![[Feedback]](/icons/banner/feedbackACT.gif){kind=icon}
![[For Subscribers]](/icons/banner/subscriberACT.gif){kind=icon}
![[Archive]](/icons/banner/archiveACT.gif){kind=icon}
![[Search]](/icons/banner/searchACT.gif){kind=icon}
![[Contents]](/icons/banner/tocACT.gif){kind=icon}
Department of Civil and Environmental Engineering, Northwestern University, Evanston, Illinois 60208,1 Department of Periodontics, The University of Oklahoma Health Sciences Center, Oklahoma City, Oklahoma 73104,2 Department of Civil, Architectural and Environmental Engineering, University of Texas at Austin, Austin, Texas 78712,3 Department of Microbiology, University of Iowa, 500 Newton Road, Iowa City, Iowa 522424
Received 12 June 2006/ Accepted 27 July 2006
 |
ABSTRACT |
|---|
 Top Abstract IntroductionMaterials and MethodsResults and DiscussionReferences |
|---|
|
INTRODUCTION |
|---|
|
TopAbstractIntroductionMaterials and MethodsResults and DiscussionReferences |
|---|
Many heavy metals, such as Cu, iron (Fe), and zinc (Zn), are required in trace amounts for bacterial growth yet are toxic when present in excess. Toxicity in biological systems occurs through a variety of mechanisms. Heavy metals bind to free thiol groups, disrupting protein structure or function (8, 19). They can also displace essential metal cofactors in proteins (e.g., Fe in cytochromes). Some heavy metals, such as Cu, also generate reactive oxygen species (ROS) through auto-oxidation or Fenton-like reactions (52, 61). Hydrogen peroxide can be produced intracellularly through the oxidation of NADPH (equation 1) and subsequent activity of the enzyme superoxide dismutase (equation 2). Hydrogen peroxide can then interact with copper to produce ROS (equations 3 and 4).
 | (1) |
 | (2) |
 | (3) |
 | (4) |
Bacteria have several mechanisms for coping with heavy metal stress, allowing them to thrive in ecosystems contaminated with toxic levels of heavy metals (49, 58, 68). Active efflux is a key aspect of resistance, transporting metal cations out of the cytosol and periplasmic space of gram-negative species (43, 56). Previous studies indicate that a network of different families of transporters coordinates efflux (38, 43, 56). One family of transporters is soft metal P-type ATPases (55). ATP hydrolysis drives transport of heavy metal cations from the cytoplasm into the periplasmic space, such as the well-characterized CopA of Escherichia coli (which effluxes Cu+1) (54). Also contributing to metal efflux across the inner membrane are cation diffusion facilitator (CDF) transporters (25, 43, 51). The CDF family has been primarily implicated in resistance to heavy metals such as Zn, Cd, and Co (3). Finally, some members of the CBA family mediate proton-driven efflux of heavy metals from the periplasm (and possibly cytoplasm) (20, 22, 43), such as the CusABCF system in E. coli (16, 22, 42, 48). Many CBA systems consist of an inner-membrane-associated resistance nodulation cell division (RND) transporter, a membrane fusion protein, and an outer membrane protein, forming a complex that spans the inner and outer membranes.
Some species can also reduce or oxidize heavy metals to less toxic forms. Pseudomonas aeruginosa reduces cytosolic mercury (Hg) to Hg0, which then passively diffuses out of cells (71, 73). The periplasmic multicopper oxidase of E. coli, CueO, has been hypothesized to protect the cell by converting Cu+1 to its less toxic form Cu+2 in an oxygen-dependent reaction (21). Alternatively (or additionally), CueO may protect E. coli by oxidizing catecholate siderophores, such as enterobactin, which can then sequester free Cu (23). The plasmid-encoded Pco system of E. coli has a multi-Cu oxidase, PcoA, that may function in a similar manner (30).
Several recent studies have explored the global responses of the paradigm enteric species, E. coli, to elevated levels of required heavy metals using transcriptional profiling (12, 34, 37, 74). Different approaches were taken to subject bacterial populations to the stress. In some cases, the heavy metal was applied to growing cultures as a "shock," and transcriptional profiles were assayed after a short period of exposure. In other cases, profiles were determined after longer periods of exposure, i.e., after bacterial populations had adapted to the stress and were actively growing. Both approaches were informative, indicating how E. coli responds and ultimately adapts to a heavy metal stress.
However, little is known in general about the global transcriptional responses of environmental bacterial species to heavy metal stress. Other than a recent report on the transcriptional response of Bacillus subtilis to different heavy metals, little has been published on environmental species (41). Unlike E. coli, P. aeruginosa is primarily an environmental organism commonly found in soil and water. Its genome has been sequenced, and commercially available DNA microarrays make it an excellent model organism for studying global transcriptional responses to environmental stimuli (64). Indeed, this approach has been successfully used to define transcriptional profiles of P. aeruginosa cultured under conditions ranging from elevated oxidative stress to denitrifying metabolism (14, 46, 50, 59, 60, 70).
In this study we examine the transcriptional profiles of P. aeruginosa exposed to a brief pulse of excess Cu (Cu shock) and cells growing continuously in the presence of excess Cu (Cu-adapted culture). Cu shock up-regulated expression of several genes involved in active transport functions as well as genes involved in oxidative and general stress responses. In contrast, transcriptional profiles of Cu-adapted populations suggested a Cu-specific physiology, with a much narrower range of genes encoding putative Cu-specific transport functions differentially expressed. An oxidative stress response was no longer apparent in the Cu-adapted population. Continuous Cu exposure also resulted in a general down-regulation of porins belonging to the OprD family. Iron homeostasis appeared to be particularly fine-tuned in Cu-adapted cultures. Genes encoding biosynthesis and transport of the iron-scavenging siderophore pyochelin were down-regulated in Cu-adapted populations, while genes encoding production of another siderophore, pyoverdin, were up-regulated. Strains bearing mutations in key copper-responsive genes identified in our study were then analyzed for Cu and Zn sensitivity. Our results suggest that two gene clusters homologous to RND family members, a P-type ATPase, a two component regulatory system, and a periplasmic system involved in proper disulfide bond formation all play important roles in tolerating elevated levels of Cu.
|
MATERIALS AND METHODS |
|---|
|
TopAbstractIntroductionMaterials and MethodsResults and DiscussionReferences |
|---|
Standard methods were used to manipulate plasmids and DNA fragments. Restriction endonucleases and DNA modification enzymes were purchased from New England Biolabs (Beverly, MA). Chromosomal DNA was isolated using DNeasy Tissue kits (QIAGEN, Valencia, CA), and plasmid isolations were performed using QIAprep spin miniprep kits (QIAGEN). DNA fragments were purified using QIAquick mini-elute PCR purification kits (QIAGEN), and PCR was performed using the Expand Long Template PCR System (Roche, Indianapolis, IN). DNA sequencing was performed by automated sequencing technology either using the University of Oklahoma Health Science Center sequencing core facility or the University of Iowa DNA sequencing core.
A selected group of transposon mutants in PAO1 were obtained from the University of Washington Genome Center and screened for Cu sensitivity (31). The following strains were obtained from the Washington collection: PTL361, PTL396, PTL1518, PTL3663, PTL4456, PTL4590, PTL4823, PTL5121, PTL5506, PTL5729, PTL7641, PTL7760, PTL7968, PTL8385, PTL8750, PTL9314, PTL9648, PTL13061, PTL14395, PTL16055, PTL16683, PTL17602, PTL17732, PTL18229, PTL19549, PTL20046, PTL21569, PTL22097, PTL22200, PTL30399, PTL35346, PTL36065, PTL36283, PTL47603, PTL12403, PTL44858, PTL48465, PTL20694, PTL9334, PTL14744, PTL6615, PTL8802, PTL17142, PTL31953, PTL7848, PTL42174, PTL8844, and PTL12392. There are several strains that have transposon insertions in the same gene. See Table 4 for data from representative strains that have been confirmed by PCR.
|
For construction of the reporter plasmid pAG100, a 496-bp DNA fragment containing the PA3920 promoter was amplified by PCR with the primers PA3920promoter-for (5'-CGGGATCCCGA TCACCTTCTGCTAAGGGCCTG-3') and PA3920promoter-rev (5'-GGGGTACCCC TCCTTGGAATGGACAGAGTGG-3') (restriction sites are underlined). The resulting PCR product was digested with the restriction enzymes BamHI/KpnI and then ligated into the BamHI/KpnI digested lacZ reporter plasmid pQF50, creating a transcriptional fusion of the PA3920 promoter with lacZ.
PCR validation of mutant strains. Strains harboring transposon mutations were verified using PCR. A DNA segment was amplified from the chromosome of mutant strains using primers internal to the transposon (either Hah-138, 5'-CGGGTGCAGTAATATCGCCCT, or LacZ-148, 5'-GGGTAACGCCAGGGTTTTCC) paired with a primer complementary to native chromosomal sequence close to the transposon insertion point. An amplified band of DNA of the correct size indicated the presence of the transposon at the expected location. The following primers were used to confirm mutant strains: for PA2809, 5'-CCTGCTTCGCCAGTTTCTCC; for PA3920, 5'-GGCAGCATCCAGTGCAG; for PA3521, 5'-CCTGTAGCTGGAACCAGGACATG; for PA0397, 5'-TCGACCACACATGCAGGTCG; for PA1435, 5'-GTTGGCGTGACACCGCACGC; for PA1436, 5'-TGAACTGCAGGGTCAGGGTGG; for PA1549, 5'-ATGGACCGCATCCTGCGCTGG; for PA2064, 5'-CGATCCAGCCGCTGGCGTCC; for PA2065, 5'-CGCCGTGGCCGAGACCTACG; for PA2476, 5'-GCAACCTGATCACCCTTGGC; for PA2477, 5'-AGGTAGCGCGCCAGGCTGCC; for PA2505, 5'-GTAGCGTCGACCTGCTGCCC; for PA2521, 5'-TGCGCTGGTCGGACAACTGC; for PA2522, 5'-GGCCAGCGCGGGGAAACGC; for PA2807, 5'-GAAGGTCCAGGTCAGCTCGG; for PA2810, 5'-CGACCATATGGCGATGTCCG; for PA3690, 5'-GATGGACTGCCCGACCGAG. The PCR protocol is available at the University of Washington website (http://www.genome.washington.edu/UWGC/pseudomonas/).
To further verify the accuracy of transposon insertion strains, PCR products generated from the reactions described above were subjected to sequencing to verify the location of the transposon insertion. Either the Hah-138 or the LacZ-148 primers were used in sequencing reactions. The following strains were confirmed in this manner: PTL12403, PTL16055, PTL20694, PTL17142, PTL12392, and PTL6615.
Transcriptional profiling experiments. For Cu shock conditions, PAO1 was inoculated at an optical density at 600 nm (OD600) of 0.05 in 50 ml of LBM in 250-ml flasks and cultured at 37°C until early logarithmic phase, corresponding to an OD600 of 0.2, at which point 10 mM CuSO4 was added to half of the flasks. The other half of the flasks were treated with a small amount of 1 M HCl to lower the pH to a similar level to that of the Cu-treated flasks (6.95 ± 0.04 for Cu-treated versus 7.01 ± 0.07 for the pH control flasks). After 45 min of treatment, cells were harvested and added to the RNA stabilization solution RNAlater (Ambion Inc., Austin, TX). Cu-adapted cells were subjected to 10 mM CuSO4 at inoculation. These cultures were grown at 37°C until early logarithmic phase (OD600 of 0.2), and then cells were harvested and added to RNAlater. The control flasks were grown at 37°C until early log phase (OD600 of 0.2). For each condition, all samples were grown in triplicate, and RNA from the different flasks was combined. In addition, biological replicates of each condition were performed on a separate day and run on a separate microarray chip.
RNA was isolated from cells using RNeasy columns (QIAGEN). An additional, off-column DNase treatment with RQ1 RNase-free DNase (Promega Corp., Madison, WI) was performed and then subsequently purified on the RNeasy column. RNA quality was assessed by agarose gel electrophoresis. Synthesis of cDNA and application to the microarray were performed following the instructions of the manufacturer of the GeneChip P. aeruginosa Genome Array (Affymetrix Inc., Santa Clara, CA) with some modifications. In the cDNA synthesis, 12 µg of RNA and semirandom hexamer primers with an average G+C content of 75% were used instead. S. Lory provided the spike transcript of control transcripts from six B. subtilis genes to monitor the labeling, hybridization, and staining efficiency of cDNA.
Analysis of DNA microarray data was first performed using Affymetrix Suite version 5.0, and further analysis was done using the online program CyberT (http://visitor.ics.uci.edu/genex/cybert/). CyberT utilizes a Bayesian statistical framework to calculate a more robust estimate of variance for a small number of experiments. The Bayesian prior estimate takes into consideration the background variance of the genes with similar expression levels as the gene of interest within a certain range, based on the theory that genes with similar expression levels should also have similar variance (27, 39). A sliding window size of 100 genes and a Bayesian prior estimate of 10 were used to calculate the P value. The posterior probability of differential expression [PPDE(<P)] was also calculated. PPDE(<P) ranges from zero to 1, and a low P value corresponds to a PPDE(<P) closer to 1. Genes were considered differentially regulated if the P value was <0.001, PPDE(<P) was >0.996, and the absolute change (n-fold) was >3.
Real-time reverse transcription-PCR (RT-PCR). Cultures of PAO1 were grown in 250-ml baffled flasks containing 20 ml of LBM at 37°C and 150 rpm. Triplicate cultures were grown for the control (no Cu addition), Cu shock (10 mM CuSO4 addition when the OD600 reached 0.2), and Cu-adapted (10 mM CuSO4 addition at inoculation) experiments. The pH was consistent among the three experimental conditions. RNA was isolated from the control experiments when the OD600 was approximately 0.5. The Cu-shocked cells were treated with copper for 45 min; the OD600 was approximately 0.5 at the end of the treatment, and RNA was isolated immediately. The Cu-adapted cells also were grown to an OD600 of approximately 0.5, at which time the RNA was isolated immediately. One RNA extraction was performed for each flask, and the RNA was pooled for each condition (control, Cu-shocked, and Cu-adapted cultures). RNA was isolated with an RNeasy kit (QIAGEN), and residual DNA was removed with DNase I treatment (RiboPureTM-Bacteria kit, Ambion, Austin, TX).
Duplicate cDNA synthesis reactions were performed with random hexamer primers and avian myeloblastosis virus reverse transcriptase (Roche, Germany). Four micrograms of RNA was mixed with 2 µl of the primers (10 µM) and nuclease-free water (Ambion) to a final volume of 10 µl. The RNA mixture was denatured at 70°C for 5 min and placed on ice. Four microliters of 5x First-Strand Buffer, 2 µl of deoxynucleoside triphosphate mix (a 10 mM concentration of each nucleoside; New England Biolabs, Beverly, MA), 2 µl of water, and 2 µl of avian myeloblastosis virus reverse transcriptase were added to each reaction on ice. Each tube was incubated at 42°C for 1.5 h in a Primus 25 Personal Cycler with a heated lid (MWG Biotech, High Point, NC). Negative controls for RT-PCR were prepared by omitting the reverse transcriptase in the cDNA synthesis reactions and were used to verify that contaminating genomic DNA was absent from the RNA preparations. Duplicate cDNA reactions were pooled and used in triplicate real-time PCRs. Real-time PCRs were run on an Applied Biosystems 7900HT Real-Time PCR System (Applied Biosystems, Foster City, CA) using Power Sybr Green PCR Master Mix (Applied Biosystems) according to the manufacturer's instructions. The final concentration of each primer was 0.3 µM for all reactions. The following thermocycler program was used: 95°C for 10 min followed by 40 cycles of 95°C for 15 s and 60°C for 60 s. Real-time PCR product dissociation curves were used to verify the specificity of the amplified product. Fivefold serial dilutions of cDNA, prepared from a mixture of equal aliquots of cDNA from each experimental condition, were used to construct a standard curve for each gene tested (PA2477, PA2505, PA2522, PA3521, PA3920, PA4067, and PA4946). The cycle number at which fluorescence crossed a selected threshold value during exponential amplification was correlated to a relative quantity. The relative quantity of a gene transcript was normalized to the relative quantity of the housekeeping gene (PA4946) transcript for the same experimental condition. The relative change in expression (n-fold) was calculated as the relative quantity of the target gene transcript under Cu shock or Cu-adapted conditions (normalized to PA4946), divided by the relative quantity of the target gene transcript under the control condition (normalized to PA4946).
Pyoverdine and pyochelin measurements. Cultures were grown in a modified Jensens medium at 37°C with glucose as the sole carbon source. Pyoverdine concentrations were calculated from the absorbance at 403 nm of cell-free, diluted culture supernatants as previously described (46). Purified pyoverdine was used to generate a standard curve. A modification of a previous protocol was used to measure pyochelin levels. The dried organic fraction of ethyl acetate-extracted cell-free culture supernatants was resuspended at 10x concentration, and pyochelin levels were determined as previously described (69). Purified pyochelin was used to generate a standard curve.
MIC determination for mutant strains. The MIC of Cu was the lowest concentration at which there was no growth after 16 h of exposure. In a 96-well microtiter plate format, PAO1 and mutant strains were subjected to an array of heavy metal concentrations in LBM. Each well contained 100 µl of LBM plus CuSO4 concentrations ranging from 1 mM to 16 mM. Each well was inoculated with a 10-µl aliquot of cells from an actively growing culture at an OD600 of 0.1. The microtiter plate was incubated at 37°C with gentle shaking for 15 s every 30 min. Growth was monitored by reading the OD600 every 30 min using a microtiter plate spectrophotometer (Genios microplate reader [Tecan Research Triangle Park, NC] or Synergy HT [Bio-Tek Instruments, Winooski, VT]), and the MIC was determined after 16 h. Each experiment was performed in triplicate.
Disk sensitivity assay. Tested strains were grown to late logarithmic phase in LB at 37°C, and then 500 µl of cell suspension was added to a cooled, molten soft LB agar (0.5% agar). The agar was then poured into petri plates and allowed to solidify. A sterilized Whatman filter disk (catalog no. 2017013) incubated overnight in either 500 mM CuSO4 or 500 mM ZnSO4 was added to the center of the agar plate after it had solidified. The plate was then incubated at 37°C overnight, and the zone of clearance surrounding the disk was measured.
Copper viability analysis. Tested strains were grown in 250-ml baffled flasks containing 50 ml of LBM until mid-logarithmic phase (OD600 of 0.5). Cells were harvested, centrifuged, and resuspended and washed in the minimal medium MSVG. Aliquots were subjected to an array of Cu concentrations in MSVG and then placed at 37°C. After 5 h, the samples were sonicated for 30 s (3 s on at an output of 7 and 1 s off for 10 cycles) (Ultrasonic liquid processor XL; Misonix Inc., Farmingdale, NY), and viable plate counts were performed on serial dilutions.
ß-Galactosidase assays. ß-Galactosidase activity was monitored colorimetrically as described by Zubay et al. (77). To examine the expression of PA3920 in response to increasing copper levels, overnight cultures of P. aeruginosa containing pAG100 were diluted in Trypticase soy broth to an OD600 of 0.05 and incubated with shaking (250 rpm) at 37°C for 1 h. Increasing concentrations of CuSO4 (50 µM to 1.5 mM) were then added to individual cultures; cultures were incubated for 45 min, and ß-galactosidase activities were determined.
Microarray accession number. Array data were submitted to the Gene Expression Omnibus website (http://www.ncbi.nlm.nih.gov/geo/) under accession number GSE4152.
|
RESULTS AND DISCUSSION |
|---|
|
TopAbstractIntroductionMaterials and MethodsResults and DiscussionReferences |
|---|
A range of CuSO4 was added to actively growing P. aeruginosa cultures in order to define acute stress conditions. A buffered complex medium (LBM) was used to maximize biomass yield. A Cu concentration was chosen (10 mM) that impacted P. aeruginosa growth rate but was not lethal, as determined by viable plate counts. Forty-five minutes after the addition of 10 mM Cu, only minimal growth had occurred in the treated culture, while growth had occurred in the control culture (OD600 of 0.310 versus 0.396, respectively) (Fig. 1).
|
Figure 2, Tables 1 to 3, and Table S1 in the supplemental material feature genes that were up- and down-regulated in shocked and adapted populations. Genes were considered differentially regulated if the relative change (n-fold) was >3.0, the P value was <0.001, and the PPDE(<P) was >0.996. Cu shock resulted in differential expression of 405 genes. Most of these genes were up-regulated. Although more genes would have qualified as repressed, they did not meet the P value and PPDE stringency cutoffs. In the case of Cu-adapted cultures, 331 genes were either up- or down-regulated. To further validate the array data, real-time RT-PCR was used to determine the relative quantities of selected transcripts under control and test culturing conditions. These are indicated in a separate column in Tables 1 and 3.
|
|
|
Active efflux is a key aspect of copper tolerance. RND, CDF, and P-type ATPase efflux transporters have been suggested to constitute an efficient, integrated network controlling heavy metal levels in the cytoplasm and periplasm for some gram-negative bacterial species. Several studies have shown that transporters belonging to these families are induced in the presence of elevated levels of their heavy metal substrates.
Cu stress up-regulated expression of genes encoding predicted active transport functions (Table 1). The P. aeruginosa chromosome has seven genes homologous to P-type ATPases (43). One, PA3920, was up-regulated in both shocked and adapted populations. In addition, a few RND family transporters were up-regulated under both conditions. The P. aeruginosa chromosome has 12 loci harboring genes homologous to RND family members (35, 43). Three were up-regulated in response to Cu exposure (PA1436, PA2520, and PA3522). The most prominently up-regulated locus is PA2520 to PA2522, which encodes the czrCBA transport system, previously shown to be important for Cd, Zn, and Co tolerance (26, 57). An adjacent two-component regulatory system (encoded by PA2523 and PA2524) was also up-regulated in Cu-stressed cultures. The PA3521 to PA3523 and PA1435 to PA1436 genes have no known efflux substrates (62). These loci also have no linked regulatory proteins.
A gene encoding a CDF family member, PA0397, was up-regulated in Cu-adapted populations. A homologous gene in Ralstonia metallidurans is thought to confer resistance to Cd, Zn, and Co (2). Nies suggested that the Cu ion may be too small to serve as a substrate for transport for this family (43), although Moore et al., recently reported that a CDF family member of B. subtilis conferred protection to Cu (41). The PAO1 chromosome has two other genes belonging to the CDF family, PA1297 and PA3693. PA1297 is 50% similar to the zinc transporter, ZntA, of Staphylococcus aureus and was up-regulated under Cu shock conditions.
Copper shock appears to produce an oxidative stress response. Two previously published studies characterized the transcriptional profiles of P. aeruginosa exposed to reactive oxygen species (50, 59). Comparing our data with the data of Palma et al. and Salunhke et al. suggests that Cu shock results in an oxidative stress response. Several genes identified in these studies were also induced in the Cu-shocked cultures, such as superoxide dismutase (sodM), catalase (katB), genes involved in organic hydroperoxide resistance (ohr and ahpF), and genes involved in the DNA SOS response/DNA damage and repair (dinP, recA, and recN) (Table 2; see Table S1 in the supplemental material). Comparing their data sets with ours allowed us to filter the Cu shock data set such that none of the genes presented in Tables 1 and 3 are related to an oxidative stress response. This finding is in agreement with the studies of Kershaw et al. and Egler et al., where the addition of either copper sulfate or copper-glycine to E. coli cultures was shown to produce an oxidative stress response (12, 34). In contrast, Cu-adapted populations no longer showed elevated expression of these genes, possibly due to decreased intracellular Cu following adaptation of the organism to the stress.
|
An operon encoding putative metal-binding proteins and Cu oxidases was also up-regulated in both Cu treatments (Fig. 2 and Table 3). Cu stress induced expression of the pcoAB genes (PA2065 and PA2064) which are homologous (81% and 75% similarity, respectively) to the copAB genes of Pseudomonas syringae. CopA and CopB have been demonstrated to bind Cu in the periplasm (CopA) and outer membrane (CopB) (6). PcoA has a multicopper oxidase domain (SufI domain; COG2132 [http://www.ncbi.nlm.nih.gov/COG]), similar to CueO of E. coli. This enzyme might participate in the oxidation of Cu+1 to the less toxic form Cu+2 or the oxidation of catecholate-like siderophores (although P. aeruginosa is not known to produce its own catecholate siderophore). Like Pseudomonas putida, P. aeruginosa lacks homologs of the copCD genes that are present in P. syringae (5).
Another gene, PA3520, was up-regulated by Cu and is associated with a cluster of genes that are also Cu responsive. This gene is predicted to encode a small protein that contains a metal binding motif, MXCXXC. This motif is found in a variety of proteins that bind heavy metals, such as CopZ. CopZ in Enterococcus hirae is thought to be a periplasmic Cu chaperone (40). However, the localization of the PA3520 gene product to the periplasm is uncertain. Homologous genes are found on the P. putida (PP0588) and P. syringae (PSPTO0752) chromosomes.
Regulation of outer membrane permeability appears to be a feature of adaptation to copper. Some passive transport functions were down-regulated in Cu-adapted populations. Eight different porins or predicted porins were down-regulated (Table 1). One porin, OprD, binds to basic amino acids or basic dipeptides. Interestingly, reduced OprD expression has been linked to resistance to carbapenem antibiotics, and addition of low micromolar levels of Zn to P. aeruginosa led to its reduced expression (24). Four more Cu-repressed porins, OprQ, OpdC, OpdN, and OpdP, belong to the OprD family but have no assigned uptake function (24). OprC is TonB dependent and is thought to be Cu specific. Elevated extracellular Cu was previously reported to repress its expression (76). OprG and OprE were also repressed. OprG has been suggested to be involved in low-affinity Fe uptake, and OprE is an anaerobically induced porin (18, 75). The porin encoded by PA2505 was the only porin induced by long-term Cu exposure. This gene encodes opdT, a porin that also belongs to the OprD family. Perhaps induction of opdT compensates for repression of the other OprD family members while not compromising the cell in terms of Cu permeability.
Increased porin expression has been previously reported as a feature of Cu resistance in E. coli. Egler et al. reported that OmpC expression was increased in response to Cu and controlled through the extracytoplasmic sigma factor, RpoE (12). Mutations in OmpC rendered E. coli sensitive to Cu, although the underlying mechanism was unclear.
Effects of copper stress on iron homeostasis. A distinct group of genes involved in Fe homeostasis were induced in Cu-stressed cultures. This group overlapped with genes identified in P. aeruginosa exposed to Fe-limiting conditions (46). Of the 30 genes with the highest relative induction (n-fold) in response to Fe limitation identified by Ochsner et al., 20 were also up-regulated in response to Cu stress. Many of these 20 genes have consensus binding sequences in their promoter regions for the Fe-responsive regulators, Fur and PvdS (10, 45). In line with this observation is the up-regulation of pvdS (PA2426) under both stress conditions. Another interesting similarity with the Ochsner study is that, of the eight genes homologous to extracytoplasmic function family sigma factors they reported to be induced by Fe limitation (PA2387, PA3899, PA4896, PA2468, PA3410, PA1300, PA0472, and pvdS), seven are also up-regulated in Cu-shocked populations (all of the above except for PA2387). The heavy metal induction of Fe siderophores has previously been reported for E. coli (12, 34). The siderophore enterobactin not only may play a role in Fe transport but also may protect the cell by sequestering heavy metals in the cytoplasm (23).
Many Fe-regulated genes, including pvdS, are under control of the regulatory protein Fur, which complexes Fe and represses expression of target genes under Fe-replete conditions. Heavy metals could potentially interact with Fur, displacing Fe and thus influencing its binding to target DNA sequences (9). Another possibility is that ROS generated by Cu stress could oxidize Fur-bound Fe2+, releasing Fe from Fur and relieving Fur-mediated repression.
The major difference between this study and that of Ochsner et al. is that Cu-adapted populations repressed expression of genes involved in synthesis of the siderophore pyochelin, whereas these genes were induced under Fe-limiting conditions. This is consistent with Visca et al., who previously reported that several transition metals, including Cu, repress pyochelin synthesis (67).
In contrast, several genes involved in the biosynthesis of pyoverdine were up-regulated in Cu-adapted populations (Table 3 and Fig. 2). Pyoverdine has been previously shown to be up-regulated in response to Cd and Zn (11, 29, 67). Pyoverdine might provide the bacterium with an iron-specific acquisition system, whereas pyochelin is down-regulated since its broader specificity for metal cations might allow Cu to enter the cell. The regulation of Fe acquisition appears to be carefully fine-tuned in a Cu-stressed environment.
To functionally verify transcriptional profiling data, P. aeruginosa was exposed to different levels of CuSO4, and pyoverdine and pyochelin levels were measured. Pyoverdine levels were seen to increase more than 10-fold in cultures treated with CuSO4 for a long period of time (Fig. 3). As suggested by the array data, pyoverdine levels did not increase after short-term exposure. A mutant strain unable to biosynthesize pyoverdine, PAO1-pvdA, did not show induced pyoverdine production in the presence of Cu. The array data also predict that long-term copper exposure would result in reduced pyochelin levels. However, even in untreated cultures, pyochelin levels were below detectable limits (data not shown).
|
|
|
Several strains for which the Cu MIC was comparable to the wild-type level exhibited a lag in growth in the presence of Cu compared to the wild-type strain (Table 4). Examples of such a lag are depicted in the growth curves presented in Fig. 4. The growth of PAO1 is unaffected by added CuSO4, until MIC levels are reached (Fig. 4A). However, the mutant strains show a delay in growth and a reduced growth rate compared to PAO1 at near MIC levels of CuSO4 (e.g., 8 mM in Fig. 4 B, C, and D). In particular, the PA1435 mutant strain showed a delay in growth at a few CuSO4 concentrations. Other mutant strains that produced a lag in growth in the presence of Cu were PA2476 and PA2478 (both belong to the same operon as PA2477 and show homology to periplasmic thiol-disulfide folding proteins), PA2810 (a homolog of a two-component sensor), and PA3521 (an outer membrane protein).
The two strains judged to be most sensitive to Cu by MIC analysis, PA2809 and PA3920, were extremely sensitive to CuSO4 (Fig. 5A and B). To verify that the transposon insertion was responsible for the Cu-sensitive phenotype, these strains were complemented in trans by supplying a functional copy of the gene on the broad-host-range vector pEX1.8. The complemented mutant strains displayed almost wild-type levels of Cu-tolerance (Fig. 5C and D and Table 4).
A disk sensitivity assay was used as a complementary means to evaluate Cu sensitivity. This assay uses a Cu-impregnated disk, which is set on the surface of soft LB agar. Zones of clearance in a lawn of bacteria were measured around the disk. The wild-type strain showed a clear demarcation between copious growth and no growth on the agar plate. However, some strains produced a halo of very light growth adjacent to the zone of clearance (Table 4).
The Cu disk assay results correlated with MIC data, with PA2809 and PA3920 displaying the largest zone of clearance (Table 4). In some cases, strains showed Cu sensitivity in the disk assay and not the MIC assay. This usually occurred with mutations in a putative operon already identified as important for Cu tolerance in the MIC analysis (e.g., PA2521). There were three exceptions. A strain with a mutation in pcoA (PA2065) showed significant Cu sensitivity in the disk assay but not in the MIC analysis. A strain with a mutation in PA0397 (a CDF family homolog) and PA2807 (a gene encoding a member of the plastocyanin/azurin Cu-binding family) were also sensitive in the disk assay. Why these strains have a sensitive phenotype in one assay and not the other is unclear.
Several genes induced by Cu stress showed homology to genes involved in responses to other cations, such as Zn. In particular, the czr system (PA2520 to PA2522) is known to be important for Zn tolerance in P. aeruginosa (26). Zinc is also a required metal for growth and in excess can present the same problems as Cu to the cell. Therefore, the mutant strains were also subjected to a zinc disk sensitivity assay. This assay demonstrated the metal specificity of certain systems. A strain with a mutation in PA3920, encoding a P-type ATPase, was very sensitive to Cu but not to Zn (Table 4). On the other hand, strains with mutations in two other P-type ATPases, PA3690 (homologous to zntA, a Zn P-type ATPase in E. coli) and PA1549, were Zn sensitive yet showed no Cu phenotype. As Hassan et al. previously demonstrated, the czrABC operon (with the exception of PA2521) appeared to be important for Zn tolerance (26). Also implicated in Zn tolerance by this assay were the pcoAB genes.
Our initial analysis suggested that PA3920 and PA2809 contributed significantly to Cu tolerance in P. aeruginosa. Therefore, these genes were subjected to further analysis.
A two-component regulator is important for growth and survival under Cu stress. The two-component response regulator PA2809 was up-regulated in both Cu-shocked and Cu-adapted populations and may be involved in pcoAB regulation. This gene is highly homologous to copR in P. syringae (82% similar), P. putida, and Pseudomonas fluorescens and to pcoR in E. coli. The regulators in these systems control expression of the Cu-binding proteins CopAB (copR) and PcoAC (pcoR) (5, 15, 44). Due to this homology, we propose naming PA2809 copR and its adjacent sensor kinase, PA2810, copS. Two neighboring genes, PA2807 and PA2808, were also induced by Cu stress. The hypothetical protein PA2807 harbors a Cu-binding motif of the plastocyanin/azurin family. PA2807 encodes a protein similar to the Cu-binding protein Cot from P. fluorescens DF57. Cot was identified as a Cu tolerance protein that has elevated expression in response to Cu (66).
Since a mutation in PA2809 significantly influenced Cu tolerance (Table 4), this strain was examined further. The mutant strain was also tested for Zn and silver (Ag) sensitivity and found to be identical to the wild-type strain (data not shown). The toxicity of Cu (as opposed to inhibition of growth) was then tested on the PA2809 mutant and wild-type strain. Cells were exposed to a 5-h treatment of various Cu concentrations. Since complex medium has many components with a high capacity to complex Cu, this assay was performed in a minimal medium designed to minimize complexation. The PA2809 mutant strain was killed at lower Cu concentrations than wild-type (Fig. 6). Resistance to Cu was regained when the mutant strain was complemented in trans with a plasmid bearing the PA2809 gene (Table 4 and data not shown). Since PA2809 is up-regulated during both Cu treatments, it appears to be a key regulator involved in Cu tolerance in P. aeruginosa.
|
The MICs of the heavy metals Cu, Co, Zn, and Hg were assayed for a strain harboring a mutation in PA3920 and compared to PAO1. In complex and defined media, the mutant exhibited increased susceptibility to copper compared to PAO1 while no differences were observed for the other heavy metals (Table 4 and data not shown). Cu susceptibility was tested in a complex and a minimal medium, and although the MICs were significantly affected by the medium composition, the mutant was hypersusceptible to copper in all media tested (data not shown).
PA3920 expression responds to external copper. Genes encoding metal-transporting ATPases are often induced by the presence of their metal substrates (1, 47, 54). Thus, a transcriptional fusion of the PA3920 promoter region was cloned upstream of the promoterless reporter gene lacZ in the plasmid pAG101. In complex medium, transcription of PA3920 exhibits a linear response to increasing concentrations of copper up to 1.2 mM (Fig. 7). The transcriptional response is immediate, occurring within 45 min after the addition of copper.
|
Transcriptional profiling of Cu-stressed P. aeruginosa produced both expected and unexpected results. Not surprisingly, many active transport functions were up-regulated in response to Cu. Previous studies have shown that RND efflux systems, P-type ATPases, and CDF transporters play an important role for metal tolerance in other organisms. In particular, the P-type ATPase encoded by PA3920 appeared to be important for Cu tolerance. The expression of several porins appeared to be down-regulated in adapted populations. Most of these porins belonged to the OprD family, which may be involved the in passive transport of positively charged molecules.
Iron homeostasis functions were also regulated under Cu stress conditions. In general, the cells produced a response indicative of iron-limiting conditions. An interesting observation involved the two iron siderophore systems, pyoverdine and pyochelin. Pyoverdine biosynthetic genes were up-regulated in Cu-adapted populations, while pyochelin genes were down-regulated. The explanation for this observation may lie with the specificity of the siderophores. Pyoverdine is highly iron specific, while pyochelin has been shown to complex a variety of divalent metal cations, including Cu.
Finally, a variety of genes not generally associated with metal tolerance were up-regulated in the Cu-adapted population. The dsb periplasmic thiol-disulfide folding system was up-regulated and appeared to be functionally important for growth in the presence of excess Cu. Several operons containing genes of unknown functions were also up-regulated in adapted populations. Particularly interesting were the genes PA3515 to PA3520, which constitute an operon adjacent to an induced RND efflux system (PA3521 to PA3523). A few of these genes have homology to lyases, but the function of this operon is a mystery. Determining how genes such as these contribute to Cu adaptation may produce new insight into metal tolerance of microorganisms.
|
ACKNOWLEDGMENTS |
|---|
We thank Charlie Cox for the gift of purified pyoverdine and pyochelin and Blaise Boles for the PAO1-pvdA strain. We thank Ivonne Granados for help with experiments and Thomas V. O'Halloran, Mike Vasil, and Hilary Godwin for helpful discussions.
|
FOOTNOTES |
|---|
Supplemental material for this article may be found at http://jb.asm.org/.
|
REFERENCES |
|---|
|
TopAbstractIntroductionMaterials and MethodsResults and DiscussionReferences |
|---|
), while the Cu-shocked culture is indicated by filled triangles (
). The time at which Cu is added to the shocked culture or HCl is added to the pH control culture is indicated by the solid arrow. Cu-adapted cultures are indicated by the open triangles (
). In this case, 10 mM CuSO4 was present at inoculation of the culture. The untreated controls for the Cu-adapted samples is indicated by filled squares (
). The dashed arrows represent when the culture was harvested for transcriptional profiling. Dashed arrow 1 indicates when the acute stresses culture was harvested, while the dashed arrow 2 indicates when the adapted culture was harvested. The data points represent three experiments with the standard deviation bars shown.
