• Inhibition;
  • Bacterial growth;
  • Antagonistic interaction;
  • Attached bacterium;
  • Particle;
  • Wadden Sea


  1. Top of page
  2. Abstract
  3. 1Introduction
  4. 2Materials and methods
  5. 3Results
  6. 4Discussion
  7. Acknowledgments
  8. References

Marine aggregates are densely colonized by bacteria, and inter-specific interactions such as inhibition are important for colonization by aggregate-associated bacteria and thus affect the turnover of organic matter in the sea. In order to study antagonistic activities we carried out inhibition tests with 51 isolates obtained exclusively from aggregates of the German Wadden Sea. 16S rRNA gene sequences of all isolates revealed that 35% of the isolates affiliated with the Flavobacteria/Sphingobacteria group, 24% and 16% with α- and γ-Proteobacteria, respectively, 16% with the Bacillus/Clostridium group, and 10% with Actinobacteria. The relatively high percentage of Gram-positive bacteria may be related to specific features of the Wadden Sea environment. After 11 days of incubation using Burkholder agar diffusion assays the percentage of inhibitory isolates was 54.1% and this decreased to 20.7% after 20 days of incubation but it did not decline for members of the Bacillus/Clostridium group. Inhibitory activity was expressed in strain-specific patterns even though some isolates were closely related according to their 16S rRNA gene sequences. Antagonistic activity was lowest for Flavobacteria/Sphingobacteria (35%) and highest for Actinobacteria (80%). We further examined whether growth of isolates was affected when they were placed on lawns of certain other isolates. In parallel with lowest percentage of inhibitory isolates, highest growth occurred on lawns of the Flavobacteria/Sphingobacteria group whereas it was lowest on lawns of Actinobacteria and the Bacillus/Clostridium group. The high inhibitory activity of both groups of Gram-positive bacteria fits well with data from chemical screening using matrix-assisted laser desorption ionization time of flight mass spectrometry. Hence, inhibitory activity greatly influences inter-specific interactions and may impact microbial degradation and remineralization of particulate organic matter in aquatic environments.


  1. Top of page
  2. Abstract
  3. 1Introduction
  4. 2Materials and methods
  5. 3Results
  6. 4Discussion
  7. Acknowledgments
  8. References

The spatial distribution of particulate and dissolved organic matter in the sea is much more heterogeneous than previously thought and heterotrophic pelagic bacteria preferentially exploit increased concentrations of organic matter within specific microenvironments, e.g. organic aggregates [1–4]. Chemical gradients surrounding these aggregates attract chemotactic bacteria [5] and greatly influence colonization and population dynamics of aggregate-associated bacteria. Quite a few marine bacteria are capable of attaching to aggregates and surfaces, using extracellular polymers or other specialized structures, e.g. flagella [6]. Organic aggregates are colonized by diverse bacterial communities and interactions among bacteria affect microbial dynamics on these aggregates and other particles [4,7]. Bacterial colonization of organic aggregates and particles is mainly controlled by attachment and detachment, growth, and mortality [8]. Hence, inter-specific interactions of bacteria on aggregates lead to changes in the population structure of aggregate-associated bacteria and therefore are critical for microbial processing of organic matter in the sea.

In a recent study more than half of 86 isolates from the Southern Californian Bight showed antagonistic activity and especially bacteria from marine snow displayed strong antagonistic activities towards other bacteria [9]. In Scottish coastal waters 35% of surface-associated bacteria of various seaweed and invertebrate species were shown to produce antimicrobial compounds [10]. The authors speculate that competition for space and nutrients might be a powerful selective force which may have led to the evolution of a variety of effective adaptations of attached bacteria. Microbial competition for limiting natural resources within a community is thought to be an important selective force that promotes biosynthesis of antimicrobial compounds [11]. The production of antimicrobial metabolites is a complicated process involving various factors, e.g. substrate availability and the physiological state of the organism [12]. Many strains which do not produce antibiotics in pure culture can be induced to do so by exposing them to living cells, supernatants from other bacterial cultures or chemicals [10]. Induction of antibiotic production has also been shown for endosymbiotic bacteria and even for human pathogens such as Bacillus subtilis, Escherichia coli, and Pseudomonas aeruginosa[13].

Despite the many studies on marine bacteria searching for secondary metabolites by standard screening procedures [14] and the two studies mentioned above [9,10] there is a lack of research on inter-specific interactions among marine bacteria in an ecological context. Comparative studies on bacteria from various marine systems will allow a better understanding of these inter-specific interactions. This is particularly important for applied fields, e.g. optimizing bioreactor engineering and fermentation protocol design in marine bacterial antibiotic production [15].

To investigate the potential of inhibition and growth of aggregate-associated bacteria in the presence of other bacteria from the same environment we isolated 51 strains from marine aggregates of the German Wadden Sea. The Wadden Sea is a unique interface between land and sea and a highly dynamic environment characterized by strong tidal currents, high wind speeds, and a shallow water depth [16]. Aggregates in the size range of 5–200 μm are very abundant [17,18] and aggregate-associated bacteria presumably gain some advantages over their free-living competitors. For example, transport of substrates to free-living bacteria is solely based on diffusion, whereas attached bacteria greatly benefit from advection of potential substrates towards the particle surface [19,20]. On the other hand, colonization of these nutrient ‘hot spots’ by heterotrophic bacteria may be restricted due to antagonistic activities of other bacteria.

2Materials and methods

  1. Top of page
  2. Abstract
  3. 1Introduction
  4. 2Materials and methods
  5. 3Results
  6. 4Discussion
  7. Acknowledgments
  8. References

2.1Isolation and cultivation of bacteria

To isolate exclusively attached bacteria aggregates (0.1–1 mm in diameter) were collected on a plankton net (0.1 mm pore size) on 15 June and 10 October 2000 from surface waters of the German Wadden Sea (53°43′20″N, 07°43′20″E) and rinsed with sterile (0.2 μm filtered) seawater several times prior to plating the samples. Attached bacteria on the rinsed aggregates were first plated on agar plates (1–2% w/w) prepared from artificial seawater without further addition of bacterial substrates to allow growth of species adapted to natural substrate concentrations on aggregates and to prevent overgrowth by fast-growing species. Different types of colonies were selected for isolation and purified on agar plates prepared from Wadden Sea water enriched with Marine Broth (10 g l−1, MB2216, Difco, USA). Only a minor fraction (<5%) of the picked colonies was not able to grow on these nutrient-rich plates. All incubations were performed at 15°C (in situ temperature) in the dark. Single colonies were transferred at least five times until considered pure. Strain T5 was isolated from surface water using Marine Broth 2216 (Difco) for enrichment (Brinkhoff et al., submitted). Even though strain T5 was not directly isolated from an aggregate it rapidly colonizes particles [7] and was therefore included into our study. The purity of the isolates was checked by denaturing gradient electrophoresis of polymerase chain reaction (PCR)-amplified 16S rRNA gene fragments [21]. Isolates were considered pure when showing a single band on the denaturing gradient gel and were sequenced thereafter.

2.2Sequencing of 16S rRNA genes

Chromosomal DNA was extracted by at least three cycles of freezing and heating cell suspensions to 95°C for 30 min. The primers GM3F (8F) and GM4R (1492R) [22] were used to amplify almost the complete 16S rRNA genes. PCR amplifications were performed with a gradient cycler (Mastercycler, Eppendorf, Germany) as follows: 2 μl of cell suspension was added to 98 μl of PCR mixture containing 1.5 U of Sigma Red Taq polymerase and 10 μl 10×RedTaq™ PCR buffer (Sigma, Deisenhofen, Germany), 250 μM of each deoxynucleotide triphosphate, 2.1 μM MgCl2, and 25 pmol of each primer. The PCR protocol consisted of a denaturing step (5 min at 95°C), followed by 30 cycles of denaturing (1 min at 95°C), annealing (1 min at 40°C), and extension (3 min at 72°C). A final extension was performed for 10 min at 72°C. Aliquots (4 μl) of the amplification products were analyzed by electrophoresis in 2% (w/v) agarose gels and stained with ethidium bromide (1 μg ml−1) [23]. PCR products were purified using the QIAquick 250 purification kit (Qiagen, Hilden, Germany) and sequenced using the DYEnamic Direct cycle sequencing kit (Amersham Life Science) and a Model 4200 Automated DNA Sequencer (LI-COR). Sequencing primers were GM3F and 798R (5′-TGG GTA TCT AAT CCT-3′) labeled with IRDye™800. In many cases we obtained sequences of high quality and longer than 550 bp by sequencing one DNA strand. If quality or length of a sequence was not appropriate for phylogenetic analysis (e.g. HP33, HP39, HP32b, and HP43, for which at least 340 bp were determined) the sequencing was done bidirectionally. Sequences were compared with those of reference organisms by BLAST search ( The taxonomy browser of the NCBI server was used for determination of family affiliation. Sequences for which we found the same related sequences by BLAST analysis were checked for similarity and were subsequently treated as identical or highly similar when they showed a similarity higher than 99%. Ambiguities and single deletions/insertions were not included in the similarity calculations since they might be due to sequencing errors.

2.3Accession numbers of sequenced isolates

The nucleotide sequences of the isolates sequenced in this study have been added to the GenBank database:

The following accession numbers were given: AY239003–AY239013 to organisms belonging to α-Proteobacteria, except T5 (AY177712); AY241547–AY241554 to γ-Proteobacteria; AY241555–AY241571 to bacteria of the Flavobacteria/Sphingobacteria group; AY177726–AY177730 to Actinobacteria; AY172662–AY172669 to organisms of the Bacillus/Clostridium group and AY241546 to the cyanobacterium.

2.4Screening for antagonistic interactions among isolates

We performed the Burkholder agar diffusion assay [24] to screen for inhibitory activity. A matrix of the target isolate was prepared from 2.5 ml molten 1% agar enriched with 10 g l−1 Marine Broth (MB2216, Difco) and 50 μl of isolate suspension (ca. 108 cells ml−1). Agar and cell suspension were mixed by vortexing for ca. 1 s and thereafter poured onto a MB2216 agar plate. After solidification, 10 μl of nine test strains (ca. 108 cells ml−1) were applied onto the lawn. Bacteria growing in a lawn are defined as bacterial lawn and bacteria applied onto the bacterial lawn are called applied bacteria. All 51 strains were tested twice against each other resulting in a total of 568 agar plates which were incubated in the dark for 20 days at 15°C. The plates were examined daily for inhibition zones and for growth of the applied bacteria. Inhibition was recorded when the inhibition zone was at least 4 mm greater than the diameter of the applied colony in both parallels. When parallels showed different results an additional assay was performed to re-assess inhibitory activity.

2.5Chemical screening

Several strains from the Burkholder agar diffusion assays were selected for chemical screening. The isolates were pre-incubated for 96 h at 20°C in 50 ml of liquid medium M-246 (10 g peptone, 10 g meat extract, 250 ml double-distilled water, and 750 ml artificial seawater) using a rotary shaker at 180 rpm. The cultures were then transferred to two different media: 5294-M (10 g starch, 10 g glucose, 10 g glycerol, 2 g peptone, 2 g yeast extract, 2 g corn steep liquor, 2 g peptone, 3 g CaCO3, 250 ml double-distilled water, and 750 ml sterile North Sea water) and MB2216 (37.4 g Difco Marine Broth in 1000 ml double-distilled water). All incubations were performed under the same conditions as described above. Extraction of metabolites was carried out by adding the same amount of ethylacetate to the culture (mycelium and culture broth) and by subsequent vigorous shaking for 30 min. After centrifugation (1700×g) the upper phase was collected, concentrated in an vacuum evaporator, and re-suspended in ethylacetate. An aliquot (100:1 concentrated) was subjected to MALDI-TOF MS (matrix-assisted laser desorption ionization time of flight mass spectrometry) analysis (Voyager-DE PRO, Applied Biosystems, reflector mode, matrix CHCA).

MALDI-TOF MS is a technique in which a co-precipitate of a UV light-absorbing matrix and a biomolecule is irradiated by a nanosecond laser pulse. Most of the laser energy is absorbed by the matrix, which prevents unwanted fragmentation of the biomolecule. The ionized biomolecules are accelerated in an electric field. During the flight in a tube, different molecules are separated according to their mass to charge ratio and reach the detector at different times yielding a distinct signal for each molecule. MALDI-TOF MS is used for detection and characterization of biomolecules, such as proteins, peptides, oligosaccharides and oligonucleotides, with molecular masses between 400 and 350 000 Da. It is a very sensitive method, which allows the detection of low (10−15–10−18 mol) quantities of sample with an accuracy of 0.1–0.01%.

The Chapman and Hall/CRC Chemical Database (Dictionary of Natural Products; version 11.2) was used for the identification of extracted metabolites.


  1. Top of page
  2. Abstract
  3. 1Introduction
  4. 2Materials and methods
  5. 3Results
  6. 4Discussion
  7. Acknowledgments
  8. References

3.1Phylogeny of the isolates

Fifty-one isolates were characterized by phylogenetic analysis of their 16S rRNA gene sequences (Table 1). Some isolates were closely related to each other or even identical (>99% similarity) on the basis of their 16S rRNA genes. For example, 12 isolates of the Flavobacteria/Sphingobacteria group were closely related to Flexibacter aggregans. In our experiments, however, all isolates were treated as distinct entities since they differed in their individual inhibition and/or growth patterns. Thirty-five percent of all isolates affiliated with the Flavobacteria/Sphingobacteria group of the Bacteroidetes phylum, 24% and 16% with α- and γ-Proteobacteria, respectively, 16% with the Bacillus/Clostridium group of Firmicutes and 10% with Actinobacteria. Even though all incubations were performed in the dark one isolate was a cyanobacterium. Within the various divisions we found great differences in diversity. All isolates of the α- and γ-Proteobacteria belonged to four different families each, whereas isolates of the Flavobacteria/Sphingobacteria group and Actinobacteria clustered into two families each. All isolates of the Bacillus/Clostridium group, however, fell into a single family.

Table 1.  Identification, next relative in GenBank, 16S rRNA gene sequence similarity, family affiliation, and inhibitory activity of 51 bacterial strains isolated from organic aggregates of the German Wadden Sea
  1. Inhibitory activity is given by the number of strains against which the formation of inhibition zones occurred. Gray boxes indicate closely related or even identical isolates on the basis of 16S rRNA gene sequence which were listed as distinct strains because they expressed differences in inhibition and/or growth pattern.

inline image

3.2Inhibition patterns

Within the different phylogenetic groups the fraction of isolates showing inhibition zones varied greatly (Fig. 1, Table 1). After 11 days of incubation inhibition zones existed with 80% of Actinobacteria, 67 and 63% of α- and γ-Proteobacteria, 50% of the Bacillus/Clostridium group, and 35% of the isolates of the Flavobacteria/Sphingobacteria group. After 20 days of incubation, however, the percentage of isolates showing inhibition zones was much lower in all phylogenetic divisions, except in the Bacillus/Clostridium group (Fig. 1). Formation of inhibition zones by γ-Proteobacteria and the Flavobacteria/Sphingobacteria appeared only sporadically after 20 days of incubation, and isolates of Actinobacteria only inhibited members of γ-Proteobacteria and the Flavobacteria/Sphingobacteria group (Fig. 2). The capability of α-Proteobacteria to form inhibition zones was also reduced after 20 days of incubation but it remained highest compared to γ-Proteobacteria, Flavobacteria/Sphingobacteria, and Actinobacteria.


Figure 1. Percentages of isolates of various phylogenetic groups exhibiting inhibition zones in Burkholder agar diffusion assays. ALF=α-Proteobacteria, GAM=γ-Proteobacteria, Flavo/Sphing=Flavobacteria/Sphingobacteria group of Bacteroidetes, ACT=Actinobacteria, BAC/CLOS=Bacillus/Clostridium group of Firmicutes. Percentages were determined after 11 and 20 days of incubation.

Download figure to PowerPoint


Figure 2. Percentages of target isolates of various phylogenetic groups which were inhibited by bacteria of specific phylogenetic groups (abbreviations as in Fig. 1). Percentages were determined after 11 and 20 days of incubation.

Download figure to PowerPoint

Three isolates (isolates 9, 10, and 11) affiliating to three different families of α-Proteobacteria inhibited 20 or more of all other isolates and included two of the most inhibitory isolates (isolates 9 and 11). For isolate 9 (T5) inhibitory activity against other marine bacteria has been described before [47], however, it was not tested against such a variety of bacteria. Patterns of inhibition were highly variable among individual isolates of even the same phylogenetic group (Table 1). Surprisingly, isolates with a 16S rRNA gene sequence similarity of >99%, e.g. isolate 17 and 18, showed great differences in inhibitory activities. While isolate 17 did not show any inhibitory activity, isolate 18 inhibited nine other isolates.

Organisms of different phylogenetic groups showed highly variable inhibitory activities against isolates from other phylogenetic groups (Fig. 2). For example, members of α-Proteobacteria expressed highest inhibitory activity against isolates of the Flavobacteria/Sphingobacteria group and Actinobacteria, whereas isolates of the Actinobacteria expressed highest inhibitory activity against those of α- and γ-Proteobacteria and of the Flavobacteria/Sphingobacteria group. Even though more than 25% of all isolates of the Flavobacteria/Sphingobacteria group formed inhibition zones (Fig. 1), they showed the lowest inhibitory activity of all phylogenetic groups (Fig. 2).

3.3Growth patterns

Since isolates used as bacterial lawns did not allow growth of all applied bacteria this aspect was examined more specifically (Fig. 3). Growth of many isolates was strongly suppressed by isolates affiliated with α- and γ-Proteobacteria, with Actinobacteria, and in particular with the Bacillus/Clostridium group after 11 days of incubation. Surprisingly, the percentage of isolates which did not grow in the presence of other isolates was higher than that indicated by inhibition zones (<20% in maximum). This notion suggests that different antagonistic effects occurred, depending on whether the test isolates were present in the form of a lawn or a single colony. The highest numbers of isolates grew on lawns of Flavobacteria/Sphingobacteria whereas the lowest grew on lawns of Actinobacteria and the Bacillus/Clostridium group. In contrast, the percentage of Flavobacteria/Sphingobacteria growing on lawns of other phylogenetic groups was relatively low. Growth of almost all isolates was lower after 20 days than after 11 days of incubation (data not shown).


Figure 3. Percentages of isolates of specific phylogenetic groups which were growing on bacterial lawns of other phylogenetic groups (abbreviations as in Fig. 1). Percentages were determined after 11 days of incubation.

Download figure to PowerPoint

Linear correlation analysis between inhibition of a particular isolate when growing as a lawn and its ability to grow on other bacterial lawns was performed for day 11 of incubation. Isolates being inhibited by other strains also had a lower ability to grow on bacterial lawns. Hence, there was a negative correlation between the number of isolates being inhibited and the number growing on bacterial lawns within a given phylogenetic group, except for Flavobacteria/Sphingobacteria (α-Proteobacteria: r2=0.87, n=12, γ-Proteobacteria: r2=0.73, n=8, Actinobacteria: r2=0.97, n=5, Bacillus/Clostridium group: r2=0.85, n=8). Vice versa, the capability of isolates to form inhibition zones was positively correlated with their ability to grow on bacterial lawns. Significant correlations (P<0.05) occurred after 11 days of incubation for: α-Proteobacteria (r2=0.69, n=12), γ-Proteobacteria (r2=0.57, n=9), Flavobacteria/Sphingobacteria (r2=0.56, n=16), and Actinobacteria (r2=0.90, n=5). These correlations were much weaker after 20 days of incubation. For isolates of the Bacillus/Clostridium group, however, this correlation was not significant even at day 11 (r2=0.39, n=8).

In general, individual isolates showed specific patterns of inhibition (Table 1) as well as of growth and these patterns were highly variable, even for isolates of the same sub-group. The number of isolates inhibiting the greatest number of target isolates was highest for α-Proteobacteria and the Bacillus/Clostridium group (Table 1). None of the applied isolates, however, could grow on all bacterial lawns and only a few isolates of different phylogenetic groups (i.e. isolates 11, 12, 20, 23, 38, and 42) were inhibited by eight or fewer isolates (<15.8%).

3.4Detection of chemical compounds

For selected isolates used in the Burkholder agar diffusion assays several compounds with and without inhibitory activity could be identified (Table 2). The presence of substances with inhibitory activity agrees well with the isolates’ ability to form inhibition zones (see Table 1).

Table 2.  Identification of natural secondary metabolites extracted from strains used in inhibition assays
  1. ALF=α-Proteobacteria, GAM=γ-Proteobacteria, Flavo/Sphing=Flavobacteria/Sphingobacteria group of Bacteroidetes, ACT=Actinobacteria, BAC/CLOS=Bacillus/Clostridium group of Firmicutes.

IsolatePhylum or groupSubstances with inhibitory activityOther compoundsStudy
HP1GAMGlaukothalin [48]
HP5ACTNonactin, dinactin This study
HP8BAC/CLOSSurfactin This study
HP9wBAC/CLOSSurfactin, bacircine This study
HP20ACTPossibly hygrolidin K2 or carbomycin This study
HP42ACTNonactin, dinactin This study
HP44Flavo/Sphing 3-(4′-Hydroxyphenyl)-4-phenylpyrrole-2,5-dicarboxylic acid[47]
T5ALFTropodithietic acid [46,47]


  1. Top of page
  2. Abstract
  3. 1Introduction
  4. 2Materials and methods
  5. 3Results
  6. 4Discussion
  7. Acknowledgments
  8. References


Isolates of the Flavobacteria/Sphingobacteria group and of α- and γ-Proteobacteria constituted the majority of isolates obtained from organic aggregates of the Wadden Sea. These findings are in agreement with other studies analyzing either the composition of bacterial communities on marine snow [4,26,27] or the total bacterial community [25] from various habitats by culture-independent approaches. However, besides culture-dependent drawbacks [28] our results may be biased by the low number of isolates and the fact that we distinguished between isolates showing different inhibition and growth patterns but having similar or even identical 16S rRNA genes (>99% similarity). For instance, the predominance of isolates of the Flavobacteria/Sphingobacteria group which are closely related to F. aggregans may reflect this bias. On the other hand, DNA–DNA hybridization analysis showed that different strains or even species can have identical or almost identical 16S rRNA gene sequences [29,30]. For example, closely related isolates such as 24, 25, and 30 showed extreme differences in patterns of inhibition and growth reflecting differences in their physiology. Thus, our isolates are presumably more diverse than solely indicated by 16S rRNA gene sequences. We cannot certainly exclude that free-living bacteria were also among our isolates. Since we carefully rinsed all our aggregates with sterile water used for cultivation we presume that the vast majority of our isolates were indeed attached to aggregates.

Actinobacteria and Firmicutes occurred frequently among our isolates from Wadden Sea aggregates but they were not detected in other studies on marine snow (see above). Even though Actinobacteria are seldom found in marine pelagic environments, they appear to be an important structural component of limnetic and estuarine bacterial communities [31–35]. Furthermore, Gram-positive bacteria such as Actinobacteria are common antibiotic producers in soil and marine sediments [36] but not in the water column. The high number of isolated Gram-positive bacteria is, indeed, unusual for marine pelagic habitats and might be a distinct feature of the Wadden Sea, which is characterized by shallow waters, high sediment resuspension, and substantial terrestrial and freshwater inputs [17,37]. Since we did not use quantitative methods such as fluorescence in situ hybridization to study bacterial diversity we cannot rule out that cultivation and isolation favored Gram-positive bacteria. However, our Gram-positive isolates were able to grow at the high salt concentrations we used for cultivation and hence we assume that they are able to efficiently compete with other marine bacteria. Thus, their high fraction in our study may reflect a unique bacterial community which is well adapted to the special environmental conditions prevailing in the Wadden Sea.

4.2Antagonistic properties

Antimicrobial activity was a common feature of our isolates. Flavobacteria/Sphingobacteria expressed least inhibitory activity but even >25% of these isolates formed inhibition zones. On average, more than 50% of all isolates inhibited at least one other strain. Number and pattern of inhibition, however, were different for each individual isolate. On the other hand, all groups included isolates which did not show any antagonistic activity at all. Classification of inhibitory activity of bacteria according to phylogenetic groups may be limited by the above-mentioned restrictions due to cultivation biases. Isolates inhibiting many isolates did not exclusively inhibit those of a particular phylogenetic division. Patterns of inhibition and also of growth were highly variable among isolates independent of their phylogenetic relationship. Closely related isolates, however, never inhibited each other.

All isolates occurred in the same environment (Wadden Sea aggregates) and thus presumably compete for resources such as space and nutrients. Furthermore, they are subject to the same controlling environmental and biotic conditions. Differences in antagonistic properties may reflect the ability of individual isolates to interact and compete with other bacteria for limiting resources on the aggregates. Despite high individual variability we also observed some common patterns of inhibition and growth among isolates of the same phylogenetic group. For example, both groups of Gram-positive bacteria and α-Proteobacteria constituted a high percentage of antagonistic isolates whereas the Flavobacteria/Sphingobacteria group accounted for the lowest percentage of antagonistic isolates. The latter group also had the lowest percentage of isolates growing on bacterial lawns.

These results are in line with previous studies which found that antimicrobial activities among aggregate-associated bacteria frequently occur in the sea [10,38]. Furthermore, Long and Azam [9] found a high potential to produce inhibitory molecules among isolates of various phylogenetic groups, except the Bacteroidetes phylum. This particular group was the only group which did not show any relation between the number of inhibited isolates and the number of isolates growing on bacterial lawns. Isolates of the Flavobacteria/Sphingobacteria group were highly inhibited even without visible inhibition zones. This may indicate that their growth is not exclusively inhibited by antibiotics but also by additional mechanisms, e.g. substrate availability. All tests for inhibition and growth were done on Marine Broth 2216 (Difco) which is very different in composition from natural substrates. Flavobacteria/Sphingobacteria preferentially utilize highly polymeric substrates [39] which constitute a minor fraction of Marine Broth 2216. Hence, inhibition and reduced growth among Flavobacteria/Sphingobacteria may be compensated by a different substrate availability in the natural environment.

Alteromonadales and Vibrionales of γ-Proteobacteria were the dominant producers of antibiotics in the study of Long and Azam [9]. We only found five isolates belonging to the Alteromonadales, of which only three isolates (isolates 18–20) formed inhibition zones and only one isolate (isolate 20) expressed relatively high antibacterial activities (Table 1). In our study, Gram-positive bacteria of the Actinobacteria and Bacillus/Clostridium groups but also α- and γ-Proteobacteria expressed high antimicrobial activities. And even one isolate of the Flavobacteria/Sphingobacteria group (isolate 30) showed inhibition zones in the presence of 16 other isolates. Hence, there was no clear difference in inhibitory activity between the phylogenetic groups.

However, for isolates of any phylogenetic group except for Flavobacteria/Sphingobacteria a high ability to form inhibition zones co-occurred with the ability to grow on bacterial lawns of other target isolates. This correlation was relatively weak for members of the Bacillus/Clostridium group which was outstanding in its strong ability to prevent bacterial growth even without formation of inhibition zones. Suppression of growth leading not necessarily to cell death may be due to substances which affect bacterial communication [40] or metabolism and physiology [41].

The observed patterns of inhibition fit well with the results of our chemical screening (Table 2) using MALDI-TOF MS. Actinobacteria (HP5, HP20, HP42) and members of the Bacillus/Clostridium group (HP8, HP9w) are able to produce a variety of well-known antibiotics [42–47]. The production of the highly inhibitory tropodithietic acid has been earlier described for the α-Proteobacterium T5 (isolate 9 [47]). The only examined γ-Proteobacterium, characterized by the production of a hitherto unknown natural product (glaukothalin), also expressed inhibitory activity [48]. Even though none of the screened Flavobacteria/Sphingobacteria was characterized by the presence of inhibitory substances, HP2 and HP11 produced two unknown compounds which could not be further characterized. In addition, they produced a new natural product (3-(4′-hydroxyphenyl)-4-phenylpyrrole-2,5-dicarboxylic acid) without any inhibitory activity and a so far unknown ecological function [47]. Since isolation and chemical characterization of natural products are extremely time- and labor-intensive only strains with an eye-catching inhibitory behavior were used for these analyses and many of the tested strains with inhibitory activity remained unstudied.

4.3Inter-species specific interactions

Members of all phylogenetic groups only grew more weakly on lawns of Gram-positive bacteria (Actinobacteria and the Bacillus/Clostridium group) than on lawns of the Flavobacteria/Sphingobacteria group (Fig. 3). Frequent formation of inhibition zones by Gram-positive bacteria corresponds with their ability to reduce bacterial growth. In addition, the percentage of bacteria of the Bacillus/Clostridium group causing inhibition zones remained high even after 20 days of incubation. This may have several reasons: (a) production of strong bacteriostatic chemicals by members of the Bacillus/Clostridium group (see Table 2), (b) reduced resistance of other phylogenetic groups, and (c) reduced degradability of antibiotics by bacteria of other phylogenetic groups [12].

Differences between formation of inhibition zones by applied bacteria and their potential to prevent growth of target isolates may be indicative of the isolates’ ability to grow as a lawn or as a colony on a lawn. Initial colonization of aggregates by bacteria is mainly determined by their attachment and detachment probabilities [8] and motile and highly chemotactic bacteria have great advantages in initial colonization of fresh and nutrient-rich particulate organic matter [5,49]. Thus, isolates rapidly growing to a dense biofilm may have great advantages in nutrient supply when anticipating aggregate colonization of other bacteria.

Enhanced cell densities may lead to increased production of signal molecules which may be linked to antimicrobial activity of these isolates [50]. Several isolates of the α-Proteobacteria (isolates HP12, 30, 32, and 37) showed production of acylated homoserine lactones (AHL) under laboratory conditions [51]. Interactions among Gram-negative bacteria are frequently controlled by their ability to communicate by using chemical signals such as AHLs [52]. AHLs allow a bacterial population to sense its own density and to express target genes only at particular (high) cell densities including those of enzymes for antibiotic production [53]. But also isolates with or without relatively low inhibitory activity such as HP12, HP30 and HP37, respectively, were able to produce AHLs suggesting that production of inhibitory substances is not necessarily linked to that of AHLs.

In the natural environment frequent resuspension, exposure to UV radiation, changes in salinity, and nutritional stress are additional factors which influence production of antibiotics and/or resistance against inhibitory chemicals [54–56]. In our agar diffusion assays parameters such as salinity and temperature were kept constant. Inhibitory activities and growth of attached bacteria are certainly much more complex in nature than in our experiments, which focused on antagonistic properties of only two strains at a time. Changes in inhibition and growth throughout the incubation period may indicate changes in physiology and may affect antagonistic and growth properties of a particular strain. Within natural bacterial communities physiological changes are probably much more diverse and may result in much more variable antagonistic properties of individual strains than observed in the present study. Consecutive studies on antagonistic interactions between marine bacteria should focus more on in situ conditions and should also include spatial and temporal changes.


  1. Top of page
  2. Abstract
  3. 1Introduction
  4. 2Materials and methods
  5. 3Results
  6. 4Discussion
  7. Acknowledgments
  8. References

The authors thank Ariane Zwintscher and two anonymous reviewers for their valuable assistance. This work was supported by the Volkswagen Foundation within the Lower Saxonian priority program ‘Marine Biotechnology’.


  1. Top of page
  2. Abstract
  3. 1Introduction
  4. 2Materials and methods
  5. 3Results
  6. 4Discussion
  7. Acknowledgments
  8. References
  • [1]
    Azam, F. (1998) Microbial control of oceanic carbon flux: The plot thickens. Science 280, 694696.
  • [2]
    Grossart, H.P., Ploug, H. (2001) Microbial degradation of organic carbon and nitrogen on diatom aggregates. Limnol. Oceanogr. 46, 267277.
  • [3]
    Kiørboe, T., Ploug, H., Thygesen, U.H. (2001) Fluid motion and solute distribution around sinking aggregates. 1. Small-scale fluxes and heterogeneity of nutrients in the pelagic environment. Mar. Ecol. Prog. Ser. 211, 113.
  • [4]
    Simon, M., Grossart, H.P., Schweitzer, B., Ploug, H. (2002) Microbial ecology of organic aggregates. Aquat. Microb. Ecol. 26, 175211.
  • [5]
    Grossart, H.P., Riemann, L., Azam, F. (2001) Bacterial motility in the sea and its ecological implications. Aquat. Microb. Ecol. 25, 247258.
  • [6]
    Heissenberger, A., Leppard, G.G., Herndl, G.J. (1996) Ultrastructure of marine snow. 2. Microbiological considerations. Mar. Ecol. Prog. Ser. 135, 299308.
  • [7]
    Grossart, H.P., Kiørboe, T., Tang, K., Ploug, H. (2003) Bacterial colonization of agar spheres: growth and inter-specific interactions. Appl. Environ. Microbiol. 69, 35003509.
  • [8]
    Kiørboe, T., Grossart, H.P., Ploug, H., Tang, K. (2002) Mechanisms and rates of bacterial colonization of sinking aggregates. Appl. Environ. Microbiol. 68, 39964006.
  • [9]
    Long, R., Azam, F. (2001) Antagonistic interactions among marine pelagic bacteria. Appl. Environ. Microbiol. 67, 49754983.
  • [10]
    Burgess, J.G., Jordan, E.M., Bregu, M., Mearns-Spragg, A., Boyd, K.G. (1999) Microbial antagonism: a neglected avenue of natural products research. J. Biotechnol. 70, 2732.
  • [11]
    Slattery, M., Rajbhandari, I., Wesson, K. (2001) Competition-mediated antibiotic induction in the marine bacterium Streptomyces tenjimariensis. Microb. Ecol. 41, 9096.
  • [12]
    Madigan, M.T., Martinko, J.M. Parker, J. (2000) Brock, Biology of Microorganisms, 9th edn. Prentice-Hall, Upper Saddle River, NJ.
  • [13]
    Samac, D.A., Willert, A.M., McBride, M.J., Kinkel, L.L. (2003) Effect of antibiotic-producing Streptomyces on nodulation and leaf spot in alfalfa. Appl. Soil Ecol. 22, 5566.
  • [14]
    Engel, S., Jensen, P.R., Fenical, W. (2002) Chemical ecology of marine microbial defense. J. Chem. Ecol. 28, 19711985.
  • [15]
    Marwick, J.D., Wright, P.C., Burgess, J.G. (1999) Bioprocess intensification for production of novel marine bacterial antibiotics through bioreactor operation. Mar. Biotechnol. 1, 495507.
  • [16]
    Poremba, K., Tillmann, U., Hesse, K.-J. (1999) Tidal impact on planktonic primary and bacterial production in the German Wadden Sea. Helgol. Mar. Res. 53, 1927.
  • [17]
    McCandliss, R.R., Jones, S.E., Hearn, M., Latter, R., Jago, C.F. (2002) Dynamics of suspended particles in coastal waters (southern North Sea) during a spring bloom. J. Sea Res. 47, 285302.
  • [18]
    Lunau, M., Grossart, H.P. and Simon, M. (2003) A new sampling device for aggregates in shallow and turbid marine environments. Limnol. Oceanogr. Methods (submitted).
  • [19]
    Karp-Boss, L., Boss, E., Jumars, A. (1996) Nutrient fluxes to planktonic osmotrophs in the presence of fluid motion. Oceanogr. Mar. Biol. Annu. Rev. 34, 71107.
  • [20]
    Kiørboe, T. (1992) Turbulence, phytoplankton cell size, and structure of pelagic food webs. Adv. Mar. Biol. 29, 171.
  • [21]
    Brinkhoff, T., Muyzer, G. (1997) Increased species diversity and extended habitat range of sulfur-oxidizing Thiomicrospira spp. Appl. Environ. Microbiol. 63, 37893796.
  • [22]
    Muyzer, G., Teske, A., Wirsen, C.O., Jannasch, H.W. (1995) Phylogenetic relationships of Thiomicrospira species and their identification in deep-sea hydrothermal vent samples by denaturing gradient gel elctrophoresis of 16S rDNA fragments. Arch. Microbiol. 164, 165172.
  • [23]
    Sambrook, J., Fritsch, E.F. and Maniatis, T. (1989) Molecular Cloning: A Laboratory Manual, 2nd edn. Cold Spring Harbor Laboratory Press, Cold Spring Harbor, NY.
  • [24]
    Burkholder, P., Pfister, R., Leitz, F. (1966) Production of a pyrrole antibiotic by a marine bacterium. Appl. Microbiol. 14, 649653.
  • [25]
    Cottrell, M.T., Kirchman, D.L. (2000) Community composition of marine bacterioplankton determined by 16S rRNA gene clone libraries and fluorescence in situ hybridisation. Appl. Environ. Microbiol. 66, 51165122.
  • [26]
    DeLong, E.F., Franks, D.G., Alldredge, A.L. (1993) Phylogenetic diversity of aggregate-attached vs. free-living marine bacterial assemblages. Limnol. Oceanogr. 38, 924934.
  • [27]
    Grossart, H.P., Ploug, H. (2000) Bacterial production and growth efficiencies directly measured on single pelagic aggregates. Limnol. Oceanogr. 45, 436445.
  • [28]
    Eilers, H., Pernthaler, J., Aman, R. (2000) Succession of pelagic marine bacteria during enrichment: a close look at cultivation-induced shifts. Appl. Environ. Microbiol. 66, 46344640.
  • [29]
    Amann, R.I., Lin, C., Key, R., Montgomery, L., Stahl, D.A. (1992) Diversity among Fibrobacter strains: towards a phylogenetic classification. Syst. Appl. Microbiol. 15, 2331.
  • [30]
    Fox, G.E., Wisotzkey, J.D., Jurtshuk, P.Jr. (1992) How close is close: 16S rRNA sequence identity may not be sufficient to guarantee species identity. Int. J. Syst. Bacteriol. 42, 166170.
  • [31]
    Glöckner, F.O., Zaichikov, E., Belkova, N., Denissova, L., Pernthaler, J., Pernthaler, A., Amann, R. (2000) Comparative 16S rRNA analysis of lake bacterioplankton reveals globally distributed phylogenetic clusters including an abundant group of Actinobacteria. Appl. Environ. Microbiol. 66, 50535065.
  • [32]
    Jaspers, E., Nauhaus, K., Cypionka, H., Overmann, J. (2001) Multitude and temporal variability of ecological niches as indicated by the diversity of cultivated bacterioplankton. FEMS Microbiol. Ecol. 36, 153164.
  • [33]
    Selje, N., Simon, M. (2003) Composition and dynamics of particle-associated and free-living bacterial communities in the Weser estuary, Germany. Aquat. Microb. Ecol. 30, 221237.
  • [34]
    Zwisler, W., Selje, N., Simon, M. (2003) Seasonal patterns of the bacterioplankton community composition in a large mesotrophic lake. Aquat. Microb. Ecol. 31, 211225.
  • [35]
    Hahn, M., Lünsdorf, H., Qinglong, W., Schauer, M., Höfle, M.G., Boenigk, J., Stadler, P. (2003) Isolation of novel ultramicrobacteria classified as Actinobacteria from five freshwater habitats in Europe and Asia. Appl. Environ. Microb. 69, 14421451.
  • [36]
    Jensen, P. and Fenical, W. (2000) Marine microorganisms and drug discovery: current status and future potential. In: Drugs from the Sea (Fusetani, N., Ed.), pp. 6–29. Karger, Basel.
  • [37]
    Behrens, B., Liebezeit, G. (1998) Particulate amino acids in Wadden Sea waters – seasonal and tidal variations. J. Sea Res. 41, 141148.
  • [38]
    Nair, S., Simidu, U. (1987) Distribution and significance of heterotrophic marine bacteria with antibacterial activity. Appl. Environ. Microbiol. 53, 29572962.
  • [39]
    Kirchman, D. (2002) The ecology of Cytophaga-Flavobacteria in aquatic environments. FEMS Microbiol. Ecol. 1317, 110.
  • [40]
    Pesci, E.C., Milbank, J.B.J., Pearson, J.P., McKnight, S., Kende, A.S., Greenberg, E.P., Iglewski, B.H. (1999) Quinolone signaling in the cell-to-cell communication system of Pseudomonas aeruginosa. Proc. Natl. Acad. Sci. USA 96, 1122911234.
  • [41]
    Wilcox, M.H., Finch, R.G., Smith, D.G.E., Williams, P., Denyer, S.P. (1991) Effects of carbon dioxide and sub-lethal levels of antibiotics on adherence of coagulase-negative staphylococci to polystyrene and silicone rubber. J. Antimicrob. Chemother. 27, 577587.
  • [42]
    Orlova, T.I. (1993) Macrotetrolide antibiotics from recombinant strain Streptococcus chrysomallus. Antibiot. Khimioter. 38, 36.
  • [43]
    Smith, W.C., Xiang, L., Shen, B. (2000) Genetic localization and molecular characterization of the nonS gene required for macrotetrolide biosynthesis in Streptomyces griseus DSM40695. Antimicrob. Agents Chemother. 44, 18091817.
  • [44]
    Kawada, M., Usami, I., Ohba, S., Romeno, T., Kim, J., Hayakawa, Y., Nose, K., Ishizuka, M. (2002) Hygrolidin induces p21 expression and abrogates cell cycle progress at G1 and S phases. Biochem. Biophys. Res. Commun. 298, 178183.
  • [45]
    Carrillo, C., Teruel, J.A., Arranda, F.J., Oritz, A. (2003) Molecular mechanism of membrane permeabilization by the peptide antibiotic surfactin. Biochim. Biophys. Acta 1611, 9197.
  • [46]
    Brinkhoff, T., Bach, G., Heidorn, T., Liang, L., Schlingloff, A. and Simon, M. (2003) Antibiotic production by a Roseobacter clade affiliated species from the German Wadden Sea and its antagonistic effects on indigenous isolates. Appl. Environ. Microbiol. (in press).
  • [47]
    Liang, L. (2003) Investigation of Secondary Metabolites of North Sea Bacteria: Fermentation, Isolation, Structure Elucidation and Bioactivity. PhD Thesis, University of Göttingen.
  • [48]
    Grossart, H.P., Thorwest, M., Simon, M., Zeeck, A. and Brinkhoff, T. (2003) Production of a new organic blue pigment (Glaukothalin) by the recently described marine bacterial genus Rheinheimera baltica. Appl. Environ. Microbiol. (submitted).
  • [49]
    Blackburn, N., Azam, F., Hagström, Å. (1997) Spatially explicit simulations of a microbial food web. Limnol. Oceanogr. 42, 613622.
  • [50]
    Fuqua, C., Greenberg, E.P. (2002) Listening in on bacteria: acyl-homoserine lactone signaling. Nat. Rev. 3, 685695.
  • [51]
    Gram, L., Grossart, H.P., Schlingloff, A., Kiørboe, T. (2002) Possible quorum sensing in marine snow bacteria: production of acylated homoserine lactones by Roseobacter strains isolated from marine snow. Appl. Environ. Microbiol. 68, 41114116.
  • [52]
    Fuqua, C., Winams, S.C., Greenberg, E.P. (1996) Census and consensus in bacterial ecosystems: the LuxR-LuxI family of quorum-sensing transcriptional regulators. Annu. Rev. Microbiol. 50, 727751.
  • [53]
    Eberl, L. (1999) N-acyl-homoserine lactone-mediated gene regulation in gram-negative bacteria. Syst. Appl. Microbiol. 22, 493506.
  • [54]
    Hermansson, M., Jones, G.W., Kjelleberg, S. (1987) Frequency of antibiotic and heavy metal resistance, pigmentation and plasmids in bacteria of the marine air-water interface. Appl. Environ. Microbiol. 53, 233823342.
  • [55]
    Barry, K.J., Wainwright, N.R. (1997) Biosynthetic induction of a secondary metabolite by a marine bacterium under nutritional stress: potential role of the incomplete oxidation of an organic acid. Biol. Bull. 193, 274275.
  • [56]
    Kitano, M., Nishio, T., Kawai, S., Morishita, H. (1986) Effects of salt environment on the drug resistance of marine bacterium, M-32. Ann. Rep. Osaka City Inst. Public Health Environ. Sci. 49, 8792.