E¡ects of nutrient availability and Ochromonas sp. predation on
size and composition of a simpli¢ed aquatic bacterial community
Gianluca Corno1,2
1
CNR-Institute of Ecosystem Study, Verbania-Pallanza, Italy; and 2Max Planck Institute for Limnology, Plön, Germany
Correspondence: Gianluca Corno, CNRInstitute of Ecosystem Study, Largo Tonolli 50,
28922 Verbania-Pallanza, Italy. Tel.: 139
0323 518321; fax: 139 0323 556513;
e-mail: [email protected]
Received 7 December 2005; revised 14 April
2006; accepted 2 June 2006.
First published online 25 July 2006.
DOI:10.1111/j.1574-6941.2006.00185.x
Editor: Riks Laanbroek
Keywords
grazing on bacteria; prey–predator interaction;
phenotypic adaptation; filament; microcolony.
Abstract
Predation and competition are two main factors that determine the size and
composition of aquatic bacterial populations. Using a simplified bacterial community, composed of three strains characterized by different responses to predation, a short-term laboratory experiment was performed to evaluate adaptations
and relative success in communities with experimentally controlled levels of
predation and nutrient availability. A strain with a short generation time
(Pseudomonas putida), one with high plasticity in cell morphology (Flectobacillus
sp. GC5), and one that develops microcolonies (Pseudomonas sp. CM10), were
selected. The voracious flagellate Ochromonas sp. was chosen as a predator. To
describe adaptations against grazing and starvation, abundance, biomass and
relative heterogeneity of bacteria were measured. On the whole, the strains in the
predation-free cultures exhibited unicellular growth, and P. putida represented the
largest group. The presence of Ochromonas strongly reduced bacterial abundance,
but not always the total biomass. The activity of grazers changed the morphological composition of the bacterial communities. Under grazing pressure the
relative composition of the community depended on the substrate availability. In
the presence of predators, P. putida abundance declined in both high and low
nutrient treatments, and Pseudomonas CM10 developed colonies. Flectobacillus
was only numerically codominant in the nutrient-rich environments.
Introduction
The complexity of interrelations among trophic levels of the
microbial food web, the heterogeneity of prey and predators,
and the fast adaptation to change, result in a system with huge
intricacy. The impact of flagellate grazing on bacteria (topdown) often results in the development of numerous resistance strategies, including morphological adaptations,
toxicity, motility, cells communication, and exopolymer
production (reviewed by Pernthaler, 2005; and Matz &
Kjelleberg, 2005). Several studies have focused on the
different morphological adaptations of bacteria against grazing, and the development of resistant morphotypes has been
described for several strains (e.g. filamentous forms, Hahn
et al., 1999; Corno & Jürgens, 2006; microcolonies, Matz
et al., 2002).
The diversity of reactions to the presence of predators,
based on the resistance of different bacterial strains, can be
summarized by applying the concept of ‘prey heterogeneity’
to microbial communities (McCauley et al., 1988; Leibold,
1989, 1996; Abrams, 1993). Many strains can change their
2006 Federation of European Microbiological Societies
Published by Blackwell Publishing Ltd. All rights reserved
c
shape and size in order to become resistant to predation,
whereas only a few bacterial strains resist predation by
growing in an organized structure.
Bacterial community size and structure, however, may
not only be determined by top-down factors. The availability of nutrients is also recognized as a basic factor controlling
bacterial community characteristics (bottom-up control).
Higher amounts of available nutrients normally result in
higher total bacterial biomass and in a modification of the
size-structure of the community (Samuelsson et al., 2002),
in agreement with both theoretical and earlier empirical
studies (Thingstad & Sakshaug, 1990; Kivi et al., 1996;
Duarte et al., 2000). Due to higher biovolume-specific
uptake rates, small cells are expected to out-compete larger
forms in nutrient-poor conditions (Legendre & Rassoulzadegan, 1995). However, larger forms can establish when
small forms are saturated or limited by predation (Thingstad & Sakshaug, 1990).
Bacteria also exhibit morphological changes in response
to different nutrient conditions, as demonstrated in studies
on single strains (Holmquist & Kjelleberg, 1993) and natural
FEMS Microbiol Ecol 58 (2006) 354–363
355
Competition and predation on a microbial community
communities (Tuomi et al., 1995), where they can favour
different survival mechanisms and adaptive responses of
bacteria to the presence of protistan grazers (Jürgens &
Matz, 2002; Matz & Jürgens, 2003).
When grazing-resistant strains were compared based on
their competitiveness against strains unable to produce
resistant forms (Matz et al., 2002; Matz & Jürgens, 2003),
there was a clear advantage for the edible strains in the
absence of predators, usually due to a smaller size and
shorter generation time.
Despite huge scientific attention, there seems to be no
general consensus about how ‘top-down’ and ‘bottom-up’
factors interact to control population dynamics (Pernthaler,
2005).
In order to evaluate capabilities and relative ecological
success, three diverse bacterial strains (prey) were exposed to
predation by a single heterotrophic nano-flagellate (HNF)
strain in high and low productivity systems. The impact of
grazing on similar populations, subjected to different substrate conditions, was tested.
The physiology and the morphological abilities of the
selected strains were all very well described from previous
studies and, because of this knowledge, it was possible to
settle on a system that generates a true strategy competition.
Pseudomonas sp. CM10 (described by Matz et al., 2002),
usually lives in single small cells (cocci or rods with a
maximum diameter of 2–3 mm, Fig. 1a), and becomes
resistant to Ochromonas sp. grazing (the selected predator)
by developing clusters (several tens of cells) in the presence
of predators (Fig. 1d). The second strain, Flectobacillus sp.
GC5 (described by Corno & Jürgens, 2006) can become
resistant to grazing. Its usual shape is a C or an S composed
of one or two rods of 3–5 mm (shown also in Fig 1b) but, in
the presence of HNF, inedible filaments and chains of several
cells (for a final length of 10–40 mm, Fig. 1e) appear. The
third strain, Pseudomonas putida MM1 (used in Matz et al.,
2002; Matz & Jürgens, 2005; Corno & Jürgens, 2006), is a
completely edible strain. It is usually used as ideal food in
laboratory systems for several HNF species. Its shape is a
typical free-living coccus, it is small (2–3 mm, Fig. 1c and f),
and has a very fast generation time.
These three strains are ‘masters’ in their respective
abilities: making microcolonies, developing filaments, or
simply growing extremely fast without making any morphological defences against protist grazing.
Another ‘specialist’ in the field was chosen as a predator:
the ‘interception feeder’ Ochromonas sp. (Salcher et al., 2005,
following the description given by Fenchel, 1987), which is a
voracious mixotrophic bacterivorous nanoflagellate (Caron,
1987; Sanders, 1991; Posch et al., 1999) is readily used as a
predator in many studies on bacterial plasticity in laboratory
experimental systems (summarized in Boenigk & Arndt,
2002), often feeding on the same bacterial strains used in
FEMS Microbiol Ecol 58 (2006) 354–363
(a)
(d)
(e)
(b)
(c)
(f)
Fig. 1. Photomicrographs of DAPI-stained cells of the three bacterial
strains used in this study: Pseudomonas sp. CM10 (a, d), Flectobacillus
sp. GC-5 (b, e), Pseudomonas putida (c, f). Scale bar = 20 mm. The first
column refers to strains grown free from predators. On the second
column bacterial strains are under high grazing pressure by Ochromonas
sp.
this study (Hahn et al., 1999; Boenigk et al., 2001; Matz et al.,
2002; Corno & Jürgens, 2006).
The relative competitiveness of these three bacterial
strains was tested in short-term semi-continuous cultures
in the laboratory, in the presence or absence of Ochromonas
sp. in order to test the effectiveness of different grazing
resistance strategies with different amounts of available
substrate, and to describe the impact of the limiting factors
on the composition of this extremely simplified bacterial
community.
Materials and methods
Flectobacillus sp. GC5 was isolated from a continuous
culture system, enriched with HNF, and inoculated with a
mixed bacterial community from Lake Shöhsee (Germany).
A fragment of its 16S rRNA gene is 98% identical with
Flectobacillus mayor, the type species of the genus Flectobacillus (Larkin & Borrall, 1984). The accession number of the
2006 Federation of European Microbiological Societies
Published by Blackwell Publishing Ltd. All rights reserved
c
356
G. Corno
nearly full-length 16S rRNA gene sequence is DQ145723
(Corno & Jürgens, 2006). Pseudomonas sp. CM10 was
isolated from Lake Shöhsee by C. Matz; its 16S rRNA gene
sequence is deposited in GenBank (AF380369 and
AF380370, Matz et al., 2002). Flectobacillus sp. GC5 and
Pseudomonas sp. CM10 are available in the bacterial strain
collection of the CNR-Institute of Ecosystem Study (Italy).
Pseudomonas putida MM1, originally derived from rhizosphere of barley (Christoffersen et al., 1997), is available in
the bacterial strains collection of the Max Planck Institute
for Limnology (Germany), and its 16S rRNA gene sequence
is also deposited in GenBank (AY623928, Matz & Jürgens,
2005).
The strains were grown separately on agar plates and then
transferred to WC liquid medium (Guillard & Lorenzen,
1972) enriched with 50 mg L1 glucose. Clonal cultures of all
the strains were obtained from single colonies. Axenic
cultures of Ochromonas sp. isolated always from Lake
Shöhsee were maintained on suspensions of heat-killed P.
putida MM1 (Corno & Jürgens, 2006). From these cultures,
batches and semi-continuous experimental treatments were
inoculated.
Batch culture experiment
Batch culture experiments were performed in order to
compare the specific growth rates of the three selected
bacterial strains. A set of three replicates for each monoculture was performed. The flasks (100 mL) were carried out
on WC medium supplemented with 25.0 mg glucose L1 as
additional carbon source. A sample of 10 mL from the
precultures described above was inoculated in each replicate
obtaining three monocultural series of three replicates each.
All cultures were incubated in the dark for 6 days on a rotary
shaker at 15 1C and sampled every 12 h.
Fed batch culture experiment
The relative competitiveness of the three bacterial strains
was tested in semi-continuous cultures free from predators
(the flagellate Ochromonas sp.) in treatments called
‘ GRAZ’ and, at the same time, in other cultures under
grazing pressure (treatments ‘1GRAZ’). The bottles
(250 mL each) were carried out on WC medium enriched
with 2.5 or 25.0 mg glucose L1 (treatments ‘ GLC’ and
‘1GLC’ respectively). Twelve cultures, three for each possible combination of treatments, were established. Every
bottle was inoculated at day 0 with all bacterial strains to
obtain an initial population of a total 3 106 bacteria mL1
(1 106 bacteria mL1 for each strain) in all the treatments.
Where present, Ochromonas sp. concentration at day 0 was
500 cells mL1. The experimental duration was 9 days, during which time the microbial assemblages were incubated in
the dark at 15 1C and shaken for 15 min every 2 h. Substrate
was continuously provided from a reservoir with a dilution
rate of 0.02 h1 to maintain a relatively constant volume and
compensate for daily sampling (Matz et al., 2002).
The substrate availability for bacterial cells, and for units
of bacterial biovolume (mm3) during the period between day
3 and day 9 of the run was calculated daily by dividing the
amount of glucose supplied for the corresponding total
number of bacteria and total bacterial biovolume (Table 1).
Determination of bacterial and protistan
abundance, biovolume and biomass
Bacterial and Ochromonas sp. abundances were determined
daily using an epifluorescence microscope and staining
Table 1. Comparison of bacterial populations during growth in high (1GLC) and low ( GLC) productivity semi-continuous cultures in the presence
(1GRAZ) and in the absence ( GRAZ) of the bacterivorous flagellate Ochromonas sp
Treatments
GRAZ
1GRAZ
Bacterial community parameter
GLC
1GLC
GLC
1GLC
Total abundance (106 cells mL1)
Total biovolume (106 mm3 mL1)
Mean cell biovolume (mm3)
Total biomass (108 mg C L1)
Proportion of inedible bacteria (%)
Substrate availability (fg Glc bact1)
Substrate availability (fg Glc mm3)
Grazing pressure (HNF edible bact1)
4.98 0.11
2.58 0.02
0.51 0.10
6.17 0.81
1.39 0.44
0.49 0.03
0.96 0.12
—
22.21 1.32
12.87 0.11
0.58 0.02
30.11 2.22
2.10 0.51
1.13 0.06
1.97 0.14
—
1.60 0.29
1.14 0.41
0.71 0.15–
2.60 0.73
84.08 5.46
1.62 0.32
2.22 0.57
0.0094 0.0008
12.45 1.87
14.20 4.19–
1.14 0.57
33.22 9.8287.50 7.87
2.21 0.52
1.76 0.52
0.0072 0.0005
The period between days 3 and 9 of the run was analysed. Values, means of three replicates per treatment, are expressed as mean SD, or simple
arithmetical mean, or relative proportion.
Substrate availability for treatments 1GRAZ does not take into account the eventual substrate released by Ochromonas sp. Grazing pressure is
calculated as number of Ochromonas sp. cells per edible bacterial cell. Paired treatments (‘ GRAZ GLC’ against ‘1GRAZ GLC’ and
‘ GRAZ1GLC’ against ‘1GRAZ 1GLC’) were statistically compared ( –, P 4 0.05; , P o 0.01).
2006 Federation of European Microbiological Societies
Published by Blackwell Publishing Ltd. All rights reserved
c
FEMS Microbiol Ecol 58 (2006) 354–363
357
Competition and predation on a microbial community
formalin-fixed samples with DAPI (4 0 ,6-diamidino-2-phenylindole, Porter & Feig, 1980). Cell size measurements were
taken from DAPI-stained samples using an automated
image analysis system (Image-Pro Plus 5.1, Media Cybernetics). Area and perimeter of 300–500 cells were measured
for each sample according to the algorithms given in
Massana et al. (1997). Accurate counting for filaments and
microcolonies was done in order to reduce the risk of
underestimation of abundance and biovolume due to the
presence of extracellular polymeric substances (EPS), indistinct cell boundaries and 3D structure of these forms (Matz
et al., 2002). For filamentous forms, measurements of cell
dimensions were taken for at least 100 randomly selected
units per filter with the help of an ocular grid (Corno &
Jürgens, 2006). At least 30 microcolonies per sample were
measured for size and abundance (Hahn et al., 2000).
Bacterial cell carbon content (in fg C cell1) and bacterial
biomass (in mg C L1) were calculated following LofererKrößbacher et al. (1998).
Specific abundances of the three bacterial strains in
competition were determined using strain-specific polyclonal antibodies produced from rabbits immunized with
Flectobacillus sp. GC-5 and Pseudomonas sp. CM10 (Eurogentec, Herstal, Belgium). Staining with primary and secondary antibodies, and the assessment by epifluorescence
microscopy, were done with a modification of the procedure
described in Christoffersen et al. (1997). In order to properly
evaluate specific abundances, and to avoid possible overlapping, two different filters for each sample were prepared,
each one stained with a single specific antibody. Cross
reactivity between the two strains and with P. putida MM1
was not observed.
Definitions
Suspended, floclike structures consisting of three or more
bacterial cells were termed ‘microcolonies’ (see Fig. 1d;
Hahn et al., 2000). In order to assess properly the impact of
predation on the heterogeneous prey populations of bacteria, an appropriate definition of the edibility of the various
bacterial morphologies was required. By selecting Ochromonas sp. as predator, it was possible to define a size limit of
7 mm, below which bacteria were considered edible (regardless of the strain). When bacteria developed filamentous
forms, or chains of cells, longer than 7 mm, they appeared to
be resistant to predation (Corno & Jürgens, 2006). A similar
threshold was also identified for microcolonies: clusters of
each strain composed of more than 10 cells were considered
inedible for flagellates. The use of prey size limits for the
identification of the degree of edibility was questioned by
Wu et al. (2004), studying contradictory effects of the
predation of an Ochromonas sp. strain on several filamentous bacteria; on the other hand, prey size and morphology
FEMS Microbiol Ecol 58 (2006) 354–363
was recognized as a decisive factor in determining the
edibility of a bacterial prey for a class of predators in several
studies. The results of this study will be discussed in the light
of the problematic nature of the imposed size limit.
Statistical analyses
Bacterial and protistan dynamics and their relative proportions in all treatments were tested, between day 0 and day 9,
for significance using two-way repeated measures ANOVA,
with Bonferroni corrected t-tests posthoc comparisons,
except for the comparison of single-strain growth dynamics
in batch cultures, which were tested with a paired t-test.
Statistical analyses were performed with SigmaStat 3.0
packed with SigmaPlot 9.0 (Systat Software, Inc.).
Results
Single -strain batch cultures
The specific growth dynamics for each of the three strains
used in this study was considered in batch cultures, during
the phase of exponential growth (Fig. 2). Pseudomonas
putida reached the top of its phase of growth after 48 h, with
24.31 2.23 106 cells mL1 and a total biovolume of
12.03 1.10 106 mm3 mL1. Flectobacillus sp. GC5 grew
up to 10.43 1.35 106 cells mL1 (total biovolume of
14.84 1.91 106 mm3 mL1), but the top was reached after
60 h. Pseudomonas sp. CM10 reached the highest abundance
(21.77 2.56 106 cells mL1) and total biovolume (13.81 1.52 106 mm3 mL1) after 72 h.
The comparison of single-strain trends during the phase of
exponential growth (between 12 and 48 h after the inoculum)
showed that the number of cells for P. putida increased
significantly faster than the other two strains (P = 0.007
against Flectobacillus sp. GC5 and P = 0.026 against Pseudomonas sp. CM10, respectively, Fig. 2a), whereas comparing
Flectobacillus sp. GC5 and Pseudomonas sp. CM10 no significant advantage was noticed (P 4 0.05). Comparing population biovolumes (Fig. 2b) the situation was different:
growth rates measured for the three strains, even during the
exponential phase of growth, did not show any significant
difference (P 4 0.05 for all possible combinations).
General dynamics of bacterial and protistan
populations in semi-continuous cultures
Trends in the abundance and total biomass of bacterial and
protistan populations in treatments were considered to start
from day 3. This approach was required in order to skip the
huge fluctuations measured in the first 3 days, when
bacterial populations were exponentially growing and the
added flagellates were breaking previous bacterial trends to
equilibrium. Starting from day 3 it was possible to identify
2006 Federation of European Microbiological Societies
Published by Blackwell Publishing Ltd. All rights reserved
c
G. Corno
mL−1---------------------------------------------------------->
(a)
25x106
20x106
15x106
10x106
5x106
0
(b)
15x106
10x106
5x106
Pseudomonas putida
Flectobacillus GC5
Pseudomonas CM10
0
0
20
40
60
80
100
120
<---------------------------- Time (h) --------------------------->
<--------------------------------------------------cells
<------------------- µm3 mL–1------------------->
<-------------------cells mL–1------------------>
358
40x106
(a)
–GLC +GRAZ
–GLC –GRAZ
+GLC +GRAZ
+GLC –GRAZ
30x106
20x106
10x106
0
(b)
–GLC
+GLC
30000
20000
10000
0
0
1
2
3
4
<-------------------------------- Time
Fig. 2. Growth of Pseudomonas sp. CM10, Flectobacillus sp. GC-5, and
Pseudomonas putida in batches, without predators. (a) Shows abundances, while (b) shows corresponding biovolumes (means of three
replicates SD).
more stable dynamics in every treatment, and thus to
discuss them and to compare abundances and tendencies.
Bacterial abundances (Fig. 3a) were significantly positively
correlated with glucose concentration, in absence
(P o 0.001) and presence of Ochromonas (P o 0.01). Bacterial populations in treatments GLC and GRAZ
reduced fluctuations in number by the second day, after
which their growth rate was constant and the population
abundance remained between 5.67 106 and 4.72 106 cells mL1. The presence of Ochromonas sp.
(2.35 0.30 103 cells mL1, Fig. 3b) in treatments GLC
resulted in a reduction of bacterial abundance and a relative
reduction in the stability of the population, which fluctuated
between 2.01 106 and 1.20 106 cells mL1. Bacterial
abundances in bottles 1GLC were higher than those from
low productivity treatments in the absence and presence of
predators: for treatments GRAZ, a minimum of 20.99 106 cells mL1 and a maximum of 24.34 106 cell mL1 were
counted from days 3 to 9, at which point these populations
reached a clear equilibrium. In the presence of predators
(11.03 1.09 103 cells mL1, Fig. 3b) bacterial abundances
reduced by almost one-half, ranged between 16.68 106 and
6.58 106 cells mL1 and followed inconsistent trends.
2006 Federation of European Microbiological Societies
Published by Blackwell Publishing Ltd. All rights reserved
c
5
6
7
8
9
(d) ---------------------------------->
Fig. 3. Bacterial and flagellate development between days 0–9. Bacterial abundances (a), in high (1GLC) and low ( GLC) productivity systems,
in the presence (1GRAZ) or absence ( GRAZ) of predators. Ochromonas
sp. abundances (b) are related to treatments 1GRAZ. Values are
expressed as means of three replicates SD.
There were very high abundances in the first period of the
experiment, followed by dramatic reductions and large
subsequent oscillations.
A comparison between the means of the period from days
3 to 9 (Table 1) shows that the biggest bacterial assemblage
grew in the treatment 1GLC and GRAZ. Populations
from treatments 1GRAZ were always lower in number than
the corresponding GRAZ populations (P o 0.01).
The mean values for total bacterial biovolume and
biomass in 1GLC and 1GRAZ treatments were slightly
higher (P 4 0.05, Table 1) than the populations with the
same substrate concentration but without predators. This
result is related to a development of larger cells for bacterial
populations subjected to grazing pressure. It was not noticed
for GLC and 1GRAZ treatments, where the difference in
total biovolume and biomass between the two series was
statistically significant (P o 0.01) and comparable to the
difference measured in abundances. In GRAZ treatments
the impact of the amount of substrate supplied influenced
the average size of bacterial cells: smaller forms were noticed
in GLC treatments when compared with bacteria from
1GLC treatments (average biovolume per cell of 0.51 0.10
vs. 0.58 0.02, P o 0.01).
FEMS Microbiol Ecol 58 (2006) 354–363
359
Competition and predation on a microbial community
Assuming that bacterial populations from different treatments kept fairly constant growth rates during the period
from days 3 to 9 (Fig. 2), a calculation of the substrate
available for a single bacterial cell (and bacterial unit of
biovolume) was proposed (Table 1). Bacterial populations
in GLC treatments, in the absence of predators, obtained
less than 50% of the substrate available (both per cell and
per unit of biovolume) compared to bacteria in treatments
1GLC. In 1GRAZ treatments the substrate available per
cell was high, was unlikely to be limiting, and was in
comparable amounts between treatments (P 4 0.05 for both
measurements, per cell or per unit of biovolume).
Ochromonas sp. populations in 1GLC treatments were
subject to large fluctuations: the mean abundance during
the period from days 3 to 9 was 1.10 0.87 104 cells mL1
with a minimum of 0.45 104 cells mL1 at day 5
and a maximum of 2.87 104 cells mL1 at day 3. Grazers
in GLC treatments during the same period had a mean of
0.24 0.10 104 cells mL1, a minimum of 0.12 104
cells mL1 at day 7, and a maximum of 0.44 104 cells mL1
at day 3 (Fig. 3b).
(Fig. 4a and b). The opposite situation was observed in the
presence of grazers: the populations started with 95% edible
cells, switched to a majority of inedible cells by day 3 in
1GLC (74%, Fig. 4c) and in GLC (63%, Fig. 4d).
Following this trend, both treatments became dominated
by inedible forms with percentages 4 90% after day 5 (see
Table 1 for relative means).
Analysis of single -strain dynamics
By measuring the abundance of single strains and their
relative success in diverse populations, it is possible to link
them to the conditions of the treatment. The relative
proportion of every strain at time 0 was 33.3% by number,
in every bottle. In GRAZ treatments, P. putida overtook
the other two strains, independent of substrate supply. In
GLC treatments it reached 50% of the total population by
day 3 and 90% by day 7 (Fig. 5b). In 1GLC its success was
even faster, and P. putida reached 70% after 24 h (Fig. 5a). In
these treatments the other two strains were relegated to
percentages usually lower than 10–15%, and it became
impossible to determine which was better adapted for any
specific substrate conditions. There was a constant increment in the relative proportion of the strain P. putida in
both treatments, and this tendency held throughout the
whole experiment.
Similar proportions were not seen in 1GRAZ treatments,
and the development of the bacterial communities was
completely different. Pseudomonas putida seems to suffer
particularly in the presence of grazers, and in both conditions it was reduced to less than 10% of the total bacterial
population by day 4. Under conditions of nutrient limitation ( GLC, Fig. 5d) Pseudomonas sp. CM10 became
dominant by day 4 (60% of the total population), and by
the end of the experiment made up 90% of the total bacterial
Effectiveness of different prey classes in
different conditions
In order to obtain a clear description of developmental
tendencies, and thus of the relative success of the diverse
morphological classes, a description of class frequencies,
instead of absolute values, was chosen (Fig. 4).
At day 0, the percentage of edible cells was 92% of
the total number (87% of total biomass) for all
treatments. Populations from GRAZ, in 1GLC or GLC
treatments, did not change their morphological composition (P 4 0.05): the relative proportion of the mean abundance of edible forms by day 3 for these treatments was 95%
1
0.8
(a)
(b)
(c)
(d)
0.6
0.4
0.2
0
1
Fig. 4. Development of different bacterial
morphologies (edible: bacterial cells shorter
than 7 mm or clusters composed by fewer than
10 cells; and inedible), measured on bacterial
abundances, in the absence (a, b) or presence
(c, d) of Ochromonas sp. and during the period
of days 0–9. (a) And (c) refer to treatments 1GLC
(enriched with 25.0 mg glucose L1), (b) and
(d) to GLC (enriched with 2.5 mg glucose L1).
FEMS Microbiol Ecol 58 (2006) 354–363
0.8
0.6
0.4
0.2
0
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
2006 Federation of European Microbiological Societies
Published by Blackwell Publishing Ltd. All rights reserved
c
360
G. Corno
1
0.8
(a)
(b)
0.6
–Graz
+Glc
–Graz
–Glc
0.4
Flectobacillus sp. GC5
Pseudomonas sp. CM10
P. putida
0.2
0
1
0.8
(c)
0.6
+Graz
+Glc
(d)
+Graz
–Glc
0.4
0.2
0
0
1
2
3
4
5
6
7
8
9
0
1
2
3
population. Pseudomonas sp. CM10 also grew well in 1GLC
(Fig. 5c), where it reached about 50% of the total population, but was unable to outcompete Flectobacillus sp. GC5,
which grew in the very same proportion with long filaments
and chains of rods.
Discussion
The relative impact of predation by HNF (top-down control) and nutrient availability (bottom-up control) on
abundance and biomass of natural and artificial bacterial
communities was measured in several studies, but there are a
few generally accepted concepts that govern microbial
dynamics (summarized in Pernthaler, 2005). In this study,
the presence of predators led to a reduction of bacterial
abundance in both high (1GLC) and low ( GLC) productivity systems (Fig. 3), and the bacterial populations in
1GLC were more abundant than in GLC, regardless of
the presence or absence of predators.
Moreover, when evaluating the shift in bacterial biomass
due to the presence of predators, it was clear that grazing
had contradictory effects on prey populations (Table 1): in
GLC grazers reduced total bacterial biomass, but in
1GLC biomass of bacterial communities was unaffected by
predation, even if the number of bacteria was half that in
GRAZ. This was not unexpected and can be explained by
the additional source of organic carbon represented by
Ochromonas sp. exudates (Pernthaler et al., 1997) that are
directly available for bacteria, and by the resistance strategy
(high phenotypic plasticity, used to develop inedible long
filaments or chains of cells) of Flectobacillus sp. GC5 (further
described in detail) which results in an increase of its single
cell biomass under grazing pressure (Corno & Jürgens,
2006).
The small amount of substrate available for bacterial
populations in GLC treatments, drives the development
2006 Federation of European Microbiological Societies
Published by Blackwell Publishing Ltd. All rights reserved
c
4
5
6
7
8
9
Fig. 5. Relative abundance of the three bacterial
strains, during the period of days 0–9. Treatments free from predators are represented in
(a) and (b); (c) and (d) refer to populations
subjected to Ochromonas sp. grazing. (a) and
(c) refer to treatments 1GLC (enriched with
25.0 mg glucose L1), (b) and (d) to GLC
(enriched with 2.5 mg glucose L1).
of small forms: increasing the amount of substrate supplied
(treatments 1GLC) increased the average bacterial cell size
significantly. On the other hand, species composition and
relative abundances were not influenced by the amount of
substrate supplied. In 1GRAZ treatments the average cell
size was bigger, and the impact of substrate supplied less
evident, also because the amount of nutrients available per
cell were much higher and probably has less impact on
bacterial populations than predation.
Since Ochromonas sp. is considered an ‘interception
feeder’, calculating the chance that an edible bacterial cell
will meet the predator (Table 1) is a valid index of the
grazing pressure on the different communities: the possibility for a flagellate to encounter an inedible filament, or a
cluster of cells is not increasing the risk for the bacterial
community, and thus is not considered. Grazing pressure
was slightly higher (P o 0.05) in GLC treatments where
this impact, coupled with the low amount of substrate
supplied, resulted in a reduction of more than 50% of the
total bacterial biomass.
Predation and nutrient availability both control the
bacterial community, reciprocally increasing or reducing
their respective impact. This confirms the speculations of
Gasol (1994) and Pace & Cole (1994) that ‘top-down’
control by protistan grazing or ‘bottom-up’ control by the
availability of organic carbon and nutrients might be related
to overall ecosystem productivity both in marine and freshwater microbial food webs.
Moreover, these findings are consistent with comparative
analyses (Gasol et al., 2002) and theoretical models (Thingstad & Lignell, 1997), which show that bacteria are more
intensely controlled by protistan predation in nutrient-poor,
oligotrophic systems than in more productive environments.
The morphological composition of the heterogeneous
prey community was highly controlled by grazing (Fig. 4).
FEMS Microbiol Ecol 58 (2006) 354–363
361
Competition and predation on a microbial community
By simply dividing the community into edible and inedible
bacteria, it was possible to identify two contrasting patterns
of development in the bacterial populations: (1) any increase, neither with higher population growth rates, of the
proportion of costly resistant morphologies in GRAZ
treatments, and (2) fast shifts to inedible forms in 1GRAZ
treatments. The prey size limits used for the definition of the
degree of edibility of bacterial cells (in this study simply
‘edible’ or ‘inedible’) were respected by the well known
predilection of nanoflagellates as Ochromonas sp. for medium-sized bacteria (González et al., 1990; Šimek & Chrzanowski, 1992). It was shown that even filamentous bacteria
are not absolutely resistant to HNF grazing but have a
selective advantage due to a strongly reduced ingestion
efficiency by bacterivorous flagellates (Wu et al., 2004). The
results obtained confirm the validity of these thresholds, as
already demonstrated (Matz et al., 2002; Corno & Jürgens,
2006) and point out a clear morphological advantage for
cells exceeding the assigned limits. Nevertheless, in this
study any direct observation of the strategy of predation
was made and thus it is possible only indirectly to confirm
that cell size and morphology represent a refuge against
HNF grazing.
In any case, even if the evaluation of prey edibility is a
useful tool (especially for HNFs), for an appropriate assessment of the effective success of the single bacterial strains it
was necessary to observe their development in more detail.
In this experiment it was possible to trace the bacterial
strains throughout the duration of the experiment, and,
finally, to confirm that, despite the fact that the three species
selected were true ‘masters’ in their specialty, it was impossible to find a ‘winning strain’ that was able to outcompete the competitors independently of the environmental conditions.
For the first time in this study the exponential growth
phases of the three well known bacterial strains used were
compared. Simply considering the dynamics, measured on
bacterial abundances (Fig. 2a), it was possible to confirm
that P. putida is definitely the strain with faster generation
time in our systems. Because of this ability, P. putida
absolutely dominated the predator-free communities (Fig.
5) generally developing in small free-living cocci. Pseudomonas putida controlled those communities independently by
the amount of substrate available: they did not seem to
suffer at all the presence of other competitors growing with
very similar abundances and dynamics in treatment 1GLC
with the other strains, and in the preliminary batch cultures
as single strain. As a result, the other strains were relegated
to a small minority of the total population. This ‘excluding
trend’ imposed by P. putida may have resulted in the
extinction of the other two strains in a longer experiment.
Pseudomonas putida was not able to survive grazing by
Ochromonas sp. and in fact, the occurrence of predators
FEMS Microbiol Ecol 58 (2006) 354–363
‘killed the potential winner’, as theorized by Thingstad
(2000) for virus predation and by Beardsley et al. (2003)
for protists. In the 1GRAZ treatments the ecological niches
left by P. putida were thus available for strains able to survive
to the new limiting factor: predation by Ochromonas. With
little substrate supply ( GLC), the competition was rapidly
won by Pseudomonas sp. CM10. In 1GLC treatments
Flectobacillus sp. GC5 was able to compete with Pseudomonas sp. CM10, and, by the end of the experiment, its relative
proportion was comparable with the CM10 strain. The
different evolution of the bacterial communities observed
in the presence of predators can be explained by the different
strategies developed by the two strains. It is obvious that
aggregation costs in terms of the effectiveness of nutrient
uptake (Matz et al., 2002), but even a reduced uptake ability
is enough to outcompete Flectobacillus sp. GC5. When
subjected by high grazing pressure by HNF this strain
develops long filaments and chain of cells: these morphologies are known not to be very competitive, in conditions of
low or normal nutrient availability, and without grazing
pressure. This is due to the huge cell size (filaments up to
40 mm long) and to the long intracellular distances that
raising the nutrient demand and reducing the velocity of
biological answers respectively, results in a largely disadvantaged morphology (Koch, 1996). The situation changes with
high level of nutrient availability and with high grazing
pressure by HNF: in this case, the strategy of Flectobacillus
sp., filament formation, becomes efficient: these findings
seem to limit its success to environments where HNF
predation is coupled with high amounts of available nutrients.
Because of the reductionistic nature of this approach, and
of the many simplifications introduced in the system, this
study is only a simple model of real communities and the
natural environment. For this reason future studies, using
new tools [such as catalyzed reporter desposition-fluorescence in situ hybridization (CARD-FISH) coupled with
microautoradiography and single cell sorting], are required
to enlarge the number of tracked bacterial species, the
diversity of predators and of their trophic levels.
Acknowledgements
The author is grateful to Carsten Matz and Klaus Jürgens for
providing Pseudomonas sp. CM10. This work benefited from
the insight and technical assistance of the staff of the
microbial ecology laboratories of the Max Planck Institute
for Limnology and of the CNR-Institute of Ecosystems
Study. The author is also grateful to Peter Deines, Cristiana
Callieri, Carsten Matz, Blake Matthews and to the anonymous reviewers for valuable comments and suggestions on
the manuscript.
2006 Federation of European Microbiological Societies
Published by Blackwell Publishing Ltd. All rights reserved
c
362
References
Abrams PA (1993) Effect of increased productivity on the
abundances of trophic levels. Am Nat 141: 351–371.
Beardsley C, Pernthaler J, Wosniok W & Amann R (2003) Are
readily cultured bacteria in coastal North Sea waters
suppressed by selective grazing mortality? Appl Environ
Microbiol 69: 2624–2630.
Boenigk J & Arndt H (2002) Bacterivory by heterotrophic
flagellates: community structure and feeding strategies.
Antonie van Leeuwenhoek 81: 465–480.
Boenigk J, Matz C, Jürgens K & Arndt H (2001) The influence of
preculture and food quality on the ingestion and digestion
process of three species of heterotrophic nanoflagellates.
Microb Ecol 42: 168–176.
Caron DA (1987) Grazing of attached bacteria by heterotrophic
microflagellates. Microb Ecol 13: 203–218.
Christoffersen K, Nybroe O, Jürgens K & Hansen M (1997)
Measurement of bacterivory by heterotrophic nanoflagellates
using immunofluorescence labelling of ingested cells. Aquat
Microb Ecol 13: 127–134.
Corno G & Jürgens K (2006) Direct and indirect effects of protist
predation on population size structure of a bacterial strain
with high phenotypic plasticity. Appl Environ Microbiol 72:
78–86.
Duarte C, Agustı́ S, Gasol JM, Vaqué D & Vazquez-Dominguez E
(2000) Effects of nutrient supply on the biomass structure on
planktonic communities: an experimental test on an
Mediterranean coast community. Mar Ecol Prog Ser 206:
87–95.
Fenchel T (1987) Ecology of Protozoa. The Biology of Freeliving
Phagotrophic Protists, pp. 32–52. Science Tech. Publishers,
Madison, WI.
Gasol JM (1994) A framework for the assessment of top-down vs
bottom-up control of heterotrophic nanoflagellate abundance.
Mar Ecol Prog Ser 113: 291–300.
Gasol JM, Pedros-Alio C & Vaqué D (2002) Regulation of
bacterial assemblages in oligotrophic plankton systems: results
from experimental and empirical approaches. Antonie Van
Leeuwenhoek 81: 435–452.
González JM, Sherr EB & Sherr BF (1990) Size-selective grazing
on bacteria by natural assemblages of estuarine flagellates and
ciliates. Appl Environ Microbiol 56: 583–589.
Guillard R & Lorenzen C (1972) Yellow-green algae with
chlorophyllide c. J Phycol 8: 10–14.
Hahn MW, Moore ERB & Höfle MG (1999) Bacterial filament
formation, a defense mechanism against flagellate grazing, is
growth rate controlled in bacteria of different phyla. Appl
Environ Microbiol 65: 25–35.
Hahn MW, Moore ERB & Höfle MG (2000) Role of microcolony
formation in the protistan grazing defense of the aquatic
bacterium Pseudomonas sp. MWH1. Microb Ecol 39: 175–185.
Holmquist L & Kjelleberg S (1993) Changes in viability,
respiratory activity and morphology of the marine Vibrio sp.
strain S14 during starvation for individual nutrient and
subsequent recovery. FEMS Microbiol Ecol 12: 215–224.
2006 Federation of European Microbiological Societies
Published by Blackwell Publishing Ltd. All rights reserved
c
G. Corno
Jürgens K & Matz C (2002) Predation as a shaping force for the
phenotypic and genotypic composition of planktonic bacteria.
Antonie van Leeuwenhoek 81: 413–434.
Kivi K, Kuosa H & Tanskanen S (1996) An experimental study on
the role of crustacean and microprotozoan grazers in the
planktonic food web. Mar Ecol Prog Ser 136: 59–68.
Koch AL (1996) What size should a bacterium be? A question of
scale. Annu Rev Microbiol 50: 317–348.
Larkin J & Borrall R (1984) Family I. Spirosomaceae. Bergey’s
Manual of Systematic Bacteriology, Vol. 1, pp. 125–132. (Krieg
N & Holt J, eds). Williams and Wilkins, Baltimore, MD.
Legendre L & Rassoulzadegan F (1995) Plankton and nutrient
dynamics in marine waters. Ophelia 41: 153–172.
Leibold MA (1989) Resource edibility and the effects of predators
and productivity on the outcome of trophic interactions. Am
Nat 134: 922–949.
Leibold MA (1996) A graphical model of keystone predators in
food webs: trophic regulation of abundance, incidence and
diversity patterns in communities. Am Nat 147: 784–812.
Loferer-Krößbacher M, Klima J & Psenner R (1998)
Determination of bacterial cell dry mass by transmission
electron microscopy and densitometric image analyses. Appl
Environ Microbiol 64: 688–694.
Massana R, Gasol JM, Björnsen PK, Blackburn N, Hagström A,
Hietanen S, Hygum BH, Kuparinen J & Pedrós-Alió C (1997)
Measurement of bacterial size via image analysis of
epifluorescence preparations – description of an inexpensive
system and solutions to some of the most common problems.
Sci Mar (Barc) 61: 397–407.
Matz C & Jürgens K (2003) Interaction of nutrient limitation and
protozoan grazing determines the phenotypic structure of a
bacterial community. Microb Ecol 45: 384–398.
Matz C & Jürgens K (2005) High motility reduces grazing mortality
of planktonic bacteria. Appl Environ Microbiol 71: 921–929.
Matz C & Kjelleberg S (2005) Off the hook – how bacteria survive
protozoan grazing. Trends Microbiol 13: 302–307.
Matz C, Deines P & Jürgens K (2002) Phenotypic variation in
Pseudomonas sp. CM10 determines microcolony formation
and survival under protozoan grazing. FEMS Microbiol Ecol
39: 57–65.
McCauley E, Murdoch WW & Watson S (1988) Simple models
and variation in plankton densities among lakes. Am Nat 132:
383–403.
Pace ML & Cole JJ (1994) Comparative and experimental
approaches to top-down and bottom-up regulation of
bacteria. Microb Ecol 28: 181–193.
Pernthaler J (2005) Predation on Prokaryotes in the water
column and its ecological implications. Nat Rev Microbiol 3:
537–546.
Pernthaler J, Posch T, Šimek K, Vrba J, Amann R & Psenner R
(1997) Contrasting bacterial strategies to coexist with a
flagellate predator in an experimental microbial assemblage.
Appl Environ Microbiol 63: 596–601.
Porter KG & Feig YS (1980) The use of DAPI for identifying and
counting aquatic microflora. Limnol Oceanogr 25: 943–947.
FEMS Microbiol Ecol 58 (2006) 354–363
363
Competition and predation on a microbial community
Posch T, Šimek K, Vrba J, Pernthaler J, Nedoma J, Sattler B, Sonntag
B & Psenner R (1999) Predator-induced changes of bacterial
size-structure and productivity studied on an experimental
microbial community. Aquat Microb Ecol 18: 235–246.
Salcher MM, Pernthaler J, Psenner R & Posch T (2005) Succession
of bacterial grazing defense mechanisms against protistan
predators in an experimental microbial community. Aquat
Microb Ecol 38: 215–229.
Samuelsson K, Berglund J, Haeck P & Andersson A (2002)
Structural changes in an aquatic microbial food web caused by
inorganic nutrient addition. Aquat Microb Ecol 29: 29–38.
Sanders RW (1991) Trophic strategies among heterotrophic
flagellates. The Biology of the Free-Living Heterotrophic
Flagellates (Patterson D & Larsen J, eds), pp. 21–38. Clarendon
Press, Oxford.
Šimek K & Chrzanowski TH (1992) Direct and indirect evidence
of size-selective grazing on pelagic bacteria by freshwater
nanoflagellates. Appl Environ Microbiol 58: 3715–3720.
FEMS Microbiol Ecol 58 (2006) 354–363
Thingstad TF (2000) Elements of a theory for the mechanisms
controlling abundance, diversity, and biogeochemical role of
lytic bacterial viruses in aquatic systems. Limnol Oceanogr 45:
1320–1328.
Thingstad TF & Sakshaug E (1990) Control of phytoplankton
growth in nutrient recycling ecosystems. Theory and
terminology. Mar Ecol Prog Ser 63: 261–272.
Thingstad TF & Lignell R (1997) Theoretical models for the
control of bacterial growth rate, abundance, diversity and
carbon demand. Aquat Microb Ecol 13: 19–27.
Tuomi P, Fagerbakke KM, Bratbak G & Heldal M (1995)
Nutritional enrichment of a microbial community: the effects
on activity, elemental composition, community structure and
virus production. FEMS Microbiol Ecol 16: 123–134.
Wu QL, Boenigk J & Hahn MW (2004) Successful predation of
filamentous bacteria by a nanoflagellate challenges current
models of flagellate bacterivory. Appl Environ Microbiol 70:
332–339.
2006 Federation of European Microbiological Societies
Published by Blackwell Publishing Ltd. All rights reserved
c