FEMS Microbiology Ecology 26 (1998) 297^310
Evaluation of microscopic techniques to observe iron precipitation
in a natural microbial bio¢lm
D. Ann Brown a; *, Terry J. Beveridge b , C. William Keevil c , Barbara L. Sherri¡
b
a
a
Department of Geological Sciences, University of Manitoba, Winnipeg, Man., R3T 2N2, Canada
Department of Microbiology, College of Biological Science, University of Guelph, Guelph, Ont., N1G 2W1, Canada
c
Center for Applied Microbiology and Research, Porton Down, Salisbury, Wilts., SP4 0JG, UK
Received 6 March 1998; revised 9 May 1998; accepted 9 May 1998
Abstract
Iron biomineralization in a microbial biofilm consortium from Canadian Shield groundwaters has been investigated with
different microscopic techniques. The advantages and disadvantages of the different methods of observing a biofilm growing
on an opaque mineral surface are discussed. Scanning electron microscopy was able to show the initial attachment and
dispersion of bacteria on the mineral surfaces, whereas transmission electron microscopy gave greater detail and revealed the
precise location of the iron precipitation on cell surfaces, including S-layers, and also throughout the extrapolymeric slime of
the biofilm. Episcopic Nomarski differential interference contrast microscopy allowed direct observation of biofilm dynamics
and confirmed the precipitation of iron directly onto certain bacteria, which were then specifically ingested by protozoa. This
novel ingestion of iron-coated bacteria by protozoans essentially eliminated iron from solution and trapped it within the
biofilm. Over time in the natural environment, this iron, enmeshed within a biofilm, may become incorporated into iron-rich
sediments. z 1998 Federation of European Microbiological Societies. Published by Elsevier Science B.V. All rights reserved.
Keywords : Microscopic technique; Di¡erential interference contrast microscopy ; Bio¢lm ; Protozoa; Iron biomineralization
1. Introduction
Terrestrial groundwaters are now understood to
contain a variety of microorganisms that help determine the local ecosystem through their metabolic
activities. Natural waters are oxidizing when in contact with the atmosphere, but when isolated they
become reducing, as oxygen is consumed by hydro-
* Corresponding author. Tel.: +1 (204) 474 9786;
Fax: +1 (204) 474 7623; E-mail: [email protected]
chemical and biochemical reactions. When oxygen is
absent bacteria use inorganic electron acceptors to
catalyze the oxidation of organic matter. This oxidation is the cause of the reducing activity of most
surface and ground waters [1]. Furthermore, the actual presence of microorganisms is able to remove
metal ions from solution through adsorption onto
their membrane surfaces.
In oligotrophic terrestrial groundwaters, bacteria
generally ¢nd it more pro¢table to live as a consortium in a bio¢lm. This consists of an envelope of
slime made up of extracellular polymeric substances
0168-6496 / 98 / $19.00 ß 1998 Federation of European Microbiological Societies. Published by Elsevier Science B.V. All rights reserved.
PII: S 0 1 6 8 - 6 4 9 6 ( 9 8 ) 0 0 0 4 4 - 0
FEMSEC 930 3-8-98
298
D.A. Brown et al. / FEMS Microbiology Ecology 26 (1998) 297^310
(EPS), often polysaccharides, that surrounds and is
excreted by the consortium of microorganisms that
makes up the bio¢lm. Bio¢lms are universal in the
aqueous environment, and form preferentially at
rock/water interfaces where nutrients tend to concentrate, thus providing a richer food source than that
of the bulk £uid. Individuals of the consortium are
arranged within the slime layer so that each organism is most favorably placed in the bio¢lm ecosystem. As a result, few free-living microorganisms are
to be found in natural groundwaters.
Bacterial cell walls and EPS are generally negatively charged, which allows positively-charged metal
ions to be adsorbed [2,3]. These metal ions may then
act as nucleation sites for the deposition of more
metal from solution [4] as can be seen in acidic
and hot spring sediments [5,6] as well as in microbial
mats in hydrothermal waters [7]. Certain bacteria
may be surrounded by an S-layer, a paracrystalline
proteinaceous structure that envelopes the cell, and
where its peripheral location allows contact with the
environment but prevents direct contact with the cell
membrane. The S-layer can also form a template for
mineralization by providing discrete, repeating nucleation sites [8]. An example has been reported from a
lake biotherm where, depending on pH, ¢ne-grained
gypsum or calcite is deposited on a cyanobacterial Slayer [9,10]. That bacteria can bind metals is perhaps
not surprising, but the amount of metal they have
been shown to accumulate, especially from dilute
solutions, is remarkable and may provide a major
mechanism for purging metal ions from groundwater
[3].
In our present work, we examine a bio¢lm consortium that was initially discovered in the Underground Research Laboratory (URL) excavated in
the Lac du Bonnet batholith, Manitoba, Canada
[11]. The organisms originated in Boggy Creek [12],
which drains the surface water from the batholith
into the Winnipeg River, and which was the source
of mine water for the laboratory. The consortium
contains a variety of bacteria with di¡ering morphologies as well as an aggressive protozoan. Very little
investigation has been done on protozoa in uncontaminated natural waters as their numbers are usually small [13]. However, selective grazing of bio¢lms
where attached bacteria are concentrated provide an
important food source [14]. Active biomineralization
by our consortium is demonstrated by the precipitation of iron in several forms, the reduction of ferric
to ferrous iron, as well as the leaching of iron from
iron minerals in the granite host rock and alteration
of magnetite to hematite over a period of several
weeks [12,15,16].
In this paper we discuss the advantages and disadvantages of various microscopic techniques that
we have used to study the spatial relationships of
the bio¢lm consortia, as well as the location of the
iron precipitation within the bio¢lm itself. We used
light microscopy both with bright-¢eld and phase
contrast formats, as well as Nomarski and scanning
confocal laser microscopy (SCLM); electron microscopy was used in both the scanning (SEM) and
transmission (TEM) modes. Light microscopy depends on the ability of an object to absorb (colour)
and refract (density) light rays, whereas SEM depends on the ability of the specimen to emit secondary electrons when it is `excited' by a powerful electron beam, TEM on the other hand depends on the
ability of the mass of the specimen to scatter the
electrons of a powerful electron beam as the electrons pass through it. Each type of microscopy
thus provides di¡erent information about a single
sample, and it is the integration of this whole set
of microscopies that provides an insight into the
structure, composition and physical properties of
the sample. Our observations from these various microscopic techniques have been used in the present
study to describe the role of iron in the bio¢lm.
2. Materials and methods
2.1. Sample source and preparation
Two cultures of the microbial consortium were
obtained, one from Boggy Creek (BC), and the other
from the water within the fractures that comprise
Fracture Zone 2 (FZ2) at the URL site which is
intersected by the shaft [12]. The consortia (BC
and FZ2) are maintained in laboratory bioreactors on a medium of 5.0 g ferric ammonium citrate,
0.5 g K2 HPO4 , 0.2 g MgSO4 .7H2 O, and O.01 g
CaCl2 .2H2 O per liter of deionized water at pH 7.0
[17]. When samples are required for cultures, media
are added and recirculated in the bioreactor through
FEMSEC 930 3-8-98
D.A. Brown et al. / FEMS Microbiology Ecology 26 (1998) 297^310
a column of granite chips, that provide the necessary
trace minerals.
Laboratory incubations have shown that the reactions mediated by both these microbial communities
are similar. However, after a period of refrigerated
storage, the BC culture was found to contain only
bacteria, having apparently lost all viable protozoa,
while the other culture, FZ2, maintained a community of both bacteria and protozoa. This was useful
since it allowed the observation of bacterial mineralization both in the presence and absence of the protozoa, as it is extremely di¤cult to prevent the
growth of protozoa in iron-containing media once
the protozoa have become encysted.
Flask cultures for scanning electron microscopy
(SEM) contained 1 g of crushed rock (1^5 mm diameter) of granite, magnetite or microcline in 50 ml
of media where the ferric ammonium citrate was
replaced with ammonium citrate, to compel the consortium to extract iron from solid minerals in a manner nearer to that which occurs in nature.
Bio¢lms from these consortia form rapidly on the
granite, but a considerable time is necessary for the
organisms to extract much iron from minerals in the
granite [16]. We have therefore found it expedient,
when growing bio¢lms for observation of iron precipitation, to supplement the media with iron chelated by ammonium citrate. The bio¢lms for di¡erential interference contrast (DIC) microscopy
examinations were grown on granite billets
(30U40U10 mm) that were placed in plastic containers, just covered with the above medium, and inoculated from one of the laboratory reactors. These
were gently rocked for a week, during which time
observations were made of the bio¢lm on the surface
of each billet, as well as on samples from the medium.
2.2. Microscopic techniques
The standard light microscope used was a Wild
M21 with a Nikon phase contrast attachment. The
unstained bio¢lm was also viewed by Nomarski DIC
microscopy with a Nikon Labophot microscope (Nikon, Telford, UK) which was equipped for both
transmitted visible light and episcopic UV illumination.
Nomarski DIC microscopy is superior to phase
299
contrast microscopy and is especially suitable for zdirection detection. This technique can examine difî , which allows the detecferent levels, less than 10 A
tion of high extinction values by increasing the image
contrast and brightness, and de¢ning the spatial relationships more clearly. The images obtained in real
time appear three-dimensional without the need to
compile a series of thin optical slices, and view the
images with red/green lenses as required for scanning
confocal laser microscopy (SCLM). DIC microscopy
is commonly used with transmitted light and is therefore unsuitable for viewing bio¢lms on opaque materials such as rocks or plumbing tubes. However,
Keevil and Walker [18] were able to adapt and recon¢gure a Nikon Labophot-2 microscope to combine the normal epi£uorescence attachment with
episcopic DIC illumination above the light stage.
Key features of this modi¢cation are the use of a
common halogen light source for both UV and higher wavelength illumination, and the repositioning of
the DIC polarizer and 1/4 wave plate in one of the
¢lter block cubes in the carriage above the microscope stage. An adjustable analyzer was ¢tted into
the main body between the DIC block and the eyepiece. This analyzer is withdrawn for epi£uorescence
work and an EFD ¢lter block (containing the excitation ¢lter, dichroic mirror and barrier ¢lter) moved
into the light path.
The long focal length, non-contact metallurgical
objectives allow depth observation throughout the
bio¢lm, and avoid the need for coverslips or oil contact with the bio¢lms which might distort their structure and interfere with the re£ected episcopic light.
The objectives used were M Plan 150/0.95 210/0, M
Plan 100/0.80 ELWD 210/0, and M Plan 40/0.5
ELWD 210/0 (Nikon). Removable DIC prisms are
incorporated within the objectives so these can be
simply withdrawn for epi£uorescence work. Cells
were stained with propidium iodide (1 mg ml31 )
for viewing under UV light.
Observations in real time under ambient temperature and pressure using DIC microscopy permitted
the observation of motile bacteria and protozoa. The
images obtained were captured using a CCD camera
mounted on a trinocular head and transferred to a
video recorder or a computer for image analysis. The
DIC block had been found suitable for di¡erentiating gold-labeled bacteria which appeared much
FEMSEC 930 3-8-98
300
D.A. Brown et al. / FEMS Microbiology Ecology 26 (1998) 297^310
brighter than unlabeled cells [19]. This aided the current observations in real time of the ecosystem of
cells that were naturally metal-encrusted.
SCLM images were obtained with an MRC-500
Lasersharp £uorescence scanning confocal microscope (Bio-Rad Microscience, Toronto, Canada) in
conjunction with a Zeiss Photomicroscope III (Oberkochen, Germany). An argon laser was used as the
excitation source for the £uorophores [20].
The SEM images were obtained by a Cambridge
Instruments model TL2097, equipped with both
backscatter and secondary detectors at 27 kV. Images were stored on a Kontron Ibas image analyzer
system. Samples from laboratory £ask incubations
were removed from the media for examination, allowed to air-dry over night and were then gold/palladium coated.
The TEM was carried out using a Philips EM300
electron microscope at 60 kV. For energy dispersive
X-ray spectroscopy (EDS), a Philips EM 400T
equipped with a Link eXL multichannel analyzer
and L2-5 detector (Link Analytical, London, UK)
was used. In these analyses electron beam spot sizes
of 2.0 Wm or less for 300 s (live time) were employed
on unstained specimens. For whole mounts and thin
sections the specimens, both from bio¢lms and £ask
grown cultures, were ¢xed in 2% (v/v) glutaraldehyde, en bloc stained in 2% (w/v) uranyl acetate as
necessary, dehydrated through an ethanol-to-propylene oxide series, and then embedded in either LR
white or Epon resins.
3. Discussion of results
Bio¢lms that grow on solid surfaces, such as rock
faces, are amongst the most di¤cult samples to accurately monitor by microscopy. Yet, it is important
to detect and understand the distribution of microorganisms throughout the bio¢lm matrix, their juxtaposition with one another, their attachment to
mineral surfaces, and if applicable, the e¡ect of protozoan grazing on bio¢lm bacteria. Clearly, as bio¢lms develop and mature, microbial strati¢cation occurs, a¡ecting both redox potential and pH within
the matrix; this can alter the speciation of inorganic
electrolytes as well as the development of mineral
precipitates [21]. This was nicely demonstrated by
the formation of both ferric iron as ferrihydrite
and ferrous iron as siderite in the original URL bio¢lm [12].
No single microscopic technique will provide all
the information that is required for the analysis of
such a bio¢lm. However, in this article we describe
the di¡erent approaches, ranging from light- to electron-based microscopies, that we have used to develop a comprehensive set of observations of the precipitation of iron within our bio¢lm. From the
images that we have obtained we discuss the activity
and spatial presence of the organisms, and how this
a¡ects the precise location of the iron precipitate
within the bio¢lm matrix, and the almost complete
removal of iron from solution.
3.1. Light microscopy
Bacteria from these oligotrophic groundwaters are
extremely small, often between 1.0 and 0.5 Wm in
diameter; they are also quite transparent and are
di¤cult to distinguish visually from the surrounding
aqueous milieu. This is especially so when they are
encased in a hydrated gelatinous matrix of EPS such
as found in our bio¢lm and other cultures. Traditional light microscopes with an oil immersion lens
can magnify up to V1200U, but the resolution is
limited by the quality of the lenses and the wavelength of light. A phase contrast system provides a
better di¡erentiation of semi-transparent microorganisms with transmitted light although the depth
of view is limited, but when the organisms are attached to opaque, solid surfaces transmitted light
cannot be used.
Cultures of the microbial consortium scraped from
the bio¢lm and viewed under phase contrast contain
bacteria with various morphologies and movements,
as well as a di£agellate protozoan that feeds on bacteria in the aerobically-grown cultures. The massive
precipitation of iron could be readily distinguished,
but the resolution was insu¤cient to provide informative photomicrographs as to the actual location of
the iron precipitate in relation to the bacterial cells,
since most became entrapped within this precipitate.
In our experiments the sterile medium was initially
yellow in color due to ferric ammonium citrate, but
after one week's incubation the whole of the bio¢lm
had become dark red-brown in color and contained
FEMSEC 930 3-8-98
D.A. Brown et al. / FEMS Microbiology Ecology 26 (1998) 297^310
301
Fig. 1. Confocal image of £uorescently stained bio¢lm growing on a granite billet optically sectioned vertically to show the varying thickness of the bio¢lm. The cells and EPS are white and the height of the bio¢lm along the Z axis is shown by the marker; bar = 100 Wm.
nearly all of iron in the sample, since the medium
was colorless and contained less than 1 mg l31 of
iron.
3.2. Scanning confocal laser microscopy
A more recently developed alternative optical
method is SCLM which uses an intense coherent
visible or UV light, that permits horizontal and vertical optical sectioning of the biomass so that threedimensional relationships can be determined within
the bio¢lm. This method has been used successfully
with UV light to obtain an understanding of the
large scale morphology of bio¢lms, and the distribution of the constituent microorganisms when they
are tagged with £uorescent markers [20]. Using this
system the BC culture exhibited columns, or pillars,
that rise to a height of 60 Wm above the bio¢lm base
(Fig. 1), but since iron is not visible under UV illumination, the iron precipitate within the bio¢lm
could not be located by this technique.
3.3. Scanning electron microscopy
SEM gives better resolution than the previous two
types of light microscopy. The narrow excitation
beam of electrons from the electron gun of the
SEM induces secondary electron emission from a
small surface area, which allows good resolution
for magni¢cations up to 30 000U. Conventional
SEM is operated in a high vacuum, which requires
the specimen to be dried and then coated with gold,
to both disperse surface charges and to optimize sec-
ondary electron emission. Under these conditions we
found SEM to be an excellent technique to visualize
the bacteria when they ¢rst became attached to the
substratum, and before production of EPS was established and hid the bacteria from view.
During the ¢rst few days of incubation we were
able to see an increase in the number and types of
bacteria adhering to the di¡erent mineral surfaces
such as granite, magnetite and microcline. Various
forms of cocci and rods made up the population
(Fig. 2A); a greater number of bacteria were found
on the fractured edges of minerals than on their
cleavage faces, possibly due to a higher energy state
at the edges. After a week, the bio¢lm became thick
with EPS and metal precipitates so that on dehydration a crust was formed that made it di¤cult to observe any organism within it [16]. At this point SEM
looses its discriminative power, since topographic
analysis is no longer useful. Although the surface
of the bio¢lm can be satisfactorily imaged, it is not
possible to determine the spatial organization within
the depth of the bio¢lm itself.
3.4. Transmission electron microscopy
TEM had the highest resolving power that we used
on our bio¢lm samples. Even though high voltages
(60^100 kV) are used, the penetrating power of the
electron beam is poor, so only relatively thin samples
can be observed, and these cannot be attached to an
opaque mineral surface. Yet, unstained specimens
reveal the magnitude and location of the iron minerals because of the strong electron scattering power
FEMSEC 930 3-8-98
302
D.A. Brown et al. / FEMS Microbiology Ecology 26 (1998) 297^310
Fig. 2. Di¡erent bacterial morphologies. A: SEM micrograph of bacteria attached to a magnetite chip after 4 days growth ; B: stained
whole mount TEM micrograph after £ask incubation for 8 days; bar = 1 Wm.
of the metal. This method produced excellent images,
both with bulk samples and with sections of individual organisms. Since the elements in the specimen
emit X-rays of distinct spectral energy upon excitation with the electron beam, EDS allows determina-
tion of the elemental composition of minerals deposited in the bio¢lm [22].
Our initial TEM of the URL bio¢lm showed that
the iron precipitated around the cells as well as
throughout the bio¢lm matrix [12]. In this study,
FEMSEC 930 3-8-98
D.A. Brown et al. / FEMS Microbiology Ecology 26 (1998) 297^310
303
Fig. 3. TEM images of thin sections showing: A: increasing biomineralization on cells (arrows), and B: precipitation around cells (C), on
the S-layer (S) and lysed cells (L) ; bars = 0.5 Wm.
FEMSEC 930 3-8-98
304
D.A. Brown et al. / FEMS Microbiology Ecology 26 (1998) 297^310
Fig. 4. DIC images of FZ2 culture. A: Bright iron-coated bacteria on the granite billet after one day incubation. B: Initiation of bio¢lm
(arrows) directly above the bacteria shown in A.
stained whole mounts show that our consortium
contained several distinct morphological types of
bacteria (Fig. 2B), which can be compared to those
in the SEM image (Fig. 2A). Thin sectioning gives a
more precise location of the iron precipitation, and it
is apparent that there was a step-wise accumulation
of metal on the cell surfaces, as bacteria with increasing amounts of metal deposited around the cells can
be seen (Fig. 3A). In some cases the metal encrusting
some of the cells was so great that it caused the cells
to lyse (Fig. 3B). In contrast the bacteria that produce S-layers were able to prevent the precipitation
FEMSEC 930 3-8-98
D.A. Brown et al. / FEMS Microbiology Ecology 26 (1998) 297^310
305
Fig. 5. DIC image of the granite billet showing the beginning of bio¢lm EPS formation (arrows) at top left, with bacteria and protozoa
(P) on the rock surface bottom right; bar = 10 Wm.
of iron directly onto the cell membrane which allowed the cells to remain viable. Precipitated iron
is also disseminated throughout the EPS. These
TEM images give only a very small sample area
so it is di¤cult to relate the thin sections to the
bulk sample. Although it is possible to see the details of iron precipitated around the cells and
throughout the EPS, it is not possible to ascertain
their relative position within the bulk of the bio¢lm
matrix.
Fig. 6. TEM image after 5 days £ask incubation showing the paracrystalline structure of the S-layer (arrows) that has separated from a
bacterium ; bar = 200 nm.
FEMSEC 930 3-8-98
306
D.A. Brown et al. / FEMS Microbiology Ecology 26 (1998) 297^310
3.5. Di¡erential interference contrast microscopy
Our observations using DIC microscopy during
bio¢lm development showed a progressive precipitation and mineralization of iron within the bio¢lm.
Patches of brightly re£ecting bacteria can be seen,
and by altering the microscope focus, they were
found to be within cracks of the granite billet (Fig.
4A), possibly where the bacteria had come into direct contact with some iron mineral. The initiation of
bio¢lm growth is shown by a small patch of EPS
slime on the surface of the billet directly above these
Fig. 7. DIC images of protozoa in the FZ2 bio¢lm. A: Protozoa (P) containing iron-coated bacteria (I). B : Encysting protozoa (P) containing many iron-coated bacteria; bars = 10 Wm.
FEMSEC 930 3-8-98
D.A. Brown et al. / FEMS Microbiology Ecology 26 (1998) 297^310
cells (Fig. 4B). In the FZ2 culture, the edge of the
EPS slime layer can be easily seen on the billet (Fig.
5). A portion of slime-forming EPS can be seen at
top left, while the surface of the granite billet itself,
with bacteria and protozoa but little EPS, is clearly
visible at bottom right. This DIC image can be compared to the optical vertical sectioning through the
bio¢lm shown in the SCLM image (Fig. 1), where
the bio¢lm thickness varies from 25 to 85 Wm.
The most noteworthy observation with the DIC
technique was that an actual site of iron precipitation was revealed. Some of the bacteria are very
small cocci with a diameter of less than 0.5 Wm.
On incubation these bacteria became very much
brighter under re£ected light than other components
of the bio¢lm; our interpretation is that these bacteria had become coated with iron, quite possibly on
the S-layer, which re£ects the light. This re£ection is
similar to the e¡ect seen when Legionella pneumophila was tagged with immunogold [19]. The fact that
light is re£ected indicates that the iron must be crystallographically organized in some way, since the
iron precipitated within the EPS, which our X-ray
di¡raction analysis has shown to be amorphous,
307
does not re£ect light under DIC conditions. This
organization of the iron quite possibly occurs on
the strikingly regular hexagonal array of the S-layer
(Fig. 6), and is possibly similar to the ordering of
particles of gypsum and calcite that has been reported for a cyanobacterial S-layer found on a biotherm [8,9].
We were surprised to discover that the preferred
prey of the protozoa appeared to be these bright,
iron-coated bacteria, rather than those bacteria without such a coating. Initially, individual iron-coated
bacteria could be seen within the protozoa (Fig. 7A);
by the sixth day of incubation the protozoa were
tightly packed with these brightly re£ecting organisms (Fig. 7B). Shortly afterwards the protozoa began encysting, possibly because there were no more
iron-coated bacteria on which to feed, although there
were plenty of other bacteria available.
The ingestion of the bacteria by the protozoa has
been con¢rmed by TEM; Fig. 8 shows a protozoan
actively engul¢ng a bacterium. Thin sections of this
sample revealed that most protozoa contained granules of iron suggesting that this internalization of the
iron was due to the protozoa engul¢ng the iron-
Fig. 8. TEM image from the FZ2 culture of a bacterium (B) being ingested by a protozoan (P). Iron is both disseminated throughout the
cytoplasm of the protozoa, and precipitated in clusters in the bio¢lm (arrows); bar = 2 Wm.
FEMSEC 930 3-8-98
308
D.A. Brown et al. / FEMS Microbiology Ecology 26 (1998) 297^310
Fig. 9. DIC image of crystallites formed in BC culture ; bar = 10 Wm.
loaded bacteria. The bacteria are broken down and
digested within the phagosomes, leaving iron particles distributed throughout the cytoplasm. When
EDS was performed on these protozoa, high levels
of iron and lower levels of phosphorus (nucleic acids
and phospholipids), magnesium and calcium (both
common cytoplasmic electrolytes) were seen. Compared to the BC culture the protozoa-containing
FZ2 culture contains much less iron associated
with the bio¢lm bacteria since the protozoa have
incorporated most of the metal internally. It appears
that the eukaryotic protozoa, by ingesting the ironcoated bacteria, are `sweeping' the iron from the
prokaryotic system.
We have no explanation why the protozoa in our
consortia should prefer these iron-coated bacteria to
the other types present. Although iron is an essential
trace nutrient for growth, and protozoa are known
to grow better in its presence, the quantity visible
within the protozoa would seem to indicate that
they were consuming far more than was necessary
for mere dietary requirements. It is also unusual
that the iron is disseminated throughout the cytoplasm rather than being localized with a speci¢c cytoplasmic region of the protozoan, although possibly
iron is not toxic in this particulate form. However,
the protozoa encystment, so soon after this meal,
when there was still plenty of other bacterial food
available, would seem to suggest that this amount of
iron was not, in fact, advantageous for them in the
short term. This observation does suggest, though,
an interesting new process for the concentration and
deposition of iron within the natural environment.
In the BC culture, which did not contain protozoa,
the iron-coated bacteria continued to act as nucleation sites for further precipitation, which in some
cases formed radiating crystallites (Fig. 9). Under
DIC microscopy, bacteria could be seen actually attached to these crystallites, although the resolution
of the micrograph is insu¤cient to show this.
4. Conclusions
The iron-precipitating bio¢lm formed by our microbial consortia has been investigated by a variety
of light and electron microscopic techniques. No one
microscopic method by itself was able to provide
complete information, but the integration of light
and electron microscopies complement each other
to give a more complete analysis than any one procedure by itself. The active organisms in hydrated
samples can be observed by light microscopy, particularly DIC, to give a unique view of the dynamic
FEMSEC 930 3-8-98
D.A. Brown et al. / FEMS Microbiology Ecology 26 (1998) 297^310
biomineralization occurring within the bio¢lm. However, the resolution obtainable from light microscopy
is inadequate to determine the precise site on the
bacteria of this iron precipitation, which is better
observed by the TEM technique. SEM has proved
to be useful to view the details of the initial bacterial
attachment to the mineral surface before the bio¢lm
EPS becomes established.
The bio¢lm consortia were able to precipitate iron
chelated in solution. Observations through microscopic techniques have enabled us to decipher the
location of this precipitated iron. The iron precipitate is disseminated throughout the EPS, and there is
more speci¢c precipitation on cell walls and S-layers
of many of the bacteria as well. Cell walls of certain
bacteria are coated with iron, which is visible by
DIC through light re£ected from the iron. These
bacteria became prey for the protozoa, who speci¢cally ingested them before encysting. Where no protozoa are present in the bio¢lm the iron is precipitated in a similar manner, but the nucleation sites are
now the bacteria themselves that initiate further biomineralization; in some cases this includes the development of crystallites.
We have found that cultures, both with and without protozoa, are able to e¤ciently remove iron
from solution and to sequester it within the bio¢lm.
The source of our consortia was typical surface
water draining a Canadian Shield granitic batholith.
These results suggest that this mechanism for removing metal ions from dilute aqueous solutions could
be widespread, both now and in the past. In our case
the metal was iron, which is widely available from
basement rocks throughout the Shield. This bacterial
sequestering could thus provide the iron for incorporation into iron-rich sediments, these may then
undergo diagenetic geochemical processing and be
transformed into sedimentary minerals.
Acknowledgments
This research was funded by an NSERC Collaborative Grant to B.L.S., T.J.B. and M.B. McNeil.
Many thanks are due to John Lawrence, of the National Hydrology Research Institute, Saskatoon for
the SCLM image, and to Sergio Mejia of the University of Manitoba for the SEM images. We would
309
like to thank Richard Sparling for discussion and
practical advice, as well as two anonymous reviewers
for their comments.
References
[1] Stumm, W. and Morgan, J.J. (1981) Aquatic Chemistry, 748
pp. John Wiley, New York, NY.
[2] McLean, R.J.C., Fortin, D. and Brown, D.A. (1996) Microbial metal-binding mechanisms and their relation to nuclear
waste disposal. Can. J. Microbiol. 42, 392^400.
[3] Schultze-Lam, S., Thompson, J.B. and Beveridge, T.J. (1993)
Metal ion immobilization by bacterial surfaces in freshwater
environment. Water Pollut. Res. J. Can. 28, 51^81.
[4] Beveridge, T.J. and Fyfe, W.S. (1985) Metal ¢xation by bacterial cell walls. Can. J. Earth Sci. 22, 1893^1898.
[5] Ferris, F.G., Beveridge, T.J. and Fyfe, W.S. (1986) Iron-silica
crystallite nucleation by bacteria in a geothermal sediment.
Nature 320, 609^611.
[6] Ferris, F.G., Fyfe, W.S. and Beveridge, T.J. (1987) Manganese oxide deposition in a hot spring microbial mat. Geomicrobiol. J. 5, 33^42.
[7] Konhauser, K.O. and Ferris, F.G. (1996) Diversity of iron
and silica precipitation by microbial mats in hydrothermal
waters, Iceland ; implications for Precambrian iron formations. Geology 24, 323^326.
[8] Beveridge, T.J. (1994) Bacterial S-layers. Curr. Opin. Struct.
Biol. 4, 204^212.
[9] Schultze-Lam, S., Harauz, G. and Beveridge, T.J. (1992) Participation of a cyanobacterial S layer in ¢ne-grain mineral
formation. J. Bacteriol. 174, 7971^7981.
[10] Schultze-Lam, S. and Beveridge, T.J. (1994) Physicochemical
characteristics of the mineral-forming S-layer from the cyanobacterium Synechococcus strain GL-24. Can. J. Microbiol. 40,
216^223.
[11] Brown, A., Soonawala, N.A., Everitt, R.A. and Kamineni,
D.C. (1989) Geology and geophysics of the Underground Research Laboratory site, Lac du Bonnet Batholith, Manitoba.
Can. J. Earth Sci. 26, 404^425.
[12] Brown, D.A., Kamineni, D.C., Sawicki, J.A. and Beveridge,
T.J. (1994) Minerals associated with bio¢lms occurring on
exposed rock in a granitic Underground Research Laboratory.
Appl. Environ. Microbiol. 60, 3182^3191.
[13] Novarino, G., Warren, A., Butler, H., Lambourne, G., Boxshall, A., Bateman, J., Kinner, N.E., Harvey, R.W., Mosse,
R.A and Teltsch, B. (1997) Protistan communities in aquifers :
a review. Microbiol. Ecol. 20, 261^275.
[14] Sibbald, M.J. and Albright, L.J. (1988) Aggregated and free
bacteria as food sources for heterotrophic micro£agellates.
Appl. Environ. Microbiol. 54, 613^616.
[15] Brown, D.A., Liwanag, J. and Sherri¡, B.L. (1997) Alteration
of iron minerals by a microbial consortium. Proc. GAC/MAC
Annual Meeting, May 19^21, Abst. p. A-19.
[16] Brown, D.A., Sherri¡, B.L. and Sawicki, J.A. (1997) Micro-
FEMSEC 930 3-8-98
310
D.A. Brown et al. / FEMS Microbiology Ecology 26 (1998) 297^310
bial transformation of magnetite to hematite. Geochim. Cosmochim. Acta 61, 3341^3348.
[17] Brown, D.A. and Hamon, C. (1993) Initial investigation of
groundwater microbiology at AECL's Underground Research
Laboratory. Atomic Energy of Canada Limited technical record TR-608, COG-93^171. SDDO, Chalk River Laboratories, AECL, Chalk River, Ont., Canada.
[18] Keevil, C.W. and Walker, J.T. (1992) Nomarski DIC microscopy and image analysis of bio¢lms. Binary 4, 93^95.
[19] Rogers, J. and Keevil, C.W. (1992) Immunogold and £uorescein immunolabelling of Legionella pneumophila within an
aquatic bio¢lm visualized by using episcopic di¡erential inter-
ference contrast microscopy. Appl. Environ. Microbiol. 58,
2326^2330.
[20] Lawrence, J.R., Korber, D.R., Hoyle, B.D., Costerton J.W.
and Caldwell, D.E. (1991) Optical sectioning of microbial bio¢lms. J. Bacteriol. 173, 6558^6567.
[21] Beveridge, T.J., Makin, S.A., Kadurugamuwa, J.L. and Li, Z.
(1997) Interactions between bio¢lms and the environment.
FEMS Microbiol. Rev. 20, 291^243.
[22] Beveridge, T.J., Hughes, M.N., Lee, H., Leung, K., Poole,
R.K., Savvaidis, I., Silver, S. and Trevors, J.T. (1996) Metal-microbe interactions: contemporary approaches. Adv. Microb. Physiol. 38, 178^243.
FEMSEC 930 3-8-98