Candidate Stress Genes of Nitrosomonas europaea for Monitoring Inhibition of Nitrification by Heavy Metals

Abstract

Heavy metals have been shown to be strong inhibitors of nitrification in wastewater treatment plants. In this research, the effects of cadmium, copper, and mercury on Nitrosomonas europaea were studied in quasi-steady-state batch reactors. When cells were exposed to 1 μM CdCl₂, 6 μM HgCl₂, or 8 μM CuCl₂, ammonia oxidation rates were decreased by about 90%. Whole-genome transcriptional and proteomic responses of N. europaea to cadmium were used to identify heavy metal stress response genes. When cells were exposed to 1 μM CdCl₂ for 1 h, 66 genes (of the total of 2,460 genes) were upregulated, and 50 genes were downregulated more than twofold. Of these, the mercury resistance genes (merTPCADE) averaged 277-fold upregulation under 1 μM CdCl₂, with merA (mercuric reductase) showing 297-fold upregulation. In N. europaea cells exposed to 6 μM HgCl₂ or to 8 μM CuCl₂, merA showed 250-fold and 1.7-fold upregulation, respectively. Cells showed the ability to recover quickly from Hg²⁺-related toxic effects, apparently associated with upregulation of the mercury resistance genes and amoA, but no such recovery was evident in Cd²⁺-exposed cells even though merTPCADE were highly upregulated. We suggest that the upregulation of merA in response to CdCl₂ and HgCl₂ exposure may provide a means to develop an early-warning indicator for inhibition of nitrification by these metals.

Nitrifying bacteria, such as Nitrosomonas europaea (ATCC 19718), are important in the removal of nitrogen in wastewater reclamation plants. N. europaea obtains essential reductants for energy and biosynthesis from the oxidation of ammonia (NH₃⁺) to nitrite (NO₂⁻) and uses CO₂ as its carbon source (1). Oxidation of NH₃⁺ by N. europaea is a two-step reaction catalyzed by ammonia monooxygenase ([AMO] a membrane-bound enzyme) and hydroxylamine oxidoreductase ([HAO] a periplasmic protein), generating NO₂⁻ as the final product (1). Nitrifying bacteria are sensitive to various environmental contaminants and generally have low growth rates, with doubling times of about 8 to 12 h (46), making them the critical step in biological nitrogen removal (BNR). Heavy metals, such as Cd²⁺ (12, 28), Hg²⁺, and Cu²⁺ (30), are extensively used in industry (e.g., in the fabrication of pigments, batteries, and electronics), and improper disposal of the metals or their by-products tends to contaminate the environment (40). These metals may inhibit nitrification in the reclamation of wastewater (4, 5, 27, 44).

Inhibition of nitrification has been best documented in the model bacterium N. europaea. N. europaea is sensitive to inorganic compounds such as Cd²⁺ (7, 21) and organic compounds such as chlorinated aliphatic hydrocarbons (23) and to pH shifts (14), among other factors. With the sequencing of the N. europaea genome, transcriptomics studies to characterize N. europaea stress responses to inhibitors are now possible. Exposure to Zn²⁺ caused inhibition of ammonia oxidation and, concomitantly, expression of specific genes encoding membrane transporter and putative metal resistance proteins (38). Exposure to chlorinated aliphatic hydrocarbons increased the expression of genes encoding heat shock proteins, sigma factors of the extracytoplasmic function (ECF) subfamily, and toxin-antitoxin loci (15). Such genes potentially may be useful as early-warning toxicity indicators to prevent excessive inhibition of nitrification and in BNR processes.

Failure or inhibition of BNR processes caused by heavy metals, such as Cd²⁺, Hg²⁺, and Cu²⁺, can lead to high ammonia discharges and contribute to eutrophication of water bodies (44). Because current methods to monitor nitrification are laborious and cannot identify causes of inhibition, quick and sensitive methods are required for early detection of nitrification inhibition and for identification of its causes. Such methods could rely on “sentinel genes” that are highly upregulated in the presence of a certain heavy metal or group of heavy metals. In a previous study, merTPCAD were identified as putative sentinel genes in N. europaea exposed to Zn²⁺ (38). The main objective of the current study was to look for similar sentinel genes by examining whole-genome transcriptional changes in N. europaea exposed to Cd²⁺. For comparison, transcriptional changes of selected genes were also determined in cells exposed to Hg²⁺ and Cu²⁺.

MATERIALS AND METHODS

Batch reactor experiments.

N. europaea cells were grown in batch cultures with 25 mM (NH₄)₂SO₄ as described previously (38), harvested in mid- to late-exponential phase (optical density at 600 nm [OD₆₀₀] of ≅0.07), and washed two times with 40 mM NaH₂PO₄ (pH 7.8). Washed cells were resuspended in 1 liter of 50 mM HEPES buffer (pH 7.8) containing (NH₄)₂SO₄ (2.5 mM). The 1-liter cell suspension was evenly divided into two, gas-tight, 1.67-liter reactor vessels (Wheaton double-sidearm cell stir), and experiments were conducted without further additions of either buffer or ammonia. Cell suspensions prepared in this way sustained a quasi-steady-state condition during 4-h incubations comparable to N. europaea batch reactor experiments previously reported using phosphate-buffered medium (15). For the experiments, cells were stirred for 1 h to allow them to reach quasi-steady state, at which point either 1 μM CdCl₂, 6 μM HgCl₂, or 8 μM CuCl₂ was injected into the treatment reactor vessels.

Oxygen uptake measurement and nitrite assay.

The ammonia-dependent specific oxygen uptake rate (A-SOUR), hydrazine-dependent specific oxygen uptake rate (H-SOUR), and NO₂⁻ production rates were tested every 30 min during the incubations, as described previously (23). An aliquot of 1 ml drawn from a batch reactor was centrifuged immediately and analyzed to determine the NO₂⁻ concentration (22). The A-SOUR was measured in a 1.8-ml glass water-jacketed reaction vessel at 30°C using a heated circulating water bath (10). The H-SOUR was then determined by blocking ammonia-dependent oxygen uptake with 100 μM allylthiourea and adding 750 μM hydrazine as an alternative substrate for HAO. The SOUR of the cells was calculated based on the saturated oxygen concentration in water (10).

Affymetrix microarray experiments.

A volume of 180 ml of N. europaea culture (OD₆₀₀ of ≅0.07) was obtained from three independent experiments (three control reactors and three Cd²⁺-treatment reactors). Trizol (Ambion Inc., Austin, TX) was used to extract total RNA from the suspension, following the manufacturer's instructions. After purification with an RNeasy Mini kit (Qiagen Inc., Valencia, CA), the quality and quantity of the RNA samples were measured using an RNA 6000 Nano LabChip kit on an Agilent Bioanlyzer 2100 (Agilent Technologies, Inc., Palo Alto, CA). All of the 2,460 annotated N. europaea genes were represented on the custom Nimble Express made-to-order arrays (NimbleGen Systems, Inc.). cDNA synthesis, labeling, and hybridization were carried out by the Center for Genome Research and Biocomputing Core Laboratories at Oregon State University. GeneSpring software (version 7.2; Silicon Genetics) was used to identify genes that were upregulated or downregulated more than twofold under Cd²⁺, applying an unpaired two-sample t test with a cutoff P value of 0.05.

qRT-PCR.

Quantitative reverse transcription-PCR (qRT-PCR) was used to evaluate the transcriptional levels of selected up- or downregulated genes identified in the microarrays and to determine transcriptional levels of merA in N. europaea cells exposed to Cu²⁺ and Hg²⁺. cDNA was synthesized using an IScript cDNA synthesis kit (Bio-Rad Laboratories, Inc., Hercules, CA) with total RNA extracted from cells treated with or without Cd²⁺, Cu²⁺, or Hg²⁺, as described previously (38). qRT-PCR was performed on an ABI 7500 instrument (Applied Biosystems, Foster City, CA) with IQ Sybr Green Supermix (Bio-Rad). Primers were designed using Primer3 software and manufactured commercially (Invitrogen, Carlsbad, CA). The qRT-PCR efficiency was determined using standard curves created by serial dilution of RNA samples. The relative changes to show ideal amplification efficiency were calculated using the formula 2^−ΔΔC_T (where C_T is cycle threshold) (39).

2D SDS-PAGE.

To find proteins that were differentially expressed under Cd²⁺ stress, total proteins, including membrane-bound proteins, were prepared from N. europaea cells after a 3-h incubation with and without 1 μM CdCl₂ using a ReadyPrep Sequential Extraction Kit (Bio-Rad). The extracted total proteins were then compared by two-dimensional (2D) sodium dodecyl sulfate-polyacrylamide gel electrophoresis (SDS-PAGE) as described previously (38). ReadyStrip immobilized pH gradient strips (13 cm; Bio-Rad) with a nonlinear pH gradient from 3 to 10 were used for isoelectric focusing. SDS-PAGE was carried out on 12.5% precast SDS-PAGE gels (Bio-Rad) at 200 V for 45 min. The upregulated proteins in Cd²⁺-exposed cells were excised from SDS-PAGE gels stained with Sypro Ruby (Cambrex Bioscience, Rockland, ME) and identified by nano-liquid chromatography-tandem mass spectrometry (Midwest Bio Services, LLC, Overland, KS).

Microarray data accession number.

The microarray data are available at the Gene Expression Omnibus database under accession number GSE9221 (http://www.ncbi.nih.gov/geo).

RESULTS

Nitrite production and SOURs.

To test the toxicity of the heavy metals in this study, N. europaea cells were placed in two batch reactors (a treatment reactor and a control reactor). In the control reactors, the cells consistently reached quasi-steady state in about 1 h (constant NO₂⁻ production rate of about 0.2 mM/min·OD₆₀₀) and were stable for up to 4 h (Fig. 1a). In the treatment reactors, 1 μM CdCl₂, 6 μM HgCl₂, or 8 μM CuCl₂ was added at 1 h in quasi-steady state. After 1 h of further incubation with Cd²⁺, the cells showed about an 88% decrease in the NO₂⁻ production rate (Fig. 1a) and an 82% decrease in A-SOUR (Fig. 1b). N. europaea exposed to 6 μM HgCl₂ or 8 μM CuCl₂ for 1 h showed about 90% and 82% inhibition, respectively, of A-SOUR (Fig. 1b), while their corresponding nitrite production rates decreased by about 88% and 76%, respectively.

FIG. 1.

N. europaea nitrite production rate (a) and A-SOUR (b) under 1 μM CdCl₂, 6 μM HgCl₂, or 8 μM CuCl₂ treatments for 4 h. Nitrite production rate and AMO activity were normalized by cell density. Closed circles, open circles, closed triangles, and open triangles represent control condition, 8 μM CuCl₂, 6 μM HgCl₂, and 1 μM CdCl₂, respectively. The vertical dashed lines indicate the injection time of inhibitors. The error bars indicate 95% confidence intervals.

Determinations of H-SOUR showed that hydrazine-dependent oxygen uptake activity remained unaffected in N. europaea cells exposed to 1 μM CdCl₂ and 6 μM HgCl₂ but that it decreased by 71% in cells exposed to 8 μM CuCl₂ (data not shown).

Interestingly, 6 μM HgCl₂ almost completely stopped NO₂⁻ production (Fig. 1a) and ammonia-dependent oxygen uptake (Fig. 1b), but both nitrite production and A-SOUR began to rebound within 60 min of further incubation (Fig. 1a and b), suggesting that the cells were able to recover from the Hg²⁺ exposure.

Transcriptomic responses to Cd²⁺.

Using microarrays, we determined the global transcriptional changes of N. europaea exposed to Cd²⁺. The analysis revealed that 1 μM CdCl₂ caused significant changes (greater than twofold) in the transcript levels of 116 genes. The genes included 66 genes with known functions (39 upregulated and 27 downregulated), 21 open reading frames with no known functions (12 upregulated and 9 downregulated), and 29 intergenic regions (15 upregulated and 14 downregulated). The up- or downregulated genes with known function were grouped by functional classification based on the N. europaea genome database (http://genome.ornl.gov) (Tables 1 and 2). Among these, the mercury resistance genes and two genes upstream of that operon were upregulated more than 100-fold. Various genes involved in coenzyme metabolism, translation, DNA replication, and nucleotide transport showed intensities slightly above twofold over the control. The arrays also showed 50 downregulated genes slightly above twofold and included genes that function in signal transduction mechanisms and cell processes. Interestingly, the membrane-bound metallopeptidase encoded by NE2218 showed about 80-fold downregulation in response to Cd²⁺ toxicity exposure, perhaps to prevent cytotoxicity by self-digestion when the cells slow down their metabolism.

TABLE 1.

Selected upregulated genes under cadmium stress^a

Gene function and N. europaea locus tag	Gene name	Description	Fold change	P value (<0.05)
Mercury resistance pathway
NE0838	merD*	Bacterial regulatory protein (MerR family)	107.4	1.9 × 10−5
NE0839	merA*	Mercuric reductase	296.7	9.1 × 10−6
NE0840	merC*	Putative mercury transport protein	171.8	9.2 × 10−5
NE0841	merP*	Mercury scavenger protein	438.6	9.6 × 10−7
NE0842	merT*	Mercuric transport protein	370.3	4.8 × 10−5
Inorganic ion transport mechanism, NE0852	yvgQ	Nitrite and sulfite reductase	2.0	2.1 × 10−3
Efflux pump, NE1640	czcC	Outer membrane efflux protein	2.4	1.4 × 10−3
Oxidative stress, NE1034	trxA	Thioredoxin domain-containing protein	2.1	2.6 × 10−2
ABC transporter, NE1899*		ATPase component ABC-type transport system	2.4	2.5 × 10−3
Coenzyme metabolism
NE0856		Flavin adenine dinucleotide biosynthesis	2.0	4.9 × 10−3
NE0634	cobO	Cobalamine biosynthesis	2.7	2.8 × 10−4
NE0636		Outer membrane cobalamin receptor protein	2.0	2.6 × 10−4
Posttranslational modification, protein turnover, chaperones
NE0221		Organic radical activating enzymes	2.1	3.3 × 10−4
NE2206	ppiD	Peptidyl-prolyl isomerase	2.0	8.0 × 10−3
Cell envelope biosynthesis
NE0378		Sugar transferases involved in lipopolysaccharide synthesis	2.1	4.7 × 10−4
NE2279	yccZ	Periplasmic protein involved in polysaccharide export	2.0	2.6 × 10−4
Translation, ribosomal structure, biogenesis
NE2072	gatA	Amidase:glutamyl-tRNA (Gln), amidotransferase A subunit	2.9	3.0 × 10−4
NE2073	gatB	GatB:glutamyl-tRNA(Gln) amidotransferase, B subunit	2.1	4.7 × 10−4
NE0389	rnpA*	RNase P protein component	2.4	2.6 × 10−4
NE1457		Ribonucleases G and E	2.4	9.2 × 10−5
NE2363	glnS	Glutamyl- and glutaminyl-tRNA synthetases	2.1	3.3 × 10−4
DNA replication, recombination, repair
NE0835	tnpA	Transposase and inactivated derivatives TnpA family	25.5	9.1 × 10−6
NE0836	tnpR	Site-specific recombinase DNA	36.7	3.0 × 10−6
NE0837*		Domain of unknown function 2	157.7	1.4 × 10−5
NE2207	hupB	Bacterial histone-like DNA binding protein	2.1	2.5 × 10−5
Transcription
NE2324	rnc	Double-stranded-RNA-specific RNase	2.4	1.4 × 10−4
NE1035		Transcription termination factor	2.6	2.5 × 10−3
NE0854	cysB	Transcriptional regulator	2.2	5.0 × 10−3
NE0951		Predicted transcriptional regulators (MerR family)	2.0	1.2 × 10−3
Amino acid transport and metabolism
NE1005	argB	Acetylglutamate kinase	2.7	3.4 × 10−4
NE0872	hisD	Histidinol dehydrogenase	2.1	5.0 × 10−3
Nucleotide transport and metabolism, NE0277		XTP pyrophosphatase	2.2	1.4 × 10−3
Carbohydrate transport and metabolism, NE1691		Phosphogluconolactonase/glucosamine-6-phosphate isomerase/deaminase	3.8	6.2 × 10−4
Fatty acid biosynthesis, NE1646		Fatty acid synthesis	2.1	3.1 × 10−4
Signal transduction mechanism, NE0848		Phosphoglycerate mutase family	2.1	3.1 × 10−4
Others
NE1176*		Peptidoglycan binding	2.6	3.0 × 10−4
NE0090		Predicted ATPase	2.0	2.6 × 10−4
NE2325		Transmembrane protein	2.4	5.1 × 10−5
NE2326	lepB	Signal peptidase I	2.0	3.0 × 10−4
NE0218	tolB	Periplasmic component of the Tol biopolymer transport system	2.0	5.3 × 10−4

^aCommonly upregulated genes under Cd²⁺ and Zn²⁺ treatment are indicated by an asterisk.

TABLE 2.

Selected downregulated genes under cadmium stress^a

Gene function and N. europaea locus tag	Gene name	Description	Fold change	P value (<0.05)
Inorganic ion transport mechanism
NE0730		Ferric uptake regulator family	2.6	4.7 × 10−2
NE0731		TonB-dependent receptor protein	2.4	4.4 × 10−2
NE0999		Phosphate transport system permease protein	3.5	2.4 × 10−2
NE1000		ABC-type phosphate transport system permease component	2.9	4.6 × 10−2
NE1001	pstB	Phosphate transport system ATP-binding protein	2.7	2.4 × 10−2
NE1531*		TonB-dependent receptor protein	2.6	2.4 × 10−2
NE0345		Acriflavin resistance protein; heavy metal efflux pump CzcA	5.9	2.4 × 10−2
RubisCO
NE1918	cbbO*	von Willebrand factor type A domain	7.3	2.4 × 10−2
NE1919*		Nitric oxide reductase NorQ protein	5.7	2.4 × 10−2
Cell processes
NE2290		Bacterial type II secretion system protein E; GAF domain	2.0	4.0 × 10−2
NE1298*		Tetratricopeptide repeat	2.2	4.0 × 10−2
NE2315	pilN	Putative type 4 fimbrial biogenesis protein	2.2	2.7 × 10−2
NE2488	flhA	Bacterial export FHIPEP family	2.1	2.4 × 10−2
NE0346		Possible cation transporter transmembrane protein	5.2	2.4 × 10−2
NE2218*		Membrane-bound metallopeptidase	79.78	2.7 × 10−2
NE1538		Chromosome segregation ATPases	2.2	2.4 × 10−2
Transcription				2.4 × 10−2
NE2435	fecI	Specialized sigma subunits of RNA polymerase	2.2	2.4 × 10−2
NE1217		Sigma 70 factor, ECF subfamily	2.4	2.4 × 10−2
NE0533		Sigma 70 factor, ECF subfamily	2.3	2.4 × 10−2
NE1452		Transcriptional regulator	2.1	4.7 × 10−2
NE0787*		Response regulator containing a CheY-like receiver domain and a helix-turn-helix DNA-binding domain	2.1	2.7 × 10−2
Signal transduction mechanism
NE1923	cheY*	Response regulator receiver domain	3.0	2.7 × 10−2
NE0534		Transmembrane sensor	2.2	2.7 × 10−2
Posttranslational modification, protein turnover, chaperone, NE1529		Signal peptide protein	2.0	2.7 × 10−2
Others
NE0315	mnxG*	Possible multicopper oxidase	2.1	4.0 × 10−2
NE2038*		Myeloperoxidase, thyroid peroxidase, cyclooxygenase catalytic domain	2.2	4.4 × 10−2
NE0353	exbB1	MotA TolQ ExbB proton channel family	2.3	2.7 × 10−2
NE1545		Pirin-related protein	3.0	2.4 × 10−2

^aCommonly downregulated genes under Cd²⁺ and Zn²⁺ toxicity are indicated with an asterisk.

To confirm expression changes observed in the microarrays, selected genes were analyzed by qRT-PCR. The relative changes measured by qRT-PCR were consistent and in agreement with the relative changes in the microarrays (Fig. 2).

FIG. 2.

Comparison of mRNA expression levels of selected up- or downregulated genes determined by qRT-PCR (black bars) and microarray (white bars). The positive value represents upregulation, and the negative value represents downregulation. Error bars represent standard errors of the means.

Expression of merA and amoA under Cu²⁺ and Hg²⁺ toxicity.

The merTPCADE cluster was upregulated 297-fold in cells exposed to Cd²⁺ and 46-fold in cells exposed to Zn²⁺ (38). To examine whether merTPCADE respond similarly to other heavy metals, we used qRT-PCR to examine time-dependent transcriptional responses of merA to 6 μM HgCl₂ and 8 μM CuCl₂. The transcript level of merA increased 150-fold in the first 30 min in response to Hg²⁺ and continued to increase up to 250-fold upregulation (Table 3). Consistent with detoxification, during the recovery of nitrification activity, the transcript level of merA decreased to 21-fold upregulation in the span of 3 h (Table 3). In contrast, the transcript level of merA did not show a significant change in response to Cu²⁺ (Table 3) at levels that inhibited nitrification (Fig. 1).

TABLE 3.

Expression change of merA and amoA determined by qRT-PCR under Hg²⁺ and Cu²⁺ treatment

Treatment and gene	Fold change in expression after incubation fora:
	30 min	60 min	120 min	180 min
Hg2+ treatment
merA	157.1 ± 0.22	254.5 ± 0.72	115.7 ± 0.01	21.6 ± 0.26
amoA	1.0 ± 0.02	2.0 ± 0.02	1.9 ± 0.05	3.7 ± 0.12
Cu2+ treatment
merA	1.8 ± 0.50	0.9 ± 0.01	1.4 ± 0.35	1.7 ± 0.12
amoA	2.0 ± 0.07	2.0 ± 0.33	0.5 ± 0.03	0.4 ± 0.02

^aThe error (± value) indicates the 95% confidence interval.

We also examined the time-dependent transcriptional responses of amoA under Cu²⁺ and Hg²⁺ stress (Table 3). The transcript level of amoA increased twofold in Cu²⁺-exposed cells, which may have caused uptake of a small amount of the Cu²⁺ in the first 60 min, but decreased twofold by 180 min. N. europaea exposed 30 min to 6 μM HgCl₂ lost almost all AMO activity and, consequently, NO₂⁻ production (Fig. 1b). However, as the transcript level of merA increased within 1 h, AMO transcript also increased, and nitrite production rates began to rebound.

2D SDS-PAGE.

To determine changes in protein expression, we used 2D SDS-PAGE, using protein extracts from control and Cd²⁺-exposed cells (1 μM CdCl₂) taken after a 3-h incubation. Several proteins showed higher intensity in the Cd²⁺-exposed cells than in control cells. The differentially expressed proteins were excised from the gels for identification by nano-liquid chromatography-tandem mass spectrometry and to deduce the genes that encode them. Identified proteins included nitrite reductase, encoded by NE0924, Rieske iron-sulfur protein, encoded by NE1503, and two hypothetical proteins encoded by NE2057 and NE1752 (Fig. 3).

FIG. 3.

Comparison in 2D SDS-PAGE of the soluble protein fraction of N. europaea treated without (left) and with 1 μM CdCl₂ (right) for 3 h. Circles represent absent or low translation under control conditions. Arrows represent protein spots upregulated under cadmium stress.

DISCUSSION

Physiological responses to Cd²⁺, Hg²⁺, or Cu²⁺.

Our results (Fig. 1) suggest that the inhibitory effects of Cd²⁺ and Hg²⁺ exposure were confined mainly to AMO while apparently causing little or no damage to other elements of the electron transport chain. However, the decrease in H-SOUR with 8 μM CuCl₂ suggests that Cu²⁺-related inhibition was not confined only to AMO. Cu²⁺ has been reported to induce cytotoxicity by producing hydroperoxide and by causing losses of intracellular K⁺, an indication of loss of membrane integrity, in nitrifying autotrophic bacteria (20; also T. S. Radniecki and R. L. Ely, unpublished data). Decreases in both A-SOUR and H-SOUR, as observed in this study, would be consistent with compromised membrane integrity associated with Cu²⁺ exposure.

We showed previously in batch reactors that Zn²⁺ caused 50% nitrification inhibition at a concentration of 3.4 μM (38). Compared to Zn²⁺, Cd²⁺ and Hg²⁺ were more toxic to ammonia oxidizers, but Cu²⁺ was less toxic, with 1 μM Hg²⁺ or 6 μM Cu²⁺ causing 50% inhibition (data not shown) and 1 μM Cd²⁺ causing 90% inhibition. Other studies have shown the inhibition of nitrification by heavy metals in wastewater sludge (27, 33), consistent with our results showing the sensitivity of ammonia oxidizers (e.g., N. europaea) to Cd²⁺, Hg²⁺, or Cu²⁺.

Proteomic responses to Cd²⁺.

In microarray experiments (1-h incubation), transcript levels of the genes NE0924, NE1503, NE2057, and NE1752, corresponding to the overexpressed peptides (3-h incubation) (Fig. 3), did not show detectable increases. Similar discrepancies between transcript and protein levels have been seen in studies with human cells (6), Plasmodium falciparum (29), iron-regulated genes of Vibrio anguillarum (8), and with Escherichia coli (25). Posttranscriptional splicing and posttranslation modifications have been suggested as possible reasons for different responses in transcription and translation (6, 29). In our experiments, microarray data showed upregulation of the genes that encode posttranslational modification proteins, NE0221 and NE2206, consistent with the possibility that posttranslational modifications may have been important in cellular responses to Cd²⁺ toxicity. Overexpression of NirK (nitrite reductase) in N. europaea exposed to Cd²⁺ might support the metal tolerance hypothesis in soil microorganisms (16, 43). It has been reported that NirK is involved in heavy metal tolerance in denitrifying bacteria (16, 43). Therefore, in addition to conferring nitrite tolerance in N. europaea (3), perhaps NirK could play a role in Cd²⁺ tolerance as well.

Expression of merTPCADE under Cd²⁺ toxicity.

Of the 66 upregulated genes detected in the microarrays of cells exposed to 1 μM CdCl₂, the highest upregulated genes (more than 100-fold) were those encoding mercury resistance proteins (merTPCADE). This observation is consistent with higher transcript levels of the merTPCADE operon seen previously in Zn²⁺-exposed cells (38), suggesting that the mercury resistance operon in N. europaea may play an important role in protecting the cells from toxic heavy metals. In addition, the transposase encoded by tnpA (NE0835) and the resolvase encoded by tnpR (NE0836) were upregulated under Cd²⁺ (31-fold) and Hg²⁺ (54-fold) stress but not under Zn²⁺ and Cu²⁺ stress (data not shown). The transposase and resolvase are thought to form a transposase-related protein, known as the mercury resistance transposon (34, 35), that may increase antibiotic resistance in gram-negative facultative bacteria (31). In N. europaea, tnpA and tnpR are located downstream of the operon merTPCADE (Fig. 4). The role of the mercury resistance transposon with respect to Cd²⁺ toxicity in N. europaea remains unclear, as we did not observe increased Cd²⁺ tolerance or adaptation in our experiments. However, we can suggest that the upregulation of merTPCADE and tnpAR may occur in concert when exposed to highly toxic heavy metals (e.g., Cd²⁺ and Hg²⁺) but not when exposed to less toxic metals (e.g., Zn²⁺ and Cu²⁺). We also observed 2.4-fold upregulation of NE1640, which putatively encodes CzcC, the outer membrane protein of the CzcCBA efflux pump thought to be involved in Cd²⁺ detoxification in Ralstonia eutropha and other gram-negative bacteria (37). However, because N. europaea cells did not recover from Cd²⁺ treatment as they did from Hg²⁺ treatment, a role for this gene in the removal of Cd²⁺ seems unlikely. The accumulation of Cd²⁺ in the cytoplasm would eventually become deleterious to the cell and inhibit nitrification completely.

FIG. 4.

Organization of merR, merTPCADE, and transposase-related genes (tnpMRA) in N. europaea. Arrows indicate the genes and their orientation. The numbers in parentheses indicate the relative changes of upregulated merTPCADE and tnpMRA under 1 μM CdCl₂ treatments.

Transcriptional responses of merA and amoA in cells exposed to Hg²⁺ and Cu²⁺.

In cells exposed to Zn²⁺ (38), Cd²⁺, or Cu²⁺, nitrite production and ammonia-dependent oxygen uptake rates did not rebound from heavy metal toxicity within the 3-h time period used in this study even though merTPCAD were upregulated 277-fold in Cd²⁺-exposed cells, perhaps due to severe oxidative stress (7, 12). In contrast, cells exposed to Hg²⁺ did recover, with nitrite production rates rebounding from 0.02 mM NO₂⁻/min·OD₆₀₀ to 0.09 mM NO₂⁻/min·OD₆₀₀ in 90 min (Fig. 1), accompanied by increased expression of merA (Table 3). Recovery of N. europaea from Hg²⁺ toxicity is thought to be related to the mercury resistance genes (merTPCAD), in particular, to merA because the merA gene product can reduce Hg²⁺ to a volatile form (Hg⁰) (2, 17, 37), as follows: Hg²⁺ + NADPH → Hg⁰ + NADP⁺ + H⁺ (36). Hg⁰ then can be volatilized by the cells (36).

Transcript levels of amoA increased in the early stages of Cu²⁺-induced toxicity, but they did not translate into increases in A-SOUR. This indicates that Cu²⁺, although essential for AMO activity (11), can be detrimental even at relatively low concentrations.

Increases in the A-SOUR in cells exposed to Hg²⁺ were reflected in the transcript levels of amoA (Table 3). While it did not change appreciably during the first 30 min of Hg²⁺ exposure, amoA transcription increased fourfold during the time that cells showed recovery. These observations suggest also that monitoring amoA expression can be valuable for tracking cellular responses to toxicity.

Shared up- or downregulated genes in Cd²⁺ and Zn²⁺ treatment.

Several genes that may be related to detoxification of heavy metals showed transcript level changes with Cd²⁺ or Zn²⁺ exposure. NE1176, encoding peptidoglycan binding protein, known to interact with integral outer membrane proteins (13), was upregulated 2.4- to 2.6-fold. A similar observation has been reported with Caulobacter crescentus exposed to Cd²⁺ and Cr²⁺ (19). Other upregulated genes included NE1899, encoding an ATPase component ABC-type metal transporter or an arsenite resistance protein (37), and NE0389, encoding an RNase P protein that processes a precursor of tRNA (18). Genes that were downregulated included NE2218, encoding a membrane-bound metallopeptidase (80-fold under Cd²⁺ and 8-fold under Zn²⁺ treatments), which requires a small amount of a transition metal such as Zn²⁺ or Co²⁺ in its active site (26). The genes cbbQO (NE1919/1918), encoding ribulose bisphosphate carboxylase/oxygenase (RubisCO), were downregulated in both Zn²⁺ and Cd²⁺ treatments. Carbon sequestration by RubisCO is an energy-intensive process (47). Under starvation conditions in N. europaea (48) and under Cu²⁺ stress in Nitrosococcus mobilis (41), genes encoding RubisCO were downregulated, possibly to conserve energy for NH₃⁺ metabolism.

Heavy metals cause oxidative stress in gram-negative bacteria (37), and in this work, the microarrays detected changes in transcript levels of genes encoding proteins associated with oxidative stress. Myeloperoxidase, encoded by NE2038, is thought to be a reactive oxygen species-generating enzyme (24) and was downregulated, perhaps to reduce oxidative stress under heavy metal toxicity. NE0315 (mnxG), encoding a multicopper oxidase known to catalyze Mn²⁺ oxidation, was downregulated under the Zn²⁺ and Cd²⁺ treatments, again perhaps to reduce oxidative stress (9). Genes that were also downregulated under both Zn²⁺ and Cd²⁺ stress include NE1298, encoding tetratricopeptide repeat, involved in protein-protein interactions (32); NE1531, encoding a TonB-dependent protein; and NE0787, encoding CheY protein. NE1923, encoding CheY (a flagellar protein) in N. europaea, was previously seen to be downregulated in response to NO, as NO promoted the formation of biofilm, and mobility was no longer necessary (42). It could be that the expression of CheY decreases during metal stress to promote the formation of cell agglomerates to protect some of the cells from further exposure.

Candidate genes to detect Cd²⁺ stress.

Of the N. europaea genes that were up- or downregulated under Cd²⁺ treatment, some potentially may serve as specific indicators of Cd²⁺ exposure. Because Cd²⁺ causes oxidative stress by producing reactive oxygen species that deplete glutathione and protein-bound sulfhydryl groups (37), the upregulation of NE1034, encoding thioredoxin (disulfide reductase) (19), could help the cell to resist oxidative stresses. NE1005 (argB) and NE0872 (hisD), encoding an amino acid transport, were upregulated under Cd²⁺. Biosynthesis of amino acids would use more energy than their uptake through an amino acid transport; therefore, these genes might be upregulated to conserve energy in a toxic environment (45). NE0221, encoding an organic radical-activating enzyme, and NE2206 (ppiD), encoding a peptidyl-prolyl isomerase that belongs to a posttranslational modification protein, showed upregulation under Cd²⁺ treatment (ppiD was also upregulated under chloroform treatment) (15). NE2324 (rnc), NE1035, NE0854 (cysB), and NE0951, which relates to transcription, showed upregulation, while NE2435 (fecI), NE1217, NE0533, NE1452, and NE0787 showed downregulation (Table 2). Transcript of NE1217, belonging to the ECF family, was also increased in response to chloroform toxicity (15). NE0835, NE0836, NE0837, and NE2207 (hupB), involved in DNA replication, recombination, and repair, showed upregulation. NE0378 and NE2279 (yccZ), involved in cell envelope biosynthesis, were upregulated, while six genes in the groups were downregulated to statistically significant degrees (Table 2). Although the role(s) of these genes in heavy metal stress was not clear, they could be important for finding heavy metal toxicity mechanisms or general stress genes in the future.