Spontaneous mutation in the AgrRS two-component regulatory system of Cupriavidus metallidurans results in enhanced silver resistance

Abstract

The uncontrolled and widespread use of (nano)silver compounds has led to the increased release of these compounds into the environment, raising concerns about their negative impact on ecosystems. Concomitantly, silver resistance determinants are widely spread among environmental and clinically relevant bacteria although the underlying mechanisms are not yet fully understood. We show that Cupriavidus metallidurans is able to adapt to toxic silver concentrations. However, none of the known silver resistance determinants present in C. metallidurans are involved in the adaptative response. Instead, increased silver resistance is achieved by the concerted action of a two-component system AgrR–AgrS, previously not associated with metal resistance, and two periplasmic proteins PrsQ₁ and PrsQ₂. Both proteins belong to an unique group of small, uncharacterized, secreted proteins restricted to the genera Cupriavidus and Ralstonia. This system gives C. metallidurans the ability to withstand much higher silver concentrations. The latter could be facilitated by the accumulation of silver ions and the formation of silver nanoparticles.

---

Significance to metallomics

In this study, the molecular adaptation to silver exposure was studied for the model metal-resistant bacterium Cupriavidus metallidurans. Although it carries many known metal and silver resistance determinants, none were involved in the adaptation response. Our results showed that a single mutation resulted in the adaptation to high silver concentrations and previously uncharacterized periplasmic proteins played a crucial role in this adaptation process. The latter are part of a larger family within this and a closely related genus, suggesting a specific adaptation potential to metals in these genera that are able to live and thrive in harsh and oligotrophic environments.

Introduction

In addition to many commercial and industrial uses, silver is widely used in health and hygiene applications such as water disinfection, topical creams for the treatment of burn wounds, dental amalgams and silver-impregnated polymers for medical devices.^1–3 The rapid development of nanotechnology further expanded its use and applications and, today, silver nanoparticles (AgNPs) are the most commonly engineered nanomaterial. AgNPs are routinely incorporated in a wide range of domestic and personal care products such as laptops, refrigerators, dietary supplements, clothing, children's toys and many more.⁴ The toxicity of AgNPs can largely be attributed to the release of Ag⁺ ions.^5–7 Silver ions interact with the cytoplasmic membrane compromising electron transfer and the proton motive force, ultimately resulting in cell death. In addition, once inside the cell, Ag⁺ ions can interact with enzymes, inhibit DNA replication and is expected to promote the production of reactive oxygen species.^7–10 Due to the excessive use of silver compounds, an increased release of silver in the environment through point source direct discharge or through wastewater effluents is expected.¹¹ In wastewater treatment plants, AgNPs are believed to be retained in sewage sludge that is often used as fertilizer for agricultural soils or send to landfills. In this manner, AgNPs can be transferred via surface runoff or leaching into the aquatic environment.^12,13 Although the fate and transformations of silver compounds in the environment are largely unknown, the possible impact on the ecosystem is an emerging cause of concern.^12,13 Moreover, concerns have been raised about the potential proliferation of silver-resistant bacteria in a manner similar to antibiotic resistant bacteria, as well as the co-selection of antibiotic resistance.^14,15 In fact, silver-resistant bacteria have been isolated from both clinical and non-clinical environments.^16–19 So far, the mechanisms of silver resistance have not been fully elucidated and different preventive strategies have been suggested. Next to chemical detoxification,²⁰ bacterial silver resistance results from active efflux systems. These efflux clusters often additionally code for periplasmic binding proteins that act as chaperones or ‘metal sponges’ by specifically binding silver.^21,22

In the present study, we assessed the adaptation potential of Cupriavidus metallidurans to toxic silver concentrations. C. metallidurans is an environmental β-proteobacterium repeatedly isolated from metal-contaminated soils^23–25 as well as other anthropogenic environments not typified by metal contamination,^26–28 including medically relevant sources.^29–32 It contains a large diversity of mobile genetic elements,^33,34 facilitating horizontal gene transfer, and a number of systems involved in silver detoxification, including the silDCBA and cusDCBAF operons coding for HME-RND-driven efflux systems located on respectively pMOL30 and the chromid,^35–38 and cupRAC coding for a P_IB1-type ATPase located on the chromosome.³⁷

Results

Spontaneous silver-resistant mutants

Spontaneous silver-resistant mutants were obtained from three different C. metallidurans strains (CH34, AE104 and NA4) by exposing each strain to a toxic concentration based on the minimal inhibitory concentration of silver for each strain. In a first series of experiments, spontaneous mutants of type strain CH34 resistant to silver were selected on LB agar supplemented with respectively 0.5 and 2 mM AgNO₃ and designated as CH34S1 and CH34S2. To reduce metal complexation, a Tris-buffered mineral medium (MM284) was used in a second series of experiments with strain AE104 and NA4. Strain AE104 is a derivative of type strain CH34 that does not contain the two megaplasmids pMOL28 and pMOL30,³⁹ which carry most of the metal detoxification functions.⁴⁰ A spontaneous mutant resistant to silver was selected after incubation in MM284 supplemented with 3 μM AgNO₃ and designated as AE104S. Strain NA4 is isolated from potable water from the Russian condensate recycle system SRV-K aboard the ISS.^41,42 In contrast to CH34 where both megaplasmids have a key role in metal detoxification, most of the metal resistance determinants are located on the chromid.⁴³ For NA4, a spontaneous mutant resistant to silver was selected after incubation in MM284 supplemented with 8 μM AgNO₃ and designated as NA4S. The minimal inhibitory concentration (MIC) of AgNO₃ for CH34S1, CH34S2, AE104S and NA4S was 4, 4, ±30 and ±10 times higher than for their parental strain, respectively (Fig. 1).

Fig. 1 Minimal inhibitory concentration of AgNO₃ for the silver-resistant mutants and respective parental strains.

All silver-resistant mutants harbour a common subset of differentially expressed genes

Genome-wide expression profiling of CH34S1, CH34S2, AE104S and NA4S under non-selective growth conditions (i.e. in the absence of silver) showed a substantial number of differentially expressed genes in each silver-resistant mutant compared to their respective parental strain. In CH34S1, 301 genes were differentially expressed (133 up- and 168 down-regulated), 293 genes (158 up- and 135 down-regulated) were differentially expressed in CH34S2, 179 genes (75 up- and 104 down-regulated) in AE104S and 308 genes (204 up- and 104 down-regulated) in NA4S. The C. metallidurans systems known to be involved in silver detoxification (silDCBA and cusDCBAF operons coding for HME-RND-driven efflux systems, and cupRAC coding for a P_IB1-type ATPase) were not differentially expressed in all mutants (ESI,† Table S1). In addition, exposure to 1 μM AgNO₃ elicited a response in CH34S1 and CH34S2 similar to that in CH34 with upregulation of silDCBA, cusDCBAF and cupRAC (ESI,† Table S1). However, induction of these silver detoxification systems in the parental CH34 strain does not result in a silver resistant phenotype like that of the evolved strains, which indicated that other mechanisms are at play.

On the other hand, only seven genes were commonly upregulated in all mutants under non-selective conditions (Table 1). In addition, their expression was not induced by AgNO₃ (ESI,† Table S1). No commonly down-regulated genes were identified. The identified genes code for the response regulator AgrR, an outer membrane protein and five conserved hypothetical proteins. The response regulator AgrR, which is part of the two-component regulatory system AgrR–AgrS, has already been shown to be constitutively expressed in CH34S1.³⁵ Its cognate target, located upstream of the agrRS locus and divergently transcribed, codes for the uncharacterized RND-driven efflux system AgrABC with characteristics of heavy metal efflux (HME) as well as hydrophilic/amphiphilic compounds (HAE) RND-driven efflux systems.³⁵ Three of the conserved hypothetical proteins belong to an uncharacterized protein family of relatively small proteins, ranging in size from 69 to 165 amino acid residues. Strains CH34 and NA4 carry 22 and 26 genes encoding such proteins, respectively. Although their function remains to be elucidated, ten members of this family were found to be transcriptionally induced in CH34 in response to different heavy metals,³⁷ including the pMOL30-encoded copQ, part of a complex 33-gene region involved in copper resistance, and czcJ, part of a 25-gene cluster involved in cadmium, zinc and cobalt resistance. Based on the first observations for the cop cluster, these homologs were previously designated as CopQ-like protein family members.⁴⁴

Table 1 Common differentially expressed genes in all silver resistant mutants and their possible predicted binding site for AgrR

CH34^a	NA4^a	Function	Ident%^b	Transcription (fold change)^c				AgrR^d
				CH34S1	CH34S2	AE104S	NA4S
a CH34 (Rmet_) and NA4 (CmetNA4v1_) locus tag based on MaGe^45 annotation v1 (ESI, Table S3). b % protein identity between CH34 and NA4 copy. c Fold change of gene expression measured by microarray expression profiling for each mutant compared to its parental strain in non-selective growth conditions. d AgrR binding site prediction (Only values above 0.8 are shown).
0477	40 157	Conserved hypothetical, CopQ-like	100	9.0	9.1	3.9	15.8
1751	170 096	DNA-binding response regulator, AgrR	100	78.5	65.4	81.0	54.2	0.99
3571	10 101	Conserved hypothetical, CopQ-like, PsrQ1	100	23.8	22.0	21.2	2.9
4461	570 171	Conserved hypothetical, CopQ-like, PsrQ2	100	81	88.6	106.6	26.2	0.83
4464	570 173	Conserved hypothetical, CzcL	98	5.1	4.9	2.1	2.8	0.87
4595	570 182	Conserved hypothetical, CzcI2	100	7.7	8.2	3.1	2.7	0.84
5118	400 150	Outer membrane protein (porin), OmpC family	100	8.6	9.0	5.8	3.6

---

Whole genome sequencing of mutants corroborates the involvement of the two-component regulatory system AgrR–AgrS

Sequencing of the CH34S1, CH34S2 and NA4S genomes showed one commonly affected gene, agrS, coding for the histidine kinase of the two-component regulatory system AgrR–AgrS (full list in ESI,† Table S2). In CH34S1 and CH34S2, agrS was inactivated by insertion of Insertion Sequence (IS) element ISRme3 and IS1086, respectively.³⁴ Similarly, PCR and sequence analysis of AE104S showed an insertion of ISRme3 in agrS (data not shown). In NA4S, a C–T substitution occurred in agrS, which resulted in an arginine to cysteine substitution at residue 414 (of 469). Notably, agrA is inactivated by an IS insertion in the parental NA4 strain. In addition, CH34S2 carries a frameshift mutation in agrA (4 bp duplication) and expression of agrCBA in CH34S1 under non-selective conditions was comparable to that in the parental strain (putatively elicited by a 7 bp duplication in the promoter region of agrCBA) (ESI,† Tables S1 and S2). These results refuted a putative role of the AgrCBA RND-driven efflux system in the acquired silver resistance. However, all mutants showed an increased expression of agrR as well as a mutation in agrS, supporting an important role for this two-component regulatory system in the observed increased silver resistance. This role was further confirmed by deleting agrRS in NA4S, which decreased silver resistance to the parental level as well as by deleting agrS in NA4, which increased silver resistance to the level of NA4S (Fig. 2). Noteworthy, when grown in LB, the stationary phase cell density of NA4 ΔagrS was comparable to that of NA4S and slightly less than that of NA4 (Fig. S1, ESI†).

Fig. 2 Survival of the different strains and constructs in the presence of AgNO₃. Viable count after 3 days on LB supplemented with 0.5 mM AgNO₃ normalized to the viable count on LB agar for C. metallidurans NA4, the silver-resistant mutant NA4S, all insertional knock-out mutants of NA4 and NA4S (preceding Δ) and complementation strains (designated with pBBR). pBBR-prsQ^SDM₂: complementation of prsQ₂ with a frameshift mutation in the start codon; pBBR-prsQ^ΔSP₂: complementation of prsQ₂ without signal peptide. Values represent the average and standard deviation of three biological replicates. The dotted line corresponds to the detection limit of this technique.

Insertional inactivation and complementation elucidated an important role for prsQ₁ and prsQ₂

Next, the role of the commonly upregulated genes was identified. Because the observations were strikingly similar for the four mutants and since strain CH34 harbours almost three times more insertion sequence elements compared to NA4,⁴³ affecting genome plasticity and adaptability,⁴⁶ further work was carried out with NA4S. Insertional inactivation mutants were constructed for each gene and the absence of the corresponding mRNA confirmed (ESI,† Fig. S2). No significant differences in viable count on LB were observed (ESI,† Fig. S3). Next, silver resistance was assessed by total viable count on LB supplemented with 0.5 mM AgNO₃. Negligible growth was observed for NA4, whereas 5.81 ± 0.08 log CFU ml⁻¹ were counted for NA4S (Fig. 2). Noteworthy, CH34S1, CH34S2 and AE104S and their respective parental strains CH34 and AE104 showed comparable total viable counts on LB and LB supplemented with 0.5 mM AgNO₃ as NA4S and NA4, respectively.

The viable count on LB with 0.5 mM AgNO₃ reduced to a level similar to that of NA4 for inactivation of agrRS and two of the copQ-like genes, consequently designated prsQ₁ and prsQ₂ (for periplasmic resistance to silver related to the CopQ-like family). Furthermore, plasmid-based complementation of NA4S prsQ₁ and prsQ₂ mutants restored transcription levels (ESI,† Fig. S2) and silver resistance (Fig. 2). As the function of the CopQ-like protein family members is unstudied, we confirmed that the actual protein is involved in the acquired silver resistance by introducing a −1 frame shift in the start codon of prsQ₂ on pBBR-prsQ₂, which abolished complementation (Fig. 2). In addition, we confirmed that translocation across the inner membrane is necessary by deleting the 26-aa signal peptide (based on PrediSi and SignalP signal peptide prediction) of prsQ₂ on pBBR-prsQ₂, which again abolished complementation (Fig. 2).

In a next step, the link between the commonly affected regulatory genes, agrS and agrR, and the other identified genes was evaluated. Phylogenetic footprinting was used to discover the DNA-binding motif for AgrR (Fig. 3). The promoter region of agrR from closely and more distantly related species was compared in silico by several motif detection algorithms (MotifSampler,⁴⁷ AlignACE,⁴⁸ MEME,⁴⁹ Weeder⁵⁰ and RSAT⁵¹), followed by a Markov clustering on the outcome of the models. In this way, a possible regulatory motif for agrR was found.

Fig. 3 Sequence logo for the predicted AgrR-binding motif.

The latter was subsequently used in a genome-wide screening to identify possible AgrR target genes (Table 1). For the commonly upregulated genes, an AgrR binding motif was predicted (cut-off score 0.8) in the promoter region of agrR, prsQ₂, czcL and czcI₂. Only deletion of agrRS and prsQ₂ reduced silver resistance of NA4S to the parental level. To identify AgrR regulation on prsQ₂ expression further, a heterologous dual expression/reporter system was designed. In this Escherichia coli system, a gfp promoter probe vector to measure prsQ₂ expression levels (P_prsQ2-gfp) was combined with the regulatory genes placed under the control of the arabinose-inducible araBAD promoter. Arabinose induction of the NA4S-type agrRS (pBAD-agrRS_R414C) strongly induced prsQ₂ expression (>30-fold) in contrast to arabinose induction of native agrRS (Fig. 4). The agrRS operon in NA4S carried a C–T substitution in agrS, resulting in an arginine to cysteine substitution at residue 414. The Arg-414 residue corresponds for instance to Arg-434 of the histidine kinase PhoQ (part of the two-component regulatory system PhoP-PhoQ in E. coli), which is essential for catalysis.⁵² This indirectly indicated that AgrR most probably binds to the promoter region of prsQ₂ in its unphosphorylated state. To support this, a variant of pBAD-agrRS was constructed in which the Asp-51 codon of agrR was replaced with an alanine codon. The Asp51 residue of AgrR corresponds for instance to Asp-55 of Spo0A (part of the two-component regulatory system Spo0A-KinD in Geobacillus stearothermophilus) and Asp-57 of CheY (part of the two-component regulatory system CheY-CheA in E. coli), which are both phosphorylation sites.^53,54 Arabinose induction of agrR carrying a dephosphomimetic mutation (pBAD-agrR_D51AS) strongly induced prsQ₂ expression to levels similar to that of the arabinose-induced NA4S-type agrRS (pBAD-agrRS_R414C) (Fig. 4).

Fig. 4 Overlay histograms showing the flow cytometric distribution of prsQ₂ promoter activity (P_prsQ2-gfp) in E. coli DG1 in the presence (induced with 0.2% arabinose; dark grey histogram) or absence (uninduced; light grey histogram) of AgrRS (pBAD-agrRS), NA4S-type AgrRS (pBAD-agrRS_R414C) or AgrR with a dephosphomimetic mutation (pBAD-agrR_D51AS).

Next to prsQ₂, also deletion of prsQ₁ reduced silver resistance of NA4S to the parental level. Although genome-wide expression profiling indicated increased expression in NA4S compared to NA4 (2.9-fold; Table 1), this could not be confirmed by a plasmid-based promoter probe reporter. No significant difference in fluorescence was measured for pSCK108-P_prsQ1-gfp in NA4 and NA4S, respectively. In contrast, and in agreement with the high differential expression observed via genome-wide expression profiling, significant difference in fluorescence was measured for pSCK108-P_prsQ2-mcherry in NA4 and NA4S, respectively (data not shown).

Sequence analysis and structural prediction of PrsQ₁, PrsQ₂ and other CopQ-like family members

After confirmation that the actual PrsQ₁ and PrsQ₂ proteins are involved in the acquired resistance (see above), we further examined their protein sequence and structure to get possible information on their function. De novo prediction of sequence motifs was performed at the MEME-server⁴⁹ for the CopQ-like family member protein sets of C. metallidurans strains CH34 and NA4 (22 and 26 sequences, respectively). This indicated that they invariably contain a signal peptide and one or more typical Dx[YF]xxG domain motifs (Fig. 5). Searching for CopQ-like family members across the available protein sequences of 7927 complete genomes based on an extended 10-aa search pattern, the presence of a signal peptide and size, retrieved 787 protein hits. All hits belonged to genomes of the Cupriavidus and Ralstonia genera. Finally, all hits were used to predict a general sequence motif in CopQ-like proteins (Fig. 5).

Fig. 5 Sequence motif discovered in 787 CopQ-like proteins (left). Sequence alignment of PrsQ₁ and PrsQ₂ (right). The sequences without signal peptide (residues 1–22 and 1–26, respectively) were aligned with MUSCLE and the discovered motif sequences are upper/underlined and color-coded accordingly. The standard consensus symbols are shown between both sequences.

A preliminary prediction towards a structural model of the PrsQ₁ and PrsQ₂ proteins by using the PHYRE2 (http://www.sbg.bio.ic.ac.uk/phyre2) and RaptorX (http://raptorx.uchicago.edu/) prediction software^55,56 indicated that these proteins do not fold. Instead, they form intrinsically disordered polypeptides, which could be explained by the abundance of chain-disturbing glycines (allow tight turns) and prolines (reduce structural flexibility) that are seemingly evenly spaced. The latter may cause the aromatic rings of phenylalanines and tyrosines to be more exposed than in normally folded proteins.

Formation of silver nanoparticles

Finally, transmission electron microscopy was performed to study the location and distribution of silver. The silver-resistant mutant NA4S was grown in the presence of 0.5 mM AgNO₃ and analysis showed the formation of nanoparticles, which mostly contained Ag as indicated by EDX (Fig. 6). Neither growth nor nanoparticle formation was observed in these conditions for the parental strain NA4 (data not shown).

Fig. 6 Silver nanoparticles in NA4S (a) TEM image of NA4S grown on LB agar supplemented with 0.5 mM AgNO₃ showing the presence of nanoparticles associated with the cell. (b) EDX analysis demonstrating that the nanoparticles mostly contain silver (red dots).

Discussion

The broad-spectrum antimicrobial activities of silver (nanoparticles) resulted in the extensive use in a wide range of medical applications as well as consumer and personal care products.^1,4 Furthermore, the increase in occurrence and number of antibiotic-resistant strains contributed to the resurgence in the interest and use, including combating multi-drug resistant bacteria.^57,58 The potential of microorganisms to develop resistance to silver and AgNPs^59,60 and the incidence of silver-resistant determinants in many environments, including those where the use of silver compounds was restricted, raises the concern for their widespread use and co-selection for antibiotic resistance. Together with the health and environmental risks associated with the increased use of nanotechnology, the European Commission advised that more data is needed to better understand the bacterial response to ionic silver and silver nanoparticles.⁶¹ Since efflux-pump related silver resistance is well studied, but less is known for other resistance and adaptation mechanisms, we addressed in this study the bacterial adaptation to toxic silver concentrations. We showed that C. metallidurans was able to adapt to toxic silver concentrations. C. metallidurans was selected because it carries many distinct and well-studied heavy metal resistance determinants, including several involved in silver resistance. However, none of these systems were involved in the adaptive response to silver described here. Specific for silver resistance determinants, the HME-RND-driven efflux system encoded by the silDCBA operon and the cop cluster (for which some genes are shown to be involved in Ag⁺ resistance) are located on plasmid pMOL30^38,40 and therefore neither present in AE104 nor in its silver-resistant mutant AE104S. In addition, the HME-RND-driven efflux system encoded by the cusDCBAF operon³⁶ and cupRAC coding for a P_IB1-type ATPase³⁷ were not commonly differentially expressed (ESI,† Table S1). Furthermore, cupA in CH34S2 was inactivated by ISRme5 (ESI,† Table S2). Instead, sequencing, whole genome expression profiling and insertional mutation indicated the involvement of the AgrRS two-component system and two uncharacterized periplasmic proteins, designated PrsQ₁ and PrsQ₂ for periplasmic resistance to silver related to the uncharacterized CopQ-like protein family. Contrary to observations involving the Cus or Sil HME-RND-driven efflux systems,^59,60 for which mutations within the associated sensor kinase resulted in an increased silver resistance via increased expression of the RND-driven efflux pump, this is not the case for the AgrCBA efflux system and associated AgrRS. Mutations in a sensor kinase that result in the constitutive and stress-independent transcriptional activation of its cognate RND-driven efflux system are often at the base of an increased resistance phenotype and have also been shown for other metals and antibiotics.^62,63 However, in our study, the AgrRS two-component regulatory system increased the expression of non-cognate targets, including prsQ₂, resulting in increased silver resistance.

The pivotal mechanistic determinants mediating the increased silver resistance are the small periplasmic proteins PrsQ₁ and PrsQ₂. Both belong to a previously uncharacterized protein family whose members can only be found in the related genera Cupriavidus and Ralstonia.⁴⁴ A recent sequence analysis of more than 28 million protein sequences showed that all 787 predicted CopQ-like proteins belong to the Cupriavidus and Ralstonia genera. Structure prediction of PrsQ₁ and PrsQ₂ indicated that they could be intrinsically disordered proteins. Such disordered proteins are common in nature and include SilE and PcoE, which have been shown to be involved in silver resistance.^22,64 Both SilE and PcoE act as a ‘metal sponge’, fold upon binding to Ag⁺ and transport Ag⁺ to the associated HME-RND efflux pump.^21,22 Although the function of PsrQ₁ and PsrQ₂ still needs to be determined, they do not contain conserved histidine and methionine residues within specific sequence motifs that are involved in Ag⁺ binding like SilE and PcoE.^22,64

As the increased silver resistance is not mediated by efflux, we hypothesize that the formation of silver nanoparticles prevents ionic silver to exert its action in the cytoplasm. The latter provides C. metallidurans the ability to withstand much higher silver concentrations than efflux-mediated resistance. Nanoparticle formation has been studied in silver-resistant bacteria as a way to mitigate silver stress.^20,65,66 However, the underlying mechanisms have not yet been studied. Our results are similar to those of a synthetic silver-binding peptide that increases silver tolerance in E. coli when expressed in the periplasm but not when expressed in the cytoplasm.⁶⁷ Similarly, the synthetic peptide does not contain conserved histidine and methionine residues and transmission electron microscopy showed the presence of electron-dense silver nanoparticles.⁶⁷

A more detailed study of this protein family is required as many of the encoding genes are highly induced in response to different metals.³⁷ As such, these studies will not only gain insights in the different effects and potential adverse outcomes of ionic silver and silver nanoparticles but also in metal (silver) resistance mechanisms as well as novel routes for metal nanoparticle formation and metal processing in biotechnical and biomedical applications.

Experimental

Strains and growth conditions

Bacterial strains and plasmids used in this study are summarized in Tables 2 and 3, respectively. C. metallidurans strains were routinely cultured at 30 °C in Lysogeny Broth (LB; Thermofisher Scientific) or Tris-buffered mineral medium supplemented with 0.2% (wt/vol) gluconate (MM284).³⁹E. coli strains were routinely cultured at 37 °C in LB. For culturing on agar plates, 2% agar (Oxoid, Belgium) was added. Liquid cultures were grown in the dark on a rotary shaker at 150 rpm. When appropriate, the following chemicals were added to the growth medium at the indicated final concentrations: kanamycin sulfate (50 μg ml⁻¹ for E. coli or 1500 μg ml⁻¹ for C. metallidurans) (Km50; Km1500), tetracycline hydrochloride (20 μg ml⁻¹) (Tc20) and ampicillin sodium salt (100 μg ml⁻¹) (Ap100) (PanReac AppliChem, Germany), chloramphenicol (30 μg ml⁻¹) (Cm30), L-arabinose (0.02%) (Sigma-Aldrich, Belgium), 5-bromo-4-chloro-3-indolyl-β-D-galactopyranoside (X-Gal; 40 μg ml⁻¹) (Thermofisher Scientific, Belgium) and isopropyl β-D-1-thiogalactopyranoside (IPTG; 0.1 mM) (Thermofisher Scientific, Belgium).

Table 2 Strains used in this study

Strain	Genotype/relevant characteristics	Ref.
Ag^R (silver resistant); Km^R and km (kanamycin resistance and resistance gene); Sm^R (streptomycin resistant); Tc^R and tet (tetracycline resistance and resistance gene).
Escherichia coli
DG1	mcrA Δmrr-hsdRMS-mcrBC (rB^−mB^−) Φ80lacZΔM15 ΔlacX74 recA1 araD139 Δ(ara-leu)7697 galU galK rpsL endA1 nupG	Eurogentec, B
MFDpir	MG1655 RP4-2-Tc::[ΔMu1::aac(3)IV-ΔaphA-Δnic35-ΔMu2::zeo] ΔdapA::(erm-pir) ΔrecA	68
HB101	F^−mcrB mrr hsdS20(rB^−mB^−) recA13 leuB6 ara-14 proA2 lacY1 galK2 xyl-5 mtl-1 rpsL20 (Sm^R) glnV44 λ^−	Lab collection
Cupriavidus metallidurans
CH34^T	Isolated from a decantation tank at a Belgian zinc factory	39
AE104	Plasmidless derivative of CH34	39
NA4	Isolated from ISS SRV-K filter reactor effluent returned on Soyuz 10S (2004) (original code: 0502478-1)	69
CH34S1	Spontaneous silver resistant mutant of CH34, Ag^R	A. Bossus
CH34S2	Spontaneous silver resistant mutant of CH34, Ag^R	A. Bossus
AE104S	Spontaneous silver resistant mutant of AE104, Ag^R	This study
NA4ΔagrS	agrR::3xFLAG agrS::km, Km^R	This study
NA4S	Spontaneous silver resistant mutant of NA4, Ag^R	This study
NA4SΔ40157	40157::tet, Tc^R	This study
NA4SΔagrRS	agrRS::tet, Tc^R	This study
NA4SΔczcI2	czcI 2::tet, Tc^R	This study
NA4SΔczcL	czcL::tet, Tc^R	This study
NA4SΔompC	ompC::tet, Tc^R	This study
NA4SΔprsQ1	prsQ 1::tet, Tc^R	This study
NA4SΔprsQ2	prsQ 2::tet, Tc^R	This study

---

Table 3 Plasmids used in this study

Plasmid	Genotype/relevant characteristics	Ref.
Ap^R (ampicillin resistant); Cm^R (chloramphenicol resistant); Km^R and km (kanamycin resistant and resistance gene); Tc^R and tet (tetracycline resistant and resistance gene); PBAD (arabinose inducible promoter); GenBio (General Biosystems, Inc., Morrisville, USA).
pACYC184	p15A ori, Cm^R, Tc^R	Lab collection
pBAD33	p15A ori, PBAD, araC, Cm^R	73
pBAD33-agrRS	agrRS of NA4 in pBAD33, Km^R	This study
pBAD33-agrRD51AS	pBAD33-agrRS with 152A > C 153T > C in agrR, Km^R	This study
pBAD33-agrRSR414C	agrRS of NA4S in pBAD33, Km^R	This study
pBBR1MCS-2	pBBR1 ori, Mob^+, lacZ, Km^R	74
pBBR-prsQ1	prsQ 1 of NA4 in pBBR1MCS-2, Km^R	This study
pBBR-prsQ2	prsQ 2 of NA4 in pBBR1MCS-2, Km^R	This study
pBBR-prsQ^SDM2	pBBR-prsQ2 harbouring a frameshift in the start codon of prsQ2, Km^R	This study
pBBR-prsQ^ΔSP2	pBBR-prsQ2 with 26-bp deletion of the prsQ2 signal sequence, Km^R	This study
pK18mob	pMB1 ori, Mob^+, lacZ, Km^R	75
pK18mob-40157::tet	tet gene flanked by 1 kb DNA region up- and downstream of CmetNA4v1_40157, Km^R, Tc^R	This study
pK18mob-agrRS	agrRS gene region of NA4 in pK18mob, Km^R	This study
pK18mob-agrRS::tet	pK18mob-agrRS derivative, agrRS::tet, Km^R, Tc^R	This study
pK18mob-czcI2::tet	tet gene flanked by 1 kb DNA region up- and downstream of czcI2 NA4, Km^R, Tc^R	This study
pK18mob-czcL::tet	pK18mob-prsQ2 derivative, czcL::tet, Km^R, Tc^R	This study
pK18mob-ompC::tet	tet gene flanked by 1 kb DNA region up- and downstream of ompC NA4, Km^R, Tc^R	This study
pK18mob-prsQ1::tet	tet gene flanked by 1 kb DNA region up- and downstream of prsQ1 NA4, Km^R, Tc^R	This study
pK18mob-prsQ2	prsQ 2–czcL gene region of NA4 in pK18mob, Km^R	This study
pK18mob-prsQ2::tet	pK18mob-prsQ2 derivative, prsQ2::tet, Km^R, Tc^R	This study
pKT230	IncQ, Mob^+, Km^R	76
pRK600	ColE1 oriV; RP4 Tra^+ RP4 oriT, Cm^R; helper in triparental matings	Lab collection
pPROBE-TT	pBBR1 ori, Mob^+, promoterless gfp, Tc^R	77
pPTT-prsQ2	Promoter region of prsQ2 NA4 in pPROBE-TT, Tc^R	This study
pT18mob	pMB1 ori, Mob^+, lacZ, Tc^R	This study
pT18mob-agrS::km	km gene flanked by 1 kb DNA region up- and downstream of agrS, agrR::3xFLAG, Km^R	This study
pSCK108	IncQ, Mob^+, lacZ, Km^R	This study
pSCK108-PprsQ1-gfp	SacI/BamHI fragment of pUC57-PprsQ1-gfp in pSCK108, Km^R	This study
pSCK108-PprsQ2-mcherry	BamHI/HindIII fragment of pUC57-PprsQ2-mcherry in pSCK108, Km^R	This study
pUC57-PprsQ1-gfp	350 bp 5′ region of prsQ1 fused to gfp (synthetic construct), Ap^R	GenBio
pUC57-PprsQ2-mcherry	350 bp 5′ region of prsQ1 fused to mcherry (synthetic construct), Ap^R	GenBio

---

Generation of spontaneous silver resistant mutants

Spontaneous silver resistant mutants of C. metallidurans CH34, its plasmidless derivative AE104 and C. metallidurans NA4 were generated by exposure to a toxic concentration of AgNO₃ (Sigma-Aldrich, Belgium). To this end, CH34S1 and CH34S2 were selected on LB agar supplemented with respectively 0.5 and 2 mM AgNO₃, AE104S and NA4S were generated in MM284 supplemented with respectively 3 μM and 8 μM AgNO₃.

Whole-genome sequencing

Total DNA (10 μg) of C. metallidurans CH34S1, CH34S2 and NA4S was extracted using the QIAamp DNA Mini Kit (Qiagen). The quantity and quality of the DNA was analysed using a NanoDrop^TM 1000 spectrophotometer (Thermofisher Scientific). A 50 bp paired-end sequencing approach with an insert size of 300 bp was performed by Baseclear on the Illumina GAIIx platform. The paired-end sequencing reads were aligned to the reference genomes (i.e. the genome sequence of C. metallidurans strain CH34 and strain NA4, respectively) using the BWA-MEM algorithm version 0.7.13.⁷⁰ Genomic rearrangements were identified using an in-house developed Perl script specifically looking for wrongly mapped paired-end reads. An increase in the insert size combined with a peak in incorrectly aligned paired reads at the breakpoints indicates a deletion, where the size of the deletion can be derived from the predict insert size based on the mapping of both mapped reads. Two neighbouring peaks (<100 bp) – respectively on the forward and reverse strand – in incorrectly aligned paired-end reads indicate an insertion of a new DNA fragment at that position. SNPs and small indels were identified using the Genome Analysis Tool Kit version 3.6 using the Haplotyper command, after sorting and indexing the BWA alignment using the samtools software (version 1.3.1).⁷¹ Sequencing data are available within the Sequencing Read Archive (SRA) of NCBI using the accession numbers SRR6340090, SRR6340091 and SRR6340092 for the strains NA4S, CH34S1 and CH34S2 respectively.

Whole genome expression analysis

Gene expression in the different spontaneous silver resistant mutants was compared with the parental strain under non-selective growth conditions. To this end, three independent stationary phase cultures of C. metallidurans CH34, AE104, NA4 and the spontaneous silver resistant mutants CH34S1, CH34S2, AE104S and NA4S were diluted 1/500 in 25 ml MM284. These subcultures were allowed to grow until a cell density of 5 × 10⁸ cells per ml (OD₆₀₀ = 0.5) was reached. Each culture was subdivided in 2 ml portions and cells were harvested by centrifugation for 2 min at 10 000g. Bacterial pellets were flash frozen by immersion into liquid nitrogen and were kept frozen at −80 °C at all times. RNA extraction, labelling and hybridization, microarray spotting, scanning and data analysis was performed according to the work of Monsieurs et al.³⁷ The full description of the microarray data have been deposited at the Gene Expression Omnibus website (http://www.ncbi.nlm.nih.gov/geo/) under accession number GSE107669. Annotation of C. metallidurans NA4 was based on GenBank entry JFZE00000000.1 and MaGe⁷² annotation v1 (ESI,† Table S3).

Construction of insertional inactivation mutants

All PCR reactions were performed with Phusion High-Fidelity Polymerase (Thermofisher Scientific, Belgium) unless stated otherwise. Primers used are presented in ESI,† Table S4. The C. metallidurans NA4 agrRS and prsQ₂-czcL gene regions, including 0.5 kb up- and downstream, were amplified by PCR (DreamTaq Polymerase, Thermofisher Scientific, Belgium) using the primer pairs agrR_FW-agrS_RV and prsQ₂-czcL_FW-RV providing amplicons with a size of 3.3 kb and 3 kb, and EcoRI/HindIII and BamHI/HindIII recognition sites, respectively. The amplicons were subsequently cloned in pK18mob digested with EcoRI/HindIII and BamHI/HindIII, respectively. The gene region of C. metallidurans NA4 CmetNA4v1_570182 (czcI₂) including 1 kb up- and downstream was amplified with the primers czcI_FW-RV providing an amplicon of 2.4 kb. This amplicon was subsequently cloned in pK18mob, which was linearized with EcoRI and amplified with the primers pK18mob_FW-RV (3.6 kb), using the GeneArt™ seamless cloning and assembly enzyme mix (Thermofisher Scientific, Belgium). The resulting constructs pK18mob-agrRS, pK18mob-prsQ₂ and pK18mob-czcI₂ were transformed to E. coli DG1 chemocompetent cells (Eurogentec, Belgium).

Next, pK18mob_agrRS, pK18mob-czcI₂, pK18mob-czcL and pK18mob-prsQ₂ were used as PCR (DreamTaq Polymerase, Thermofisher Scientific, Belgium) template using the outward-oriented primers agrRS_OUT_FW-RV, czcI_OUT_FW-RV, czcL_OUT_FW-RV and prsQ₂_OUT_FW-RV, respectively. A tetracycline resistance cassette, amplified from pACYC184 with primers tet_FW-RV, was cloned as a BspTI/BcuI fragment in each BspTI/BcuI digested amplicon. The resulting constructs pK18mob-agrRS::tet, pK18mob-czcI₂::tet, pK18mob-czcL::tet and pK18mob-prsQ₂::tet were transformed to E. coli DG1.

GeneArt™ was applied to obtain different pK18mob constructs in which the 1 kb up- and downstream DNA region of prsQ₁, CmetNA4v1_40157 and ompC flanked a tetracycline cassette. To this end, the 1 kb up- and downstream region of prsQ₁, CmetNA4v1_40157 and ompC were amplified with the primers prsQ₁_5_FW-RV and prsQ₁_3_FW-RV, 40157_5_FW-RV and 40157_3_FW-RV and ompC_5_FW-RV and ompC_3_FW-RV, respectively. A tetracycline resistance cassette was amplified from pACYC184 with primers tet-5 and tet-3 and the EcoRI-linearized pK18mob was amplified with the primers pK18mob_FW-RV (3.6 kb). Assembly of the products resulted in pK18mob-prsQ₁::tet, pK18mob-40157::tet and pK18mob-ompC::tet, which were transformed to E. coli DG1. All plasmids were confirmed by sequencing with tet-5_OUT and tet-3_OUT primers.

For constructing pT18mob-agrS::km, the agrRS gene region of C. metallidurans NA4 was amplified with primer pair (agrS_5_FW/agrS_3_RV) and cloned via GeneArt™ in pT18mob (linear PCR fragment with primer pair pK18mob_FW-RV). Next, pTmob-agrS was linearized with primer pair (agrS_3FL_FW/agrS_5FL_RV) and fused via GeneArt™ to a PCR fragment of pSUB11 (primer pair 3xFLAGKM_FW-RV) creating a C-terminal 3xFLAG tagged AgrR while replacing agrS by a kanamycin resistance cassette. The resulting pT18mob-agrS::km construct was transformed to E. coli DG1.

For insertion mutagenesis, all constructs (pK18mob-40157::tet, pK18mob-agrRS::tet, pK18mob-czcI₂::tet, pK18mob-czcL::tet, pK18mob-ompC::tet, pK18mob-prsQ₁::tet, pK18mob-prsQ₂::tet, and pT18mob-agrS::km) were conjugated into C. metallidurans NA4S (or NA4 for pT18mob-agrS::tet) either with E. coli HB101 pRK600 as helper (triparental) or via E. coli MFDpir as donor (biparental). For each mutant, a tetracycline-resistant and kanamycin-sensitive exconjugant was selected (kanamycin-resistant and tetracycline-sensitive in the case of agrS::km), and confirmed being a genuine mutant by PCR and sequencing, yielding NA4SΔ40157, NA4SΔagrRS, NA4SΔczcI₂, NA4SΔczcL, NA4SΔompC, NA4SΔprsQ₁, NA4SΔprsQ₂ and NA4ΔagrS, respectively.

Construction of plasmids

PCR amplification of prsQ₁ and prsQ₂ (including their promoter region) was performed on genomic DNA from C. metallidurans NA4 with the respective primer pairs prsQ₁_FW-RV (0.6 kb) and prsQ2_FW-RV (0.6 kb). These PCR products were cloned as BamHI/HindIII fragments into pBBR1MCS-2. The resulting pBBR-prsQ₁ and pBBR-prsQ₂ plasmids from transformants selected on LB Km50 were further confirmed by sequencing with the primers pBBR_FW and pBBR_RV. Afterwards, pBBR-prsQ₁ and pBBR-prsQ₂ were transformed by electroporation to NA4SΔprsQ₁ and NA4SΔprsQ₂, respectively.

The agrRS genes from C. metallidurans NA4 and NA4S as well as agrR_D51AS (using pBBR-agrR_D51AS as template, see below) were amplified by PCR with the primer pair agrRS_BAD_FW-agrS_RV (2.1 kb), providing BamHI/HindIII recognition sites. These amplicons were subsequently cloned in pBAD33⁷⁸ and transformed to E. coli DG1, resulting in the plasmids pBAD-agrRS, pBAD-agrRS_R414C and pBAD-agrR_D51AS.

For the transcriptional prsQ₂ promoter-gfp fusion, a 0.3 kb DNA fragment upstream the prsQ₂ start codon was amplified by PCR using the primer pairs PprsQ₂_FW-RV, providing BamHI/XbaI recognition sites. This amplicon was subsequently cloned into pPROBE-TT⁷⁷ containing a promoterless gfp and transformed into E. coli DG1 competent cells, resulting in pPTT-prsQ₂.

Plasmid-based promoter probe reporters in C. metallidurans were constructed in the broad-host-range vector pSCK108. The latter was constructed by ligating a BsiWI-digested PCR fragment of pK18mob (primer pair pK18mob_inv_FW-RV) to an Acc65I-digested PCR fragment of pKT230 (primer pair KT261_FW-RV). pSCK108-P_prsQ1-gfp was constructed by cloning a SacI/BamHI fragment of pUC57-P_prsQ1-gfp, carrying a 350-bp 5′ region of prsQ₁ fused to gfp, in SacI/BamHI-digested pSCK108. pSCK108-P_prsQ2-mcherry was constructed by cloning a BamHI/HindIII fragment of pUC57-P_prsQ2-mcherry, carrying a 350 bp 5′ region of prsQ₂ fused to mcherry, in BamHI/HindIII-digested pSCK108. pUC57-P_prsQ1-gfp and pUC57-P_prsQ2-mcherry were custom synthesized by General Biosystems, Inc. (Morrisville, USA).

Plasmid pT18mob was constructed by replacing the kanamycin cassette of pK18mob with the tetracycline cassette of pACYC184 via GeneArt™ cloning (PCR fragment of pK18mob (primer pair pK18_inv_FW-RV) and PCR fragment of pACYC184 (primer pair pACYC184_tet_FW-RV)).

Site-directed mutagenesis

Site-directed mutagenesis was performed with the “Phusion Site-Directed Mutagenesis Kit” (Thermo Scientific, Belgium). Plasmid pBBR-prsQ₂ was used as a template for PCR amplification. Phosphorylated primer pairs prsQ^SDM₂_FW-RV and prsQ^ΔSP₂_FW-RV were used for constructing a frame shift mutation in the prsQ₂ start codon and a 26-bp deletion of the prsQ₂ signal sequence, respectively. Plasmid pBBR_agrRS was used as template for PCR amplification with the phosphorylated primer pair agrR_SDM_FW-RV, which introduced 152A > C 153T > C, resulting to D51A in AgrR. The resulting linear fragments were purified using the “Wizard® SV Gel and PCR Clean-Up System” (Promega, the Netherlands), subsequently self-ligated, and finally transformed to E. coli DG1. The resulting pBBR-prsQ^SDM₂ and pBBR-prsQ^ΔSP₂ plasmids from transformants selected on LB Km50 were further confirmed by sequencing prior to transformation to C. metallidurans NA4SΔprsQ₂.

cDNA amplification and RT-PCR

Total RNA was extracted from the silver-resistant mutant NA4S, all insertional inactivation mutants and complementation strains using the SV Total RNA Isolation System (Promega, the Netherlands). To this end, cultures were grown in MM284 media until an OD of 0.6–0.8 was reached. Each culture was subdivided in 2 ml portions and cells were harvested by centrifugation for 2 min at 10 000g. Bacterial pellets were flash frozen by immersion into liquid nitrogen and were kept frozen at −80 °C at all times. From each strain, single-stranded complementary DNA (cDNA) was synthesised from 1 μg total RNA by using random hexamers as primers and MultiScribe® Reverse Transcriptase (ThermoFisher Scientific, Belgium) and the quantity of synthesised cDNA was measured by a NanoDrop® 1000 spectrophotometer (ThermoFisher Scientific, Belgium). This cDNA (50 ng) was used to check the absence or presence of mRNA for inactivated or reintroduced genes by PCR amplification (ESI,† Table S4; primers with _cDNA tag).

The expression of prsQ₂ was analysed by RT-PCR using the MESA Blue qPCR MasterMix Plus for SYBR® Assay Low ROX (Eurogentec, Belgium) according to the manufacturer's protocol with primers prsQ_{1_}cDNA_FW-RV (ESI,† Table S4). Expression was compared with that of uvrD, which was amplified with primers uvrD_cDNA_FW-RV (ESI,† Table S4). RT-PCRs were performed with a 7500 Real-Time PCR System (Applied Biosystems, UK).

Total viable count

Three independent colonies of each strain were cultivated up to stationary phase in LB medium at 30 °C. Afterwards, 10⁹ CFU ml⁻¹ were pelleted, washed with MgSO₄ (10 mM) and resuspended in MgSO₄ (10 mM). Cell suspensions of 100 μl of a serial ten-fold dilution in MgSO₄ (10 mM) were spread on LB agar plates and on LB agar plates supplemented with 0.5 mM AgNO₃. The plates were incubated at 30 °C and colony forming units were counted after 3 days. To calculate the survival, the viable count on LB agar supplemented with 0.5 mM AgNO₃ was normalized to the viable count on LB agar.

Phylogenetic footprinting

A phylogenetic footprinting approach with the promoter region of agrR was performed in silico by combining the output of different motif detection algorithms (MotifSampler,⁴⁷ AlignACE,⁴⁸ MEME,⁴⁹ Weeder⁵⁰ and RSAT⁵¹) followed by a Markov clustering of the output motif models. The regulatory motif obtained in this manner was used for a genome wide screening of all upstream regions in type strain C. metallidurans CH34 to identify putative binding sites for AgrR.

In vivo cross-regulation in E. coli

Plasmid pPTT-prsQ₂ was transformed by electroporation in E. coli DG1 carrying pBAD-agrRS, pBAD-agrRS_R414C, pBAD-agrR_D51AS and pBAD-agrR, respectively. Possible cross-regulation was tested by comparing Gfp production in each construct, either un-induced or induced with 0.2% arabinose. To this end, overnight LB cultures of three independent colonies (grown in the presence of Tc (20 μg ml⁻¹) and Cm (30 μg ml⁻¹)), were diluted 10 times in 2 ml LB with the same antibiotics and 0.2% arabinose on the one hand and water on the other hand. These cultures were incubated for 6 hours at 30 °C, diluted 100-fold in 10 mM MgSO₄ (filtered with 0.2 μm filter) and analysed with a C6 Accuri™ flow cytometer (BD Biosciences, Belgium). The flow cytometer is equipped with a 20 mW 488 nm laser, two scatter detectors and four fluorescence detectors (530/30 nm, 585/40 nm, >670 nm and 675/25 nm). The flow cytometer is operated with Milli-Q water (Merck Millipore, Belgium) as sheath fluid. Gfp production was measured with filter FL1 (excitation 488 nm/emission 533/30 nm) and 100 000 cells were recorded. The respective cell populations were delimited to eliminate background signals originating from cell debris. All data analysis was performed with the CFlow and FlowJo V10 Software.

Motif discovery CopQ-like proteins

The available protein sequences of 7927 complete genomes were retrieved from the NCBI ftp server (http://ftp://ftp.ncbi.nlm.nih.gov/genomes/genbank/bacteria/ on October 4, 2017) and concatenated in one FastA file. This FastA file, containing 28 248 917 protein sequences, was queried by the multi-threaded pattern-matching tool Biogrep 1.0⁷⁹ using an extended 10-aa search pattern [GSRDEQ][GSAKRTPN]xD[PVT][YF][TS][DEQ]G[ASG] covering the Dx[YF]xxG domain motif characteristic for CopQ-like family proteins. This extended search pattern is based on a multiple alignment and additional visual inspection of the protein sequences of CopQ-like family members identified in several Cupriavidus and Ralstonia proteomes by BLASTp analysis of the non-redundant protein sequence database with PrsQ₁ as a query. When applied in Biogrep analysis, this pattern detected exactly and exclusively all 22 CopQ-like family members in strain CH34.

Transmission electron microscopy

Bacterial cells were adhered to a copper EM-grid (300 mesh) by depositing 3 μl of cell suspension on the grid and drying in an evacuated desiccator. Bacteria were observed in a JEOL JEM2100 LaB6 TEM operated at 200 kV, equipped with an Oxford Instruments (UK) EDX-system and STEM detectors. Images were recorded with a Gatan multiscan 1k × 1k pxls camera. For X-ray spectra, areas of the sample were scanned using the bright field STEM detector.

Conclusions

In summary, we showed that C. metallidurans is able to adapt to toxic silver concentrations by activating latent mechanisms that are not related to the known metal or silver resistance determinants. A single mutation was sufficient for activation and provided resistance to much higher silver concentrations. Previously uncharacterized small periplasmic proteins carrying distinctive sequence motifs played a central role in this adaptation. Moreover, they appear to be exclusively found in members of the Cupriavidus and Ralstonia genera, suggesting a specific adaptation potential to metal and putative other stressors in these genera that are able to live and thrive in harsh and oligotrophic environments. Finally, the increased resistance is most probably facilitated by the accumulation of silver ions and the formation of silver nanoparticles.

Conflicts of interest

There are no conflicts to declare.

Acknowledgements

This work was supported by the European Space Agency (ESA-PRODEX) and the Belgian Science Policy (Belspo) through the COMICS project (C90356). The authors want to thank Dr Pieter Baatsen from the VIB BIO Imaging Core electron microscopy facility (KULeuven, Belgium) for conducting the TEM and EDX analysis of the silver resistant mutant.

Notes and references

1. S. Silver, T. Phung le and G. Silver, Silver as biocides in burn and wound dressings and bacterial resistance to silver compounds, J. Ind. Microbiol. Biotechnol., 2006, 33, 627–634.
2. V. Edwards-Jones, The benefits of silver in hygiene, personal care and healthcare, Lett. Appl. Microbiol., 2009, 49, 147–152.
3. K. Chaloupka, Y. Malam and A. M. Seifalian, Nanosilver as a new generation of nanoproduct in biomedical applications, Trends Biotechnol., 2010, 28, 580–588.
4. M. E. Vance, T. Kuiken, E. P. Vejerano, S. P. McGinnis, M. F. Hochella, Jr., D. Rejeski and M. S. Hull, Nanotechnology in the real world: Redeveloping the nanomaterial consumer products inventory, Beilstein J. Nanotechnol., 2015, 6, 1769–1780.
5. N. Duran, M. Duran, M. B. de Jesus, A. B. Seabra, W. J. Favaro and G. Nakazato, Silver nanoparticles: A new view on mechanistic aspects on antimicrobial activity, Nanomedicine, 2016, 12, 789–799.
6. V. De Matteis, M. A. Malvindi, A. Galeone, V. Brunetti, E. De Luca, S. Kote, P. Kshirsagar, S. Sabella, G. Bardi and P. P. Pompa, Negligible particle-specific toxicity mechanism of silver nanoparticles: The role of Ag⁺ ion release in the cytosol, Nanomedicine, 2015, 11, 731–739.
7. K. Mijnendonckx, N. Leys, J. Mahillon, S. Silver and R. V. Houdt, Antimicrobial silver: uses, toxicity and potential for resistance, Biometals, 2013, 26, 609–621.
8. O. Gordon, T. Vig Slenters, P. S. Brunetto, A. E. Villaruz, D. E. Sturdevant, M. Otto, R. Landmann and K. M. Fromm, Silver coordination polymers for prevention of implant infection: thiol interaction, impact on respiratory chain enzymes, and hydroxyl radical induction, Antimicrob. Agents Chemother., 2010, 54, 4208–4218.
9. A. D. Russell and W. B. Hugo, Antimicrobial activity and action of silver, Prog. Med. Chem., 1994, 31, 351–370.
10. H. J. Park, J. Y. Kim, J. Kim, J. H. Lee, J. S. Hahn, M. B. Gu and J. Yoon, Silver-ion-mediated reactive oxygen species generation affecting bactericidal activity, Water Res., 2009, 43, 1027–1032.
11. E. McGillicuddy, I. Murray, S. Kavanagh, L. Morrison, A. Fogarty, M. Cormican, P. Dockery, M. Prendergast, N. Rowan and D. Morris, Silver nanoparticles in the environment: Sources, detection and ecotoxicology, Sci. Total Environ, 2017, 575, 231–246.
12. K. Schlich, T. Klawonn, K. Terytze and K. Hund-Rinke, Hazard assessment of a silver nanoparticle in soil applied via sewage sludge, Environ. Sci. Eur., 2013, 25, 17.
13. G. E. Batley, J. K. Kirby and M. J. McLaughlin, Fate and risks of nanomaterials in aquatic and terrestrial environments, Acc. Chem. Res., 2013, 46, 854–862.
14. SCENIHR, Assessment of the antibiotic resistance effects of biocides, European Commission, Brussels, 2009.
15. N. Seltenrich, Nanosilver: weighing the risks and benefits, Environ. Health Perspect., 2013, 121, A220–A225.
16. P. J. Finley, R. Norton, C. Austin, A. Mitchell, S. Zank and P. Durham, Unprecedented silver resistance in clinically isolated Enterobacteriaceae: Major implications for burn and wound management, Antimicrob. Agents Chemother., 2015, 59, 4734–4741.
17. G. L. McHugh, R. C. Moellering, C. C. Hopkins and M. N. Swartz, Salmonella typhimurium resistant to silver nitrate, chloramphenicol, and ampicillin, Lancet, 1975, 1, 235–240.
18. C. Haefeli, C. Franklin and K. Hardy, Plasmid-determined silver resistance in Pseudomonas stutzeri isolated from a silver mine, J. Bacteriol., 1984, 158, 389–392.
19. P. Choudhury and R. Kumar, Multidrug- and metal-resistant strains of Klebsiella pneumoniae isolated from Penaeus monodon of the coastal waters of deltaic Sundarban, Can. J. Microbiol., 1998, 44, 186–189.
20. T. Klaus, R. Joerger, E. Olsson and C. G. Granqvist, Silver-based crystalline nanoparticles, microbially fabricated, Proc. Natl. Acad. Sci. U. S. A., 1999, 96, 13611–13614.
21. S. Silver, A. Gupta, K. Matsui and J. F. Lo, Resistance to Ag⁺ cations in bacteria: environments, genes and proteins, Met. – Based Drugs, 1999, 6, 315–320.
22. M. Zimmermann, S. R. Udagedara, C. M. Sze, T. M. Ryan, G. J. Howlett, Z. Xiao and A. G. Wedd, PcoE - A metal sponge expressed to the periplasm of copper resistance Escherichia coli. Implication of its function role in copper resistance, J. Inorg. Biochem., 2012, 115, 186–197.
23. J. Goris, P. D. Vos, T. Coenye, B. Hoste, D. Janssens, H. Brim, L. Diels, M. Mergeay, K. Kersters and P. Vandamme, Classification of metal-resistant bacteria from industrial biotopes as Ralstonia campinensis sp. nov., Ralstonia metallidurans sp. nov. and Ralstonia basilensis, Int. J. Syst. Evol. Microbiol., 2001, 51, 1773–1782.
24. H. Brim, M. Heyndrickx, P. de Vos, A. Wilmotte, D. Springael, H. G. Schlegel and M. Mergeay, Amplified rDNA restriction analysis and further genotypic characterization of metal-resistant soil bacteria and related facultative hydrogenotrophs, Syst. Appl. Microbiol., 1999, 22, 258–268.
25. L. Diels and M. Mergeay, DNA probe-mediated detection of resistant bacteria from soils highly polluted by heavy metals, Appl. Environ. Microbiol., 1990, 56, 1485–1491.
26. M. T. La Duc, W. Nicholson, R. Kern and K. Venkateswaran, Microbial characterization of the Mars Odyssey spacecraft and its encapsulation facility, Environ. Microbiol., 2003, 5, 977–985.
27. C. M. Ott, R. J. Bruce and D. L. Pierson, Microbial characterization of free floating condensate aboard the Mir space station, Microb. Ecol., 2004, 47, 133–136.
28. P. Monsieurs, K. Mijnendonckx, A. Provoost, K. Venkateswaran, C. M. Ott, N. Leys and R. Van Houdt, Genome sequences of Cupriavidus metallidurans Strains NA1, NA4, and NE12, isolated from space equipment, Genome Announcements, 2014, 2(4), e00719–14.
29. T. Coenye, T. Spilker, R. Reik, P. Vandamme and J. J. Lipuma, Use of PCR analyses to define the distribution of Ralstonia species recovered from patients with cystic fibrosis, J. Clin. Microbiol., 2005, 43, 3463–3466.
30. R. Van Houdt, P. Monsieurs, K. Mijnendonckx, A. Provoost, A. Janssen, M. Mergeay and N. Leys, Variation in genomic islands contribute to genome plasticity in Cupriavidus metallidurans, BMC Genomics, 2012, 13, 111.
31. S. Langevin, J. Vincelette, S. Bekal and C. Gaudreau, First case of invasive human infection caused by Cupriavidus metallidurans, J. Clin. Microbiol., 2011, 49, 744–745.
32. P. Monsieurs, A. Provoost, K. Mijnendonckx, N. Leys, C. Gaudreau and R. Van Houdt, Genome sequence of Cupriavidus metallidurans Strain H1130, isolated from an invasive human infection, Genome Announcements, 2013, 1(6), e01051–13.
33. R. Van Houdt, S. Monchy, N. Leys and M. Mergeay, New mobile genetic elements in Cupriavidus metallidurans CH34, their possible roles and occurrence in other bacteria, Antonie van Leeuwenhoek, 2009, 96, 205–226.
34. K. Mijnendonckx, A. Provoost, P. Monsieurs, N. Leys, M. Mergeay, J. Mahillon and R. V. Houdt, Insertion sequence elements in Cupriavidus metallidurans CH34: Distribution and role in adaptation, Plasmid, 2011, 65, 193–203.
35. V. Auquier, Docteur en Sciences Agronomiques et Ingénierie Biologique, Université Libre de Bruxelles, 2006.
36. M. Mergeay, S. Monchy, T. Vallaeys, V. Auquier, A. Benotmane, P. Bertin, S. Taghavi, J. Dunn, D. van der Lelie and R. Wattiez, Ralstonia metallidurans, a bacterium specifically adapted to toxic metals: towards a catalogue of metal-responsive genes, FEMS Microbiol. Rev., 2003, 27, 385–410.
37. P. Monsieurs, H. Moors, R. Van Houdt, P. J. Janssen, A. Janssen, I. Coninx, M. Mergeay and N. Leys, Heavy metal resistance in Cupriavidus metallidurans CH34 is governed by an intricate transcriptional network, Biometals, 2011, 24, 1133–1151.
38. B. Bersch, K. M. Derfoufi, F. De Angelis, V. Auquier, E. Ngonlong Ekende, M. Mergeay, J. M. Ruysschaert and G. Vandenbussche, Structural and metal binding characterization of the C-terminal metallochaperone domain of membrane fusion Protein SilB from Cupriavidus metallidurans CH34, Biochemistry, 2011, 50, 2194–2204.
39. M. Mergeay, D. Nies, H. G. Schlegel, J. Gerits, P. Charles and F. V. Gijsegem, Alcaligenes eutrophus CH34 is a facultative chemolithotroph with plasmid-bound resistance to heavy metals, J. Bacteriol., 1985, 162, 328–334.
40. S. Monchy, M. A. Benotmane, P. Janssen, T. Vallaeys, S. Taghavi, D. van der Lelie and M. Mergeay, Plasmids pMOL28 and pMOL30 of Cupriavidus metallidurans are specialized in the maximal viable response to heavy metals, J. Bacteriol., 2007, 189, 7417–7425.
41. L. Bobe, A. Kochetkov, V. Soloukhin, P. Andreichuk, N. Protasov and Y. Sinyak, International conference on environmental systems, SAE International, San Francisco, CA, USA, 2008.
42. K. Mijnendonckx, A. Provoost, C. M. Ott, K. Venkateswaran, J. Mahillon, N. Leys and R. Van Houdt, Characterization of the survival ability of Cupriavidus metallidurans and Ralstonia pickettii from space-related environments, Microb. Ecol., 2013, 56, 347–360.
43. M. M. Ali, A. Provoost, L. Maertens, N. Leys, P. Monsieurs, D. Charlier and R. Van Houdt, Genomic and transcriptomic changes that mediate increased platinum resistance in Cupriavidus metallidurans, Genes, 2019, 10, E63.
44. P. J. Janssen, R. Van Houdt, H. Moors, P. Monsieurs, N. Morin, A. Michaux, M. A. Benotmane, N. Leys, T. Vallaeys, A. Lapidus, S. Monchy, C. Medigue, S. Taghavi, S. McCorkle, J. Dunn, D. van der Lelie and M. Mergeay, The complete genome sequence of Cupriavidus metallidurans strain CH34, a master survivalist in harsh and anthropogenic environments, PLoS One, 2010, 5, e10433.
45. D. Vallenet, L. Labarre, Z. Rouy, V. Barbe, S. Bocs, S. Cruveiller, A. Lajus, G. Pascal, C. Scarpelli and C. Medigue, MaGe: a microbial genome annotation system supported by synteny results, Nucleic Acids Res., 2006, 34, 53–65.
46. J. Vandecraen, M. Chandler, A. Aertsen and R. Van Houdt, The impact of insertion sequences on bacterial genome plasticity and adaptability, Crit. Rev. Microbiol., 2017, 43, 709–730.
47. M. Claeys, V. Storms, H. Sun, T. Michoel and K. Marchal, MotifSuite: workflow for probabilistic motif detection and assessment, Bioinformatics, 2012, 28, 1931–1932.
48. F. P. Roth, J. D. Hughes, P. W. Estep and G. M. Church, Finding DNA regulatory motifs within unaligned noncoding sequences clustered by whole-genome mRNA quantitation, Nat. Biotechnol., 1998, 16, 939.
49. T. L. Bailey and C. Elkan, Fitting a mixture model by expectation maximization to discover motifs in biopolymers, Proc. Int. Conf. Intell. Syst. Mol. Biol., 1994, 2, 28–36.
50. G. Pavesi, G. Mauri and G. Pesole, An algorithm for finding signals of unknown length in DNA sequences, Bioinformatics, 2001, 17(suppl. 1), S207–S214.
51. A. Medina-Rivera, M. Defrance, O. Sand, C. Herrmann, J. A. Castro-Mondragon, J. Delerce, S. Jaeger, C. Blanchet, P. Vincens, C. Caron, D. M. Staines, B. Contreras-Moreira, M. Artufel, L. Charbonnier-Khamvongsa, C. Hernandez, D. Thieffry, M. Thomas-Chollier and J. van Helden, RSAT 2015: Regulatory Sequence Analysis Tools, Nucleic Acids Res., 2015, 43, W50–W56.
52. A. Marina, C. Mott, A. Auyzenberg, W. A. Hendrickson and C. D. Waldburger, Structural and mutational analysis of the PhoQ histidine kinase catalytic domain. Insight into the reaction mechanism, J. Biol. Chem., 2001, 276, 41182–41190.
53. R. J. Lewis, J. A. Brannigan, K. N. Muchová, I. Barák and A. J. Wilkinson, Phosphorylated aspartate in the structure of a response regulator protein, J. Mol. Biol., 1999, 294, 9–15.
54. G. S. Lukat, B. H. Lee, J. M. Mottonen, A. M. Stock and J. B. Stock, Roles of the highly conserved aspartate and lysine residues in the response regulator of bacterial chemotaxis, J. Biol. Chem., 1991, 266, 8348–8354.
55. M. Kallberg, H. Wang, S. Wang, J. Peng, Z. Wang, H. Lu and J. Xu, Template-based protein structure modeling using the RaptorX web server, Nat. Protoc., 2012, 7, 1511–1522.
56. L. A. Kelley, S. Mezulis, C. M. Yates, M. N. Wass and M. J. Sternberg, The Phyre2 web portal for protein modeling, prediction and analysis, Nat. Protoc., 2015, 10, 845–858.
57. J. R. Morones-Ramirez, J. A. Winkler, C. S. Spina and J. J. Collins, Silver enhances antibiotic activity against Gram-negative bacteria, Sci. Transl. Med., 2013, 5, 190ra181.
58. J. A. Garza-Cervantes, A. Chavez-Reyes, E. C. Castillo, G. Garcia-Rivas, O. Antonio Ortega-Rivera, E. Salinas, M. Ortiz-Martinez, S. L. Gomez-Flores, J. A. Pena-Martinez, A. Pepi-Molina, M. T. Trevino-Gonzalez, X. Zarate, M. Elena Cantu-Cardenas, C. Enrique Escarcega-Gonzalez and J. R. Morones-Ramirez, Synergistic antimicrobial effects of silver/transition-metal combinatorial treatments, Sci. Rep., 2017, 7, 903.
59. J. L. Graves, Jr., M. Tajkarimi, Q. Cunningham, A. Campbell, H. Nonga, S. H. Harrison and J. E. Barrick, Rapid evolution of silver nanoparticle resistance in Escherichia coli., Front. Genet., 2015, 6, 42.
60. C. P. Randall, A. Gupta, N. Jackson, D. Busse and A. J. O'Neill, Silver resistance in Gram-negative bacteria: a dissection of endogenous and exogenous mechanisms, J. Antimicrob. Chemother., 2015, 70, 1037–1046.
61. SCENIHR, Opinion on the guidance on the determination of potential health effects of nanomaterials Used in Medical Devices, European Commission, Luxemburg, 2015.
62. K. Perron, O. Caille, C. Rossier, C. Van Delden, J. L. Dumas and T. Kohler, CzcR-CzcS, a two-component system involved in heavy metal and carbapenem resistance in Pseudomonas aeruginosa, J. Biol. Chem., 2004, 279, 8761–8768.
63. I. Marchand, L. Damier-Piolle, P. Courvalin and T. Lambert, Expression of the RND-type efflux pump AdeABC in Acinetobacter baumannii is regulated by the AdeRS two-component system, Antimicrob. Agents Chemother., 2004, 48, 3298–3304.
64. K. R. Asiani, H. Williams, L. Bird, M. Jenner, M. S. Searle, J. L. Hobman, D. J. Scott and P. Soultanas, SilE is an intrinsically disordered periplasmic “molecular sponge” involved in bacterial silver resistance, Mol. Microbiol., 2016, 101, 731–742.
65. R. Y. Parikh, S. Singh, B. L. Prasad, M. S. Patole, M. Sastry and Y. S. Shouche, Extracellular synthesis of crystalline silver nanoparticles and molecular evidence of silver resistance from Morganella sp.: towards understanding biochemical synthesis mechanism, ChemBioChem, 2008, 9, 1415–1422.
66. P. J. Finley, R. Norton, C. Austin, A. Mitchell, S. Zank and P. Durham, Unprecedented Silver Resistance in Clinically Isolated Enterobacteriaceae: Major Implications for Burn and Wound Management, Antimicrob. Agents Chemother., 2015, 59, 4734–4741.
67. R. H. Sedlak, M. Hnilova, C. Grosh, H. Fong, F. Baneyx, D. Schwartz, M. Sarikaya, C. Tamerler and B. Traxler, Engineered Escherichia coli silver-binding periplasmic protein that promotes silver tolerance, Appl. Environ. Microbiol., 2012, 78, 2289–2296.
68. L. Ferrieres, G. Hemery, T. Nham, A. M. Guerout, D. Mazel, C. Beloin and J. M. Ghigo, Silent mischief: bacteriophage Mu insertions contaminate products of Escherichia coli random mutagenesis performed using suicidal transposon delivery plasmids mobilized by broad-host-range RP4 conjugative machinery, J. Bacteriol., 2010, 192, 6418–6427.
69. K. Mijnendonckx, A. Provoost, C. M. Ott, K. Venkateswaran, J. Mahillon, N. Leys and R. Van Houdt, Characterization of the Survival Ability of Cupriavidus metallidurans and Ralstonia pickettii from Space-Related Environments, Microb. Ecol., 2013, 65, 347–360.
70. H. Li and R. Durbin, Fast and accurate short read alignment with Burrows-Wheeler transform, Bioinformatics, 2009, 25, 1754–1760.
71. H. Li, B. Handsaker, A. Wysoker, T. Fennell, J. Ruan, N. Homer, G. Marth, G. Abecasis and R. Durbin, and S. Genome Project Data Processing, The Sequence Alignment/Map format and SAMtools, Bioinformatics, 2009, 25, 2078–2079.
72. D. Vallenet, L. Labarre, Z. Rouy, V. Barbe, S. Bocs, S. Cruveiller, A. Lajus, G. Pascal, C. Scarpelli and C. Medigue, MaGe: a microbial genome annotation system supported by synteny results, Nucleic Acids Res., 2006, 34, 53–65.
73. L. M. Guzman, D. Belin, M. J. Carson and J. Beckwith, Tight regulation, modulation, and high-level expression by vectors containing the arabinose pBAD promoter, J. Bacteriol., 1995, 177, 4121–4130.
74. M. E. Kovach, P. H. Elzer, D. S. Hill, G. T. Robertson, M. A. Farris, R. M. Roop, 2nd and K. M. Peterson, Four new derivatives of the broad-host-range cloning vector pBBR1MCS, carrying different antibiotic-resistance cassettes, Gene, 1995, 166, 175–176.
75. A. Schafer, A. Tauch, W. Jager, J. Kalinowski, G. Thierbach and A. Puhler, Small mobilizable multi-purpose cloning vectors derived from the Escherichia coli plasmids pK18 and pK19: selection of defined deletions in the chromosome of Corynebacterium glutamicum, Gene, 1994, 145, 69–73.
76. M. Bagdasarian, R. Lurz, B. Ruckert, F. C. H. Franklin, M. M. Bagdasarian, J. Frey and K. N. Timmis, Specific-Purpose Plasmid Cloning Vectors 2. Broad Host Range, High Copy Number, RSF1010-Derived Vectors, and a Host-Vector System for Gene Cloning in Pseudomonas, Gene, 1981, 16, 237–247.
77. W. G. Miller, J. H. Leveau and S. E. Lindow, Improved gfp and inaZ broad-host-range promoter-probe vectors, Mol. Plant-Microbe Interact., 2000, 13, 1243–1250.
78. S. T. Cardona and M. A. Valvano, An expression vector containing a rhamnose-inducible promoter provides tightly regulated gene expression in, Burkholderia cenocepacia Plasmid, 2005, 54, 219–228.
79. K. Jensen, PhD, Massachusetts Institute of Technology, 2006.

---

Footnotes

† Electronic supplementary information (ESI) available. See DOI: 10.1039/c9mt00123a

‡ Present address: Unit Health, Flemish Institute for Technological Research (VITO), Belgium.

---