Towards elucidation of microbial community metabolic pathways: Unraveling the network of carbon sharing in a pollutant-degrading bacterial consortium by immunocapture and isotopic ratio mass spectrometry

Abstract

Although much information on metabolic pathways within individual organisms is available, little is known about the pathways operating in natural communities in which extensive sharing of nutritional resources is the rule. In order to analyse such a consortium pathway, we have investigated the flow of 4-chlorosalicylate as carbon substrate within a simple chemostat microbial community using 13C-labelled metabolites and isotopic ratio mass spectrometric analysis of label enrichment in immunocaptured member populations of the community. A complex pathway network of carbon sharing was thereby revealed, involving two different metabolic routes, one of which is completely novel and involves the toxic metabolite protoanemonin. The high stability of the community results, at least in part, from interdependencies based on carbon sharing and the rapid removal of toxic metabolites.

Full text

Towards elucidation of microbial community metabolic
pathways: unravelling the network of carbon sharing in
a pollutant-degrading bacterial consortium by
immunocapture and isotopic ratio mass spectrometry
Oliver Pelz,
†
Michael Tesar, Rolf-Michael Wittich,
Edward R. B. Moore, Kenneth N. Timmis and Wolf-
Rainer Abraham*
Division of Microbiology, GBF-National Research Centre
for Biotechnology, Mascheroder Weg 1,
D-38124 Braunschweig, Germany.
Summary
Although much information on metabolic pathways
within individual organisms is available, little is
known about the pathways operating in natural com-
munities in which extensive sharing of nutritional
resources is the rule. In order to analyse such a con-
sortium pathway, we have investigated the flow of 4-
chlorosalicylate as carbon substrate within a simple
chemostat microbial community using
13
C-labelled
metabolites and isotopic ratio mass spectrometric
analysis of label enrichment in immunocaptured
member populations of the community. A complex
pathway network of carbon sharing was thereby
revealed, involving two different metabolic routes,
one of which is completely novel and involves the
toxic metabolite protoanemonin. The high stability
of the community results, at least in part, from inter-
dependencies based on carbon sharing and the
rapid removal of toxic metabolites.
Introduction
The activities of microorganisms have been studied
traditionally in monocultures, and important progress in
the elucidation of individual metabolic pathways and
underlying mechanisms has been gained thereby. How-
ever, in nature, microorganisms live in mixed communities
of various complexities that are generally characterized by
considerable metabolic and phylogenetic diversity. Little is
known about how microbial communities function as biolo-
gical units, how their activities are regulated by ecological
interactions between the community members or about
the metabolic routes that are followed by available nutri-
tional resources within the community as an entity. If we
are to understand microbial control of environmentalqual-
ity, the role of microbes in global change, the ecological
parameters regulating algal blooms, etc., we must extend
our understanding of the cellular metabolism of individual
organisms to that of the community as a biological entity,
analyse the sharing of resources and the functional roles
of the different members and characterize the principal
parameters and interactions that regulate the activities of
the community. A major hindrance to such investigations
is the paucity of adequate methods. However, the increas-
ing availability of stable isotope-labelled nutrients and the
exceptional sensitivity of measurement of isotope enrich-
ment in cell materials by isotopic ratio mass spectrometry,
coupled with the growing number of taxon-specific ‘bio-
markers’beingidentified,beginsto provide the experimen-
tal opportunity for investigating community metabolicroutes
and the critical ecological interactions that determine and
influence them.
We report here an analysis of carbon sharing in a
carbon-limited chemostat community growing on 4-
chlorosalicylate, an intermediate in the aerobic degrada-
tion of important organic environmental pollutants such
as 3-chlorodibenzofuran (Harm
et al
., 1991) and 2-chloro-
naphthalene (Morris and Barnsley, 1982). This study has
exposedanintricate network ofcarbon sharing in the com-
munity, defined the ecological roles of its three dominant
members and revealed that the substrate is catabolized
by two completely different parallel routes, one of which
is novel and involves protoanemonin as a critical inter-
mediate, a toxic substance not previously found as a path-
way intermediate in the microbial world (Blasco
et al
.,
1995; 1997).
Results
The experimental microbial community studied here was
grown as an aerated carbon-limited continuous culture in
a phosphate-buffered mineral salts medium containing
Environmental Microbiology (1999) 1(2), 167–174
䊚1999 Blackwell Science Ltd
Received 13 September, 1998; revised 4 December, 1998; accepted
17 December, 1998. †Present address: Swiss Federal Institute of
Technology (ETH Zu
¨rich), Institute of Terrestrial Ecology, Soil Biol-
ogy, Grabenstrasse 3, CH-8952 Schlieren, Switzerland. *For corre-
spondence. E-mail wab@gbf.de; Tel. (þ49) 531 6181 419; Fax
(þ49) 531 6181 411.

5mM 4-chlorosalicylate (4CS) (Faude, 1995). The inocu-
lum for the chemostat was taken from the aerobic zone
of the sediment of the Spittelwasser, a small tributary of
the Elbe River system near Bitterfeld, Sachsen-Anhalt,
Germany, which served for decades for the disposal of
untreated waste water from the chemical industries con-
centrated in this region. Until the recent shutdown of
most of these industries, the Spittelwasser contained no
eukaryotes, but a rich diversity of prokaryotes (Mau and
Timmis, 1998). Once the microbial community in the
chemostat reached equilibrium, it consisted principally of
four organisms, which were identified by determination
of their 16S rRNA gene sequences as two different
Pseu-
domonas
spp. (MT1 and MT4), an
Alcaligenes
sp. (MT3)
and an
Empedobacter
sp. (MT2). Polyclonal rabbit anti-
sera specific for each of the members of the community
were raised and used to measure the relative abundances
inthe chemostat. These were84⫾3% for MT1, 8 ⫾4%for
MT3 and MT4 and 1% for MT2. This community composi-
tionhas been stably maintained over a periodof more than
3years. ‘Daughter’ chemostats have been subjected to a
variety of perturbations, during which the composition of
the community was observed to change substantially,
but after which it eventually returned to that of the ‘parent’
chemostat (A. Rabenau, unpublished). As can be seen in
Fig. 1, the consortium tolerated a stepwise increase in
dilution rate from 0.16 day
¹1
to 0.64 day
¹1
; at such dilution
rates, all substrate was consumed, and no secreted
metabolites were observed. However, when the dilution
rate was increased to 0.8 day
¹1
, the toxic product proto-
anemonin(Blasco
etal
.,1997)accumulatedinandpoisoned
the chemostat.
Characterization of the catabolic potential of the indi-
vidual members of the community revealed that only
MT1 is able to transform and grow with the substrate
4CS as the sole source of carbon and energy. This
explains its high abundance in the chemostat. However,
in batch culture, it only tolerated substrate concentrations
as high as 1mM, above which it did not survive. Cultivation
of MT1 as a monoculture in an identical parallel chemostat
revealed that it could grow at the low dilution rate initially
selected, but it is very sensitive to small changes, and it
䊚1999 Blackwell Science Ltd,
Environmental Microbiology
,1, 167–174
Fig. 1. Effects of increasing dilution rate in chemostats of the community and of
Pseudomonas
sp. strain MT1. The initial dilution rate of 0.16
day
¹1
was increased to 0.2 day
¹1
(both chemostats), 0.24 day
¹1
,0.32day
¹1
, 0.4 day
¹1
, 0.48 day
¹1
, 0.64 day
¹1
and 0.8 day
¹1
(community
chemostat). The chemostat outflows were monitored for microbial densities (OD
600
), viable bacteria (plate counting) (A, community chemostat;
C, MT1 chemostat) and concentrations of substrate and metabolites (B, community; D, MT1 chemostat).
168
O. Pelz
et al
.

could not tolerate an increase in dilution rate. After
inoculation and during equilibration of the chemostat, 4-
chlorocatechol appeared transiently in the medium. More-
over, once equilibrium at the initial dilution rate of 0.16
day
¹1
was reached, a further metabolite, protoanemonin,
was observed. An increase in the dilution rate of only 25%,
to 0.2 day
¹1
, resulted in the accumulation of 3-chloro-
cis
,
cis-
muconate to a level of 4.5 ␮M and of protoanemonin to
43␮M, at which point bacterial growth stopped and wash-
out occurred.
Chlorocatechols are ordinarily channelled into produc-
tive chlorocatechol (type II
ortho-
cleavage) pathways
(Blasco
et al
., 1997). In the absence of such a pathway,
they may be partially catabolized by catechol (type I
ortho-
cleavage) pathways to the dead-end intermediate
protoanemonin (Blasco
et al
., 1995). These two pathways
have been studied intensively and are readily dis-
tinguished by the substrate preferences of certain
enzymes and the presence or absence of others. In
order to gain information on the degradation route of
4CS in the community, enzyme assays were carried out
on extracts prepared from biomass obtained from the
two chemostats. No catechol-2,3-dioxygenase activity,
the key enzyme of the
meta-
cleavage pathway (Gibson
and Subramanian, 1984), was detected in either extract.
As can be seen in Table 1, both extracts contain (chloro-)
catechol-1,2-dioxygenase activity. In the case of the
extractfrom the consortium, high activities weremeasured
for both catechol and 4-chlorocatechol, whereas in the
case of the MT1 extract, a low activity was measured for
4-chlorocatechol. Given the observation that MT1 com-
prises as much as 84% of the total community, the high
activity for 4-chlorocatechol in the community extract,
which presumably comes from the other members, is par-
ticularly impressive. The low
ortho-
cleavage activity for 4-
chlorocatechol of MT1 suggests that this organism may
not have a chlorocatechol pathway, a suggestion con-
firmed by the finding that MT1 extracts contained no
maleylacetate reductase activity, an enzyme characteris-
tic of type II pathways (Blasco
et al
., 1995). In contrast,
the consortium extract exhibited high maleylacetate
reductase activity. Thus, it seems that the community
exploits two different pathways for the degradation of
4CS, namely a type II chlorocatechol pathway and
another, probably a type I, pathway. If this latter pathway
were used, and it is known that protoanemonin can be
formed from 4-chlorocatechol via type I pathways (Blasco
et al
., 1995), then protoanemonin may be a productive
intermediate in the MT1 pathway, rather than a dead-
end product, at least when 4CS substrate concentrations
are low enough to avoid its accumulation at significant
levels.
This possibility was tested by offering MT1 [U-
13
C]-
labelled protoanemonin and determining whether it is
metabolized and the labelled carbon atoms incorporated
into phospholipids of cellular material. Labelled protoane-
monin was prepared by transformation of commercially
available [U-
13
C]-4-chlorocatechol by a cell extract of
MT1 (Blasco
et al
., 1995), purified and added as a pulse
to the MT1 chemostat. The initial concentration in the
chemostat, 50nM, is extremely low [1000-fold lower than
the minimum inhibitory concentration (MIC) for MT1
(Blasco
et al
., 1995)] and has no observable effect on
the bacteria. Phospholipids were subsequently isolated
from biomass collected from the community chemostat,
and the ␦
13
C-values of their fatty acids were measured
by means of a gas chromatograph coupled via a combus-
tion interface to an isotopic ratio mass spectrometer (GC-
C-IRMS; Abrajano
et al
., 1994). Phospholipid fatty acids,
particularly the C16:0 and C16:1, were found to be signifi-
cantly enriched in
13
C with a ⌬␦
13
Cþ3⫾0.5 enrichment
in C16:1 fatty acids compared with those from a control
chemostat fed with the same amount of non-labelled pro-
toanemonin. Protoanemonin can thus be taken up, chan-
nelled into central metabolic pathways and metabolized
to cellular components. This supports the possibility that
protoanemonin may serve as a productive metabolite in
4CS degradation.
Incubation of resting cells of MT1 with 1 mM 4-
chlorocatechol in sodium phosphate buffer resulted in
the formation of 25 ␮M protoanemonin, 17 ␮M 3-chloro-
cis
,
cis-
muconate and 16␮M
cis-
acetylacrylate. It has
been demonstrated recently that protoanemonin can be
transformed, albeit inefficiently, by the dienelactone
hydrolase of
Pseudomonas
sp. B13 to
cis-
acetylacrylate
(Bru
¨ckmann
et al
., 1998). We therefore tested whether
cis-
acetylacrylate can serve as a sole source of carbon
and energy for MT1 grown in minimal medium. This was
found to be the case. Resting cells of salicylate-grown
MT1 cells metabolized 1mM
cis-
acetylacrylate within
䊚1999 Blackwell Science Ltd,
Environmental Microbiology
,1, 167–174
Table 1. Enzymatic activities in extracts of the consortium and of
Pseudomonas
sp. MT1.
Consortium
Pseudomonas
sp. MT1
Enzyme Substrate (Ug
¹1
protein) (Ug
¹1
protein)
Catechol 2,3-dioxygenase Catechol 2 1
(Chloro-)catechol 1,2-dioxygenase Catechol 2104 779
4-Chlorocatechol 947 181
Maleylacetate reductase Maleylacetate; NADH 4052 30
Network of carbon sharing in a bacterial consortium
169

24h. These results confirm that protoanemonin and
cis-
acetylacrylate can serve as carbon sources for MT1 and
strongly suggest that 4CS is metabolized via a novel path-
way with 4-chlorocatechol, 3-chloro-
cis
,
cis-
muconate,
protoanemonin and
cis-
acetylacrylate as intermediates.
Given that MT1 can productively metabolize 4CS via a
catechol pathway involving protoanemonin as intermedi-
ate,the question arose as to what roles the othermembers
of the consortium play and why the consortium is more
robust when confronted with higher substrate concentra-
tions. We therefore analysed the fate of the major inter-
mediates released by MT1, namely 4-chlorocatechol and
protoanemonin. In order to do this, we introduced
[U-
13
C]-4-chlorocatechol or -protoanemonin as pulses
into the consortium chemostat and at intervals thereafter
collected samples. To determine the contribution of the
individual strains to the degradation of these intermedi-
ates, we used the strain-specific antibodies (Faude,
1995) to separate the consortium into its component popu-
lations by immunocapture. The level of
13
C-enrichment in
the bacteria-specific (White
et al
., 1979) C16:1 fatty acids
of the phospholipids extracted from the individual popula-
tions was then determined by isotopic ratio mass spectro-
metry. Pulse feeding of the consortium with 0.6 ␮M
[U-
13
C]-4-chlorocatechol resulted in an overall
13
C enrich-
ment in the consortium of ⌬␦
13
Cþ24 at 6h. The highest
13
Cenrichment of⌬␦
13
Cþ1557wasfoundfor
Alcaligenes
sp. MT3 at 2.5 h, while that for
Pseudomonas
sp. MT1 was
only þ126 at 6h and that for
Pseudomonas
sp. MT4 was
þ24 at 6h (Fig. 2A). These measurements demonstrate
that MT3 has an exceptionally high affinity for 4-
chlorocatechol and very effectively scavenges essentially
all the 4-chlorocatechol released by MT1. The enzyme
activities reported above are consistent with the assump-
tion that MT3 metabolizes 4-chlorocatechol via a chloro-
catechol pathway.
When 1.8␮M [U-
13
C]-protoanemonin was pulsed into
the chemostat, the
13
C enrichment in the total community
was 45.7 at 6h. The
13
C enrichment in C16:1 of
Pseudo-
monas
sp. MT4 was ⌬␦
13
C 17.6 at 2.5 h and considerably
lower, ⌬␦
13
C 3.0–4.64, for fatty acids of strains MT1 and
MT3 at 2.5h and 6h, although these values rose to
⌬␦
13
C 14.9 (MT1) and 10.8 (MT3) by 24h (Fig. 2B). The
rapid and substantial
13
C enrichment in the C16:1 of
MT4 demonstrated a high affinity of this organism for pro-
toanemonin and indicated that MT4 effectively scavenges
most of this metabolite as it is released by MT1. The slow
enrichment of
13
C in MT1 and MT3 might result from the
direct but slow uptake of residual protoanemonin from
the pulse or, more likely, from the uptake of
13
C-labelled
metabolites of protoanemonin, such as
cis-
acetylacrylate,
released by MT4.
Owing to the low abundance of
Empedobacter
sp. MT2,
it was not possible to determine the isotopic ratios of fatty
acids from immunocaptured cells of this community mem-
ber. However, analysis of the individualstrains of the con-
sortium in pure culture, using unlabelled yeast extract as a
carbon source and 0.2mM chlorocatechol, partially
enriched with [U-
13
C]-4-chlorocatechol to a ␦
13
C of 500,
revealed assimilation of 4-chlorocatechol into the biomass
of MT1, MT3 and MT4, but not MT2 (Pelz
et al
., 1997).
This result and the low abundance of MT2 suggests that
it survives on other metabolites and/or perhaps on cell
debris, with no active involvement in the degradation of
4-chlorosalicylate.
Discussion
Figure 3 summarizes the flow of chlorosalicylate sub-
strate carbon through the chemostat community, as
revealed by the data presented here. MT1 is the only
member able to metabolize 4-chlorosalicylate and consti-
tutes the dominant population in the chemostat. A signifi-
cant amount, almost 10%, of the substrate carbon spills
䊚1999 Blackwell Science Ltd,
Environmental Microbiology
,1, 167–174
Fig. 2. Incorporation of stable isotope-labelled metabolites into
biomarker fatty acids of individual consortium members. Aliquots of
0.6␮M labelled 4-chlorocatechol (A) or 1.8␮M labelled
protoanemonin (B) were added as a pulse to the chemostat
consortium. Samples were collected at the indicated times, the
individual bacterial populations separated by immunocapture and,
for each population, the isotopic ratios of bacterial biomarker
phospholipid fatty acids C16:1 were determined by GC-C-IRMS.
170
O. Pelz
et al
.

out from MT1 as the first metabolite, 4-chlorocatechol, and
is very efficiently taken up by MT3 and metabolized via
what is likely to be a classical chlorocatechol pathway.
MT3 thereby protects MT1 from the toxicity that would
otherwise accrue from a build-up of 4-chlorochatechol, a
metabolite presumably reflecting a kinetic bottleneck in
the MT1 pathway. MT1 also releases small amounts (in
the range of 1% or less of total substrate carbon) of the
next metabolite in the pathway, 3-chloro-
cis
,
cis
-muconate,
which may be taken up by MT3 and/or MT4. However, it is
the next metabolite in the pathway that represents a sec-
ond important carbon spill by MT1 and provides a major
surprise. Protoanemonin is a natural product of certain
members of the
Ranunculaceae
family (Seegal and
Holden, 1945) and has broad antibiotic activity (Didry
et al
., 1991). Although not a natural product found so far
in microorganisms, protoanemonin has been shown to
be formed during microbial metabolism of chlorinated aro-
matics as a toxic dead-end product that kills the producing
strain (Blasco
et al
., 1995) and, when produced in micro-
bial communities, is assumed to kill other members.
Washout of the MT1 chemostat at an only slightly elevated
dilution rate was correlated directly with the accumulation
of protoanemonin, so this observation is consistent with
earlier findings. The microbial community, however, is
extremely stable over a long period of time over a fourfold
range of dilution rates, the higher of which is presumably
characterized by high levels of protoanemonin release
by MT1. In this case, it is MT4 that takes up and metabo-
lizes protoanemonin efficiently and thereby protects MT1
from suicide by poisoning. MT4 thus has a productive
pathway for the metabolism of toxic protoanemonin. As
(i) resting cells of MT1 form protoanemonin and
cis-
acetylacrylate when incubated with 4-chlorocatechol; (ii)
at least one dienlactone hydrolase has been shown to be
able to convert protoanemonin to
cis-
acetylacrylate; and
(iii) MT1 is able to use
cis-
acetylacrylate as a sole source
of carbon and energy, it seems likely that the primary
route of 4-chlorosalicylate metabolism in the community,
namely by MT1, involves a novel pathway with toxic proto-
anemoninasa critical metabolite that is channelledvia
cis
-
acetylacrylate by as yet unknown reactions into the Krebs
cycle. This pathway probably also operates in MT4. In
addition, the classical chlorocatechol pathway operates
in parallel in the community, although it seems to be
responsible for the metabolism of less than 10% of the
total substrate carbon. Moreover, MT3 also has the capa-
city to catabolize either protoanemonin or a metabolite
thereof productively, although quantitatively this route
may not play a significant role in the metabolism of 4-
chlorosalicylate by the community.
In conclusion, the use of
13
C-labelled substrates
coupled with isotopic ratio mass spectometry to measure
isotopic enrichment in immunocaptured individual
populations of a stable microbial consortium is shown to
be a powerful approach for exploring community
metabolism and has revealed a highly complex system
of carbon sharing involving a catabolic pathway network
consisting of at least two routes, the major one of which
is novel. The community seems to be so stable because
each member plays a crucial role, either in providing
carbon skeletons for the others (MT1) or in scavenging
toxic metabolites that inhibit the primary degrader if they
accumulate. Additional interdependencies, such as
cross-feeding of growth factors, may well exist but were
not analysed in this study. From the results presented
here, it would seem that metabolic and physiological
weaknesses of primary degraders of xenobiotics may be
effectively compensated for by recruitment of other organ-
isms with appropriate complementary physiology to build
a consortium that, as a biological unit, is robust and able
to extract the maximal metabolic benefit from the nutri-
tional opportunity.
The period since 1950 has been a golden age in the elu-
cidation of metabolic routes within individual species. The
challengenowis to elucidate metabolic networks innatural
biological assemblages (and models thereof) in which
metabolites produced via a pathway of one cell type flow
to other cells to enter new pathways. The surprises
exposed in the present study are surely only a mere
taste of what is in store! Exploration of metabolic networks
in natural assemblages will ultimately yield an understand-
ing of the functioning of such assemblages as biological
units, of the roles of the component members and of
their metabolic, physiological and energetic benefits and
sacrifices as team players, and how the assemblage as
a unit interacts with its abiotic environment. Such an
understanding will provide the ground rules for interven-
tionsto influence environmentalprocesses and to optimize
biotechnological applications in the environment, such as
in situ
bioremediation.
Experimental procedures
Chemostat conditions
The microbial community studied here was grown in a 5 l ves-
sel as an aerated 3 l carbon-limited continuous culture in a
phosphate-buffered mineral salts medium (7.8 g of Na
2
H-
PO
4
×2H
2
O, 6.8g of KH
2
PO
4
, 410mg of MgSO
4
×7H
2
O,
10mg of NH
4
-Fe-citrate, 50mg of Ca(NO
3
)
2
×4H
2
O, 85mg
of NaNO
3
,13␮l of concentrated HCl, 0.7 mg of ZnCl
2
,1mg
of MnCl
2
×4H
2
O, 0.62mg of H
3
BO
3
, 1.9mg of CoCl
2
×6H
2
O,
0.17mg of CuCl
2
×2H
2
O, 0.24mg of NiCl
2
×6H
2
O,
0.36mg of NaMoO
4
×2H
2
Oin1lofH
2
O), containing
5mM 4-chlorosalicylate (4CS). The inoculum for the
chemostat was 10ml of sediment taken from the upper
2–3mm aerobic zone of the Spittelwasser. The chemostat
was maintained at 12⬚C, which is approximately the mean
䊚1999 Blackwell Science Ltd,
Environmental Microbiology
,1, 167–174
Network of carbon sharing in a bacterial consortium
171

䊚1999 Blackwell Science Ltd,
Environmental Microbiology
,1, 167–174
172
O. Pelz
et al
.

temperature of the Spittelwasser sediment, at adilution rate of
0.16 day
¹1
.
Enumeration of bacteria
Serial dilutions were made in 0.85% NaCl. Colony-forming
units (cfu) were determined from 100 ␮l of the appropriate
dilution spread on 1/10 Luria–Bertani (LB) medium (1g l
¹1
tryptone, 1g l
¹1
NaCl, 0.5g l
¹1
yeast extract) agar plates
at room temperature and counted on at least two different
plates after 3 days.
Generation of rabbit antisera
Bacteria were grown to an OD
600
of0.3, washedthreetimesin
PBS, fixed in PBS containing 2% formaldehyde, washed
again three times in PBS, aliquoted and stored at ¹70⬚C. A
1ml aliquot corresponding to an OD
600
of 0.3 was mixed
with an equal volume of incomplete Freund’s adjuvant
(Sigma) and the emulsion injected subcutaneously into a
female rabbit. Four booster injections were given at inter-
vals of 3 weeks, and test samples were taken at the
same time. All sera were tested in immunofluorescence
using all four chemostat isolates as pure cultures. Rabbit
antisera that showed ‘ring-like’ surface labelling of cells
were tested subsequently for cross-reactivity against other
chemostat isolates.
Indirect immunofluorescence
A formaldehyde-fixed chemostat sample (5 ␮l) was diluted in
PBS and filtered through a filter sandwich (0.2 ␮m) composed
of a polycarbonate filter (Costar), followed by a nitrocellulose
filter support (Sartorius). Bacteria were incubated for 2 h with
the rabbit antiserum diluted 1:250 in PBS supplemented with
2% fetal calf serum (FCS). After three washing steps in PBS
supplemented with 0.05% Tween 20 (PBST), a Cy3-labelled
sheep anti-mouse IgG antibody (Sigma) was used according
to the instructions of the supplier. After the final washing step,
cells were stained by SYBR stain (Molecular Probes), the
polycarbonate filter was removed and mounted in glycerol
supplemented with 2.5% 1,4-diazobicyclo[2,2,2]-octane (Merck)
on a microscopic slide. A fluorescence-equipped microscope
(Axioplan; Zeiss) was applied for cell counting, using the fol-
lowingfiltersets(Zeiss):#09forSYBRgreenstaining(Molecular
Probes; excitation: 450–490nm; emission: 520nm), #03 for
Cy3 (excitation: 546nm; emission: 590 nm).
Purification of antibodies and immunocapture
Rabbit antisera (3.6 ml) were diluted 1:2 in PBS and immuno-
globulins precipitated by the addition of saturated ammonium
sulphate to a final concentration of 60% (v/v; 4.5 ml). Immuno-
globulins were dissolved in 4.5ml of PBS and 1.5 ml of purified
on a protein A column, according to the instructions of the
supplier (Pharmacia). IgG thus purified was dialysed against
PBS and analysed for purity by electrophoresis on an SDS–
PAGE gel, followed by silver staining or Western blotting
with a rabbit-specific peroxidase-labelled antiserum (Promega).
In parallel, purified IgG was analysed for specificity by indirect
immunofluorescence with various test strains.
For immunocapture (Tamura
et al
., 1984; Chapman
et al
.,
1997), microtitre plates (Nunc) were coated at 4⬚C overnight
with 5–10 ␮g/well of protein A-purified rabbit antibodies.
Unbound antibodies were removed, and the plates were
washed three times in PBS by means of an enzyme-linked
immunosorbent assay (ELISA) washer and residual binding
capacities blocked by a 2% solution of BSA (Sigma). After
washing the plates three times in PBS, 100␮l of the chemo-
stat culture was added to each well and incubated for 2h.
Unbound cells were removed from the plates by repeated
washings in PBS (six times), and bound cells were subse-
quently released by the addition of 150 ␮l/well of 100 mM gly-
cine-HCl (pH3.0). The eluate was adjusted to approximately
pH7.0 by the addition of 750␮l of 1 M Tris-HCl (pH8.0) and
frozen at ¹20⬚C until further use.
Enzyme assays
The activity of key enzymes was determined according to
the procedure described by Blasco
et al
. (1995). High-
performance liquid chromatography (HPLC) analyses were run
according to the protocol described by Armengaud
et al
. (1998).
3-Chloro-
cis
,
cis
-muconate concentrations were calculated from
measured
cis-
dienelactone concentrations because this
compound is formed quantitatively on column from 3-chloro-
cis
,
cis-
muconate under acidic HPLC conditions (Armengaud
et al
., 1998).
Lipid analysis
Lipids were extracted using a modified Bligh–Dyer protocol,
separated into polarity classes, and the fatty acids of the
phospholipid fraction were converted to their methyl esters
as described previously (Abraham
et al
., 1997). Capillary
gas chromatographic analyses were performed on a Hewlett
Packard 5890 series II gas chromatograph equipped with a
capillary column HP Ultra 2 (5% diphenyl-, 95% dimethylpoly-
siloxane; 50m; inner diameter 0.2mm; film thickness
0.11mm). The oven programme was 150⬚C for 2min, 150⬚C
to 289⬚Cat4⬚Cmin
¹1
followed by an isothermal period of
11min. Hexadecenoic methyl ester C16:1 comprised 9- and
11-hexadecenoic methyl esters (C 16:1␻9 and C 16:1␻7),
which were not completely separated by GC-C-IRMS
because of the requirement to handle very low amounts
of substance. Therefore, both were determined as a single
compound and referred to as C 16:1.
䊚1999 Blackwell Science Ltd,
Environmental Microbiology
,1, 167–174
Fig. 3. Assignment of metabolic functions and interactions of members of the 4-chlorosalicylate metabolizing consortium. All metabolites
shown for MT1 have been identified. The thick solid arrows indicate the experimentally determined release and fate of the corresponding
metabolite in the community. The thick outline arrows indicate the experimentally determined release of small amounts of 3-chloromuconate
whose fate has not been determined. The thin arrows indicate the possible release of further metabolites (see text).
Network of carbon sharing in a bacterial consortium
173

The fatty acid methyl esters were analysed for their isotopic
ratios by means of a gas chromatograph coupled via a com-
bustion/reduction interface to a Finnigan MAT 252 isotopic
ratiomass spectrometer (GC-C-IRMS). Details of theanalysis
have been described elsewhere (Abraham
et al
., 1998).
Notation
The standardnotation for theexpression ofhigh-precision gas
isotope ratio mass spectrometry results in the ␦notation,
defined as
␦ð‰Þ¼ððRFAME=RPDBÞ¹1Þ¬103
where R
FAME
and R
PDB
are the
13
C/
12
C isotope ratios cor-
responding, respectively, to the sample and to the inter-
national internal standard PeeDee belemnite, a South
Carolinian carbonate rich in
13
C. The more
13
C a com-
pound contains the higher ␦
13
C becomes.
Acknowledgements
Christian Hesse is thanked for skilfully operating the
IRMS, and Birgit Jung for technical assistance in the anti-
body experiments. The International Atomic Energy Agency,
Vienna,Austria, is acknowledgedfor providing freereference
materials for the calibration of the IRMS. This work was sup-
ported by grants from the German Federal Ministry for
Science, Education and Research (project nos 0319433B
and 0319433C). K.N.T. expresses gratitude to the Fonds
der Chemischen Industrie for generous support.
References
Abraham, W.R., Meyer, H., Lindholst, S., Vancanneyt, M.,
and Smit, J. (1997) Phospho- and sulfolipids as biomarkers
of
Caulobacter
,
Brevundimonas
and
Hyphomonas
.
Syst
Appl Microbiol
20: 522–539.
Abraham, W.R., Hesse, C., and Pelz, O. (1998) Isotopic
ratios of carbon in microbial lipids as indicator of substrate
usage.
Appl Environ Microbiol
64: 4202–4209.
Abrajano, T.A., Murphy, D.E., Fang, J., Comet, P., and
Brooks, J.M. (1994)
13
C/
12
C ratios in individual fatty acids
of marine mytilids with and without bacterial symbionts.
Org Geochem
12: 611–617.
Armengaud, J., Happe, B., and Timmis, K.N. (1998) Genetic
analysis of dioxin dioxygenase of
Sphingomonas
sp. strain
RW1: catabolic genes dispersed on the genome.
J Bacter-
iol
180: 3954–3966.
Blasco, R., Wittich, R.M., Mallavarapu, M., Timmis, K.N., and
Pieper, D.H.(1995) From xenobiotic to antibiotic. Formation
of protoanemonin from 4-chlorocatechol by enzymes of the
3-oxoadipate pathway.
J Biol Chem
49: 29229–29235.
Blasco, R., Mallavarapu, M., Wittich, R.M., Timmis, K.N., and
Pieper, D.H. (1997) Evidence that the formation of protoa-
nemonin from metabolites of 4-chlorobiphenyl degradation
negatively affects the survival of 4-chlorobi-phenyl-come-
tabolizing microorganisms.
Appl Environ Microbiol
63:
427–434.
Bru
¨ckmann, M., Blasco, R., Timmis, K.N., and Pieper, D.H.
(1998) Detoxification of protoanemonin by dienelactone
hydrolase.
J Bacteriol
180: 400–402.
Chapman, P.A., Malo, A.T., Siddons, C.A., and Harkin, M.
(1997) Use of commercial enzyme immunoassays and
immunomagnetic separation systems for detecting
Escher-
ichia coli
0157 in bovine fecal samples.
Appl Environ
Microbiol
63: 2549–2553.
Didry, N., Dubreuil, L., and Pinkas, M. (1991) Antibacterial
activityofvaporsof protoanemonin.
Pharmazie
46:546–547.
Faude, U.C. (1995) Immunchemische Analyse mikrobieller
Lebensgemeinschaften. PhD Thesis, University of Leipzig,
Germany.
Gibson, D.T., and Subramanian, V. (1984) Microbial
degradation of aromatic hydrocarbons. In
Microbial Degra-
dation of Organic Compounds
. Gibson, D.T. (ed.). New
York: Dekker, pp. 181–252.
Harms, H., Wilkes, H., Sinnwell, V., Wittich, R.M., Figge, K.,
Francke, W.,
et al
. (1991) Transformation of 3-
chlorobenzofuran by
Pseudomonas
sp. HH 69.
FEMS
Microbiol Lett
65: 25–29.
Mau, M., and Timmis, K.T. (1998) Use of subtractive
hybridization to design habitat-based oligonucleotide
probes for investigation of natural bacterial communities.
Appl Environ Microbiol
64: 185–191.
Morris, C.M., and Barnsley, E.A. (1982) The cometabolism of
1- and 2-chloronaphthalene by pseudomonads.
Can J
Microbiol
28: 73–79.
Pelz, O., Hesse, C., Tesar, M., Coffin, R.B., and Abraham,
W.R. (1997) Development of methods to measure carbon
isotope ratios of bacterial biomarkers in the environment.
Isotopes Environ Health Stud
33: 131–144.
Seegal, B.C., and Holden, M. (1945) The antibiotic activity of
extracts of
Ranunculaceae
.
Science
101: 413–414.
Tamura, G.S., Dailey, M.O., Gallatin, M.W., McGrath, M.S.,
Weissman, I.L., and Pillemer, E.A. (1984) Isolation of
molecules by monoclonal antibodies and antisera: the
solid phase immunoisolation technique.
Anal Biochem
136: 458–464.
White, D.C., Davis, W.M., Nickels, J.S., King, J.D., and
Bobbie, R.B. (1979) Determination of the sedimentary
microbial biomass by extractable lipid phosphate.
Oecolo-
gia
40: 51–62.
䊚1999 Blackwell Science Ltd,
Environmental Microbiology
,1, 167–174
174
O. Pelz
et al
.