Introduction

TonB dependent transporters (TBDTs) have been initially identified as bacterial outer membrane proteins involved in the uptake of siderophore bound iron (Andrews et al. 2003; Wiener 2005; Noinaj et al. 2010; Chu et al. 2010; Mirus et al. 2010; Braun and Hantke 2011). Siderophores are strong Fe(III)-chelators secreted by the microorganisms. Siderophore production occurs especially under iron starvation. They are generally divided into three classes, namely catecholates, hydroxamates or mixed-types that contain an additional iron complexing group (Miethke and Marahiel 2007; Krantzler et al. 2013). Among them the hydroxamate-type siderophore schizokinen was identified in Anabaena sp. PCC 6411 and Anabaena sp. PCC 7120 (Simpson and Neilands 1976; Lammers and Sanders-Loehr 1982). However, the function of schizokinen is dual; it acts as siderophore for iron acquisition and its ability to complex copper is required for alleviating copper toxicity (Lammers and Sanders-Loehr 1982; Clarke et al. 1987). Schizokinen is not the only siderophore as the genome of Anabaena sp. PCC 7120 contains multiple gene clusters encoding for proteins involved in siderophore synthesis (Jeanjean et al. 2008; Nicolaisen et al. 2008) and a second siderophore of a yet unknown nature was biochemical analyzed (Nicolaisen et al. 2010).

Siderophore uptake is best studied in E. coli and Pseudomonas aeruginosa (e.g. Faraldo-Gómez and Sansom 2003). Based on these two model systems it was established that siderophores bind with high affinity to TBDTs, which drive the transport into the periplasm. TBDTs are composed of a 22-stranded ß-barrel embedded in the membrane and an N-terminal plug domain (Chimento et al. 2005; Noinaj et al. 2010). As typical for outer membrane proteins, the membrane embedded ß-barrel is poorly conserved in sequence, while the plug domain is similar between different proteins. The substrate binding site is composed of amino acids of the plug and the barrel domain. However, these binding sites are not conserved (Chimento et al. 2005) reflecting the specificity of the transporters for different substrates (Schauer et al. 2008; Noinaj et al. 2010).

Various genes coding for putative TBDTs can be found in cyanobacterial genomes. The number can be as large as 33 different genes in one organism (Mirus et al. 2009), which led to the proposal that the number of TBDTs might be dependent on the ecological niche in which the organism thrives (Hopkinson and Morel 2009; Mirus et al. 2009; Stevanovic et al. 2012; Krantzler et al. 2013) and reflect the need for transport of different molecules as TBDTS are not sole iron transporting units (Schauer et al. 2008).

In the genome of Anabaena sp. PCC 7120 22 different genes code for putative TBDTs (Fig. 1a). Classification based on sequence motifs suggested the existence of two BtuB homologoues, the vitamin B12 transporter, and two HutA homologues, facilitating heme transport (Mirus et al. 2009). Heme and Vitamin B12 are porphyrins that are able to chelate iron or cobalt, respectively, which might be a first indication that a differential metal selectivity exists between the two respective homologues. In line, only one of the BtuB and HutA type transporters, respectively, are under the control of FurB (Napolitano et al. 2012).

Fig. 1
figure 1

Growth analysis under iron or copper stress conditions. a The accession number or standard name of the analyzed genes with exception of the fhuA-like genes (see Fig. 3) are given and the localization of the according protein is indicated. OM outer membrane, PM plasma-membrane. b The growth of Anabaena sp. PCC 7120 in BG11 (filled circle), BG11 without addition of iron (open circle), with 1 mM ferric ammonium citrate (grey circle), without addition of copper (open square) or with 1 μM CuSO4 (grey square) was analyzed by measuring the cell density at 750 nm. Lines represent (different) logarithmic regimes. c The ratio of the expression values determined for hgdD and for rnpB at the different growth stages for the different media are presented as bar diagram. Error bars indicate the standard deviation of the four independent experiments

Full size image

The schizokinen transporter SchT (alr0379) is the only TBDT in Anabaena sp. PCC 7120 for which a substrate has been defined (Nicolaisen et al. 2008), because the mutation in schT drastically reduces the uptake of schizokinen (Nicolaisen et al. 2008). The two additional transporters of the IutA-type exist in the genome of Anabaena sp. PCC 7120 (alr2209 and alr2581; Mirus et al. 2009) might thus utilize schizokinen only with low efficiency.

The notion that the TBDTs of the FhuA, IutA, ViuA, HutA and BtuB type of Anabaena sp. PCC 7120 have different substrate specificity is supported by observations for other bacterial TBDTs, which demonstrated the existence of other substrates like nickel and cobalt (Schauer et al. 2008). The hypothesis is further supported by the observed differential expression of the 22 TBDT genes under iron, copper and nitrogen limiting conditions (Mirus et al. 2009). Further analysis of iron uptake by Anabaena sp. PCC 7120 showed a relation between iron and copper uptake. The only TBDT found to be homologous to the catecholate-type transporter ViuA (all4026; Mirus et al. 2009) is annotated as iron and copper transporter IacT (Nicolaisen et al. 2010). The mutant of iacT shows reduced iron and copper levels as well as reduced transport rates for iron and copper when compared to wild-type (Nicolaisen et al. 2010).

Despite the high complexity of the outer membrane uptake system, the subsequent machinery is not as divers. In Anabaena sp. PCC 7120 four genes coding for TonB like proteins, three genes or gene cluster code for the Exb-like components and five gene cluster coding for iron transport systems of Fut, Fhu and Fec-type have been identified (Stevanovic et al. 2012). In addition, a not yet explored gene cluster codes for genes with similarity to FeoB (Fig. 1a). In here we addressed the expression of the genes most likely to be involved in the metal homeostasis of Anabaena sp. PCC 7120. We have analyzed the mRNA abundance of genes with putative function in iron transport in Anabaena sp. PCC 7120 after variation of the iron or the copper level. On the one hand, copper was included as its concentrations influences the iron uptake by Anabaena sp. PCC 7120 (Clarke et al. 1987; Nicolaisen et al. 2010). On the other hand, alterations of the mRNA levels in response to different iron levels was analyzed as the expression of e.g. tonB in E. coli is repressed at high iron levels and induced by iron starvation (Hantke 1987; Postle 1990). Similarly, exbB, exbD, fecC are upregulated upon iron starvation e.g. in Actinobacillus pleuropneumoniae or E. coli, while the fhu system is not majorly affected (Higgs et al. 2002; McHugh et al. 2003; Nielsen and Boye 2005). We further tested the mRNA levels at different growth stages of Anabaena sp. PCC 7120, because e.g. for the E.coli moderate iron starvation is observed at high culture density (Postle 1990). This approach was further inspired by the observation that the rate of uptake per cell rapidly declines after inhibition of transporter gene transcription due to factor dilution by cell division (e.g. Bassford et al. 1977). The plasma membrane anchored TonB is found to be rapidly inactivated in addition to the dilution by cell division (e.g. Bassford et al. 1977) with a half life of about 30 min (Ahmer et al. 1995). Even further, the half life of the according mRNA was determined to be in the range of 1 min (Postle and Good 1985). Thus, the point at which factors are expressed is very important for a detailed understanding of the transport system and to obtain a first insight into the complexity of the network in Anabaena sp. PCC 7120.

Materials and methods

Strains and general procedures

The study was carried out with a nucA-nuiA (all7361-62) single recombinant mutant of Anabaena sp. PCC 7120, designed by integration of the plasmid pCSEL24 (Olmedo-Verd et al. 2006). This mutant does not show any phenotype and is used for growth of Anabaena sp. in moderate antibiotic concentrations. General procedures like transformation of E. coli, isolation of plasmid DNA and conjugal transfer of plasmids into Anabaena sp. have been previously described (Elhai and Wolk 1988; Stevanovic et al. 2012).

Growth of Anabaena sp. PCC 7120

The mutant of Anabaena sp. was grown photoautotrophically at 30 °C (Nicolaisen et al. 2008, 2010) in liquid BG11 medium (Rippka et al. 1979) under constant illumination (OSRAM L 58 W/954—LUMILUX de Luxe, Daylight) at 70 μmol photons m−2 s−1. For qRT-PCR analysis Anabaena sp. was grown in media with reduced metal content, BG11 without the source of iron (no ferric ammonium citrate added; −Fe) or copper (no CuSO4 × 5 H20 added; −Cu). For the media with elevated metal content BG11Fe–Cu medium was either supplemented with 1 μM of CuSO4 × 5 H20 (+Cu) as the copper source or with 1 mM ferric ammonium citrate (+Fe) as iron source. The strain was under all conditions grown in the presence of 3 μg ml−1 streptomycin and 3 μg ml−1 of spectinomycin. For growth curve analysis 50 ml shaking cultures (100 rpm), without additional air/CO2 bubbling were grown in standard BG11 medium for 1 week. The cells were washed three times with the indicated medium, and a volume was inoculated in the according medium to produce a suspension with an OD750 of 0.05. Samples were taken and the OD750 was measured immediately after the inoculation and afterwards regularly every 8 and 12 h for 7 days. Results represent the average of biological triplicates.

Isolation of RNA from Anabaena sp. PCC 7120

Total RNA was isolated using TRIzol reagent (Invitrogen) according to the manufacturer’s instructions and treated with RNase-free DNase I (Roche). DNA contaminations were excluded by PCR in the absence of reverse transcriptase. For reverse transcription 1 μg of total RNA was used, the reaction was performed with Revert Aid Reverse Transcriptase according to the manufacturer’s protocol for random hexamer primer usage (Thermo Scientific).

qRT-PCR analysis

Quantitative real-time (qRT) PCR reaction was performed with 100 ng of cDNA in 10 μl of SYBR Green JumpStart Taq Ready Mix (Sigma-Aldrich Chemie) using an Mx3000P thermal cycler (Stratagene), as described previously (Stevanovic et al. 2012). Melting curve analysis was used to verify the specificity of fragment amplification (Bohnsack et al. 2008). The qRT-PCR results were analyzed using the Mx3000P computer software. The values are means of four samples of two biological independent experiments assayed by qRT-PCR. Expression levels are shown relative to rnpB expression. Oligonucleotides used are listed in Table S1.

Cluster analysis

The expression levels at the 14 different conditions were used to cluster these analyzed genes according to their expression pattern. A hierarchical clustering via Cluster3.0 (de Hoon et al. 2004) was created by taking the expression values for the different samples of a tab delimited file. The program clusters the different genes by their correlation of global expression level under different metal conditions. The tree was visualized using Java TreeView (Saldanha 2004).

Results and discussion

Anabaena sp. PCC 7120 growth is altered in response to metal supply

The growth of Anabaena sp. PCC 7120 under different nutrient supply was monitored by measuring the optical density at 750 nm (Fig. 1b). We observed that the reduction of the iron content reduces the growth rate from 0.039 per hour in BG11 (closed circle) to 0.027 per h (white circle). When Anabaena sp. PCC 7120 cells are grown under elevated levels of iron (grey circle) the initial growth rate is comparable to the growth in BG11 (closed circle), but in the middle of the stationary phase the rate is reduced to 0.018 per h, which is even lower than found for the growth under low iron conditions. This documents that elevated iron also exhibits an environmental stress in Anabaena sp. PCC 7120 as previously reported e.g. for E. coli (Touati et al. 1995; Keyer et al. 1995).

Anabaena sp. PCC 7120 growth in the absence of copper supply (white square) results in a lag phase of about two days. The subsequent growth rate is with 0.030 per h almost comparable to the growth in BG11. Anabaena sp. PCC 7120 grown in the presence of elevated copper levels (grey squares) shows the same growth rate as in BG11. Summarizing, we observed a difference in the growth rate of Anabaena sp. PCC 7120 after alteration of the iron supply, while the growth rate itself is not affect by alteration of the copper supply. In turn, we observe a lag phase in the presence of low copper concentrations which might reflect a more stringent dependence on reprogramming in response to copper limitation when compared to iron limitation, which might be the consequence of the tenfold lower intracellular availability of copper (Andrews 1998; Waldron and Robinson 2009; Nicolaisen and Schleiff 2010).

To probe for cell stage dependent regulation of expression we analyzed the transcript abundance (i) in cells shortly after transfer to the new growth condition (ln(x/x0) < 0.2; after ~8 h); (ii) in cells in the logarithmic phase (ln(x/x0) ~ 2.0; after 48–72 h depending on the growth condition) and (iii) in cells in the stationary phase (ln(x/x0) > 3.5; after ~120–172 h). For better readability of the manuscript the growth stages are assigned to as stage 0.1; 2.0 and 4.0 or ‘early growth stage’, ‘logarithmic phase’ and ‘stationary phase’, respectively .

We used the previously established method to analyze the transcript abundance (Stevanovic et al. 2012). This method includes the normalization of the expression values to the expression of rnpB, which enables us to compare the expression pattern of the different genes, but not of the expression level, which, however, is not intended by the study presented. Nevertheless, the expression level at different growth conditions can be compared for each gene individually. Before quantitative analysis by RT-PCR we ensured that each oligonucleotide pair chosen yields a single product (not shown).

As control we determined the transcript abundance of the gene coding for the outer membrane TolC-like protein HgdD encoded by alr2887 (Fig. 1c), as it is discussed as outer membrane protein involved in siderophore secretion (Nicolaisen et al. 2010). Cells grown in BG11 show a low expression of hgdD in the early and exponential phase and a two fold increase of expression in the stationary phase. The expression of hgdD is enhanced in cultures with medium density grown in media with reduced iron or copper, while high iron levels result in a reduction of the hgdD expression. In contrast, growth in the presence of enhanced copper concentrations does not reduce the expression of hgdD, only at medium cell densities in the logarithmic growth phase a slight reduction was observed.

The metal dependent expression of putative outer membrane localized iron transporter

As depicted in Fig. 1a, the putative components of the metal transport system in the outer membrane have been identified by bioinformatic means (Mirus et al. 2009). At first we determined the expression at mRNA level of non-FhuA type iron transporters at different cell densities of Anabaena sp. PCC 7120 grown in the standard medium BG11 (Fig. 2). The genes coding for the ViuA-like TBDT (all4026), the BtuB like TBDT (alr4028), two IutA-like TBDTs (schT and alr2209) and the HutA like TBDT (alr2153) have a basal expression at stage 0,1, while all3310 (BtuB-like) and alr3242 (HutA-like) are higher expressed at this early stage when compared to the latter stage. For schT and alr2153 we realized a moderate increase of the transcript abundance during the exponential phase only, while alr2209 appears to be higher expressed in the exponential and stationary phase and alr2581 only in the stationary phase (IutA-like TBDT).

Fig. 2
figure 2

Expression of TBDT coding genes distinct from FhuA-type. Each panel shows the result for the indicated genes as described in legend of Fig. 1c. Black bar shows the results for the early growth phase, the grey bar for the logarithmic growth stage and the dark grey bar for the stationary phase

Full size image

Next we analyzed the expression of the according genes in the media with altered iron or copper content. We observed that all4026 and all3310 are highly expressed at stage 0.1 irrespective of the stress source, but expression is significantly reduced as the cell density increases. Only in the absence of copper a slight increase of expression in the stationary phase is observed as well. For the second BtuB-like gene (alr4028) and for two IutA-like genes (alr2209, alr2581) we observed a moderate enhanced expression under iron starvation at all growth stages as well as an enhanced expression at stage 0.1 when elevated levels of iron or copper are present. For the gene coding for the schizokinen transporter (schT) and one HutA-like TBDT (alr2153) we realized an enhanced expression in the medium with reduced iron levels at stage 2.0 and 4.0 when compared to expression in BG11. For schT we further realized an enhanced expression under high copper. The gene coding for the HutA-like Alr3242 is the only one which shows a lower expression than seen when Anabaena is grown in BG11 (with the exception of stage 0.1 in the medium with high iron or copper).

While analyzing the expression of the genes coding for FhuA-like proteins (Fig. 3) we observed that the genes can be divided into four clusters according to the global expression pattern. The first cluster of genes shows a very high expression during logarithmic growth phase in the standard medium BG11 (alr2592, alr2158, alr2588, alr2626, alr2596, alr2236 and alr2211). All of these genes have further in common that they are also expressed in the early phase in the presence of high iron or copper. Only all2236 is also expressed at comparable level in the absence and presence of iron at the early phase, while all2158, alr2588, alr2626 and alr2211 are expressed to a certain extent in general under iron depletion.

Fig. 3
figure 3

Expression of TBDT coding genes of FhuA-type. Each panel shows the result for the indicated genes as described in legend of Fig. 1c

Full size image

The second cluster unifies genes which are highest expressed under iron limited conditions. The gene all4924 is highest expressed in the logarithmic phase when grown in iron limited media, but significant expression levels were also observed in the early growth phase in media with reduced copper or enhanced iron or copper content. The second gene of this cluster, all1101, is expressed highest in iron limited media in logarithmic and stationary phase, while under all other conditions only low transcript levels were detected. The last gene of this cluster, alr2175, is highest expressed in cells grown to the stationary phase in iron-limiting media, while a certain transcript level was also detected in the logarithmic state of cells grown in BG11 or iron limiting condition.

For the third cluster we observe the highest expression values in the early phase when Anabaena sp. is grown in the presence of elevated iron or copper (alr2185 and all2148). All2148, but not alr2211, is also expressed under normal growth conditions and under iron limiting conditions, here highest in the logarithmic phase. The fourth cluster includes all2674 and all2610. These genes show the lowest variation in transcript level with the exception of all2610 in strains at logarithmic state grown in BG11. Thus, it becomes obvious that not all TBDTs are expressed in an iron-starvation dependent manner, and that particularly the genes coding for FhuA-like genes might have rather housekeeping functions.

The expression of the putative plasma membrane components involved in metal transport

To complement the observations for the outer membrane components with putative function in iron transport, we also analyzed the expression of genes coding for the plasma membrane localized components identified by bioinformatic tools (Fig. 1a; Stevanovic et al. 2012). The expression of the genes coding for FeoB like proteins is generally high at low cell densities when grown under stress conditions, but at later growth stages rather repressed when compared to the situation in BG11 (Fig. 4).

Fig. 4
figure 4

Expression of genes coding for plasma-membrane components likely to be involved in metal transport. Each panel shows the result for the indicated genes as described in legend of Fig. 1c

Full size image

For tonB genes we find distinct expression patterns. As previously shown (Stevanovic et al. 2012), tonB2 is particularly expressed under high iron conditions, but we also find a significant expression level while analyzing cells at the early growth stage in BG11. For tonB3 we realized a particularly expression at the early growth stage under iron stress conditions, irrespective whether iron limitation or high iron caused the stress. However, at low iron conditions the gene is significantly higher expressed as well when compared to cells grown in BG11. An unexpected observation is the strong expression under high copper conditions, which has not been seen in the previous study, where this was only observed for tonB4. In here we confirm this observation and demonstrate that the transcript of tonB4 is more concentrated at the earliest growth stage when grown in BG11, in medium without or with enhanced iron. The tonB1 gene is highly expressed at the early growth stage in BG11 and in the medium with enhanced iron or copper.

Inspecting the exbB/exbD genes we confirmed the previous observation that each of the exbB1/exbD1 and exbB3/exbD3 cluster show comparable expression. At the early growth stage the exbB1/exbD1 cluster is expressed higher than in BG11 under all analyzed stress conditions, but most prominent in the presence of enhanced iron. It further shows a higher expression at logarithmic stage compared to cells grown in BG11 when grown in medium with reduced iron as well as enhanced or reduced copper. In turn the exbB3/exbD3 cluster is highest expressed under iron limiting conditions irrespective of the growth stage. In addition, the expression is somewhat higher under high copper conditions when compared to BG11. The exbB2 gene shows an enhanced expression at the early growth stage, but otherwise no alterations when compared to BG11.

We finally analyzed the expression of the plasma membrane permeases. The schE involved in siderophore secretion is generally expressed and we only noticed an enhanced expression at early and logarithmic growth stage under enhanced copper concentrations. The futB transcript is most abundant at early growth stage under iron limiting as well as under enhanced copper or iron conditions. It is further more abundant at logarithmic phase when Anabaena sp. is grown in the presence of high or low copper concentration when compared to the situation in cells grown in BG11. In turn, fhuB shows a higher transcript abundance under iron limiting conditions when compared to BG11, but otherwise no other significant alterations.

For fecD1 and fecD1 we noticed a higher expression at early growth stage under all stress conditions when compared to the abundance when grown in BG11, which is most pronounced under iron limiting conditions. In addition, we realized a higher transcript abundance at logarithmic growth stage in the presence of copper. fecD2 shows only an enhanced transcript abundance at early growth stage in high iron medium and at early and logarithmic growth stage in low copper medium. fecD3 is enhanced in its expression at early growth stage in any altered medium when compared to the transcript abundance when grown in BG11, and also at logarithmic phase when grown in the medium with enhanced copper. Thus, the expression pattern of the plasma membrane components becomes more complex when different growth stages are analyzed than previously reported (Stevanovic et al. 2012).

Distinct clusters of transport components according to the metal dependent expression

The results presented (Figs. 2, 3, 4) show a complex expression behavior of the components putatively involved in uptake of metals. We realized that the expression of most of the components is highest at early growth stages and lowest at the stationary stage (Figs. 1, 2, 3, 4). By clustering the results we observed 6 cluster and 4 individual genes, which did not join any cluster (Fig. 5). The 4 genes not assigned to a clusters are tonB1, tonB2, btuB1 (all3310) and the fhuA-like alr2175. The btuB1 is generally high expressed with highest mRNA abundance at early growth stages and the protein was one of the TBDTs identified by mass spectrometry in the cell wall of vegetative cells (Moslavac et al. 2005). The tonB2 was previously found to be exclusively expressed in the presence of high iron (Stevanovic et al. 2012), which could be confirmed for the logarithmic phase. However, we also observed a significant mRNA level at early growth phase in normal growth medium and in the presence of elevated iron levels. In contrast, tonB1 is highly expressed at low cell densities when Anabaena sp. PCC 7120 is grown in normal growth medium or in the presence of high iron concentrations, but at logarithmic phase in the absence of copper. The fhuA-like alr2175 shows the highest mRNA abundance in the late growth phase in the medium with low iron content.

Fig. 5
figure 5

Clustering of the genes. The left side shows the clustering of the expression profiles by Cluster 3.0. The distance concerning the global expression correlation is shown. On the right the general profile is show with the same color code as described in Fig. 2. The profile was calculated by normalizing the gene specific profile to the maximal value observed, and subsequently calculation of the mean of the profiles of all genes in one cluster

Full size image

The first cluster unifies genes which are particularly high expressed at early growth phase in the medium with elevated iron (Fig. 5). This cluster is exclusively composed of plasma membrane components, including the four feo-type genes identified in the genome of Anabaena sp. PCC7120. The second cluster contains genes coding for plasma membrane components as well. However, in contrast to genes of cluster 1 these genes show an mRNA distribution as expected for components required under iron starvation; namely they are expressed at low iron or high copper. Indeed, it was previously suggested that the three genes found in this cluster fhuB, exbB3 and exbD3 are the plasma membrane components activated under iron starvation (Stevanovic et al. 2012).

The third cluster represents an iron limitation induced cluster as well, but is composed exclusively of TBDT coding genes, namely the previously described schT (Nicolaisen et al. 2008), a hutA-like gene (alr2153) and two FhuA-like genes (all1101, all4924). Together with the single standing fhuA-like gene alr2175, which shows the highest transcript abundance at stationary phase under iron starvation, these TBDTs are the main candidates for iron transport in Anabaena sp. PCC 7120 based on the expression pattern observed. Interestingly, schT and all1101 have been described to be enhanced in FurA overexpression mutants of Anabaena sp. PCC 7120 (González et al. 2010, 2012).

The fourth cluster unifies eight FhuA-like genes as well as hutA2. All of these genes are highest expressed at logarithmic phase in normal growth medium. This observation signs remarkable and might point toward a function distinct from iron transport.

The fifth cluster is composed of two genes only (iutA2, all2610), which are generally high expressed at early growth stages irrespective of the medium used, but highest at stationary phase in normal growth media. Indeed, the mRNA level of all2610 was not found to be altered by the expression of FurA (González et al. 2012), which might hold true for iutA2 as well.

The sixth cluster contains beside TBDT coding genes the schE, which is required for schizokinen export (Nicolaisen et al. 2010). The genes are highest expressed at early growth stage when grown in the presence of enhanced iron or copper levels. The latter is consistent with the importance of e.g. SchE for secretion of schizokinen and thus, copper detoxification (Clarke et al. 1987; Nicolaisen et al. 2010). However, the genes also show a significant mRNA level when Anabaena sp. PCC 7120 in normal growth media or in the absence of iron, which in turn would be consistent with a function in iron uptake.

Concluding hypothesis

In general it is discussed that iron transporter are enhanced in their expression under low iron conditions, while factors involved in defense against iron or oxidative stress are upregulated under high iron conditions (e.g. Foster and Hall 1992; Vasil and Ochsner 1999). This was also the basic observation while analyzing the global expression of genes in Synechocystis sp. PCC 6803 in response to iron limitation by gene-chip analysis (Katoh et al. 2001; Singh et al. 2003; Hernández-Prieto et al. 2012; Supp. Figure S1). Based on this general assumption we propose that genes of cluster 2 and 3 as well as the FhuA-like alr2175 are the core units of the iron uptake system induced by iron limiting conditions, while for the genes of cluster 4, 5 and 6 a function distinct from iron transport can be expected. In turn, the strong expression of the genes of cluster one at early growth stage might reflect a reaction of a bacteria typically living in low iron environments, which depends on the ability to rapidly react to short term high iron conditions at which iron uptake would be possible. In addition, we realized that TBDT coding genes and genes of the plasma membrane localized components are not co-regulated in their expression. On the one hand, this might reflect the lower complexity of the plasma membrane components; on the other hand it might be the result of the lower stability of such components, especially of tonB genes (e.g. Bassford et al. 1977). This is consistent with the observation that most of the tonB genes are highly expressed at early growth stages in general. Furthermore, while tonB3 and tonB4 are similarly expressed, tonB1 and tonB2 show an expression not related to iron starvation. Up to now and to the best of our knowledge such behavior has only been reported for human pathogens like Neisseria gonorrhoeae or Vibrio vulnificus as response to human serum (Hagen and Cornelissen 2006; Alice et al. 2008). Thus, an alternative function or regulatory mechanism for tonB1 and tonB2 is expected and has to be further investigated. In general, the results provided in here and before (Mirus et al. 2009; Stevanovic et al. 2012) for Anabaena sp. PCC 7120 show that genes related to metal transport are regulated by culture density, metal supply and the nitrogen source.