The discovery of microbial life beneath the surface of
the continents and the seabed has shown that a large
proportion of all the bacteria and archaea on Earth
live in the deep biosphere
1
and that the energy supply
for microbial life that is present in this environment
spans many orders of magnitude. Even though so many
microorganisms reside in the deep biosphere — about
3 × 10
29
cells according to the most recent census
2
— the
energy flux available from buried organic carbon is less
than 1% of the photosynthetically fixed carbon on the
surface of our planet
1
. These microorganisms generally
have no cultivated relatives, and their biochemistry and
physiology are largely unknown. One of the greatest
enigmas, however, is how these vast communities can
subsist in conditions that provide only marginal energy
for cell growth and division and seem to barely support
the maintenance of basic cellular functions.
This Review explores the range of metabolic states in
cultivated cell populations under feast and famine con-
ditions and compares the lessons learned from labora-
tory cultures with our knowledge of energy-depleted
environ mental conditions. We present current estimates
of mean metabolic rates in natural communities and
discuss whether highly energy-starved microorganisms
have previously unexplored physiological mechanisms to
cope with low energyflux.
Energy metabolism in culture
Most of our knowledge about microbial physiology
and metabolism is derived from the study of organisms
in batch culture, but the pace of life, the physiological
states and the prominent organisms differ widely from
the Petri dish to the environment. Particularly when
considering oligotrophic natural systems such as the
deep biosphere, it is important to bear in mind that our
understanding of microbial energy metabolism stems
primarily from studying populations that are charac-
terized by rapid growth, high metabolic rates and high
cell densities — characteristics that do not apply to most
microorganisms innature.
The phases of life in batch culture. Populations in batch
culture transition through lag phase, exponential phase,
stationary phase, death phase and extended stationary
phase (FIG.1). The most heavily studied aspects of the
microbial life cycle are the growth (exponential) and
stationary phases. Growth phase is characterized by cells
operating at rates near the maximum allowed by their
enzyme kinetics (growing in exponential fashion) and
dividing with a time constant that is determined by the
physiology of the organisms concerned and the growth
conditions. Compilations across a broad range of cultured
taxa
3–7
show maximum growth rates that generally cluster
within a two orders of magnitude range, from less than
1 d
−1
to more than 30 d
−1
. At the high end of this range,
populations double in less than 30minutes, and for the
discussion that follows it is useful to recall that a typical
culture of Escherichia coli has a doubling time of about
20 minutes and transitions through the lag-to-death batch
culture life cycle within about 3days
8
.
Following stationary phase, most of the popula-
tion — typically 99% or more in batch cultures of E.coli
1
NASA Ames Research
Center, Mail Stop 239–4,
Moffett Field, California
94035-1000, USA.
2
Center for Geomicrobiology,
Institute of Bioscience,
Aarhus University,
Ny Munkegade 114,
8000 Aarhus C, Denmark.
*The authors contributed
equally to this work.
Correspondence to T.M.H.
e-mail:
tori.m.hoehler@nasa.gov
doi:10.1038/nrmicro2939
Deep biosphere
The set of ecosystems and
their organisms living beneath
the upper few metres of the
solid earth surface.
Extended stationary phase
A phase of the batch culture
life cycle characterized by the
persistence of a small fraction
of cells for months to years
beyond the death of the
majority of the culture, without
new addition of substrate.
Microbial life under extreme
energy limitation
Tori M.Hoehler
1
* and Bo Barker Jørgensen
2
*
Abstract | A great number of the bacteria and archaea on Earth are found in subsurface
environments in a physiological state that is poorly represented or explained by laboratory
cultures. Microbial cells in these very stable and oligotrophic settings catabolize 10
4
- to
10
6
-fold more slowly than model organisms in nutrient-rich cultures, turn over biomass on
timescales of centuries to millennia rather than hours to days, and subsist with energy
fluxes that are 1,000-fold lower than the typical culture-based estimates of maintenance
requirements. To reconcile this disparate state of being with our knowledge of microbial
physiology will require a revised understanding of microbial energy requirements, including
identifying the factors that comprise true basal maintenance and the adaptations that might
serve to minimize these factors.
REVIEWS
NATURE REVIEWS
|
MICROBIOLOGY VOLUME 11
|
FEBRUARY 2013
|
83
© 2013 Macmillan Publishers Limited. All rights reserved
Nature Reviews | Microbiology
10
10
Cells ml
–1
(3×10
4
d)
(2×10
5
d)
(10
5
–10
6
d)
e
a
b
c
d
f
g
h
10
8
10
6
10
4
10
2
10
–2
10
0
10
0
10
2
10
4
10
6
10
8
10
10
10
12
Duration of activity (d)
Cultures Natural populations
Y
ATP
Cellular growth yield
normalized to ATP
consumption.
— becomes non-viable
8
. The onset and extent of cell
death within a given species is reproducible
9,10
, but the
mechanisms responsible are not well understood. Finkel
8
suggests two possibilities. Cell death might simply result
from exhaustion of resources and consequent inability to
support a high-density population, or by accumulation of
inhibitory substances; the energy and nutrients provided
by the death and lysis of a fraction of the cells can then
support the remaining viable fraction. Alternatively, cell
death could be a programmed response that is induced by
the sensing of high cell density and resource limitation,
with subsequent sensing of newly sustainable conditions
following the death of most of the population inducing
the survivors to exit the programmed death pathway.
The cause and rate of cell death are crucially important
in the context of low-energy environments because the
energy required to support a steady-state population
varies dramatically depending on the level of cell turn-
over within that population. Biosynthesis is energetically
costly. The need to constantly resynthesize whole cells
(as opposed to coping with molecular attrition within a
cell) would greatly increase the effective per-cell energy
needed to support a steady-state population and might
exclude many low-energy environments from habitation.
Importantly, from this perspective, both of the mecha-
nisms envisioned to initiate cell death relate to high
population densities and resources that can be rapidly
depleted
8
. In both mechanisms, but particularly the
second, batch culture conditions are not representative of
the conditions found in low-energy environments. Thus
it is not clear whether, or to what extent, the death that
occurs in batch cultures is part of microbial life under
low energyflux.
The small fraction of cells that survive through the
death phase (which, even at a 99% attrition rate, can still
comprise cell densities of 10
6
–10
7
ml
−1
) has been shown
to persist for months or years without further addition
of substrate, forming what has been referred to as an
extended stationary phase
8,10,11
. Among the states repre-
sented in batch culture, extended stationary phase would
seem the most relevant for understanding life in low-
energy environments, but it must be borne in mind that
there are key differences in the nature of energy provision
and the relevant timescales and thus, potentially, also in
the prevailing physiological states. Extended stationary
phase in E.coli is characterized by the rapid emergence
of at least four distinct mutations that enhance the abil-
ity to catabolize amino acids
12–14
— that is, to subsist on
the remains of dead predecessors — and thereby confer
a growth advantage relative to wild-type cells or even
earlier generations of survivors
9,10
. The emergence and
succession of these growth advantage in stationary phase
(GASP) mutants drives population turnover that occurs
on monthly or shorter timescales and that can persist for
more than a year
10,15
. Such dynamic overturn in the pop-
ulation (with relatively high rates of growth replacing cell
attrition) reflects an energy-intensive state that is driven
by the initial availability of 100–1,000 dead cells’ worth
of energy for each survivor of death phase. Because it
seems unlikely that such dynamism could be supported
over the much longer timescales that are characteristic of
sustained microbial activity in natural settings
16
, caution
is warranted in considering even the extended stationary
phase as a model for life in low-energy settings.
Energy partitioning: Y
ATP
, growth and maintenance.
The energy yield of substrate catabolism is partitioned
among various cellular functions, including both those
that are related to and those that are not related to
growth
17–19
. The demonstration that energy is diverted
to processes other than growth is well illustrated by
the growth yield parameter Y
ATP
20
. Equating growth
to the catabolic yield of ATP, rather than to substrate
consumption, facilitates direct comparison of growth
yield across different taxa and substrates
18,20
, and permits
theoretical estimates of the biomass yield if ATP were
used exclusively for growth
21
. The many determinations
of Y
ATP
in cultures have demonstrated that the true yield
almost always falls significantly short of the theoretical
maximum (in E.coli, Y
ATP
is about one-third of theo-
retical maximum
21
), meaning that a substantive fraction
of energy is diverted to functions other than growth
22
.
Reported values for Y
ATP
vary by fivefold across different
organisms and growth conditions
21,23
, indicating that the
amount of energy diverted to non-growth functions is
highly variable. Some of the energy diverted away from
growth can be understood as ‘maintenance energy’ — the
flux of energy needed to sustain a steady-state population
of cells without net growth
24,25
(BOX1).
Figure 1 | Timescales, population sizes and biomass turnover times associated with
culture-based and natural environment studies. Blue points represent a hypothetical
batch culture experiment for an organism with a 1 h doubling time (after Finkel
8
). Red
points reflect measurements in natural systems. Values in parentheses represent estimates
of carbon turnover in terms comparable to doubling time. Time points of note are (a)
b
(cd
e
20 cmbsf (centimetres below sea floor)
70,71
f
(metres below sea floor)
70,71
g
65
(h
16
. The majority of our
knowledge about microbial physiology derives from the region approximated by a,
whereas a large proportion of bacteria and archaea on Earth live in circumstances
approximated by e–h.
REVIEWS
84
|
FEBRUARY 2013
|
VOLUME 11 www.nature.com/reviews/micro
© 2013 Macmillan Publishers Limited. All rights reserved
Box 1 | Maintenance energy
Nature Reviews | Microbiology
Dilution rate
Culture vessel
Stirrer
consumption rate (q)
Slope = 1/Y
G
Intercept = m
b
a
The advent of continuous culture (chemostat) tech-
niques provided a standard methodology for determin-
ing maintenance energy
26,27
. Tijhuis and colleagues
compiled more than 80 chemostat-based determinations
of maintenance energy to develop a biomass-specific
Gibbs energy consumption term
28
. The compiled values
ranged from 2.2 to 365 kJ C mol biomass
−1
h
−1
with a
suggested exponential (Arrhenius-like) dependence on
temperature. In the cell-specific units utilized below
(and assuming 1 C mol biomass = 24.6 g dry biomass
and an average dry cell mass of 10
−13
g), the derived cor-
relation predicts a maintenance energy for anaerobes at
25 °C of 3.3 × 10
−13
kJ cell
−1
d
−1
. For perspective, the pre-
dicted maintenance energy for anaerobes at human body
temperature of 37˚C, 9.7 kJ C mol biomass
−1
h
−1
, can be
compared with an approximate human dietary intake of
about 0.37 kJ C mol biomass
−1
h
−1
(for an example case
with 10,000 kJ d
−1
and 70 kg biomass consisting of 60%
“How would we express in terms of statistical theory that
marvellous faculty of a living organism, by which it delays
the decay into thermodynamic equilibrium (death)? … ‘It
feeds upon negative entropy’.” — Erwin Schrödinger
121
It has been understood for more than a century that life
must require a minimal flux of energy to support standing
biomass without net growth
122
, but the contributors to
‘maintenance energy’ are not well characterized or
quantified, and laboratory estimates exceed by many orders
of magnitude the energy fluxes that appear to support life
under conditions of low energy flux. The challenge is
summarized by van Bodegom
19
: “Many microbial ecologists
have a clear concept on the processes entangled in
physiological maintenance or endogenous metabolism, but
this is not what is measured empirically.” Pirt
24
described
maintenance as “energy consumed for functions other than
production of new cell material”, providing an operational
definition upon which to base observations (see below). But
this is more a description of what maintenance is not than
of what it is, and contributions to this pool can vary widely
depending on the physiological state of the population.
Van Bodegom defines eight distinct categories of
non-growth functions to which energy can be diverted,
and which are included in empirical determinations of
maintenance energy
19
: shifts in metabolic pathways,
energy spilling reactions, cell motility, changes in stored
polymeric carbon, osmoregulation, extracellular losses of
compounds not involved in osmoregulation, proofreading,
synthesis and turnover of macromolecular compounds
such as enzymes and RNA, and defence against O
2
stress.
It is clear that these vary in relative importance depending
on the organism in question, its environment and its
metabolic rate, and that several factors that are large
contributors at high metabolic rate are not requirements
for minimal maintenance.
Most empirical determinations of maintenance energy
have been made in chemostats, in which continuous flow
of medium supplies substrate at a constant rate and
likewise dilutes the microbial population at a constant
rate
24–26,123
. Part a of the figure shows a simplified
schematic of a chemostat. The experimenter varies the
dilution rate (medium inflow rate divided by total volume
of medium), which at steady state equals the growth rate,
μ, and observes the concentration of substrate in the
effluent (to determine substrate consumption rate) and
the standing biomass within the culture at steady state.
New cell growth balances cell removal, with a resulting
steady-state population size that is dependent on energy
flux. Operationally, population sizes are measured across
a range of dilution rates, and maintenance energies are
determined by extrapolating the results to a hypothetical
zero-dilution (zero growth) condition
18,124
. Part b of the
figure shows a Tempest-type plot for estimating the
maintenance coefficient, m, based on the relationship
q = μ/Y
G
+ m. Here, μ, the growth rate (in units of inverse
time), is equal to the experimentally varied dilution rate;
and q, the specific energy consumption rate (in units of, for
example, mol substrate g biomass
h
), is calculated from
the observed rate of substrate depletion and the steady
state biomass or population size. The slope of the plot
gives the inverse of the growth yield, Y
G
, and the intercept
(extrapolation to a hypothetical zero growth condition),
gives the maintenance coefficient, m.
The ability to achieve very low growth rates is limited in
practice, because dilution rates <0.05 h
(corresponding
can result in heterogeneity of the population and/or the
conditions within the chemostat
125
. Thus, it should be
borne in mind that chemostat-based observations always
reflect a physiological state of active growth at metabolic
rates that are much higher than those in energy-starved
natural populations. Continued development of culture
methodology might help to address these limitations.
‘Retentostats’ seek to maintain a true zero-growth
condition by retaining biomass using filter elements
32,38,39
,
and microfluidic chemostats
36,37
do the same by supplying
substrates via diffusion through chamber walls that are
impassable to cells — promising steps towards accessing
the physiological states that are relevant for understanding
life in energy-starved environments.
REVIEWS
NATURE REVIEWS
|
MICROBIOLOGY VOLUME 11
|
FEBRUARY 2013
|
85
© 2013 Macmillan Publishers Limited. All rights reserved
Basal power requirement
Energy turnover rate per cell
or per unit biomass associated
with the minimal complement
of functions required to sustain
a metabolically active state of
the cell.
Mean cell-specific metabolic
rates
Estimate of average cellular
metabolic rate among a whole
community of cells obtained by
measurement of bulk metabolic
process rates and cell numbers.
Primary productivity
The formation of living
organic biomass from carbon
dioxide through the process
of photosynthesis or
chemosynthesis.
Reaction-transport
modelling
Calculation of metabolic
process rates based on
steady-state concentration-
depth profiles and calculated
metabolite fluxes.
Power law
A mathematical relationship
between two quantities
describing how one quantity,
c, varies as a power of another
quantity, z: for example,
c = A × z
−b
, in which c could be
cell density, z sediment depth
(z>>0), and A and b constants.
water weight). The fact that the chemostat-based main-
tenance value is 26-fold higher than that for humans
— who could hardly be envisioned as functioning near
the minimal limits of required energy intake — empha-
sizes the need for caution in applying such values too
literally to the question of life in low-energy environ-
ments. Indeed, Morita
29,30
has suggested that mainte-
nance energy requirements in natural settings can be
3–6 orders of magnitude lower, and a few chemostat
and ‘retentostat’ studies
31,32
have yielded values that are
2–3 orders of magnitude lower than those seen in the
Tijhuis compilation.
In considering chemostat-based estimates of mainte-
nance energy, it should be borne in mind that ‘main-
tenance’ is operationally defined as any energy not
devoted to growth
19,24,25
, but this can comprise energy
‘wasted’ in spilling reactions
33–35
and energy spent in ‘use-
ful’ functions (for example, motility) that might become
superfluous in highly energy-limited conditions
19
. To dis-
tinguish from this operational definition and the added
energy costs it may comprise, we suggest the term basal
power requirement to describe the energy flux associated
with the minimal complement of functions required to
sustain a metabolically active state. One challenge facing
researchers in this area is the development of culture-
based approaches in which basal power requirements
can be quantified; efforts to modify continuous culture
methodology for this specific purpose
32,36–39
and to bring
microorganisms from low-energy natural settings into
continuous culture
40,41
both hold promise in this regard.
Importantly, however, although the state of basal main-
tenance might be elusive in culture, in nature it is prob-
ably expressed as the norm in low-energy communities.
Here, the challenge is to accurately characterize the
physiological states and identity of cells, along with
the often extremely low process rates and cell numbers
that are key to quantifying basal power requirements in
the environment.
Energy metabolism in the environment
Measuring mean cell-specific metabolic rates. What are
the maintenance states in energy-limited environments?
The published data on the mean cell-specific metabolic
rates of natural microbial communities, particularly
the extremely low rates calculated for deep subsurface
environments, are variable and often highly uncertain.
One of the early problems in accurately characterizing
deep communities — the retrieval of undisturbed and
uncontaminated samples — can now be solved to a
large extent by modern drilling and coring techniques
(BOX2). However, many additional challenges remain.
The determination of mean cell-specific rates requires
that bulk volumetric process rates and cell numbers be
quantified in parallel, a criterion that has not often been
met. Where it has been met, it is important to criti-
cally evaluate these results in light of the capabilities
and limitations of the methods used to measure these
parameters.
Whitman et al.
1
, who made the first bold extrapolation
of bacterial and archaeal cell numbers to global biomass
estimates, also estimated the average turnover time of
bacterial and archaeal biomass. They assumed that about
1% of total net primary productivity reaches the subsurface
and that the efficiency of cellular carbon assimilation is
0.2. This implies that the mean turn over time of bacterial
and archaeal biomass, and thus the mean generation time
of bacterial and archaeal cells, is 1,000years. From this
surprisingly long generation time the authors concluded
that most cells were probably inactive or even non-viable.
Parkes etal.
42
used incubation experiments with
35
S-labelled sulphate to measure the rate of respiratory
sulphate reduction in an ODP core from the continental
shelf off Peru. They found peak rates of 0.46 nmol
SO
4
2−
cm
−3
d
−1
at 1.5 m sediment depth with direct epi-
fluorescence cell counts showing 1.3 × 10
9
cells cm
−3
. If
it is assumed that one-tenth of the cells were sulphate
reducers with a carbon assimilation efficiency of 2 g
cellC mol
−1
substrate, then the mean rate of cell-specific
carbon metabolism (calculated by dividing the rate of
sulphate reduction per cm
3
by the number of sulphate
reducers per cm
3
) was 4 × 10
−4
fmol SO
4
2−
cell
−1
d
−1
and
the mean cell turnover was 30years. Compared to more
recent studies of sulphate reduction rates at similar sedi-
ment depth this rate is very high and might reflect the
fact that experimental measurements of the sulphate
reduction rate in deeper sediments tend to overestimate
the insiturate.
Rates of metabolism in natural systems can be quan-
tified by directly measuring chemical transformation
or by reaction-transport modelling. Incubations that track
the turnover of radio-labelled substrates can be used to
determine the rates of various processes
43,44
, but prac-
tical detection limits restrict their use in systems with
very low metabolic rates
45
. Reaction-transport model-
ling of metabolic substrates or products can extend
detectability to extremely low rates
46–48
. In systems with
slow water flow — for example, terrestrial subsurface
aquifers or mid-oceanic ridges — rates of metabolism
can be estimated from the change in water chemistry
in downstream boreholes divided by the flow velocity
between the holes
49–51
. However, uncertainties in flow
direction and velocity limit the accuracy of such esti-
mates. Below, we focus on sedimentary environments
where the restriction of mass transport predominantly
to molecular diffusion allows for accurate modelling of
substrate and product turnover.
The quantification of total microbial cell numbers
in subsurface sediments is generally done by direct
micro scopy after fluorescent DNA staining
52,53
. At very
low numbers, <10
5
cells cm
−3
, cells must first be extracted
from the sediment before microscopic counting
54
. Cell
numbers generally drop according to a power law from
around 10
9
cells cm
−3
near the sediment surface to 10
6
cells
cm
−3
hundreds of metres below
1,55
. Subsurface chemical
interfaces, such as the zone where energy-rich methane
from below diffuses up and meets sulphate-containing
sediment, can have cell numbers 10
2
–10
3
-fold above
the mean
45
. Sediment environments with extremely low
organic influx have cell numbers that can fall 10
2
–10
3
-
fold below the mean
56,57
. The practical detection limit set
by counting statistics is currently around 10
3
cells cm
−3
(REF.54).
REVIEWS
86
|
FEBRUARY 2013
|
VOLUME 11 www.nature.com/reviews/micro
© 2013 Macmillan Publishers Limited. All rights reserved
Nature Reviews | Microbiology
a
b
c
d
Bioturbated sediment
The uppermost part of the
seabed that is physically
reworked by animals.
Quantification of phylogenetically or physiologically
defined groups of microorganisms is done by fluores-
cence in situ hybridization (FISH) or catalysed reporter
deposition FISH (CARD-FISH) targeting ribosomal
RNA, or by quantitative PCR (qPCR) targeting the
16S ribosomal RNA gene or genes that are diagnostic
of distinct types of energy metabolism
58
(such as the
dsrA gene, which encodes dissimilatory sulphite
reductase, a metabolic enzyme found in sulphate-
reducing bacteria
59
). Such methods require previous
knowledge of the sequence diversity in the targeted
microorganisms and can be limited by insufficient
specificity or coverage of the hybridization probes or
primers used
60
. Sediment communities of bacteria and
archaea can also be quanti fied from their intact polar
membrane lipids
61
, assuming that these lipids originate
from viable cells. However, archaeal lipids in particular
can persist in the sediment after the death of the organ-
isms, so a proportion of the extracted intact polar lipids
can be of fossil nature
62,63
.
D’Hondt etal.
64
calculated the total rate of sulphate
reduction under 1 m
2
of the sea floor from the gradi-
ent in pore water sulphate and the resulting molecular
diffusion flux. For the open Pacific the sulphate flux
downwards into the seabed past 1.5 m depth was then
divided by the total cell number per m
2
down through
the sulphate zone. If it is assumed that one-tenth of the
cells are sulphate reducers, the mean cell-specific rate
in this deep and many million-years-old sediment
would be 4 × 10
−6
fmol SO
4
2−
cell
−1
d
−1
, and the estimated
mean cell turnover would be 3,000years. It should be
noted that this approach combines depth intervals of
very different metabolic rates and very different cell
numbers into one mean value, which will therefore
be skewed relative to the cell-specific activity at any
specificdepth.
Lomstein etal.
65
used a very different approach to
calculate microbial turnover. Their approach is based on
the built-in molecular clock in organic material in the
form of amino acids that very slowly undergo spontane-
ous racemization to the alternative stereo-isomeric form:
that is, l-amino acids to the d form and vice versa
66
. By
analysing the ratio between the l and d forms of aspar-
tate in the pool of total hydrolysable amino acids, the
authors could show that the bulk of subsurface amino
acids in >10
6
-year-old sediment were produced insitu
by now-dead bacteria (necromass). A simple steady-state
model (FIG.2) was developed to calculate the turnover
time of this bacterial necromass and thus the ~100-fold
faster turnover of the ~100-fold smaller amino acid pool
in living bacteria. The estimated turnover time of the
total microbial biomass in deep subsurface sediment was
200–2,000years, similar to the results obtained by Biddle
etal.
67
at the samesite.
Cell-specific metabolic rates. Relatively few studies have
simultaneously quantified both respiration rates and cell
numbers in deep marine sediments. Here, we examine
two such studies that reflect very different environments,
processes, timescales and methods for quantifying meta-
bolic rates. Despite these differences, the cell-specific
metabolic rates converge to similar values with increasing
depth and age in the sediment.
Sulphate reduction rates have been measured experi-
mentally, and sulphate-reducing microorganisms have
been quantified in marine sediments from the high
Arctic, the Baltic Sea and the Black Sea
68–73
. Intheupper-
most bioturbated sediment, the mean cell-specific rate of
sulphate reduction was around 0.1 fmol SO
4
2−
cell
−1
d
−1
Box 2 | Deep biosphere sampling
The retrieval of uncontaminated sample material is a great challenge for research on
the energy-deprived subsurface biosphere. Samples from deep sub-seafloor sediments
are obtained during expeditions with one of the scientific drill-ships JOIDES Resolution
or Chikyu, which are organized in international cooperation through the Integrated
Ocean Drilling Program (IODP; or its earlier phase, the ODP). The consistent recovery
and enumeration of cells in such carefully cored sediments is now accompanied by
rigorous contamination controls
55,126,127
: a dissolved perfluorocarbon tracer is
introduced into the seawater that is pumped down into the drill hole to flush out
suspended sediment and/or bacterial-size fluorescent beads are smeared around
the cored sediment to check for possible penetration of micro-particles into the
microbiology samples. Only non-contaminated sediment is used for microbiology
research.
Scientific drilling is not exactly a sterile operation. However, sediment cores with an
uncontaminated interior can now be brought rapidly on deck in transparent plastic
liners and sectioned by scientists for microbiological and biogeochemical research
(see the figure, parts a and b, which show the drilling and preparation of a 9.5 m-long
sediment core in a plastic liner in the eastern tropical Pacific Ocean during ODP
Expedition 201 on the JOIDES Resolution). Microorganisms can also be retrieved from
the more shallow subsurface by gravity coring down to many metres or even tens of
metres (see the figure, part c, which shows gravity coring in the Baltic Sea during
the ABC-2010 Expedition by the Center for Geomicrobiology, Aarhus University).
Microbiological subsampling from such cores can be done at high-depth resolution by
techniques that avoid penetration of microorganisms into the inner part of the core
(see the figure, part d, also from the ABC-2010 Expedition).
Drilling of deep ocean crust along the mid-oceanic ridges or of subsurface sediment
in tectonically active regions with flow of pore fluid poses very different demands on
sampling techniques. This is also the case for most land-based drilling
128,129
. It is difficult to
determine the cell-specific metabolic rates in such systems, which have therefore
not been included in this Review.
Photos in parts c and d
REVIEWS
NATURE REVIEWS
|
MICROBIOLOGY VOLUME 11
|
FEBRUARY 2013
|
87
© 2013 Macmillan Publishers Limited. All rights reserved
