126 ± 1131; b) D. A.
Bazylinski, R. B. Frankel in Biomineralization (Ed.: E. Baeuerlein), Wiley-VCH, Weinheim 2000, p. 41 ± 43; c) B. Devouard,
M. PÛsfai, X. Hua, D. A. Bazylinski, R. B. Frankel, P. R. Busek,
Am. Mineral. 1998, 83, 1387 ± 1398; d) K. L. Thomas-Kerpta,
D. A. Bazylinski, J. L. Kirschvink, S. J. Clemett, D. S. Mc Kay,
S. J. Wentworth, H. Vali, E. K. Gibson, Jr., C. S. Romanek,
Geochim. Cosmochim. Acta 2000, 64, 4049 ± 4081.
[5] a) R. P. Blakemore, Science 1975, 190, 377 ± 379; b) R. B.
Frankel, R. P. Blakemore, R. S. Wolfe, Science 1979, 203,
1355 ± 1356.
[6] D. Balkwill, D. Marata, R. P. Blakemore, J. Bacteriol. 1980, 141,
1399 ± 1408.
[7] a) N. Petersen, T. von Dobeneck, H. Vali, Nature 1986, 320,
611 ± 615; b) J. F. Stolz, S.-B. R. Chang, J. L. Kirschvink, Nature
1986, 321, 849 ± 850; c) S.-B. R. Chang, J. L. Kirschvink, Annu.
Rev. Earth Planet Sci. 1989, 17, 169 ± 195;d) E. I. Friedmann, J.
Wierzchos, C. Ascaso, M. Winklhofer, Proc. Natl. Acad. Sci.
USA 2001, 98, 2176 ± 2181; e) D. Sch¸ler, Dissertation, Technische Universit‰t M¸nchen, 1994; f) D. Sch¸ler, E. Baeuerlein in
Transitions Metals in Microbial Metabolism (Eds.: G. Winkelmann, C. J. Carrano), Harwood Academic Press, Amsterdam,
1997, pp. 159 ± 185.
[8] S. W. F. Grant, Am. J. Sci. 1990, 290A, 261.
[9] A. H. Knoll, S. B. Caroll, Science 1999, 284, 2129.
[10] F. J. Ayala, A. Rzhetsky, Proc. Natl. Acad. Sci. USA 1998, 95,
606.
[11] H. A. Lowenstam, L. Margulis, Biosystems 1980, 12, 27.
[12] S. Bengtson, Y. Zhao, Science 1992, 257, 367.
[13] K. Simkiss, K. M. Wilbur, Biomineralization, Academic Press,
San Diego, 1989.
[14] J. L. Kirschvink, J. W. Hagedorn in Biomineralization (Ed.: E.
Baeuerlein), Wiley-VCH, Weinheim, 2000, p. 139 ± 149.
[15] S. J. Gould, E. S. Vrba, Paleobiology 1982, 8, 4.
[16] F. F. Torres de Araujo, M. A. Pires, R. B. Frankel, C. E. M.
Bicudo, Biophys. J. 1986, 50, 375 ± 378.
[17] S. Mann, N. H. C. Sparks, M. M. Walker, J. L. Kirschvink, J.
Exp. Biol. 1988, 140, 35 ± 49.
[18] a) M. M. Walker, C. E. Diebel, C. V. Haugh, P. M. Pankhurst,
J. C Montgomery, C. R. Green, Nature 1997, 390, 371 ± 376;
b) C. E. Diebel, R. Proksch, C. R. Green, P. Neilson, M. M.
Walker, Nature 2000, 406, 299 ± 302.
[19] a) E. Holtkamp-Rˆtzler, G. Fleissner, M. Hanzlik, N. Petersen,
Ann. Geophys. 1997, 15 (Suppl. IC), 117; b) M. Hanzlik, Ch.
Heunemann, E. Holtkamp-Rˆtzler, M. Winklhofer, N. Petersen, G. Fleissner, BioMetals 2000, 13, 325 ± 331; c) D. AcostaAvalos, E. Wajnberg, P. S. Oliveira, I. Leal, M. Farina, D. S.
Esquivel, J. Exp. Biol. 1999, 202, 2687 ± 2692.
[20] a) J. L. Kirschvink, A. Kobayashi-Kirschvink, B. J. Woodford,
Proc. Natl. Acad. Sci USA 1992, 89, 7683; b) J. Dobson, P. P.
Grassi, Brain Res. Bull. 1996, 39, 255 ± 259; c) P. P. SchultheissGrassi, R. Wessiken, J. Dobson, Biochem. Biophys. Acta 1999,
1426, 212 ± 216; d) P. P. Schultheiss-Grassi, J. Dobson, BioMetals 1999, 12, 67 ± 72.
638
E. B‰uerlein
[21] T. J. Beveridge, in Bacteria in Nature (Eds: J. S. Pointdexter,
E. R. Leadbetter), Plenum Press, New York, 1989, pp. 1 ± 65.
[22] Allgemeine Mikrobiologie, (Ed.: H. G. Schlegel) 7th ed.,
Thieme, Stuttgart 1992, p. 14.
[23] a) F. G. Ferris in Metal Ions and Bacteria (Eds.: T. J. Beveridge,
R. J. Doyle), Wiley, New York, 1989, p. 295 ± 323; b) D. Fortin,
T. J. Beveridge in Biomineralization (Ed.: E. Baeuerlein),
Wiley-VCH, Weinheim 2000, p. 7 ± 22.
[24] a) T. J. Beveridge, Int. Rev. Cytol. 1981, 72, p. 229 ± 317; b) W.
Baumeister, G. Lembcke, G. D¸rr, B. Phipps in Electron
Crystallography of Organic Molecules (Eds.: J. R. Freyer, D. L.
Dorset), Kluwer Academic, Dordrecht, 1990, p. 283 ± 296;
c) M. Engelhardt, J. Petres, J. Struct. Biol. 1998, 125, 276 ±
302; d) T. J. Beveridge, P. H. Pouwels, M. Sara, A. Kotiranta,
K. Lonnatmaa, K. Kari, E. Kerosuo, M. Haapasalo, E. M.
Egelseer, I. Schocher, U. B. Sleytr, FEMS Microbiol. Rev. 1997,
20, 99 ± 149; e) U. B. Sleytr, P. Messner, D. Plum, M. Sara,
Angew. Chem. 1999, 111, 1099 ± 1120; Angew. Chem. Int. Ed.
1999, 38, 1035 ± 1054.
[25] M. Sumper, Biochim. Biophys. Acta 1987, 906, 69 ± 79.
[26] C. J. Daughney, J. B. Fein, N. Yee, Chem. Geol. 1998, 144, 161 ±
176.
[27] M. M. Urrutia, T. J. Beveridge, Chem. Geol. 1994, 116, 261 ±
280.
[28] D. Fortin, T. J. Beveridge, Geomicrobiol. J. 1997, 14, 1 ± 21.
[29] S. Schultze-Lam, G. Harauz, T. J. Beveridge, J. Bacteriol. 1992,
174, 7971 ± 7981.
[30] S. Mann in Biomimetic Materials Chemistry, VCH, New York
1996, p. 1 ± 40.
[31] a) M. Mertig, R. Kirsch, W. Pompe, H. Engelhardt, Eur. Phys. J.
D 1999, 9, 45 ± 48; b) S. R. Hall, W. Shenton, H. Engelhardt, S.
Mann, ChemPhysChem. 2001, 3, 184 ± 186.
[32] G. Southam, T. J. Beveridge, Geochim. Cosmochim. Acta 1994,
58, 4527 ± 4530.
[33] a) D. H. Nies, Appl. Microbiol. Biotechnol. 1999, 51, 730±750;
b) S. Silver, Ann. Rev. Microbiol. 1996, 50, 753 ± 789.
[34] a) H. E. Jones, P. A. Trudinger, L. A. Chambers, N. A. Pyliotis,
Z. Allg. Microbiol. 1976, 16, 425 ± 435; b) M. Vainshtein, N.
Suzina, V. Sorokin, Syst. Appl. Microbiol. 1997, 20, 182 ± 186.
[35] a) S. Glasauer, S. Langley, T. J. Beveridge, Science, 2001, 295,
117 ± 119; b) U. Schwertmann, R. M. Cornell, Iron Oxides in the
Laboratory, VCH, Weinheim, 1991.
[36] a) R. B. Frankel, R. B. Blakemore, Iron Biominerals, Plenum
Press, New York, 1991; b) S. Mann, R. B. Frankel in Biomineralization, Chemical and Biochemical Perspectives (Eds.: S.
Mann, J. Webb, R. J. P. Williams), VCH, Weinheim 1989,
pp. 389 ± 426.
[37] S. Spring, R. Amann, W. Ludwig, K.-H. Schleifer, H. van Gemerden, N. Petersen, Appl. Environ. Microbiol. 1993, 59, 2397 ±
2403.
[38] D. A. Bazylinski, R. B. Frankel, B. R. Heywood, S. Mann, J. W.
King, P. L. Donaghay, A. K. Hanson, Appl. Environ. Microbiol.
1995, 61, 3232 ± 3239.
[39] H. Petermann, Berichte Fachbereich Geowissenschaften Univ.
Bremen, Nr. 56, Bremen, 1994; Fahrten M16/2, M20/2, M23/1,
M23/2 des Forschungsschiffes Meteor.
[40] a) M. Farina, H. Lins des Barros, D. M. S. Esquival, J. Danon,
Biol. Cell. 1983, 48, 85 ± 88; b) E. F. De Long, R. B. Frankel,
D. A. Bazylinski, Science 1993, 259, 803 ± 806.
[41] K. H. Schleifer, D. Sch¸ler, S. Spring, M. Weizenegger, R.
Amann, W. Ludwig, M. Kˆhler, Syst. Appl. Microbiol. 1991, 14,
379 ± 385.
[42] R. P. Blakemore, D. Maratea, R. Wolfe, J. Bacteriol. 1979, 140,
720 ± 729.
[43] D. Sch¸ler, S. Spring, D. A. Bazylinski, Syst. Appl. Microbiol.
1999, 22, 466 ± 471.
[44] C. R. Woese, Microbiol. Rev. 1987, 51, 221 ± 271.
Angew. Chem. Int. Ed. 2003, 42, 614 ± 641
Biomineralization of Unicellular Organisms
[45] a) R. Amann, R. Rosello-Mora, D. Sch¸ler in Biomineralization (Ed.: E. Baeuerlein), Wiley-VCH, Weinheim 2000, p. 47 ±
59; b) O. Strunk, O. Gross, B. Reichel, M. May, S. Hermann, N.
Stuckmann, B. Nonhoff, M. Lenke, A. Ginhart, A. Vilbig, T.
Ludwig, A. Bode, K. H. Schleifer, W. Ludwig, http://www.mikro.biologie.tu-muenchen.de/pub/ARB, Department of Microbiology, Technische Universit‰t M¸nchen, Germany, 1999;
c) W. Ludwig, O. Strunk, S. Klugbauer, N. Klugbauer, M.
Weizenegger, J. Neumaier, M. Bachleitner, K. H. Schleifer,
Electrophoresis 1998, 19, 554 ± 568; d) R. Amann, W. Ludwig,
K. H. Schleifer, Microbiol. Rev. 1995, 59, 143 ± 169.
[46] D. A. Bazylinski, A. J. Garrett-Reed, R. B. Frankel, Microsc.
Res. Tech. 1994, 27, 389 ± 401.
[47] R. F. Butler, S. K. Banerjee, J. Geophys. Res. 1975, 80, 4049 ±
4058.
[48] Y. A. Gorby, T. J. Beveridge, R. P. Blakemore, J. Bacteriol.
1988, 170, 834 ± 841.
[49] D. Sch¸ler, M. Kˆhler, Zentralbl. Mikrobiol. 1992, 147, 150 ±
151.
[50] D. Sch¸ler, E. Baeuerlein, J. Bacteriol. 1998, 180, 156 ± 162.
[51] a) M. Gassmann, Diplomarbeit, Universit‰t T¸bingen (Germany), 1996; b) E. Baeuerlein, D. Sch¸ler, R. Reszka, S.
P‰user, DE 197 167 32 1997; E. Baeuerlein, D. Sch¸ler, R.
Reszka, S. P‰user, PCT/DE 98/00668 1998; E. Baeuerlein, D.
Sch¸ler, R. Reszka, S. P‰user, US 6251 365 2001.
[52] H. A. Lowenstam, S. Weiner, On Biomineralization, Oxford
University Press, Oxford, 1989.
[53] a) Y. Okuda, K. Denda, Y. Fukumori, Gene 1996, 171, 99 ± 102;
b) D. Sch¸ler, E. Baeuerlein, D. A. Bazylinski, Abstracts of the
General Meeting of the ASM, Miami, USA, 1997, p. 332.
[54] D. Sch¸ler, in Biomineralization (Ed.: E. Baeuerlein), WileyVCH, Weinheim, 2000, p. 109 ± 118.
[55] a) K. Gr¸nberg, C. Waver, B. M. Tebo, D. Sch¸ler, Appl.
Environ. Microbiol. 2001, 67, 4573 ± 4582; b) D. Sch¸ler, C.
Waver, K. Gr¸nberg, Jahrestagung der VAAM, 2001, Oldenburg, Abstract KSB 11.
[56] G. L. Blatch, M. Lassle, Bioassays 1999, 21, 932 ± 934.
[57] Y. Fukumori in Biomineralization (Ed.: E. Baeuerlein), WileyVCH, Weinheim, 2000, p. 93 ± 107.
[58] I. T. Paulsen, M. H. Saier, J. Membr. Biol. 1997, 156, 99 ± 103.
[59] O. H. Nies, S. Silver, J. Ind. Microbiol. 1995, 14, 186 ± 199.
[60] http://www.jgi.doe.gov/temweb/JGI microbial/html/index.html.
[61] M. J. Pallen, B. W. Wren, Mol. Microbiol. 1997, 26, 209 ± 221.
[62] L. Lenz, B. Dere, M. Bevan, Immunology 1996, 5, 63 ± 72.
[63] D. Sch¸ler, E. Baeuerlein, Arch. Microbiol. 1996, 166, 301 ± 307.
[64] a) J. B. Neilands, J. Biol. Chem. 1995, 270, 26 723 ± 26 726; b) G.
Winkelmann in CRC Handbook of Microbial Iron Chelates
(Ed.:G. Winkelmann), CRC Press, Boca Raton, 1991, pp. 65 ±
106.
[65] C. Nakamura, T. Sakaguchi, S. Kudo, I. G. Burgess, K. Sude, T.
Matsunaga, Appl. Biochem Biotechnol. 1993, 39/40, 169 ± 171.
[66] A. Hartmann, V. Braun, Arch. Microbiol. 1981, 130, 353 ± 356.
[67] a) T. S. Magnuson, A. L. Hodges-Myerson, D. R. Lovely,
FEMS Microbiol. Lett. 2000, 185, 205 ± 211; b) D. R. Lovely,
J. D. Coates, D. A. Saffarini, D. J. Lonergan in Transition Metals
in Microbial Metabolism (Eds.: G. Winkelmann, C. J. Carrano),
Harwood Academic Press, Amsterdam, 1997, pp. 187 ± 215;
c) S. A. B. Knight, E. Lesuisse, R. Stearman, R. D. Klausner, A.
Dencis, Microbiology 2000, 148, 29 ± 40; d) M. Fontecave, J.
Covõs, J. L. Pierre, BioMetals 1994, 7, 3 ± 8; e) L. C. Paoletti,
R. P. Blakemore, Curr. Microbiol. 1998, 17, 339 ± 342.
[68] a) Y. Noguchi, T. Fujiwara, K. Yoshimatsu, Y. Fukumori, J.
Bacteriol. 1999, 181, 2142 ± 2147; b) M. Abe, Y. Tamaura, Jpn. J.
Appl. Phys. 1983, 22, L511; c) T. Yamazaki, H. Oyanagi, T.
Fujiwara, Y. Fukumori, Eur. J. Biochem. 1995, 233, 665 ± 671.
Angew. Chem. Int. Ed. 2003, 42, 614 ± 641
Angewandte
Chemie
[69] a) D. Sch¸ler, R. Uhl, E. Baeuerlein, FEMS Microbiol. Lett.
1995, 132, 139 ± 145; b) E. Baeuerlein in Biomineralization
(Ed.: E. Baeuerlein), Wiley-VCH, Weinheim, 2000, pp. 61 ± 79.
[70] D. Grimm, A. J. Koster, U. Ziese, D. Typke, W. Baumeister, J.
Microsc. 1996, 183, 60 ± 68.
[71] R. P. Blakemore, N. A. Blakemore in Iron Biominerals (Eds.:
R. Frankel, R. P. Blakemore), Plenum Press, New York 1991,
pp. 51 ± 67.
[72] R. E. Dunin-Borkowski, M. R. Mc Cartney, R. B. Frankel,
D. A. Bazylinski, M. PÛsfai, P. R. Busek, Science 1998, 282,
1868 ± 1870.
[73] a) D. A. Bazylinski, R. B. Frankel in Biomineralization (Ed.:
Baeuerlein), Wiley-VCH, Weinheim 2000, p. 36 ± 37; b) R. B.
Frankel, D. Bazylinski, M. S. Johnson, B. L. Taylor, Biophys. J.
1997, 73, 994 ± 1000.
[74] R. Wiltschko, W. Wiltschko, Magnetic orientation in animals,
Zoophysiology, Vol. 33, Springer, Berlin, 1995.
[75] J. C. Weaver, T. E. Vaughan, D. Astumian, Nature 2000, 405,
707 ± 709.
[76] a) V. P. Shcherbakov, M. Winklhofer, Eur. Biophys. J. 1999, 28,
380 ± 392; b) M. Winklhofer, Modelle hypothetischer Magnetfeldrezeptoren auf der Grundlage biogenen Magnetits, M.
Leidorf GmbH, Rahden, 1999; c) M. Winklhofer, E. Holtkamp-Rˆtzler, M. Hanslik, G. Fleissner, N. Petersen, Eur. J.
Mineral. 2001, 13, 659 ± 669.
[77] a) J. C. Green, K. Perch-Nielsen, P. Westbroek in Handbook of
Protoctista (Eds.: L. Margulis, J. O. Corliss, M. Melkonian, D. J.
Chapman), Jones & Bartlett, Boston, 1990, pp. 293 ± 317; b) C.
van den Hoek, H. M. Jahns, D. G. Mann, Algen, 3rd ed.,
Thieme, Stuttgart, 1993, pp. 167 ± 179.
[78] a) H. Trappan, Paleobiology in Plant Protists, W. H. Freeman
and Co., San Francisco, 1980; b) A. S. Jafar, Neues Jahrb. Geol.
Palaeontol., Abh., 1983, 166, 218 ± 258.
[79] a) P. R. Brown, Calcareous Nanofossils Biostratigraphy, British
Micropaleontology Society Series, Chapman & Hall, London,
1998; b) L. Cros, A. Kleijne, A. Zeltner, C. Billiard, J. R.
Young, Mar. Micropaleontol. 2000, 39, 1 ± 34; c) J. R. Young,
S. A. Davis, P. R. Brown, S. Mann, J. Struct. Biol. 1999, 126,
195 ± 215.
[80] J. D. Rowson, B. S. C. Leadbeater, J. C. Green, Br. Phycol. J.
1986, 21, 359 ± 370.
[81] a) M. E. Marsh in Biomineralization (Ed.: E. Baeuerlein),
Wiley-VCH, Weinheim, 2000, pp. 251 ± 268; b) M. E. Marsh,
Protoplasma 1999, 207, 54 ± 66; c) I. R. Young, I. M. Didymus,
P. R. Brown, N. Prins, S. Mann, Nature 1992, 356, 516 ± 518; d) P.
van der Wal, E. W. de Jong, P. Westbrock, W. C. de Bruijn,
A. A. Hulder-Stapel, J. Ultrastruct. Res. 1983, 85, 139 ± 158.
[82] a) E. L. Gonzalez, in Biomineralization (Ed.: E. Baeuerlein),
Wiley-VCH, Weinheim, 2000, pp. 269 ± 282; b) Y. Araki, PhD
thesis, University of California, Los Angeles, 1997.
[83] M. E. Marsh, D. P. Dickinson, Protoplasma 1997, 199, 9 ± 17.
[84] a) M. E. Finbow, M. A. Harrison, Biochem. J. 1997, 324, 697 ±
712; b) N. Nelson, Organellar Proton-ATPases, Springer, Heidelberg, 1995.
[85] J. M. Ward, A. Reinders, H.-T. Hsu, H. Sze, Plant Physiol. 1992,
99, 161 ± 169.
[86] P. L. A. M. Corstjens, Y. Araki, E. L. Gonzales, J. Phycol. 2001,
37, 71 ± 78.
[87] a) Y. Araki, E. L. Gonzalez, J. Phycol. 1998,34, 79 ± 88; b) D.-K.
Kwon, E. L. Gonzalez, J. Phycol. 1994, 30, 689 ± 695.
[88] A. A. Israel, E. L. Gonzalez, Mar. Ecol. Prog. Ser. 1996, 137,
243 ± 250.
[89] M. E. Marsh, Protoplasma 1994, 177, 108 ± 122.
[90] a) M. E. Marsh, D. K. Chang, G. C. King, J. Biol. Chem. 1992,
267, 20 507 ± 20 512; b) P. van der Wal, E. W. de Jong, P. Westbrock, W. C. de Bruiyn, in Environmental Biogeochemistry,
639
Reviews
[91]
[92]
[93]
[94]
[95]
[96]
[97]
[98]
[99]
[100]
[101]
[102]
[103]
[104]
[105]
[106]
[107]
[108]
[109]
[110]
[111]
[112]
640
Vol. 35 (Ed.: R. Hallberg), Swedish Research Council, Stockholm, 1983, pp. 251 ± 258.
a) S. L. Lee, A. Veis, T. Gloneck, Biochemistry 1977, 16, 2971 ±
2979; b) W. G. Stetler-Stevenson, A. Veis, Calcif. Tissue Int.
1987, 40, 97 ± 102.
M. E. Marsh, R. L. Sass, J. Exp. Zool. 1983, 226, 193 ± 203.
M. T. Dimuzo, A. A. Veis, Calcif. Tissue Res. 1978, 25, 169 ± 178.
M. E. Marsh, Biochemistry 1989, 28, 339 ± 345.
a) H. Miyamoto, T. Miayashita, M. Okushima, S. Nakano, T.
Morita, A. Matsushiro, Proc. Natl. Acad. Sci. USA 1996, 93,
9657 ± 9660; b) B. L. Smith, T. Sch‰ffer, M. Viani, J. B. Thompson, N. A. Frederick, J. Kindt, A. M. Belcher, G. D. Stucky,
D. E. Morse, P. K. Hansma, Nature 1999, 399, 761 ± 763; c) M.
Kono, N. Hayashi, T. Samata, Biochem. Biophys. Res. Commun.
2000, 269, 213 ± 218; d) A. M. Belcher, E. E. Gooch in Biomineralization (Ed.: E. Baeuerlein), Wiley-VCH, Weinheim,
2000, pp. 221 ± 249; e) K. Mann, I. M. Weiss, S. Andrÿ, H.-J.
Gabius, M. Fritz, Eur. J. Biochem. 2000, 267, 5257 ± 5264;
f) I. M. Weiss, W. Gˆring, M. Fritz, K. Mann, Biochem. Biophys.
Res. Commun. 2000, 285, 244 ± 249.
M. E. Marsh, A. L. Ridall, P. Azadi, P. J. Duke, J. Struct. Biol.
2003, 139, 39 ± 45.
M. E. Marsh, Protoplasma 1996, 190, 181 ± 188.
a) C. van den Hoek, D. G. Mann, H. M. Jahns, Algae, Cambridge University Press, Cambridge, 1998; b) F. E. Round,
R. M. Crawford in Handbook of Protoctista (Eds.: L. Margulis,
J. O. Corliss, M. Melkonian, D. J. Chapman), Jones & Bartlett,
Boston, 1990, pp. 574 ± 596.
a) R. Tacke, Angew. Chem. 1999, 111, 3197 ± 3200; Angew.
Chem. Int. Ed. 1999, 38, 3015 ± 3018; b) P. Trÿguer, D. M.
Nelson, A. J. van Bennekom, D. J. DeMaster, A. Leynaert, B.
Quÿguiner, Science 1995, 268, 375 ± 379.
a) L. K. Medlin, W. H. C. F Kooistra, R. Gersonde, P. A. Sims,
U. Welbrock, Nova Hedwigia 1997, 65, 1 ± 11; b) W. H. C. F.
Kooistra, L. K. Medlin, Mol. Phylogenet. Evol. 1996, 6, 391 ±
407.
A. Rothpletz, Z. Dtsch. Geol. Ges. 1986, 48, 854 ± 914.
a) D. H. Erwin, Rev. Ecol. Syst. 1990, 21, 69 ± 91; b) D. H.
Erwin, Nature 1994, 367, 231 ± 236; c) P. R. Renne, Z. Zichao,
M. A. Richards, M. T. Black, A. R. Basu, Science 1995, 269,
1413 ± 1416; d) G. J. Retallack, Science 1995, 267, 77 ± 80.
M. Hildebrand, D. R. Higgins, K. Busser, B. E. Volcani, Gene
1993, 132, 213 ± 218.
N. Krˆger, C. Bergsdorf, M. Sumper, EMBO J. 1994, 13, 4676 ±
4683.
a) J. D. Pickett-Heaps, A.-M. M. Schmid, L. A. Edgar in
Progress in Phycological Research, Vol. 7 (Eds.: F. E. Round,
D. J. Chapman), Biopress, Bristol, 1990, p. 2 ± 168; b) N. Krˆger, M. Sumper, Protist 1998, 149, 213 ± 219; c) R. Wetherbee, S.
Crawford, P. Mulvaney in Biomineralization (Ed.: E. Baeuerlein), Wiley-VCH, Weinheim, 2000, p. 189 ± 206.
A. S. Crawford, J. M. Higgins, P. Mulvaney, R. Wetherbee, J.
Phycol. 2001, 37, 543 ± 554.
C. van den Hoek, D. G. Mann, H. M. Jahns, Algen, 3rd ed.
Thieme, Stuttgart, 1993, p. 115.
L. A. Edgar, J. D. Pickett-Heaps in Progress in Phycological
Research, Vol. 3, Biopress, Bristol, 1984, pp. 47 ± 88.
N. Krˆger, M. Sumper in Biomineralization (Ed.: E. Baeuerlein), Wiley-VCH, Weinheim, 2000, p. 151 ± 170.
R. W. Drum, H. S. Pankratz, J. Ultrastruct. Res. 1964, 10, 217 ±
223.
a) M. Hildebrand, B. E. Volcani, W. Gassmann, J. I. Schrˆder,
Nature 1997, 385, 688 ± 689; b) M. Hildebrand, K. Dahlin, B. E.
Volcani, Mol. Gen. Genet. 1998, 260, 480 ± 486.
a) M. Hildebrand in Biomineralization (Ed.: E. Baeuerlein),
Wiley-VCH, Weinheim, 2000, pp. 171 ± 188; b) H. L. Conway,
E. B‰uerlein
[113]
[114]
[115]
[116]
[117]
[118]
[119]
[120]
[121]
[122]
[123]
[124]
[125]
[126]
[127]
[128]
[129]
[130]
[131]
[132]
[133]
[134]
[135]
P. J. Harrison, C. O. Davies, Mar. Biol. 1976, 35, 187 ± 199;
c) H. L. Conway, P. J. Harrison, Mar. Biol. 1977, 43, 33 ± 43.
M. A. Brzezinski, R. J. Olson, S. W. Chisholm, Mar. Ecol. Prog.
Ser. 1990, 67, 83 ± 96.
V. Martin-Jezequel, M. Hildebrand, M. A. Brzezinski, J.
Phycol. 2000, 36, 821 ± 840.
C. W. Sullivan, J. Phycol. 1977, 13, 86 ± 91.
a) J. Reizer, A. Reizer, M. J. Saier, Jr., Biochim. Biophys. Acta
1994, 1197, 133 ± 166; b) Y. Deguchi, I. Yamoto, Y. Anraku, J.
Biol.Chem. 1990, 265, 21 704 ± 21 708.
a) T. Caspari, R. Stadler, N. Sauer, W. Tanner, J. Biol.Chem.
1994, 269, 3498 ± 3502; b) A. D. Due, Q. Zhinchao, J. M.
Thomas, A. Buchs, A. C. Powers, J. M. May, Biochemistry
1995, 34, 5462 ± 5471; c) H. Katagiri, T. Asano, H. Ishihara, K.
Tsukuda, J.-L. Lin, K. Inukai, M. Kikuchi, Y. Yazaki, Y. Oka, J.
Biol. Chem. 1992, 267, 22 550 ± 22 555.
C. W. Sullivan, J. Phycol. 1979, 15, 210 ± 216.
R. K. Iler, The Chemistry of Silica: Solubility, Polymerization,
Colloid and Surface Properties, and Biochemistry, Wiley, New
York, 1979.
a) D. Werner, Arch. Microbiol. 1966, 55, 278 ± 308; b) F. Azam,
B. B. Hemmingsen, B. E. Volcani, Arch. Microbiol. 1974, 97,
103 ± 114.
P. Bhattacharyya, B. E. Volcani, Biochem. Biophys. Res.
Commun. 1983, 114, 365 ± 372.
H. Drechsel, G. Winkelmann in Transition Metals in Microbial
Metabolism (Eds.: G. Winkelmann, C. J. Carrano), Harwood
Academic Press, Amsterdam, 1997, pp. 1 ± 51.
A.-M. Schmid, D. Schulz, Protoplasma 1979, 100, 267 ± 288.
M. Lee, C.-W. Li, Bot. Bull. Acad. Sin. 1992, 33, 317 ± 325.
J. D. Pickett-Heaps, A. M. Schmid, L. A. Edgar, in Progress in
Phycological Research, Vol. 7, Biopress, Bristol, 1990, p. 127 ±
133.
a) C.-W. Li, B. E. Volcani, Philos. Trans. R. Soc. London, Ser. B
1984, 304, 519 ± 528; b) C.-W. Li, B. E. Volcani, Protoplasma
1985, 124, 10 ± 29; c) D. Schulz, G. Drebes, H. Lehmann, R.
Jank-Ladwig, Eur. J. Cell. Biol. 1984, 33, 43 ± 51.
C. W. Mehard, B. E. Volcani, Cell Tissue Res. 1976, 166, 255 ±
263.
a) B. E. F. Reimann, J. C. Lewin, B. E. Volcani, J. Phycol.
1966,2, 74 ± 84; b) A. M. Schmid, M. A. Borowitzka, B. E.
Volcani in Cytomorphogenesis in Plants (Ed.: O. Kiermayer),
Cell. Biol. Mon. 1981, 8, 63 ± 97; c) R. M. Crawford, A. M.
Schmid in Biomineralization in Lower Plants and Animals
(Eds.: B. S. C. Leadbeater, R. Riding), The Systematics Association special volume 1986, 30, 291 ± 314.
E. Schnepf, G. Deichgr‰ber, G. Drebes, Plant Syst. Evol. 1980,
135, 265 ± 277.
a) B. E. F. Reimann, Exp. Cell Res. 1964, 34, 605 ± 608; b) E. F.
Stoermer, H. S. Pankratz, C. C. Bowen, Am. J. Bot. 1965, 52,
1067 ± 1078; c) J. Coombs, I. A. Lauritis, W. M. Darly, B. E.
Volcani, Z. Pflanzenphysiol. 1968, 59, 124 ± 152; d) P. A.
Dawson, J. Phycol. 1973, 9, 353 ± 365.
a) C. W. Li, S. Chu, M. Lee, Protoplasma 1989, 151, 158 ± 163;
b) M. A. Brzezinski, D. J. Conley, J. Phycol.1994, 30, 45 ± 55;
c) E. G. Vrieling, W. W. C. Gieskes, T. P. M. Beelen, J. Phycol.
1999, 35, 548 ± 559.
N. Krˆger, C. Bergsdorf, M. Sumper, Eur. J. Biochem. 1996, 239,
259 ± 264.
a) N. Krˆger, G. Lehmann, R. Rachel, M. Sumper, Eur. J.
Biochem. 1997, 250, 99 ± 105; b) N. Krˆger, R. Wetherbee,
Protist 2000, 151, 263 ± 273.
a) N. Krˆger, R. Deutzmann, M. Sumper, Science 1999, 286,
1129 ± 1132; b) N. Krˆger, R. Deutzmann, M. Sumper, J. Biol.
Chem. 2001, 276, 26 066 ± 26 070.
a) N. Krˆger, R. Deutzmann, C. Bergsdorf, M. Sumper, Proc.
Natl. Acad. Sci. USA 2000, 97, 14 133 ± 14 138; b) D. R. Morris,
Angew. Chem. Int. Ed. 2003, 42, 614 ± 641
Angewandte
Chemie
Biomineralization of Unicellular Organisms
[136]
[137]
[138]
[139]
[140]
[141]
[142]
[143]
[144]
L. J. Marton, Polyamines in Biology and Medicine, Dekker,
New York, 1981.
T. Nakajima, B. E. Volcani, Biochem. Biophys. Commun. 1970,
39, 28 ± 33.
H. Sch‰gger, G. von Jagow, Anal. Biochem. 1987, 166, 368 ± 379.
A. M. M. Schmid, B. E. Volcani, J. Phycol. 1983, 19, 387 ± 402.
T. F. Schultz, L. Egerton-Warburton, S. Crawford, R. Wetherbee, Protist 2001, 152, 315 ± 327.
a) C. T. Kresge, M. E. Leonowicz, W. J. Roth, J. C. Vartuli, J. S.
Beck, Nature 1992, 359, 710 ± 712; b) S. Mann, G. A. Ozin,
Nature 1996, 382, 313 ± 318; c) H. P. Lin, C. H. Mou, Science
1996, 273, 765 ± 767; d) H. Yang, N. Coombs, G. A. Ozin, Nature
1997, 386, 692 ± 695.
a) S. S. Kim, W. Z. Zhang, T. J. Pinnavia, Science 1998, 282,
1302 ± 1305; b) P. T. Tauer, T. J. Pinnavia, Science 1996, 271,
1267 ± 1269.
a) F. Caruso, R. A. Caruso, H. Mˆhwald, Science 1998, 282,
1111 ± 1114; b) B. T. Holland, C. F. Blanford, T. Do, A. Stein,
Chem. Mater. 1999, 11, 795 ± 805.
a) J. N. Cha, G. D. Stucky, D. E. Morse, T. J. Demning, Nature
2000, 403, 289 ± 292; b) C. Gˆltner, S. Henke, M. C. Wei˚enberger, M. Antonietti, Angew. Chem. 1998, 110, 633 ± 636;
Angew. Chem. Int. Ed. 1998, 37, 613 ± 616; c) D. Y. Zhao, J. J.
Feng, Q. S. Huo, N. Melosh, G. H. Frederickson, B. F. Chmelka,
G. D. Stucky, Science 1998, 279, 548 ± 552; d) M. Templin, A.
Frank, A. DuChesne, H. Leist, Y. M. Zhang, R. Ulrich, V.
Schadler, U. Wiesner, Science 1997, 278, 1795 ± 1798.
a) P. Schmidt-Winkel, W. W. Lukens, Jr., P. Yang, D. I. Margolese, I. S. Lettow, J. Y. Ying, G. D. Stucky, Chem. Mater. 2000,
12, 686 ± 696; b) S. Schacht, Q. Huo, I. G. Voigt-Martin, G. D.
Angew. Chem. Int. Ed. 2003, 42, 614 ± 641
[145]
[146]
[147]
[148]
[149]
[150]
[151]
[152]
Stucky, F. Sch¸th, Science 1996, 273, 768 ± 771; c) A. Imhof, D. J.
Pine, Nature 1997, 389, 948 ± 951.
a) E. G. Vrieling, T. P. M. Beelen, R. A. van Santen, W. W. C.
Gieskes, Angew. Chem. 2002, 114, 1613 ± 1616; Angew. Chem.
Int. Ed. 2002, 41, 1543 ± 1546; b) A. R. Hemsley, P. C. Griffith,
Philos. Trans. R. Soc. London, Ser. A 2000, 358, 547 ± 564; c) R.
Gordon, R. W. Drum, Int. Rev. Cytol. 1994, 150, 243 ± 372; d) J.
Parkinson, Y. Brechet, R. Gordon, Biochim. Biophys. Acta
1999, 1452, 89 ± 102; e) A. M. M. Schmid, Protoplasma 1994,
181, 43 ± 60; f) A. M. M. Schmid, R. K. Eberwein, M. Hesse,
Protoplasma 1996, 193, 144 ± 173.
a) M. Sumper, Science 2002, 295, 2430 ± 2433; b) D. Volkmer, S.
Tugulu, M. Fricke, T. Nielsen, Angew. Chem. 2003, 115, 60 ± 64;
Angew. Chem. Int. Ed. 2003, 42, 38 ± 61; c) N. Krˆger, S. Lorenz,
E. Brunner, M. Sumper, Science 2002, 298, 584 ± 586.
a) A. van Blaaderen, R. Ruel, P. Wiltzins, Nature 1997, 385,
321 ± 324; b) A. McPherson, Crystallization of Biological Macromolecules, Cold Spring Harbor Laboratory Press, Cold
Spring Harbor, New York 1999.
a) E. Epstein, Annu. Rev. Plant Physiol. Plant Mol. Biol. 1999,
50, 641 ± 644; b) A. J. Milligan, F. M. M. Morel, Science 2002,
297, 1848 ± 1850.
R. D. Adam, Clin. Microbiol. Rev. 2001, 14, 447 ± 475.
L. L. Brott, R. R. Naik, D. J. Pikas, S. M. Kirkpatrik, D. W.
Tomlin, P. W. Whitlock, S. J. Clarson, M. O. Stone, Nature 2001,
413, 291 ± 293.
a) S. Brown, Nat. Biotechnol.1997, 15, 269 ± 274; b) S. R.
Whaley, D. S. English, E. L. Hu, P. F. Barbara, A. M. Belcher,
Nature 2000, 405, 665 ± 668.
R. R. Naik, L. L. Brott, S. J. Clarson, M. O. Stone, J. Nanosci.
Nanotechnol. 2002, 2, 95 ± 100.
641
Документ
Категория
Без категории
Просмотров
2
Размер файла
907 Кб
Теги
production, inorganic, organisms, unicellular, nano, biomineralization, biochemistry, unusual, membranes, microstructure
1/--страниц
Пожаловаться на содержимое документа