Increasing the copper sensitivity of microorganisms by restricting iron supply, a strategy for bio-management practices.

Pollution by copper (Cu2+ ) extensively used as antimicrobial in agriculture and farming represents a threat to the environment and human health. Finding ways to make microorganisms sensitive to lower metal concentrations could help decreasing the use of Cu2 + in agriculture. In this respect, we showed that limiting iron (Fe) uptake makes bacteria much more susceptible to Cu2 + or Cd2+ poisoning. Using efflux mutants of the purple bacterium Rubrivivax gelatinosus, we showed that Cu+ and Cd2+ resistance relies on the expression of the Fur-regulated FbpABC and Ftr iron transporters. To support this conclusion, inactivation of these Fe-importers in the Cu+ or Cd2+ -ATPase efflux mutants gave rise to hypersensitivity towards these ions. Moreover, in metal overloaded cells the expression of FbpA, the periplasmic iron-binding component of the ferric ion transport FbpABC system was induced, suggesting that cells perceived an 'iron-starvation' situation and responded to it by inducing Fe-importers. In this context, the Fe-Sod activity increased in response to Fe homoeostasis dysregulation. Similar results were obtained for Vibrio cholerae and Escherichia coli, suggesting that perturbation of Fe-homoeostasis by metal excess appeared as an adaptive response commonly used by a variety of bacteria. The presented data support a model in which metal excess induces Fe-uptake to support [4Fe-4S] synthesis and thereby induce ROS detoxification system.

Introduction

Although copper serves as a catalytic cofactor to drive a variety of biochemical processes including respiration and photosynthesis (Andreini et al., 2008), excess Cu2+, exceeding cellular needs, is toxic. Copper is thus the main toxic component in the ‘Bordeaux Mixture’, an effective bactericide and fungicide used for decades in agriculture to control diseases of vine fruits, olive groves, ornamental plants and fruit orchards. The extensive use of Cu2+ as a fungicide against mildew in vineyards or in farming, for example, is a source of soils and groundwater contamination. Furthermore, extensive use of metals at high concentrations appears to promote co‐occurrence and co‐selection of antibiotic resistance genes with metal resistance gene (Baker‐Austin et al., 2006; Rensing et al., 2018; Asante and Osei Sekyere, 2019).

A recent study reported the factors influencing copper distribution in agricultural lands at the European scale and highlighted the importance of land management practices in copper concentration and the strong correlation between topsoil copper and vineyards. (Ballabio et al., 2018). Moreover, the increased copper concentration in soil over a long period was shown to negatively affect bacterial richness and evenness (Nunes et al., 2016). The European Commission also pinpointed the environmental and health risk associated with high copper concentration use in agriculture and the urgent need for more sustainable ‘metal‐based antimicrobial treatments’ management to limit the spread of copper and its adverse effects on ecosystems and living organisms.

https://ec.europa.eu/environment/integration/research/newsalert/pdf/agricultural_management_practices_influence_copper_concentrations_european_topsoils_518_na2_en.pdf

Both prokaryotes and eukaryotes deal with metals such as Cu+, Zn2+ or Fe2+ and maintain optimal cytoplasmic concentration either by storing the excess using specific proteins and compartments or by blocking the import using regulators to repress gene expression or expelling the excess using the efflux systems. Copper balance in bacteria relies mainly on efflux systems. When the homoeostasis system is dysregulated, accumulation of Cu+ could directly lead, through Fenton‐like chemistry, to hydroxyl radical generation (Gunther et al., 1995). Cu+ can also displace iron from proteins or damage exposed [4Fe‐4S] clusters resulting in released Fe atoms (Macomber and Imlay, 2009; Barwinska‐Sendra and Waldron, 2017) that can induce iron‐based Fenton chemistry and reactive oxygen species (ROS) production. As with Cu+, Cd2+, a non‐biological and non‐redox active metal, can also trigger iron‐based Fenton chemistry. This supports the idea that excess Cd2+ could give rise to intracellular mismetallation of proteins and release of ‘free iron’ (Xu and Imlay, 2012). One may expect that under such conditions, cells will repress iron uptake to limit the harmful effects of excess iron and ROS. Inconsistently, however, several transcriptomic studies in bacteria, yeast or plants reported that excess Cu+, Cd2+or Co+ induced iron uptake gene expression (Gross et al., 2000; Stadler and Schweyen, 2002; Teitzel et al., 2006; Yoshihara et al., 2006; Houot et al., 2007; Chillappagari et al., 2010). In Escherichia (E.) coli, expression of some genes involved in the synthesis and uptake of siderophore or iron was induced when cells were exposed to excess Cu+, Zn2+, Ni2+ or Co2+ (Kershaw et al., 2005; Fantino et al., 2010; Macomber and Hausinger, 2011; Xu et al., 2019). Similarly, in Pseudomonas (P.) aeruginosa, several genes involved in iron uptake, usually induced under iron‐limiting condition, were also upregulated in Cu+ stressed cells (Teitzel et al., 2006). Interestingly, in Bacillus subtilis, microarray data indicated that Cu+ stress‐induced genes were required for iron uptake, whereas induction of the Cu+ efflux system CopZA in the ΔcsoR Cu+‐sensing transcriptional repressor mutant prevented upregulation of these Fur‐regulated genes (Chillappagari et al., 2010).

In the yeast Saccharomyces cerevisiae, it was also shown that elevated amount of Cu+ or Co+ induced the expression of the iron regulon (Fet, Ftr), thereby increasing the intracellular iron level (Gross et al., 2000; Stadler and Schweyen, 2002; Alkim et al., 2013). On the other hand, Cd2+ was shown to upregulate the genes involved in Fe acquisition, in the cyanobacterium Synechocystis PCC6803, in the green alga Chlamydomonas reinhardtii or in plants (Rubinelli et al., 2002; Yoshihara et al., 2006; Houot et al., 2007) but not in E. coli (Helbig et al., 2008). In agreement with these reports, it was shown that co‐incubation of Cu+ stressed hepatocytes with the iron chelator deferoxamine significantly inhibited ROS production and prevented cell death. This suggested an increased iron uptake under Cu+ excess stress in hepatocytes (Krumschnabel et al., 2005). Although these studies provided indirect evidence for a central role of iron homoeostasis to cope with excess metal, an understanding of the underlying processes at a molecular level is still lacking.

In the context of metal stress, exposure of the efflux mutants ΔcopA (the Cu+‐efflux ATPase) or ΔcadA (the Zn2+/Cd2+ efflux ATPase) to elevated Cu+ or Cd2+ level resulted in coproporphyrin III accumulation in the purple non‐sulphur photosynthetic bacterium Rubrivivax (R.) gelatinosus (Azzouzi et al., 2013; Steunou et al., 2020a) and in Neisseria gonorrhea (Djoko and McEwan, 2013), likely denoting an effect on [4Fe‐4S] clusters. R. gelatinosus can grow either by respiration or by photosynthesis and tolerate high concentration of Cu2 + and Cd2+ (Azzouzi et al., 2013; Steunou et al., 2020a). To better decipher the consequences of excess Cu+ or Cd2+ in R. gelatinosus, we used transposon mutagenesis to select and characterize hypersensitive mutants to Cu+ and to Cd2+. Unexpectedly, most isolated mutations were found within the fbpA gene, encoding the periplasmic iron‐binding protein (FbpA) component of the ferric iron FbpABC system. It is interesting that both Cu+ and Cd2+ elicited similar phenotypes, that is [4Fe‐4S] degradation in E. coli or in R. gelatinosus. This provided opportunity to address the impact of redox‐active and non‐active metals on iron homoeostasis in bacteria, with a focus on the events that followed [4Fe‐4S] clusters degradation after Cu+ or Cd2+ stress. Based on previous work from various groups, our study allowed us to draw a model on how excess Cu+ or Cd2+ would poison cells, starting with protein mismetallation to ROS generation. The central role of Fe–uptake, likely to maintain Fe‐S clusters synthesis, in response to Cu+ or Cd2+ homoeostasis dysregulation in bacteria and eukaryotes is also discussed. Metal‐based antimicrobial strategies using copper or cadmium have potential applications in many fields; nevertheless, sustainable practices to avoid metal pollution must be found (Turner, 2017). Here, we demonstrated that limiting iron uptake is an effective way to inhibit bacterial growth with very low copper concentration.

Results

Iron uptake provided a survival advantage during copper or cadmium stress

A genetic approach using Tn5 random mutagenesis system in R. gelatinosus was applied to select mutants with increased sensitivity to Cu+ and Cd2+ in the double mutant ΔcopRcadR background, in which both metal regulators CopR and CadR were inactivated. CopR activates the expression of cop operon in response to excess Cu+, and CadR activates cadA expression in presence of Cd2+ (Azzouzi et al., 2013; Steunou et al., 2020a). Under photosynthesis condition, this mutant can still grow in presence of 400 µM CuSO4 and CdCl2, because the Cu+‐exporting ATPase CopA and the Cd2+ efflux pump CadA are still expressed, albeit at lower level. Upon screening for transposon mutants unable to survive on 50 µM CuSO4, 5 out of 10 hypersensitive ΔcopRcadR::Tn mutants had the transposon inserted within the fbpA gene at different positions (Fig. S1). These mutants were also sensitive to 50 µM CdCl2. Strikingly, fbpA gene is predicted to play a role in iron uptake. fbpA encodes the periplasmic substrate‐binding protein component of the membrane ABC‐type iron (Fe3+) transporter, FbpABC (Fig. S2; Parker Siburt et al., 2012). Unlike many other bacteria, R. gelatinosus fbpA and fbpBC genes are not organized in a single operon but split in the genome. Nonetheless, putative Fur boxes were identified in the promoter of fbpA and fbpBC genes suggesting that this system could be induced under iron limitation (Fig. S1). All of the five ΔcopRcadR‐fbpA::Tn mutants exhibited comparable growth in medium containing increased concentrations of CuSO4 or CdCl2 (Fig. S1).

Dose–response growth experiments confirmed that, in contrast to the parental ΔcopRcadR strain, the transposon triple mutant ΔcopRcadR‐fbpA::Tn was more sensitive to copper and cadmium (Fig. 1A and B). Hereafter, we used ΔcopRcadR‐fbpA2::Tn5 insertion mutant, in which Tn5 was found at position 477 in the fbpA gene, for further investigation. These data suggested that FbpA plays an important role in the tolerance mechanisms towards Cu+ and Cd2+.

Fig. 1

FbpA is involved in copper and cadmium tolerance. Growth inhibition of the ΔcopRcadR and ΔcopRcadR‐fpbA2::Tn5 mutant in malate medium supplemented with increasing CuSO4 (A) or CdCl2 (B) concentrations under photosynthesis condition. Cells were inoculated with an OD680 of 0.02 and grown overnight for 18 h at 30°C before OD680nm measurement. Results are the average of 3 independent experiments.

Hypersusceptibility of the ATPase‐deficient mutants upon inactivation of the periplasmic iron‐binding protein FbpA

To confirm the transposon mutant phenotype, fbpA2::Tn5 allele was transferred in the wild‐type , in the Cu+‐ATPase (copATp) or in the Cd2+‐ATPase (ΔcadATp) deficient mutants. Their sensitivity to Cu+ or Cd2+ under photosynthesis condition was examined. In contrast to the wild‐type and the fbpAKm mutant that can tolerate up to 400 µM CuSO4 in the medium, growth of the copATp mutant was affected by increasing concentration of CuSO4 and inhibited beyond 200 µM CuSO4 (Table S1). Inactivation of the fbpA gene in the copATp background led to a hypersensitive strain. Indeed, growth was drastically decreased even in malate medium that contained only 1.6 µM CuSO4 and was completely inhibited when the medium was supplemented with 50 µM CuSO4 (Fig. 2A). This Cu+‐sensitive phenotype was even more pronounced than the phenotype observed for ΔcopRcadR‐fbpA::Tn mutant, likely because the latter still expressed CopA and expelled some Cu+ outside the cytoplasm. Similarly, the ΔcadAfbpA mutant was more sensitive to Cd2+ than the single ΔcadATp and fbpAKm mutants, confirming that fbpA was somehow also involved in Cd2+ tolerance (Fig. 2B). Altogether, these data indicated that fbpA expression and presumably iron uptake were required for Cu+ and Cd2+ tolerance in the absence of the detoxification efflux systems.

Fig. 2

FbpA is required for copper and cadmium tolerance when the efflux systems are defective. Growth inhibition of the copAfbpA − (A), ΔcadAfbpA (B) in comparison with the wild‐type and the single mutants in malate medium supplemented with increasing CuSO4 or CdCl2 concentration under photosynthesis condition. Cells were grown overnight for 18 h at 30°C before OD680nm measurement. Results are the average of 3 independent experiments. C. Quantification of coproporphyrin III in the medium was performed by absorbance (maxima at 390 nm), and values were normalized with the absorbance of the culture at 680 nm. The mean and standard deviation of four independent experiments were shown. D. Absorbance spectra of the spent medium of wild‐type and mutant strains. Representative spectra curves of four independent experiments are shown.

Hypersusceptibility to Cu⁺ is associated with increased coproporphyrin III production in the absence of FbpA

Toxicity of copper in the ATPase‐deficient mutant copA − of R. gelatinosus and Neisseria gonorrhea was related to an impaired porphyrin biosynthesis illustrated by the release of coproporphyrin III (oxidized coproporphyrinogen III) in the medium. Under this condition, excess copper is likely to damage the iron–sulphur cluster of the coproporphyrinogen III oxidase HemN (Azzouzi et al., 2013; Djoko and McEwan, 2013). Combination of severe decrease of porphyrins/increased release of coproporphyrin III and impaired iron uptake might explain the Cu+ sensitivity phenotype of copAfbpA − strain. To establish a correlation between Cu+ concentration, coproporphyrin III release and fbpA gene disruption, the growth of wild‐type, copATp, fbpAKm and copAfbpA − cells was challenged with very low CuSO4 concentration ranging from 2 to 6 µM. As discussed above, the growth of copAfbpA − mutant was affected even at very low concentration of CuSO4. Under these conditions, concentration of coproporphyrin III released in the culture supernatant was spectroscopically analysed. As shown in Fig. 2C and D, no coproporphyrin III could be detected in the medium of wild‐type or fbpAKm mutant cells regardless of the copper concentration. In the copATp cells, coproporphyrin III could only be detected in the medium containing at least 5 µM CuSO4. Noticeably, for the copAfbpA − mutant, higher amount of coproporphyrin III was detected even in the growth medium containing only 2 µM CuSO4. The amount of released coproporphyrin III was significantly increased in the medium containing 3 µM CuSO4 (Fig. 2C and D). These data showed that the effect of copper on porphyrin biosynthesis was more pronounced when both the ATPase CopA and the iron uptake system Fbp were inactivated.

Excess Cu⁺ or Cd²⁺ induces the expression of the periplasmic iron‐binding protein FbpA

Previous transcriptomic studies in bacteria have suggested an induction of different iron uptake systems in response to excess copper (Kershaw et al., 2005; Teitzel et al., 2006; Chillappagari et al., 2010). Our genetics data strongly suggest the expression of the iron uptake FbpABC system in R. gelatinosus under Cu+ or Cd2+ excess stress. Moreover, semi‐quantitative RT‐qPCR showed a ~ 2‐fold induction of fbpA transcripts in response to iron starvation and to Cu+ or Cd2+ excess (Fig. S3). In order to follow FbpA protein expression under different metallic stress conditions, the fbpA gene was substituted by a histidine‐tagged copy (fbpAH) on the chromosome under its own Fur‐regulated promoter. This strain was subjected to elevated concentration of CuSO4 or CdCl2 either in an iron‐containing or in iron‐limited medium and the amount of FbpAH6 in the periplasmic fraction was assessed on Western blots (Fig. 3). In the iron‐containing medium with basal 1.6 µM CuSO4, FbpAH6 was not detected. On the contrary, in the medium supplemented with 250 or 500 µM CuSO4 a gradual increase in the amount of FbpAH6 was observed in the periplasm. This increase correlated with the gradual increase in the amount of CopI, the periplasmic Cu+‐induced protein (Fig. 3A), which can be also detected by the HisProbe due to its histidine‐rich motif (Durand et al., 2015). Under iron‐limited condition, FbpAH6 was detected in the periplasmic fraction even in the malate medium containing 1.6 µM CuSO4. This expression was very likely related to iron limitation in the medium. As expected, addition of 250 or 500 µM CuSO4 to this medium resulted in a substantial induction of FbpAH6 and CopI. Similar induction of FbpAH6 and CopI was obtained with addition of high concentrations of CdCl2 to the growth medium (Fig. 3B), confirming the induction of FbpA when excess Cd2+ was present in the medium.

Fig. 3

Expression of the periplasmic iron‐binding protein FbpA is induced under excess copper and cadmium or under iron limitation. fbpAHis6 cells were grown overnight for 18 h in the presence of increasing concentration of CuSO4 (A) or CdCl2 (B), in malate medium (M) or iron‐depleted malate medium (M‐Fe). The presence of FbpAH6 and CopI was revealed in the periplasmic fractions using the HRP‐HisProbe. CopI is induced by copper and cadmium and served as an internal control. * 19 kDa protein induced in iron‐depleted medium. C. Cu+ or Cd2+ excess affects the intracellular iron content in the ATPase efflux mutants. Intracellular content of iron in the WT, copATp and ΔcadATp was measured by ICP‐MS. The values were normalized by the culture absorbance at 680 nm. The results are expressed as the mean ± SD (error bars). Significance of variation between wild‐type in malate medium and samples were determined by one‐way ANOVA with Dunnett's multiple comparison test. ns, non‐significant; **P < 0.01.

These data argued in favour of an induction of the iron transporter FbpA by excess Cu+ or Cd2+, and interestingly, FbpA induction underlied an iron ‘depletion‐like’ situation in cells facing excess metal. To test this assumption, we also investigated the effect of excess Cu+ or Cd2+ stress on total iron content in cells grown in medium supplemented or not with excess metal. ICP‐MS analyses showed a decrease in the amount of total iron content in the copATp cells challenged with 100 µM CuSO4 and ΔcadATp cells challenged with 100 µM CdCl2 (Fig. 3C). In the wild‐type, the amount of total iron remained comparable under both stress conditions confirming that Cu+ or Cd2+ excess altered Fe2+ homoeostasis in the efflux mutant that accumulated metals within the cytoplasm.

Hypersusceptibility of the ATPase‐deficient mutants upon disruption of the iron uptake system fbpBC

fbpABC is proposed to encode a periplasmic binding protein‐dependent ABC transport system that enables iron transport in Gram‐ bacteria. The fbpBC genes encode the cytoplasmic membrane‐associated proteins FbpB and FbpC that act together with FbpA for the transport of iron into the cell. Although the involvement of FbpA as a periplasmic binding protein in the Cu+ or Cd2+ response and resistance is now well established, the involvement of FbpBC proteins in Cu+ or Cd2+ response remains to be demonstrated to unequivocally state that the FbpABC iron transport system is involved in metal excess tolerance. To this aim, the fbpBC locus was deleted in the wild‐type and efflux mutants, copAKm and ∆cadAKm strains. The tolerance of the double mutant towards Cu+ or Cd2+ excess was assessed and compared to that of the wild‐type and the single mutants (Fig. 4). As expected, while the wild‐type and the ΔfbpBC mutant tolerated up to 400 µM CuSO4 in the medium, growth of the copAKm mutant was affected by increasing concentration of CuSO4 and inhibited at 200 µM CuSO4 (Fig. 4B and C). Deletion of the fbpBC genes in the copAKm background gave a hypersensitive ΔcopAfbpBC strain, in which growth was dramatically decreased by excess Cu+ and completely inhibited in medium containing 50 µM CuSO4 (Fig. 4B and C ). This Cu+‐sensitive phenotype was comparable to that observed for the copAfbpA − mutant demonstrating that, as for FbpA, the membrane cytoplasmic FbpBC transporter, which uses ATP hydrolysis to drive iron transport into the cytoplasm, was also required for copper tolerance when the Cu+‐efflux system was missing. Similarly, the ΔcadAfbpBC mutant was more sensitive to Cd2+ than the single mutants (Fig. 4D), confirming that fbpBC was also involved in Cd2+ tolerance.

Fig. 4

The iron importer FbpBC is required for copper and cadmium tolerance when the efflux system is defective. Organization and deletion of fbpBC genes. A trimethoprim cassette was inserted in the Eco47III restriction enzyme sites deleting fbpB and fbpC genes (A). Growth of the WT, ΔfbpBC, copAKm and the double mutant ΔcopAfbpBC cells under photosynthesis in presence of increasing CuSO4 concentrations after 18 h (B). Growth inhibition curves of the ΔcopAfbpBC (C), ΔcadAfbpBC (D) in comparison with the WT and the single mutants in malate medium supplemented with increasing CuSO4 or CdCl2 concentrations under photosynthesis condition. Cells were inoculated at 0.02 and grown overnight for 18 h at 30°C before OD680nm measurement. Results are the average of 3 independent experiments.

Altogether, these data strongly indicated that the iron acquisition FbpA/FbpBC system and very likely iron uptake were required for Cu+ and Cd2+ tolerance in the absence of the efflux detoxification systems.

The Ftr iron import system is also required for metal tolerance

The fbpAKm and ΔfbpBC mutants still grew in the malate medium with excess Cu+ or Cd2+ suggesting the presence of other iron uptake systems. To analyse the R. gelatinosus response to iron limitation, the wild‐type and fbpAKm strains were cultured in parallel either in iron‐containing or in iron‐depleted media and periplasmic fractions were analysed on SDS‐PAGE. This analysis revealed a strong induction of a 19 kDa protein (Figs 3A and 5A). The search for periplasmic protein‐encoding genes within the R. gelatinosus genome database suggested that this protein might correspond to FtrA (also annotated as P19), involved in Fe3+ uptake. To ascertain that the induced protein under iron‐limiting condition corresponded to FtrA, the ftrA gene was inactivated in the wild‐type strain and the periplasmic protein content was compared under iron‐rich and iron‐depleted condition. The analysis confirmed that the identified band corresponded to FtrA as it was absent in the ΔftrATp strain (Fig. 5A). FtrA/P19 is the periplasmic iron‐binding protein of the tripartite FtrABC (EfeUOB) and P19‐Ftr1P system identified in Escherichia coli strain O157:H7 and Campylobacter jejuni respectively (Cao et al., 2007; Liu et al., 2018). FtrA encoding gene in R. gelatinosus was found within a putative Fur‐regulated operon of five genes, ftrAPBCD, also encoding an outer membrane protein (FtrP), a putative periplasmic Cu‐oxidase protein (FtrB), a permease (FtrC) and a membrane polyferredoxin (FtrD) (Fig. S2). To assess whether this iron uptake system was required for excess metal resistance in R. gelatinosus, the ftrA gene was also inactivated in copAKm and ΔcadAKm mutant. Analyses of growth inhibition in the ΔcopAftrA or ΔcadAftrA mutants showed that these mutants were more sensitive to excess metal than the single mutants copA − (Fig. 5B) and ΔcadA (Fig. 5C) and revealed that, as FbpA, FtrA was also involved in the tolerance to excess Cu+ or Cd2+. Semi‐quantitative RT‐PCR also confirmed the induction of ftrA under iron‐depleted condition and under excess Cu+ or Cd2+ (Fig. S3).

Fig. 5

Ftr iron import is also required for copper and cadmium tolerance when the efflux system is defective. SDS‐PAGE showing the absence of FtrA (P19) in the periplasmic fraction of the corresponding mutant (A). Cells were grown in malate (M) or in iron‐depleted malate medium (‐Fe). FtrA is also involved in copper and cadmium tolerance. Growth inhibition of the ΔcopAftrA and ΔcadAftrA in malate medium supplemented with increasing CuSO4 (B) or CdCl2 (C) concentrations under photosynthesis condition. Growth inhibition of the ΔftrAfbpA in malate medium supplemented with increasing CuSO4 (D) or CdCl2 (E) concentrations under photosynthesis condition, in comparison with the wild‐type and the single mutants. Cells were grown overnight for 18 h at 30°C under PS condition, before OD680nm measurement. Results are the average of 3 independent experiments.

Given the evidence that both iron uptake systems Fbp and Ftr contributed to the Cu+ and Cd2+ resistance, we anticipated that inactivation of the two iron uptake systems may display increased sensitivity to excess Cu+ or Cd2+. We therefore generated a double mutant ΔftrAfbpA and analysed its ability to grow in the presence of excess Cu+ or Cd2+. The ΔftrAfbpA growth was not affected in malate medium, probably thanks to the presence of other iron transporters like FeoAB. Nonetheless, growth of the ΔftrAfbpA mutant was more affected under iron‐depleted condition suggesting that FeoAB is not sufficient under iron‐limiting condition. As shown in Figure 5D and E, the ΔftrATp or fbpAKm single mutants behave as the wild‐type and tolerate metal excess in the medium. In sharp contrast, the ΔftrAfbpA double mutant displayed sensitivity to both Cu+ and Cd2+. Taken together, these data indicated that iron import through Fbp and Ftr systems (Fig. S2) was required to face excess Cu+ or Cd2+ in the cytoplasm.

Iron import is also required for Cu+ tolerance in Vibrio cholerae

Vibrio cholerae, an aquatic bacterium that can infect human intestine to cause diarrhoeal diseases, has a copper efflux system comparable to that of R. gelatinosus. Both include the Cu+‐ATPase CopA and the periplasmic CopI protein induced by excess copper, while lacking the Cus system. Taking into account the results presented here for R. gelatinosus, we assumed that inactivation of the iron transporter FbpA in a copA‐deficient background would lead to a hypersensitive strain to excess copper in V. cholerae as well. To test this assumption, the effect of CuSO4 on the growth in liquid LB medium of the ΔcopAfbpA disruption strain was compared with the ΔfbpA, ΔcopA and wild‐type strains (Fig. 6). As expected, the ΔcopA strain exhibited decreased resistance to CuSO4. Yet, growth inhibition by CuSO4 was more pronounced in the double mutant ΔcopAfbpA, confirming the role of iron in Cu+ tolerance in this bacterium. We should stress out that in contrast to R. gelatinosus, V. cholerae possesses a wide battery of iron transport systems that could help the bacterium to face excess copper in order to occupy different niches (Payne et al., 2016).

Fig. 6

FbpA is also involved in copper tolerance in V. cholerae. Growth inhibition of the ΔcopAfbpA mutant in comparison with the WT, ΔcopA and the ΔfbpA challenged with increasing CuSO4 concentrations under aerobic condition. Cells were grown overnight for 16 h at 37°C before OD600nm measurement. Results are the average of 3 independent experiments.

Under Cu+ excess stress, Fe‐Sod activity correlates with an iron dysregulation status in E. coli, but not in R. gelatinosus and V. cholerae

It is well established that the expression of superoxide dismutases (SOD), Mn‐Sod SodA and Fe‐Sod SodB is regulated by iron status in bacteria. This regulation involves the Fur repressor and in some species the sRNA RyhB that downregulate nonessential iron‐containing proteins when iron is limited (Masse and Gottesman, 2002; Troxell and Hassan, 2013; Imlay, 2019). In E. coli, under iron‐limiting condition, the Fe‐Sod expression was repressed, whereas the Mn‐Sod was induced to convert superoxide into H2O2 and protect the cell from oxidative stress (Carlioz and Touati, 1986). Fe‐Sod and Mn‐Sod activities could thus reflect the iron status within the cells. Our results, showing that excess Cu+ induced the induction of iron transporter and likely elicited iron limitation, prompted us to check the activity and expression of the superoxide dismutases in R. gelatinosus, V. cholerae and E. coli in response to excess Cu+.

In contrast to E. coli, in which both the Fe‐Sod and Mn‐Sod are active, R. gelatinosus genome encodes only the cytosolic Fe‐Sod superoxide dismutase. To understand how the bacterium controls the expression of the Fe‐Sod to deal with excess Cu+, soluble fractions from wild‐type and copA − mutant grown in presence of excess CuSO4 were analysed. In‐gel SOD activity assay and Western blot analyses showed no differences in the Fe‐Sod activity in the wild‐type samples in response to Cu+ (Fig. 7A). On the contrary, in the copA − mutant, exposure to Cu+ resulted in a ~ 2‐fold increase of the SodB activity and amount in response to excess Cu+. In V. cholerae, genes encoding the Mn‐Sod (vc2696 or sodA) and the Fe‐Sod (vc2045 or sodB) are found in the genome. However, in‐gel SOD activity assay indicated that only SodB was active in our condition (Fig. 7B). As for the effect of copper in R. gelatinosus, both the activity and amount of SodB were induced (around 2‐ and 6‐fold, respectively) in V. cholerae copA − cells, when challenged with CuSO4 (Fig. 7B). The periplasmic copper protein CopI (Durand et al., 2015) is noteworthy induced under copper stress in R. gelatinosus and V. cholerae in response to Cu+. Cadmium stress in R. gelatinosus and V. cholerae also resulted in the induction of SodB in the CadA ATPase‐deficient mutants (accompanying paper (Steunou et al., 2020b)).

Fig. 7

Effect of excess copper on SOD activity. Induction of Fe‐Sod activity and expression in response to excess CuSO4 in R. gelatinosus wild‐type and copATp mutant (A). Induction of Fe‐Sod activity and expression in V. cholerae wild‐type and ΔcopA mutant (B). Induction of Mn‐Sod activity and expression in the soluble fractions from the WT, ΔcopA, ΔcusAcueO and ΔcopAcusAcueO strains of E. coli (C). Proteins were labelled according to their size and expression profile. All cells were grown overnight in appropriate medium supplemented or not with 100 µM CuSO4 (Cu2+).

Together, both R. gelatinosus and V. cholerae showed induction of SodB under Cu+ stress. Importantly, SodB is the only functional SOD in these bacteria as R. gelatinosus lacks the Mn‐Sod and Mn‐Sod is not expressed or not functional in V. cholerae under our condition. Therefore, R. gelatinosus and V. cholerae can only express the Fe‐Sod to deal with excess metal and superoxide.

We also analysed the SOD activity in response to excess CuSO4 in E. coli for a set of Cu+ efflux mutants. As clearly shown in Figure 7C, addition of copper to the growth medium strongly induced the activity of the Mn‐Sod in ΔcopA, ΔcusAcueO and ΔcopAcusAcueO mutants but not in the wild‐type, demonstrating that accumulation of CuSO4 induced the Mn‐Sod. Concomitant to SodA induction, a drastic decrease in the Fe‐Sod activity was observed. To further support these results, the expression level of the Mn‐Sod was assessed on Western blots using the HisProbe that reacted with the 5 histidines in the N‐ter of the E. coli Mn‐Sod (Fig. 7C). The data confirmed the increased amount of Mn‐Sod in the cytosolic fractions of the copper efflux‐deficient mutants. Similar experiments were conducted using the ZntA mutant of E. coli to check the effect of Cd2+ on SOD activity. Likewise, excess Cd2+ in the medium resulted in an induced activity of the Mn‐Sod and decreased activity of the Fe‐Sod only in the ZntA mutant (accompanying paper (Steunou et al., 2020b)). Altogether, these results showed that the accumulation of copper affected the expression and activity of superoxide dismutases presumably because of iron homoeostasis dysregulation linked to copper stress.

Discussion

The interplay between copper and iron dates back to the middle of the 19th century when copper was used as a therapeutic agent to treat anaemia. Copper was later shown to indirectly enhance haemoglobin formation by increasing iron absorption. This was brilliantly documented in the review by Paul Fox in 2003 (Fox, 2003). Another copper/iron interplay occurs in macrophages in the immune system. It was proposed that overloading the phagosome with toxic metal such as copper and limiting the availability of essential ions like iron are used to poison intracellular pathogens (Hood and Skaar, 2012; Neyrolles et al., 2015). Several other indirect lines of evidence, mainly transcriptomics, support the involvement of iron in response to heavy metal excess (Gross et al., 2000; Stadler and Schweyen, 2002; Teitzel et al., 2006; Yoshihara et al., 2006; Houot et al., 2007; Chillappagari et al., 2010); however, this has not been directly tested. In this study, using a random mutagenesis approach, we focused on metal excess‐induced toxicity and response and demonstrated that iron transport/uptake plays a key role in Cu+ as well as Cd2+ excess‐induced stress.

Previous studies in bacteria and eukaryotes showed that exposed [4Fe‐4S] clusters are susceptible to damage by metals such as Cu+, Ag+ and Cd2+ (Macomber and Imlay, 2009; Xu and Imlay, 2012; Vallieres et al., 2017). It was suggested that this led to the accumulation of ‘free iron’ and potentially increased ROS stress via Fe‐catalysed Fenton chemistry. However, beyond the generated oxidative stress that can be scavenged by the ROS detoxification system, this situation of [4Fe‐4S] cluster degradation and loss of key metabolic enzymes will presumably force bacteria to react quickly and repair these clusters to survive. The nature of the iron source used to rebuild these [4Fe‐4S] clusters is an opened question. Keyer and Imlay elegantly showed that upon exposure to peroxynitrite in E. coli, the [4Fe‐4S] cluster of dehydratases was degraded and the ‘released iron’ originating from [4Fe‐4S] degradation was rapidly sequestered and no more available for cellular processes (Keyer and Imlay, 1997). Therefore, for [4Fe‐4S] cluster repair, most of the iron was imported from the external medium. Iron uptake is thus required to supply sufficient iron to the Fe‐S machinery in response to peroxynitrite stress (Keyer and Imlay, 1997). Here, our study showed that similar events might take place in the case of Cu+ and Cd2+ stress, for which excess metal damage exposed [4Fe‐4S], thus causing the release of iron. Paradoxically, despite the presence of Fe in the medium, cells might perceive the situation as an ‘iron‐starvation’ situation and respond to it by inducing the expression of iron uptake systems to enhance Fe‐import (Fig. 8). This raises the question of what happens to released iron. In addition to its possible sequestration by iron storage proteins, released iron could also be exported out of the cells by Fe2+‐efflux transporters to prevent iron overload and related damages. Our ICP‐MS data did show a decrease in the amount of total iron content in copATp and ΔcadATp cells in agreement with a putative Fe2+‐efflux in these mutants, as in other species (Pi and Helmann, 2017). An additional stress imposed by Cu+ and Cd2+ excess was proposed to occur on the Fe‐S cluster biogenesis (ISC) machinery. Indeed, Cu+ and Cd2+ appeared to directly bind and inhibit components of the E. coli ISC machinery (Chillappagari et al., 2010; Tan et al., 2014, 2017; Roy et al., 2018). Inhibition of the ISC machinery will increase the demand for iron and would thus contribute to the activation of iron uptake systems.

Fig. 8

Interplay between metal efflux, iron uptake and ROS detoxifying system. In the absence of the efflux ATPases CopA or CadA accumulation of Cu+ or Cd2+ in the cytosol led to the degradation of [4Fe‐4S] clusters. ‘Released iron’ originating from this degradation could rapidly be sequestered or exported out of the cells, thus generating an iron‐depleted status in the poisoned cells. Consequently, iron uptake is induced to rebuild [4Fe‐4S] clusters and the superoxide dismutases are induced. It was supposed that released iron could generate ROS, but oxidative stress may be further exacerbated by the induction of Fe‐uptake in response to damaged Fe‐S clusters. ETC: electron transfer chains (respiration or photosynthesis) are also poisoned by excess metal. ETC generate ATP for the ATPases but can also generate superoxide.

In contrast to our results, Helbig et al. reported that iron uptake was downregulated when E. coli cells were exposed to Cd2+ (Helbig et al., 2008). Nevertheless, these experiments were performed with strains with an affective Cd2+ efflux system (ZntA) that was able to expel Cd2+ from the cytoplasm and allow normal cellular growth. Indeed, a 10‐min exposure to 100 µM CdCl2 resulted in a significant induction of zntA expression (Helbig et al., 2008). On the contrary, in a recent study, Xu et al. reported that Fe‐uptake was required to maintain cell fitness during Zn2+ excess in E. coli and that excess Zn2+ led to a transient dysregulation of iron uptake (Xu et al., 2019). They showed that iron uptake was upregulated by excess Zn2+ and once bacteria were adapted to excess Zn2+, the system was switched off. Obviously, if the efflux system is efficient, it will provide sufficient protection against excess metal, as the need of upregulation of defence, iron uptake or Fe‐S repair systems would no longer be justified or necessary to sustain growth. Thus, because the activity of the metal detoxification pumps can hide metal toxicity effects, mutants that lack the efflux pumps are suitable to test elevated metal stress and its consequences on iron homoeostasis. In P. aeruginosa, excess Cu+, Zn2+, Co2+, Ni2+ or Cd2+ affects siderophore synthesis (Visca et al., 1992; Teitzel et al., 2006; Schalk et al., 2020). It was shown that genes encoding synthesis of pyoverdine were upregulated in response to Cu+, Zn2+ and Cd2+ suggesting that pyoverdine may protect the cells by sequestering heavy metals. Nevertheless, although able to chelate Cu+, Zn2+ and Cd2+, pyoverdine is highly iron specific and its affinity to iron is much more higher, indicating that it might rather provide the bacterium with iron under excess metal (Visca et al., 1992; Teitzel et al., 2006; Schalk et al., 2020). In contrast, pyochelin has a broader specificity for cations and its synthesis was repressed by Cu+, Cd2+, Co2+ and Ni2+. Inhibition of its synthesis under Cu+ excess would protect bacteria from Cu+ poisoning (Teitzel et al., 2006). In the oysters’ pathogenic Vibrio tasmaniensis LGP32 bacterium, comparative transcriptomic showed that the Cu+‐efflux ATPase CopA, as well as many iron (including FbpABC), siderophore uptake systems and Fe‐S biogenesis genes were induced in the phagosomes (Vanhove et al., 2016).

The mechanisms by which Cu+ or Cd2+ induces iron uptake are not yet well studied. Metals such as Co2+, Zn2+ or Cu+ can displace or replace Fe2+ in the ferric uptake regulator Fur and may therefore affect Fur for DNA binding (Adrait et al., 1999; Mills and Marletta, 2005; Vitale et al., 2009). The facts that (i) induction of iron uptake system occurs with a variety of cations or superoxide generators that target [4Fe‐4S], (ii) the demand for iron for the [4Fe‐4S] regeneration is high and (iii) the observation that released iron is rapidly sequestered rather argue for a mechanism in which stressed cells respond to iron or Fe‐S depletion. While the response seems to be Fur‐independent for peroxynitrite stress (Keyer and Imlay, 1997), the involvement of regulatory factors, including Fur, remains to be investigated for the metallic stress.

Concomitant to the degradation of [4Fe‐4S] clusters and iron homoeostasis dysregulation by excess Cu+ or Cd2+, stressed cells also induce the expression of superoxide dismutases SodA or SodB. One may ask whether these enzymes are solely induced because of iron dysregulation or because they are required under Cu+ or Cd2+ stress. The induction of SODs is not fortuitous but functional. Indeed, mutants in which both the efflux system CopA or CadA and the superoxide dismutase SodB are missing are also extremely sensitive to Cu+ and to Cd2+ stress (accompanying paper (Steunou et al., 2020b)).

The overall findings and the bacterial model depicted in Figure 8 to rationalize metal excess toxicity in bacteria could also apply to eukaryotes. It was recently shown that excess copper targets Fe‐S clusters and ferredoxins in yeast (Vallieres et al., 2017) and various studies reported that Cu+, Cd2+, Ni2+, Cr3+ or Co2+ exposure in yeast (Stadler and Schweyen, 2002; Alkim et al., 2013; Foster et al., 2014; Johnson et al., 2016), in plants (Yoshihara et al., 2006) or in human (Fox, 2003; Davidson et al., 2005) triggers iron deficiency responses and the induction of iron uptake systems. Together, these data point to the essential role of iron homoeostasis in response to excess heavy metal in eukaryotes as well, although other hypotheses including competition between metals and iron uptake may account for the iron deficiency status (Rubinelli et al., 2002; Davidson et al., 2005).

Extensive use of antibiotics in health care and agriculture has led to an increase in antibiotic resistance (Asante and Osei Sekyere, 2019). Metals that exhibit antimicrobial properties are used as alternatives to antibiotics in farming and agriculture. Nevertheless, the use of metals at high concentrations contributes to environment contamination and to co‐selection of antibiotic resistance genes (Baker‐Austin et al., 2006; Purves et al., 2018; Rensing et al., 2018; Bischofberger et al., 2020). The combination of excess Cu2+, at very low concentration, and iron limitation poses a serious challenge to bacteria as shown in this work. This combination could be exploited in the course of metal‐based antimicrobial treatments in agriculture, farming and drug design strategies.

Experimental procedures

Bacterial strains and growth

R. gelatinosus was grown at 30°C, in the dark aerobically (high oxygenation: 250 ml flasks containing 20 ml medium) or under light microaerobically (photosynthetic condition, in filled tubes with residual oxygen in the medium) in malate growth medium. E. coli and V. cholerae were grown overnight at 37°C in LB medium. Antibiotics (50 µg ml−1), kanamycin (Km), spectinomycin (Sp), streptomycin (Sm) and trimethoprim (Tp), were added when appropriate.

Bacterial strains and plasmids are listed in Table S1. For growth inhibition under photosynthetic condition, strains were grown overnight in filled tubes and OD680nm was measured. For V. cholerae OD600nm was measured after overnight growth using the Tecan Infinite M200 luminometer (Tecan, Mannedorf, Switzerland).

Transposon mutagenesis and mutant selection

R. gelatinosus ΔcopRcadR mutant was mutagenized using the EZ‐Tn5‐KAN‐2‐Tnp Transposome Kit (Epicentre) and following the manufacturer's protocol. Cells were transformed by electroporation (Steunou et al., 2013). Transformants were selected by plating cells onto malate plates containing Km, Tp and Sp. The colonies were first transferred on plates containing 50 µM of CuSO4 to identify colonies sensitive to CuSO4. Cu+‐sensitive clones were then screened for their Cd2+ sensitivity on 50 µM of CdCl2. Among 4000 transformants screened, 10 clones were confirmed as sensitive to both Cu+ and Cd2+. Transposon insertion site of each clone was determined by sequencing the flanking region of the transposon.

Gene cloning, plasmid constructions and mutant strain construction

Standard methods were performed according to Sambrook et al. (1989) unless indicated otherwise. KS‐fbpAKm was obtained from DNA isolated from ΔcopRcadR‐fbpA2::Tn5 digested by SgrAI and cloned into Bluescript KS+. This plasmid was used to inactivate fbpA by electroporation in the wild‐type, copATp, ∆cadATp and ∆ftrATp mutants of R.gelatinosus. The fbpBC DNA fragment (2208 bp) was cloned into pGEM‐T by PCR amplification using the primers fbpBC‐For and fbpBC‐Rev (Table S2). The resulting plasmid pGfbpBC was digested with Eco47III to delete 1088 fragment. The Tp cassette from p34S‐Tp was inserted into the Eco47III site within fbpBC to create pGfbpBC::Tp. This plasmid was used to inactivate fbpBC in the wild‐type, in copAKm and ∆cadAKm mutants of R.gelatinosus.

To inactivate ftrA, a 1 kb fragment was amplified using the primers ftrA‐For and ftrA‐Rev and cloned into the PCR cloning vector pGEM‐T to give pGftrA. A 0.2‐kb StuI and MscI fragment was deleted and replaced by the 0.7‐kb Tp resistance cassette to disrupt the ftrA gene. The resulting recombinant plasmid was designated pGftrATp. This plasmid was used to inactivate ftrA by electroporation in the wild‐type, in copAKm, ∆cadAKm and fbpAKm mutants in R. gelatinosus. Transformants were selected on malate plates supplemented with the appropriate antibiotics under aerobic condition. Genomic DNA was prepared from the ampicillin‐sensitive transformants, and confirmation of the presence of the antibiotic resistance marker at the desired locus was performed by PCR. The pET28bFbpAH6 plasmid was generated by cloning fbpA in the pET‐28b plasmid. The fbpA gene was amplified by PCR from R. gelatinosus DNA using the fbpA‐NcoI and fbpA‐HindIII primers and cloned in pET‐28b plasmid. The plasmid was integrated at the fpbA locus on the chromosome of R. gelatinosus by selecting for kanamycin resistance. The integration of this plasmid at the fbpA locus was confirmed by PCR on genomic DNA.

To construct ∆copA in V. cholerae, ~ 600 bp fragments upstream and downstream to the copA (vc2215) were amplified using primers oYo848 and oYo849, and oYo850 and oYo851 respectively. Resulting fragments were cloned into SmaI site of pCVD442 vector (Donnenberg and Kaper, 1991) using Gibson Assembly (Gibson et al., 2009), resulting in pEYY345. Similarly, to construct ∆fbpA in V. cholerae, ~ 600 bp fragments upstream and downstream to the fbpA (vc0608) were amplified using primers oYo854 and oYo855, and oYo856 and oYo857, respectively, and cloned into the pCVD442 vector, resulting in pEYY346. Subsequently, plasmid was transferred by conjugation to introduce mutation in V. cholerae by allelic exchange (Donnenberg and Kaper, 1991). DNA oligonucleotides used in this study are listed in Table S2.

SOD in‐gel activity assay on non‐denaturing gel electrophoresis

20 µg of soluble proteins was separated on a 10%non‐denaturing polyacrylamide gel and stained for SOD activity as described in Weydert et al. (48), with minor modifications. Incubation with TEMED (0.85%) and Riboflavin‐5‐Phosphate (56 µM) was performed for 15 min at light and room temperature (RT), followed by the addition of nitroblue tetrazolium (2 mg ml−1) and a 15 min incubation in the dark at RT. Gel was washed twice in ddH2O and left in ddH2O at RT on a light box until SOD‐positive staining appeared.

Western blot and HisProbe‐HRP detection

Equal amount of soluble proteins (20 µg) or periplasmic fractions was separated on SDS‐PAGE and transferred onto a Hybond ECL Polyvinylidene difluoride membrane (GE Healthcare). Membrane was then probed with the HisProbe‐HRP (horseradish peroxidase, from Pierce) according to the manufacturer's instruction. Positive bands were detected using a chemiluminescent HRP substrate according to the method of Haan and Behrmann (Haan and Behrmann, 2007). Image capture was performed with a ChemiDoc camera system (Bio‐Rad).

Inductively coupled plasma mass spectrometry (ICP‐MS) measurements

The concentrations of total iron in cells were measured by ICP‐MS as described in Grassin‐Delyle et al. (2019). Pellets of overnight grown cells in the presence or absence of 100 µM CuSO4 or 100 µM CdCl2 were washed with cold PBS buffer five times and stored at −80°C prior to ICP‐MS analyses. Metal concentrations in the samples were calculated according to the standard curves. The values were normalized by the culture absorbance at 680 nm.

Periplasmic fraction preparation

R. gelatinosus cells were washed twice with 50 mM Tris‐HCl (pH 7.8) and resuspended in the same buffer in the presence of 0.45 M sucrose, 1.3 mM EDTA and 0.6 mg ml−1 lysozyme. After 1h of incubation at 30°C with soft shaking, the extract was centrifuged for 15 min at 6000 r.p.m. The supernatant corresponds to the periplasmic fraction (Durand et al., 2015).

mRNA preparation and RT‐PCR

Total RNA was purified from wild‐type cells grown in photosynthesis condition in malate medium (M), in an iron‐depleted malate medium (‐Fe) or in malate medium supplemented with 1 mM of CuSO4 or 1 mM of CdCl2 as described in (Steunou et al., 2013). For semi‐quantitative RT‐PCR, cDNA was generated from 1 µg of total RNA with random hexamers using the Superscript IV (Invitrogen) and by following the manufacturer's protocol. PCR was done with 2 µl of cDNA with specific primers (Table S2) to amplify fragments of 16S, pucA (encoding the α subunit of LH2), sodB (encoding superoxide dismutase), ftrA and fbpA genes. Amplified products were analysed on a 1.4%agarose gel, and the bands were quantified using imageJ program. The relative amount was calculated based on the signal obtained in malate medium.

Conflict of interest

The authors declare that they have no conflicts of interest with the contents of this manuscript.

Author contributions

AS.S., A.D., M.B., S.L. and S.O. designed research; AS.S., ML.B., M.B., Y.Y. and S.O. performed research; AS.S., A.D., M.B., S.L., ML.B. and S.O analysed data; AS.S., M.B. and S.O. wrote the paper.

Table S1 . Bacterial strains and plasmids used in this work.

Table S2 . DNA oligonucleotides used in this work.

Fig. S1 . Tn mapping within fbpA and phenotype of the selected fbpA::Tn mutants. A. Tn mutagenesis mapping. Position of Tn within fbpA gene are indicated. B. Growth of ΔcopRcadR (1) and ΔcopRcadR‐fbpA::Tn5 mutants (2‐6) on agar plates in malate medium (M) or malate supplemented with CuSO4 or CdCl2.

Fig. S2 . A. Organization of the gene clusters and iron transport systems involved in iron acquisition in R. gelatinosus. Fur box sequences within the promoters of the identified clusters are shown. Fur boxes from fecI and fhuA iron regulated genes were used in the alignement. B. R. gelatinosus iron transport systems. FbpABC and FtrAPBCD are the major inner membrane iron transporters identified in this study. Genes encoding other systems including TonB‐dependent receptor family and siderophores importers (Fec, Fhu, Fpv…) are also present in the bacterium genome.

Fig. S3. Expression profiles (semi‐quantitative RT‐PCR) of fbpA and ftrA genes under various metal stress condition. A. Expression profiles (semi‐quantitative RT‐PCR) of fbpA, ftrA genes in WT cells grown under photosynthesis in Malate (M) medium, iron depleted Malate medium (‐Fe), or Malate medium supplemented with 1 mM CuSO4 or CdCl2. The pucA gene encoding the light harvesting II α‐subunit, sodB encoding the superoxide dismutase, and the 16S Rrna were used as references to normalize the relative expression of the induced genes. B. The results are expressed as fold changes of the expression of target genes in modified malate medium (−Fe, +Cu+, +Cd2+) relative to their expression in malate medium (M).

Click here for additional data file.