Introduction

Interactions among microorganisms constitute a fundamental characteristic of the ecosystem. These interactions include various kinds of relationships that promote and inhibit the proliferation and/or expression of certain microbial functions. A detailed characterization of such interactions is expected to provide significant insights into the foundation of microbial ecology and biotechnology; however, these barely visible interactions have not been effectively observed.

Previously, we identified desferrioxamine (DF), a siderophore widely produced by Streptomyces, to be responsible for increased growth and/or effective development in some Streptomyces spp.1 Originally, our extensive screening for interaction between different Streptomyces spp. showed that antibiotic production and/or morphological differentiation in a wide variety of strains was stimulated by cocultivation with different strain. Of these, we observed that cocultivation of Streptomyces griseus with Streptomyces tanashiensis caused marked growth of the latter due to a diffusible metabolite produced by the former. Our study revealed that the growth-promoting substance produced by S. griseus was DFE.2 Furthermore, we discovered that exogenous supply of this type of siderophore affected various characteristics of Streptomyces spp., including growth promotion, cell differentiation and/or secondary metabolite formation. This suggested that DF’s are widely utilized by this group of bacteria to activate specific cellular functions.

It is generally understood that siderophores are involved in the acquisition of ferric ions.3 Despite its wide occurrence in the natural environment, the insoluble nature of iron is responsible for its limited availability to cells. Microorganisms acquire ferric iron by secreting a siderophore, and subsequently uptaking it as a complex with ferric iron. To date, various siderophores and membrane translocators involved in the uptake of those siderophores have been identified.3 On the other hand, the above-mentioned dependence of S. tanashiensis on exogenous DF for growth implies that some microbial populations lack the ability to synthesize siderophores, but retain the mechanism to utilize the siderophores produced by other microbes.2

In this study, we screened for bacterial strains whose growth characteristics were affected by the exogenous supply of DFB, the medically utilized siderophore produced by Streptomyces pilosus.4 The group of siderophore has long been known as a growth factor for some specific bacterial strains such as Arthrobacter spp.,5 but its diverged effect on microbial physiology has not yet been fully studied. The fact that the uptake of the siderophore takes place even in the organisms that do not have the ability to synthesize it2 makes us think of the possibility that the siderophore produced by Streptomyces significantly affects the biological activities in the natural environment. Here, we found that DFB had various effects not only on growth but also on specific functions in taxonomically divergent bacteria. This evidence implies that this type of siderophore serves as an association factor that widely affects microbial population dynamics and functional expression.

Materials and methods

Bacterial strains, plasmids and culture conditions

DFB-dependent bacteria were isolated from soil samples collected at Fujisawa City (Kanagawa, Japan). Appropriately diluted soil suspensions were first inoculated onto 1/100 Luria–Bertani (LB) medium containing 0.01% Bacto Tryptone (Difco, Detroit, MI, USA), 0.005% Bacto Yeast extract (Difco), 0.005% NaCl, 10 μM FeCl3 (pH was adjusted to 7.2 by NaOH), 0.1 mM DFB (Sigma-Aldrich, Tokyo, Japan; if required) and 1% agar (chemicals were purchased from Kokusan, Tokyo, Japan, if not indicated otherwise) and cultured for 3–10 days at 28 °C. The use of diluted LB medium enabled isolation of diverged bacteria, as it reduced the growth rate of rapid growers. Resultant bacterial colonies were then plated in duplicate onto 1/100 LB medium or LB agar medium containing 1% Bacto Tryptone, 0.5% Bacto Yeast extract, 0.5% NaCl, 10 μM FeCl3 (pH 7.2) and 1% agar, supplemented with or without 0.1 mM DFB, and cultured for 3–10 days at 28 °C. Isolates whose growth or some other property was affected by the presence of DFB were selected, and the dependence was confirmed by repeated cultivation. Cells were inoculated by spreading and streaking on the agar medium, and the condition effectively showing the dependence was used. To determine the minimum DFB concentration for growth promotion, KW016 cells were precultured in LB liquid medium containing 1 μM DFB by shaking reciprocally at 300 r.p.m. at 28 °C for 3 days, washed twice with LB medium, and inoculated at 1% (vol/vol) into fresh LB liquid medium supplied with various concentrations of DFB. To assess for IC50 values, the strains of complete inhibition group were precultured in LB liquid medium by shaking reciprocally at 300 r.p.m. at 28 °C for 3 days and inoculated at 1% (vol/vol) into fresh LB liquid medium supplied with various concentrations of DFB. Growth was studied by measuring OD (λ=600).

The following bacterial strains were obtained from National Bioresource Center, NITE, Japan: Chromobacterium violaceum NBRC 12614, Janthinobacterium lividum NBRC 12613, Collimonas spp. NBRC 3740, Microbacterium flavescens NBRC 15039, Microbacterium testaceum NBRC12675, Microbacterium trichothecenolyticum NBRC15077, Microbacterium foliorum NBRC12124, Microbacterium oxydans NBRC15586, Microbacterium barkeri NBRC15036, Microbacterium arabinogalactanolyticum NBRC14344, Microbacterium chocolatum NBRC3758 and Microbacterium imperiale NBRC12610. Escherichia coli JM109 and pUC19 (Takara-shuzo, Kyoto, Japan) were used as the host and vector for general DNA manipulation. Conditions for the culture of E. coli were previously described.6

Taxonomic characterization

To study the 16 S rRNA gene sequence of the isolate, DNA was extracted using a bacterial genomic DNA purification kit (Edge Biosystems, MD, USA). Nearly complete 16 S rRNA gene fragments were amplified by PCR using primers B8F (5′-AGAGTTTGATCCTGGCTCAG; nt 8–27 based on Escherichia coli numbering) and B1492R (5′-GGTTACCTTGTTACGACTT; nt 1508–1484 based on E. coli numbering). PCR was performed on a T1 Thermocycler (Biometra, Gottingen, Germany), with Ex Taq polymerase (Takara-shuzo). The PCR protocol included an initial denaturation period of 4 min at 94 °C; 30 cycles of 94 °C for 1 min, 55 °C for 1 min and 72 °C for 1 min; 3 min at 72 °C; and incubation at 4 °C until further processing. The 16 S rRNA gene fragment was directly sequenced by using the BigDye terminator v3.1 cycle sequencing kit on an ABI 3130 genetic analyzer (Applied Biosystems Japan, Tokyo, Japan). The 16 S rRNA sequences were compared with those from the GeneBank/EMBL/DDBJ nucleotide sequence database, by using the BLAST program (http://www.ncbi.nlm.nih.gov/BLAST/).

Scanning electron microscopy

The DA027 strain, cultured at 28 °C for 3 days on LB medium, was observed using scanning electron microscopy. To prepare specimens, agar blocks were fixed with 2% osmium tetroxide for 30 h, and then dehydrated by freeze-drying. Each specimen was sputter coated with palladium/gold by using an E-1010 ion sputter (Hitachi, Tokyo, Japan) and scanned under a VE8800 scanning electron microscope (Keyence, Tokyo, Japan).

Pigment extraction

The DA027 cells, grown at 28 °C for 3 days on LB agar medium containing 0.1 mM DFB, were collected from the medium surface and used to isolate the purple pigment. The pigment was extracted by treating the cells with diethylether. After concentration by evaporation, the extract was dissolved in appropriate amount of ethanol and applied onto a thin layer chromatography plate (C18; Merck Japan, Tokyo). Violacein similarly extracted from the cells of C. violaceum NBRC 12614 grown at 28 °C for 3 days on LB agar medium was used as a standard. The thin layer chromatography plate was developed with ethanol:water=6:4.

Results

Approximately 8000 bacterial strains isolated from soil samples using diluted LB medium were studied for their dependence on the exogenous supply of DFB (materials and methods). The divergent effects of DFB included both promotion and inhibition of growth, as well as alteration of specific properties such as colony morphology and pigment formation (appearances of the representative colonies of DFB-dependent isolates are shown in Figure 1). The DFB-dependent bacteria were assessed for their taxonomic affiliation by 16 S rRNA gene-based phylogenetic analysis. The results are summarized in Table 1.

Figure 1
figure 1

Typical DFB-dependent bacterial phenotypes observed in this study. Representative colony of strains were developed by cultivating them for 3 days at 28 °C on LB medium containing 10 μM FeCl3 without (−) and with (+) the supply of 0.1 mM DFB. Bar, 5 mm. A full color version of this figure is available at The Journal of Antibiotics journal online.

Full size image
Table 1 DFB-dependent isolates obtained in this study
Full size table

All three isolates that were completely dependent on DFB for growth belonged to Microbacterium (Table 1, a representative strain KW016 is shown in Figure 1). On the basis of this finding, nine species of Microbacterium were obtained from NBRC culture collection (see materials and methods), and studied for their dependence on DFB. The results show that all bacteria, except M. flavescens, effectively grew on medium without DFB (data not shown). M. flavescens NBRC 15039 exhibited marked dependence on DFB, similar to what the three isolates obtained in this study exhibited (data not shown). The minimal concentration of DFB for the growth promotion in KW016 strain was determined to be ca. 0.1 μM.

Other four diverged bacterial isolates affiliating with Burkholderia, Gordonia and Paenibacillus (Paenibacillus DA010 is shown in Figure 1), grew on the medium without DFB, but their growth was significantly promoted by the presence of DFB. In contrast, the supply of DFB completely or partially inhibited the growth of many divergent bacteria; (for example, KW009 and KW112, Figure 1 and Table 1) including both gram-negative and –positive bacteria. We studied the IC50 values with respect to the strains of complete inhibition group, and found that the values were between 50 and 70 μM.

The effects of DFB included not only growth promotion/inhibition but also alteration of colony morphology and induction of pigment production (Table 1). DA021 strain closely related to Paenibacillus formed a diffusible colony in the DFB-containing medium, probably because of promotion of motility (Figure 1). Similar stimulation of motility was observed with respect to KW133, DA003 and DA018 (Supplementary figure S1). The colony of KW103 exhibited a smooth surface on DFB-plus medium although it showed rough surface on DFB-minus one. KW128, a Streptomyces strain, abundantly formed aerial mycelium and spores in DFB-plus condition as observed previously with regard to some strains of the same bacterial group2 (Supplementary figure S1).

DA027 strain belonging to J. lividum effectively produced purple pigment in the DFB-containing medium (Figure 1). DFB promoted production of a yellow intracellular pigment in strains KW001 and 003, and a purple pigment in KW122 strain (Supplementary figure S1).

Figure 2 shows the stimulation of production of purple pigment in DA027 and related strains obtained from the culture collection center. Similar promotion of pigment production was observed in J. lividum NBRC12613 and Collimonas spp. NBRC3740 (Figure 2a). Scanning electron microscopy revealed that the DA027 cells grown in the presence of DFB effectively formed pili (Figure 2b). Thin layer chromatography analysis of the purple pigment extracted from the cells of DA027 strain identified it to be violacein, a pigmented antibiotic produced by some gram-negative bacteria including C. violaceum (data not shown).

Figure 2
figure 2

DFB-induced violacein production (a) and pilus formation (b) in DA027. (a) Filter discs containing 0 (−) and 2.8 mg (+) of DFB were placed into the LB agar inoculated with DA027 and related bacterial strains, and the bacteria were incubated for 3 days at 28 °C. (b) Scanning electron micrographs of the DA027 strain grown under the same conditions as in (a). The photograph of the DFB(+) condition shows the appearance of the violacein-producing cells. A full color version of this figure is available at The Journal of Antibiotics journal online.

Full size image

Discussion

Many microorganisms cope with extreme iron limitations in the natural environment by synthesizing siderophores.3 The producer organisms secrete the siderophore(s) and uptake the substance(s) in ferric-bound form, via specific transport mechanisms. These microbes carry out this energy-wasting process because of the necessity of acquiring ferric iron for cellular ferric-dependent functions.

This study showed that siderophore production is not just a self-supporting process in a specific microorganism, but an event that affects the physiology of other microorganisms too. Although the wide distribution of siderophore transporter genes suggests that uptake of the siderophores produced by other bacteria generally occurs in bacteria,7 the actual effects have not yet been extensively surveyed. The divergent effects caused by DFB, including stimulation and inhibition of growth and induction of specific functions, suggest that the availability of siderophores significantly affects the dynamics of the microbial community in nature.

The isolates belonging to Microbacterium strictly depend on the supply of DFB for growth. M. flavescens (formerly described as Arthrobacter flavescens), with which the isolates obtained in this study (KW111 and KW080; Table 1) were closely related, has been known for its strict growth dependence on a supply of ferrichrome or nocardamine (synonym of desferrioxamine).8 The DFB-independent growth of many other species of Microbacterium indicates that the requirement is a specific feature of the limited taxonomic group of the genus. Probably, these bacteria lack the ability to synthesize siderophores, but they are capable of uptaking and utilizing ferrioxamines. This suggests that the bacteria syntrophically grow by associating with the microbial community producing DFB and related siderophores in the natural environment.

This strict dependence observed with respect to the Microbacterium spp. indicates that some siderophores should be added to culture medium in order to isolate a particular type of bacteria. Recently, D’Onofrio et al.9 reported that supply of the siderophore acyl-DF enabled cultivation of previously uncultured isolates. This as well as our current evidence strongly reinforces the view that the conditions of ferric supply are an important factor with respect to microbial screening.

In contrast to the positive effects, DFB inhibited the growth of a wide variety of bacteria. Presumably, this inhibition is due to the sequestration of ferric iron by DFB. It is supported by the relatively high level of IC50 values (50–70 μM). Since the dose is probably higher than the actual content of this kind of siderophore in the natural environment, the growth inhibition observed in our studies may not be a naturally occurring phenomenon. However, available evidence supports the view that these isolates do not contain the uptake system for DFB, and that they obtain ferric iron on the basis of the function of another type(s) of siderophore and/or a different uptake mechanism. This further makes us speculate that the microbial population in the environment consists of multiple siderophore-utilizing groups, and that the commensalism within each group as well as the competition between different groups serves as a basis for microbial population dynamics in terms of ferric iron utilization.

The addition of DFB caused divergent effects not only on growth but also on specific phenotypes of the bacterial isolates. The colony morphology of some isolates was affected by DFB, possibly due to the stimulation of motility. In Vibrio cholerae, a ferric-depleted condition induces the transcription of genes, including those involved in motility and chemotaxis, on the basis of the function of the ferric-dependent repressor Fur and small regulatory RNA RyhB.10 Although the mechanism of DFB-dependent motility observed in this study is not yet clear, it is possible that depletion of ferric iron due to the sequestration by DFB induces the tactic behavior in some bacteria. Another possibility is that the availability of ferric-DFB complex serves as a positive factor inducing motility in some bacteria. Future characterization is required to understand the mechanism of DFB-dependent motility.

Notable stimulation was observed with respect to violacein production in Janthinobacterium and related gram-negative bacteria. Our preliminary study on the expression of violacein biosynthesis (vio) genes, indicated that the addition of DFB does not affect the transcription level of vio genes in J. lividum, indicating that the supply of DFB affects some post-transcriptional step for violacein synthesis (our unpublished observation). One of the biosynthetic enzymes, VioB, which catalyzes tryptophan dimerization, is a heme-containing enzyme.11 Acquisition and supply of ferric iron to this biosynthetic enzyme based on the availability of DFB could be the reason for DFB-dependent production of violacein.

We discovered that the supply of DFB-induced marked pilus formation in DA027 strain (Figure 2b). To date, the physiological role of pilus formation in this gram-negative bacterium is not known. However, we can reasonably speculate that frequent pilus formation facilitates cell-to-cell and cell-to-surface associations in this organism. The availability of abundant ferric iron due to the presence of DFB may serve as a positive signal for colony development based on the cellular interactions in some bacteria. Future detailed analyses with respect to the effects of DFB and other siderophores will provide useful information regarding the social behavior of microorganisms.