Search tips
Search criteria 


Logo of jbacterPermissionsJournals.ASM.orgJournalJB ArticleJournal InfoAuthorsReviewers
J Bacteriol. 2001 August; 183(16): 4866–4875.

Oxidation of Phenolate Siderophores by the Multicopper Oxidase Encoded by the Escherichia coli yacK Gene


A gene (yacK) encoding a putative multicopper oxidase (MCO) was cloned from Escherichia coli, and the expressed enzyme was demonstrated to exhibit phenoloxidase and ferroxidase activities. The purified protein contained six copper atoms per polypeptide chain and displayed optical and electron paramagnetic resonance (EPR) spectra consistent with the presence of type 1, type 2, and type 3 copper centers. The strong optical A610610 = 10,890 M−1 cm−1) and copper stoichiometry were taken as evidence that, similar to ceruloplasmin, the enzyme likely contains multiple type 1 copper centers. The addition of copper led to immediate and reversible changes in the optical and EPR spectra of the protein, as well as decreased thermal stability of the enzyme. Copper addition also stimulated both the phenoloxidase and ferroxidase activities of the enzyme, but the other metals tested had no effect. In the presence of added copper, the enzyme displayed significant activity against two of the phenolate siderophores utilized by E. coli for iron uptake, 2,3-dihydroxybenzoate and enterobactin, as well as 3-hydroxyanthranilate, an iron siderophore utilized by Saccharomyces cerevisiae. Oxidation of enterobactin produced a colored precipitate suggestive of the polymerization reactions that characterize microbial melanization processes. As oxidation should render the phenolate siderophores incapable of binding iron, yacK MCO activity could influence levels of free iron in the periplasm in response to copper concentration. This mechanism may explain, in part, how yacK MCO moderates the sensitivity of E. coli to copper.

Multicopper oxidases (MCOs), a diverse family of metalloenzymes widely distributed among eukaryotes, are characterized by distinctive structural, spectroscopic, and enzymatic properties (45). One of the best-known member of this class of enzymes, laccase, also known as p-diphenol:O2 oxidoreductase (EC, was among the first enzymes recognized to require metal for activity (7, 29). Being the simplest enzyme that combines all three known organic Cu(II) magnetic types in a single molecule, laccase has been particularly well studied with respect to its intramolecular electron transfer reactions (44). Other well-known MCOs include ascorbate oxidase (EC, cytochrome c oxidase (EC, and ceruloplasmin (sometimes referred to as ferroxidase; EC However, in recent years, the MCO family has grown rapidly through the addition of new enzymes, such as phenoxazinone synthase (22), bilirubin oxidase (42), and dihydrogeodin oxidase (27). Most of these new MCOs are of fungal origin, and in these organisms they often catalyze oxidative steps in the biosynthesis of secondary metabolites, including antibiotics of commercial interest. A notable recent addition to the MCO family, however, is the FET3 protein of Saccharomyces cerevisiae, which acts as a ferroxidase and is a critical component of a high-efficiency iron uptake system (3, 16, 47). Studies of the FET3 system have had a profound impact on our understanding of iron metabolism across the spectrum of eukaryotes, clarifying some of the physiological roles played by other MCOs, such as ceruloplasmin, hephaestin, and cartilage matrix glycoprotein (2, 4, 21, 25, 49).

In contrast to the abundance of well-characterized MCOs in eukaryotes, the first, and for a long time the only, report of a potential MCO in a prokaryotic system was that describing an ascorbate oxidase activity in a poorly characterized isolate of Aerobacter aerogenes (48). Unfortunately, no evidence was presented in that report to link the membrane-bound activity to a copper-containing enzyme and no further studies were reported. A laccase-like phenoloxidase activity was found in a melanizing isolate of Azospirillum lipoferum (23), but characteristics recently reported for the purified protein are not consistent with its identification as an MCO (17). Proteins that are structurally homologous to MCOs with respect to the canonical copper-binding sites, such as the CopA protein from Pseudomonas syringae (35) or the PcoA protein from Escherichia coli (13), have been shown to be important for bacterial copper resistance, but there have been no reports as to whether or not these proteins possess oxidase activity. In contrast, transposon mutagenesis has been used to identify bacterial genes encoding MCOs responsible for Mn2+ oxidation in Pseudomonas putida (12) and melanization (phenol oxidation) in a marine bacterium, Marinomonas mediterranea (43). Furthermore, as noted by Alexandre and Zhulin (1), microbial genomes contain numerous genes that could encode MCOs, suggesting that these enzymes may play important roles in bacterial metabolism.

For this study, we cloned and expressed the putative MCO encoded by the yacK gene of E. coli. The expressed enzyme contained the predicted copper centers, harbored both phenoloxidase and ferroxidase activities, and demonstrated changes in its physical and enzymatic characteristics upon addition of exogenous copper. This study presents the first evidence of a physiological rationale for the maintenance of oxidative activity against both metal ions and phenolic compounds by MCOs.



Synthetic enterobactin was provided by Carlos Gutierrez (41). All of the other compounds and reagents used in these studies were of American Chemical Society reagent grade or better and were used without further purification.

Cloning of the yacK gene.

E. coli genomic DNA was isolated from strain C600 by standard procedures (5). Oligonucleotide primers were synthesized to match the 5′ and 3′ ends of the E. coli yacK coding sequence as it appears in the genomic sequence (8). The 5′ primer (5′-GGAATTCAGGAAATAACTATGCAACGTCG-3′) contained an additional EcoRI recognition site at the 5′ terminus and spanned both the apparent Shine-Dalgarno site and the initiating ATG codon. The 3′ primer (5′-GGATCCGAATACGGTCTTTTTATACCG-3′) contained a terminal BamHI recognition site and spanned the apparent termination codon. PCRs contained 10 ng of genomic DNA as template, and the predicted 1.58-kbp product was amplified by using standard reaction conditions. The amplimer was cloned into pCR 2.1-TOPO (Invitrogen) to generate plasmid pWLFO-4, and the sequence identity of the insert was verified by using dye terminator sequencing on an Applied Biosystems ABI 377XL automated DNA sequencer. Restriction digests were performed on pWLFO-4 using EcoRI and BamHI, and the insert DNA was purified by using a QiaQuick gel extraction kit (Qiagen, Valencia, Calif.). The purified fragment was subcloned into the isopropyl-β-d-thiogalactopyranoside (IPTG)-inducible expression plasmid pKEN-2 (20) to yield pWLFO-5.

Expression and purification of yacK MCO.

E. coli strain TG-1 [supE hsdΔ5 thiΔ (lac-proAB) F′ (traD36 proAB+ lacIq lacZΔM15)] containing pWLFO-5 was grown aerobically on Luria-Bertani (LB) medium at 37°C in a 250-liter fermentor containing ampicillin (100 μg/liter) and CuSO4 (1 mM), and the cells were harvested at an optical density of 1.8 (mid- to late-log phase). Approximately 2 g (wet weight) of cells was obtained per liter of culture. Cells resuspended in 2 volumes (wt/vol) of 50 mM Tris-HCl buffer (pH 8)–50 mM NaCl–1 mM CuSO4 were lysed by sonication after incubation at 37°C for 30 min with lysozyme (1 mg/g of cells) and DNase (50 μg/g of cells). Cell extract was obtained by centrifugation for 1 h at 10,000 × g in an SLA 3000 rotor (Ivan Sorvall, Inc., Norwalk, Conn.). The crude extract was heated in a water bath for 5 min at 70°C, cooled, and centrifuged at 10,000 × g for 30 min. To the supernatant, dry ammonium sulfate was slowly added to 35% saturation at 25°C and allowed to equilibrate for 1 h. After centrifugation at 10,000 × g for 30 min, the phenoloxidase activity remained in the supernatant, which was subsequently dialyzed extensively against 50 mM Tris-HCl buffer, pH 8. The dialyzed enzyme was loaded at 5 ml/min onto a Q-Spherilose column (5 by 12 cm; ISCO, Lincoln, Nebr.) equilibrated with 50 mM Tris buffer, pH 8. After washing with 2 column volumes of equilibration buffer, the enzyme was eluted from the column with a 1,400-ml linear gradient of NaCl (0 to 0.7 M) in equilibration buffer. Fractions containing phenoloxidase activity, eluting near 0.1 M NaCl, were pooled and dialyzed against 50 mM Na-acetate buffer, pH 5. Dialyzed enzyme was loaded at 2 ml/min onto an SP-Spherilose column (2 by 15 cm; ISCO) equilibrated with acetate buffer. The column was washed with 2 column volumes of the same buffer, and the enzyme was eluted by using a 400-ml linear gradient of NaCl (0 to 1.0 M) in acetate buffer. The fractions containing enzyme, which eluted at ca. 0.2 M NaCl, were pooled and brought to 0.8 M with solid (NH4)2SO4. The enzyme was loaded at 3 ml/min onto a phenyl-Spherilose column (2 by 18 cm; ISCO) that was equilibrated with 50 mM Tris-HCl, pH 8, containing 0.8 M (NH4)2SO4. Enzyme was eluted with a 600-ml linear gradient of (NH4)2SO4 (0.8 to 0 M). Active fractions [ca. 0.4 M (NH4)2SO4] were dialyzed extensively against 50 mM Tris-HCl, pH 5, and then concentrated in a stirred-cell ultrafiltration device against a 30-kDa molecular mass cutoff membrane (PM 30; Amicon, Beverly Mass.).

Protein elution from the chromatography columns was monitored by measuring the A280, while elution of yacK MCO was specifically detected by monitoring of the A610 due to the type 1 (blue) copper. Specific enzyme activity and mobility during sodium dodecyl sulfate (SDS)-polyacrylamide gel electrophoresis (PAGE) were used to evaluate the purity of yacK MCO at each step of the purification procedure.

Enzyme assays and kinetics.

The phenoloxidase activity of yacK MCO was measured by using 2,2′-azinobis(3-ethylbenzthiazolinesulfonic acid) (ABTS) as the electron donor. The assay mixture (1 ml) contained 50 mM Na-acetate buffer, pH 5, and 0.3 to 1.5 μg of the enzyme. After preincubation for 5 min at 30°C, the reaction was started by addition of ABTS to 3 mM and its oxidation was monitored by measuring the increase in absorbance (Ε420 = 36 mM−1 cm−1). In studies where the activity was also measured in the presence of added copper, the enzyme was first incubated for 1 min in an appropriate concentration of CuSO4, followed by addition of the Na-acetate reaction buffer containing 3 mM ABTS. Control reactions were run by using heat-denatured enzyme. Specific activities are expressed as units of activity per milligram of protein, where 1 activity unit represents 1 μmol of ABTS oxidized per min.

Ferroxidase activity was determined by using two different assays. For the greatest sensitivity, assays were routinely performed at 25°C in a microtiter plate format by using ferrous sulfate as the electron donor and 3-(2-pyridyl)-5,6-bis(4-phenylsulfonic acid)-1,2,4-triazine (ferrozine) as a chelator to specifically detect the ferrous iron remaining at the end of the reaction. Each assay mixture (0.2 ml) contained 50 mM Na-acetate buffer, pH 5, and 0.1 to 0.3 μg of the enzyme, and the reaction was started by addition of FeSO4 to 0.2 mM. Samples were quenched at time intervals by adding ferrozine to 3.75 mM, and the rate of Fe(II) oxidation was determined by measuring the absorbance of residual Fe(II)-ferrozine (Ε570 = 7.26 mM−1 cm−1). To ensure that interactions between ferrozine and exogenous copper were not the source of the apparent increase in ferroxidase activity in response to copper addition, an independent assay measuring the increased optical absorbance (Ε315 = 2.20 mM−1 cm−1) due to the production of Fe3+ from Fe2+ was used to confirm the results of the ferrozine assay. When assays were performed in the presence of added copper, the enzyme was incubated with an appropriate amount of CuSO4 and the reaction was started by addition of Na-acetate buffer containing 0.2 mM FeSO4. Control reactions and specific activities were as described for the phenoloxidase reactions.

Assays to determine the kinetic constants for ABTS, p-phenylenediamine (p-PD), 2,6-dimethoxy phenol (2,6-DMP), syringaldazine, or FeSO4 [Fe(II)] were all performed at 40°C in 50 mM Na-acetate buffer, pH 5, containing 1 mM CuSO4. The molar absorption coefficients used for these substrates were as follows: ABTS, Ε420 = 36,000 M−1 cm−1; p-PD, Ε487 = 14,685 M−1 cm−1; 2,6-DMP, Ε477 = 14,800 M−1 cm−1; syringaldazine, Ε525 = 65,000 M−1 cm−1; Fe(II), Ε315 = 2,200 M−1 cm−1. The rates of oxidation of 2,3-dihydroxybenzoic acid (2,3-DHB), 3,4-dihydroxybenzoic acid (3,4-DHB), and enterobactin [cyclic trimer of N-(2,3-dihydroxybenzoyl)-l-serine] to the corresponding quinoline compounds were measured by monitoring the A400. The oxidation of 3-hydroxyanthranilic acid (3-HAA) was measured by determining the A445. The aqueous molar absorption coefficients of the oxidation products of 2,3-DHB, 3,4-DHB, enterobactin, and 3-HAA were determined by measuring absorbance after the substrates were completely oxidized by yacK MCO. The values were as follows: 2,3-DHB, Ε400 = 2,328 M−1 cm−1; 3,4-DHB, Ε400 = 2,328 M−1 cm−1; enterobactin, Ε400 = 12,248 M−1 cm−1; 3-HAA, Ε445 = 3,790 M−1 cm−1.

Protein and copper content determinations.

The concentration of protein in enzyme pools was routinely determined by the method of Bradford using bovine serum albumin as the standard (10). The amount of yacK MCO in the pool recovered after the final purification step was determined by video densitometry using a ChemImager 4000 (AlphaInnotech, San Leandro, Calif.) to quantitate the polypeptide resolved by SDS-PAGE. The yacK MCO and bovine serum albumin standards were stained using Sypro Orange fluorescent protein dye (Molecular Probes, Eugene, Oreg.) and quantitated by using excitation and emission wavelengths of 485 and 590 nm, respectively. Copper content was determined by inductively coupled plasma mass spectrometry using a PlasmaQuad 3 spectrometer (MicroMass, Beverly, Mass.) operated by the University of Georgia Research Services Chemical Analysis Laboratory.

SDS-PAGE and zymograms.

SDS-PAGE was performed as previously described (30). Ferroxidase zymograms were performed essentially as described by Yuan et al. (51). Briefly, samples from each stage of purification (0.5 to 20 μg of protein) were mixed with SDS-PAGE sample buffer lacking β-mercaptoethanol and subjected to electrophoresis on a 10% polyacrylamide gel without prior heat denaturation. The gel was then incubated in a solution of 10% glycerol containing 0.05% Triton X-100 to remove SDS and stabilize the enzyme. Ferroxidase activity was detected by soaking the gel in 100 mM Na-acetate (pH 5)–0.2 mM FeSO4 for 1 h, after which the gel was removed to moistened filter paper. Ferrozine (15 mM) was dropped onto the gel and allowed to react for 10 to 15 s before being gently rinsed off. The gel was then incubated under humid conditions for 3 to 4 h until the cleared bands were fully visible. Phenoloxidase activity was detected by soaking the glycerol-stabilized gel in 50 mM Na-acetate buffer, pH 5, containing 2 mM 1,8-diaminonaphthalene for 5 min and incubating the gel under humid conditions until the bands of oxidized substrate were fully visible (26). The reaction was stopped by soaking the gel in 10% trichloroacetic acid.


UV-visible light spectra of the enzyme were collected by using a diode array spectrophotometer (model 8452A; Hewlett-Packard, Co., Palo Alto, Calif.). X-band (~9.6 GHz) electron paramagnetic resonance (EPR) spectra were recorded with a Bruker (Billerica, Mass.) ESP 300E spectrometer equipped with a dual-mode ER-4116 cavity and an Oxford Instruments (Oxford, England) ESR-9 flow cryostat (4.2 to 300K). Frequencies were measured with a BEI Systron-Donner (Concord, Calif.) 6054B frequency counter, and the magnetic field was calibrated with a Bruker ER 035 M gaussmeter.

Determination of thermal stability.

Enzyme stability was determined by measuring residual phenoloxidase activity after incubation in 50 mM Tris-HCl, pH 8.0, at each test temperature. At time intervals, aliquots were removed, cooled on ice, and assayed immediately for phenoloxidase activity in the presence of 1 mM CuSO4 as previously described.

Computer modeling.

By using default parameters for the GAP alignment function of the GCG Sequence Analysis Software Package, version 3.2 (Pharmacopeia, Inc., Madison, Wis.), the yacK protein sequence, less the signal peptide, was aligned with that of the laccase from Coprinus cinereus, for which a crystal structure is available (18). This alignment was subsequently submitted for online processing by SWISS-MODEL via the ExPASy Proteomics Server at the Swiss Institute of Bioinformatics ( Energy minimization of the resultant structural model of yacK MCO was performed with SYBYL 6.6 (Tripos Inc., St. Louis, Mo.) by the Powell method in conjunction with Tripos Force Field parameters and the Kollman all-atom charge set. The gradient was set to 0.025 kcal mol−1−1 and carried through 200 iterations. All other parameters were left at the default settings. Four copper atoms corresponding to those in the laccase crystal structure were added manually after energy minimization by aligning the model with the template structure and editing the .pdb file directly.


Identification of an E. coli MCO. A BLAST search using the yeast FET3 sequence to identify potential MCOs in the E. coli K-12 genome (8) identified yacK, an open reading frame at about 3 min on the chromosome map. The GenBank entry noted the open reading frame as having an unknown function but strong homology to eukaryotic MCOs. Homologous genes containing the four canonical copper-binding domains that characterize MCOs were also identified in BLAST searches of the genomes of several other bacterial species ( Figure Figure11 shows an amino acid alignment of the four predicted copper-binding domains from the E. coli yacK gene and genes from a selection of other bacteria. Analysis of the coding sequence using the SignalP web server (36) suggested that the protein should be secreted with cleavage of a 28-amino-acid leader sequence. That E. coli secretes the yacK gene product to the periplasm and cleaves the protein at the predicted site was previously demonstrated by Link et al. (34). The signal peptide contains the amino acid motif for secretion via the twin-arginine translocation pathway (6), and Stanley et al. (46) have shown that the yacK-encoded protein is, in fact, processed by this system.

FIG. 1
Amino acid sequence alignments for the four copper-binding domains from a selection of bacterial MCOs. The numbers across the top refer to the sequence positions in the E. coli protein. Boldface lettering indicates amino acids that act as copper ligands ...

Expression and purification of yacK MCO.

Phenoloxidase activity was undetectable in crude extracts of untransformed E. coli grown in LB medium through late log-phase and only barely detectable in stationary-phase extracts. However, this activity was readily detectable in crude extracts of log-phase cells that had been transformed with an expression vector carrying the yacK gene. Although the gene was placed under control of the sterically repressed lacZ promoter (srp) of pKEN2 (20), which is supposed to significantly minimize leaky expression, phenoloxidase activity conferred by this construct remained essentially constitutive in a variety of host backgrounds whether or not IPTG was present. Similar constitutive expression was seen when the yacK gene was placed in other inducible vectors and may be a result of the extensive secondary structure predicted for the coding sequence by the DNA mfold server (courtesy of Michael Zucker;

Total phenoloxidase activity was greatest when transformed cells were grown in LB medium supplemented with 1 mM CuSO4, suggesting that copper can become limiting when this enzyme is overexpressed. Limiting copper levels in growth media have been shown to have a similar effect on the specific activity of MCOs in several eukaryotic systems (9, 40, 50).

yacK MCO was fully soluble in crude extracts, as >95% of the phenoloxidase activity detectable by SDS-PAGE remained in the supernatant after centrifugation of the cell lysate. Quantitation was highly variable for either enzyme activity (phenoloxidase or ferroxidase) in samples taken prior to the ammonium sulfate precipitation step, presumably due to the presence of interfering compounds, likely reductants, in the crude extracts. This conclusion was supported by the observation that crude extracts, which were distinctly blue when first prepared, turned yellow-brown upon standing. The blueness quickly reappeared when the extracts were shaken, suggesting that the color change was due to reduction and subsequent reoxidation of the type 1 “blue” copper centers within the enzyme. The enzyme preparation resulting from the complete purification procedure retained a fully oxidized blue color indefinitely. Table Table11 shows the complete course of purification for the MCO phenoloxidase activity, but because of the presence of interfering compounds, the activity values for the first two enzyme pools (crude extract and heat-treated fraction) are of marginal use for comparison with the enzyme recovered in subsequent steps. The yacK phenoloxidase and ferroxidase activities copurified with final yields and purification factors of 25% and 2.4-fold for phenoloxidase and 27% and 2.9-fold for ferroxidase, based on the ammonium sulfate pool activities as a starting point.

Purification of E. coli yacK MCO (phenoloxidase activity)

Zymogram analyses.

The protein composition of each pool resulting from yacK MCO purification was assessed by SDS-PAGE analysis of heat-denatured (Fig. (Fig.2A)2A) and nondenatured (Fig. (Fig.2B)2B) samples. The purified protein migrated as a single band with a molecular mass of ~54 kDa, in close agreement with the size (53.4 kDa) calculated for the mature protein. Phenoloxidase (Fig. (Fig.2C)2C) and ferroxidase (Fig. (Fig.2D)2D) activities comigrated with the predominant protein band noted when an identically loaded gel was stained for protein (Fig. (Fig.2B).2B). A minor phenoloxidase band with mobility faster than that of yacK MCO was detected in crude extracts when zymograms were incubated overnight (data not shown), but as the protein samples were not heat denatured prior to loading onto the gel, it is not possible to say whether this band represented a different enzyme or was just a conformer of yacK MCO. The band was not detected in extracts from untransformed cells. In any case, no other ferroxidase activities were noted in the zymogram assays.

FIG. 2
SDS-PAGE analysis of yacK MCO at various stages of purification. Protein samples were either heat denatured (A) or not heated (B to D) prior to electrophoresis. Panels A and B show gels that were stained with Coomassie dye to detect protein, while the ...

It should also be noted that with the zymograms, a low level of phenoloxidase activity from yacK MCO could be detected in crude extracts from untransformed cells grown in LB medium containing added copper but not from cells grown in medium without added copper (data not shown). This would be expected from the work of Outten et al. (39) demonstrating that the endogenous yacK gene is induced by copper. In untransformed cells, ferroxidase activity remained below detectable levels, even with copper supplementation.

Enzyme copper content.

The yacK MCO purification protocol was developed by monitoring the enzyme activity in conjunction with denaturing SDS-PAGE. However, notable in the protein-stained gel containing samples that had not been heat treated (Fig. (Fig.2B)2B) was a cluster of contaminating protein bands at ~62 kDa. These bands were absent from gels in which the protein samples were heat denatured prior to electrophoresis, suggesting that the bands represented a protein complex. Because of the contaminating protein that remained after the hydrophobic-interaction chromatography step, video densitometry was used to specifically quantitate the yacK-encoded protein resolved by SDS-PAGE (no heat denaturation step) and stained with a fluorescent dye. Based on this quantitation for the MCO polypeptide and the results from inductively coupled plasma mass spectrometry, the copper content of the enzyme was calculated to be 6.2 Cu atoms per polypeptide chain. The densitometric protein quantitation was also used to calculate extinction coefficients of Ε610 = 10,890 M−1 cm−1 and Ε330 = 6,980 M−1 cm−1 for the type 1 and type 3 copper centers, respectively. A similar calculation yielded an extinction coefficient of Ε280 = 66,500 M−1 cm−1 for the protein. As accurate protein quantitation is generally the most problematic part of calculating metal stoichiometry in metalloenzymes, it should be noted that the copper content was calculated to be only 4.5 Cu atoms per polypeptide chain when quantitative densitometry was used to analyze protein subjected to heat-denatured SDS-PAGE or when the protein concentration in the final pool was estimated by the Bradford assay.

Substrate kinetics and effects of added metals.

yacK MCO oxidized a variety of phenolic substrates, as well as ferrous iron. The enzyme displayed roughly equivalent kinetic parameters for its reactions with ABTS, p-PD, and 2,6-DMP (Table (Table2).2). However, the Km and Vmax values for syringaldazine were approximately 2 orders of magnitude lower than those for the other phenolic substrates. With respect to Fe(II) oxidation, the enzyme's Km value for the substrate (70 μM) was in the same range as that measured for syringaldazine, while the Vmax (41 U mg−1) was closer to that seen for the other phenolic substrates.

Substrate specificity and kinetic parameters of yacK MCOa

A variety of metals—Co(II), Cu(I), Cu(II), Fe(II), Fe(III), Mg(II), MoO4, Mn(II), Ni(II), and Zn(II)—were examined for their effects on yacK MCO phenoloxidase activity. Whereas Fe(II) reduced the apparent phenoloxidase activity, ostensibly due to competition with the phenolic substrate, Cu(II) enhanced this activity six- to sevenfold. Copper addition also led to comparable increases in apparent phenoloxidase activity when p-PD, 2,6-DMP, and syringaldazine were used as substrates (data not shown). This activity enhancement by copper was fully reversible and could be detected over a range of 5 orders of magnitude (0.1 μM to 10 mM) of CuSO4 concentrations (Fig. (Fig.3).3). There was no uptake of oxygen when either Cu(I) or Cu(II) was added to the enzyme in the absence of Fe(II) or a phenolic substrate (data not shown). Thus, copper enhancement of enzyme activity does not come about through a redox cycle involving free copper. None of the other metals had any effect on the reaction.

FIG. 3
Copper activation of yacK MCO. The specific phenoloxidase activity of the enzyme is shown versus the concentration of copper added to the reaction mixture. The concentration of the enzyme in the assay was 0.006 μM; therefore, a 1 μM [CuSO ...

yacK MCO ferroxidase activity was also stimulated upon inclusion of copper in the appropriate assay mixtures (Table (Table3).3). In fact, the effect of copper was more dramatic with respect to ferroxidase activity, which increased nearly 400-fold, than to phenoloxidase activity upon the addition of 1 mM CuSO4 (1.6 × 105 molar excess of Cu2+ over the MCO polypeptide). Control reaction mixtures containing either no enzyme or the heat-denatured enzyme showed neither ferroxidase nor phenoloxidase activity, with or without the addition of copper. In addition to these controls, two different iron oxidation assays were used in order to ensure that the addition of copper did not in some way interfere with the assay, i.e., through the formation of a colored Cu2+-ferrozine chelate. Ferroxidase activities in the presence of 1 mM CuSO4 were 19.7 and 18.6 U/mg, as determined by the ferrozine and direct absorbance assays, respectively, compared to less than 0.1 U/mg when either assay was conducted in the absence of exogenous copper.

Effect of copper on substrate oxidation by yacK MCO

Spectral properties.

The absorbance (Fig. (Fig.4)4) and EPR spectra (Fig. (Fig.5)5) of yacK MCO exhibited features typical of three types of copper centers that exist in MCOs. The protein displayed a strong A610 and the characteristic EPR spectra with a narrow hyperfine splitting (g[perpendicular] = 2.05, g|| = 2.24, A|| = 66 G) due to the presence of type 1 copper. An EPR signal characteristic of type 2 copper (g|| = 2.26, A|| = 152 G) was also apparent. Although the binuclear type 3 copper pair does not yield a detectable EPR signal, the protein did have a strong absorbance at 330 nm, as would be expected from a type 3 center.

FIG. 4
Optical absorbance spectra of yacK MCO. The lines represent yacK MCO in 50 mM Tris buffer, pH 8 (solid line); yacK MCO in 50 mM Tris buffer, pH 8, containing 1 mM CuSO4 (dotted line); and 50 mM Tris buffer, pH 8, containing 1 mM CuSO4 with no enzyme (dashed ...
FIG. 5
EPR spectra of yacK MCO. Line A shows the spectrum recorded for yacK MCO as isolated, while line B shows the spectrum of yacK MCO in buffer containing 1 mM CuSO4. Absorbance bands from the type 1 (I) and type 2 (II) copper centers are indicated. The inset ...

Addition of 1 mM Cu2+ to the enzyme elicited changes in both absorbance and EPR spectra. First, not only was the A330 increased, but light absorbance was also shifted slightly toward longer wavelengths with no significant change in the A610 (Fig. (Fig.4).4). With respect to the EPR spectrum, additional signals at 2,750 and 2,890 G appeared when copper was added (Fig. (Fig.5B).5B). Although the optical absorbance maximum for the type 3 copper was slightly shifted, neither the line shape nor the intensity of the EPR signals for type 1 or type 2 copper centers was changed by the addition of copper.

Thermal stability.

yacK MCO was surprisingly stable at temperatures of 70°C or lower, with a half-life at 70°C of >5 h (Fig. (Fig.6A).6A). Furthermore, the activity actually increased 33% after incubation at 60°C for 2 h. However, when subjected to the same temperature conditions in the presence of 1 mM CuSO4, the enzyme was significantly less stable, with a half-life at 70°C of ~40 min (Fig. (Fig.6B).6B). Whether the principal mode for this shift in enzyme stability derives from a conformational change after copper binding to the protein or from the consequent enhancement of oxidative activity is unknown.

FIG. 6
Thermal stability of yacK MCO. The enzyme was incubated at the desired temperatures in the absence (A) or presence (B) of 1 mM CuSO4. Residual phenoloxidase activity was subsequently determined in the presence of 1 mM CuSO4. The incubation temperature ...

Oxidation of iron siderophores.

In addition to oxidizing several synthetic phenoloxidase substrates, yacK MCO oxidized phenolate siderophores used by E. coli for iron uptake, namely, enterobactin and 2,3-DHB. The enzyme also oxidized 3,4-DHB and 3-HAA, the latter of which is a metabolite that can function as an iron siderophore for yeast (31). Figure Figure77 shows time courses for oxidation of these compounds by yacK MCO. Reaction rates were determined from the absorbance increase recorded during the first 2 min, after which linearity was lost due to precipitation of the oxidized products. Oxidation of enterobactin by yacK MCO led to rapid precipitation of a flocculate that was distinctly gray-green. At a substrate concentration of 0.18 mM, the specific activities of the enzyme for enterobactin and 2,3-DHB were 7.50 and 1.42 U/mg, respectively. The enzyme had a much greater affinity (low Km) for these compounds than for nearly every synthetic substrate tested. Siderophores were oxidized by the enzyme at the same rate whether or not iron was present in the reaction mixture, but none of the compounds were oxidized when the heat-denatured enzyme was used in the assays.

FIG. 7
Time course of the UV-visible light spectra during the oxidation of 2,3-DHB (A), enterobactin (B), and 3-HAA (C) by yacK MCO. The 1-ml reaction volume contained 50 mM Na-acetate buffer (pH 5), 1 mM CuSO4, and 5 (enterobactin and 3-HAA) or 15 (2,3-DHB) ...


Bacterial proteins resembling eukaryotic MCOs, including those encoded by the copA gene in P. syringae (35) and the pcoA gene in E. coli (13), were identified more than a decade ago as important components of bacterial copper resistance systems. However, a mechanism for their protective action has never been clearly identified, although there has been a general presumption that the proteins might act as copper sinks that bind and sequester excess copper. Results presented here demonstrate that the yacK MCO from E. coli contains copper coordinated in the type 1, type 2, and type 3 centers anticipated on the basis of sequence analysis. Evidence suggests that the enzyme has a copper content of six copper atoms per polypeptide chain, with three of the copper atoms bound in type 1 centers. Although this would be a novel structure for a microbial MCO, a mammalian MCO, ceruloplasmin, contains six covalently bound copper atoms, three of which are in type 1 centers (33). Molecular modeling of yacK MCO suggests that most of the 14 histidine residues not involved in binding of the four copper atoms predicted from the laccase crystal structure are clustered near the trinuclear center, which contains the type 2 and type 3 copper sites (Fig. (Fig.8).8). If these histidine residues participate in the formation of additional type 1 sites, they appear to be localized where the liganded copper atoms could participate in redox reactions catalyzed by the enzyme. However, the model shows no cysteine residues in this region of the protein. As cysteines generally provide the third ligand (in addition to two histidines) that characterizes type 1 copper sites, it is not obvious where additional type 1 centers might form in the enzyme. No doubt X-ray crystallography will clarify the location of copper atoms bound in this enzyme.

FIG. 8
Computer model showing relative positions of histidine residues not involved in the binding of the four copper atoms corresponding to the copper atoms in the fungal laccase crystal structure. Clustering of these residues near the known copper-binding ...

Noting a potential copper response element in the promoter region of yacK, Outten et al. (39) demonstrated that the gene was induced by copper and proposed that it be renamed cueO for Cu efflux oxidase by virtue of its apparent homology with other MCOs. However, no evidence was presented to demonstrate an oxidase activity associated with the gene product. Disruption of the yacK gene made E. coli slightly more sensitive to copper, from which Grass and Rensing (24) hypothesized that yacK MCO might function to oxidize Cu(I) to Cu(II), thereby preventing uptake of excess copper via the system encoded by cusCBA. We were unable to detect oxygen uptake by yacK MCO when either Cu(I) or Cu(II) was added to the enzyme; thus, it is unlikely that protection from excess copper entails redox reactions involving free copper. Other bacterial MCOs have been shown to oxidize Mn(II) (12, 43), but E. coli yacK MCO did not catalyze this reaction either (Table (Table2).2). So, how does the yacK-encoded protein function to protect E. coli against copper?

As pointed out by Grass and Rensing (24), given the limited number of copper atoms that can physical associate with yacK MCO, any protective mechanism seems most likely to be related to a catalytic activity. Despite obvious structural similarities between the bacterial copper resistance proteins and eukaryotic MCOs, there have been no published reports indicating whether the bacterial proteins also possess the oxidase activities commonly associated with eukaryotic enzymes. Our biochemical characterization of yacK MCO clearly demonstrates that the enzyme possesses both the phenoloxidase and ferroxidase activities commonly associated with eukaryotic MCOs. Thus, it seems appropriate to consider how one or the other of these catalytic activities could protect the bacterium from an elevated copper concentration.

Actually, it is something of a puzzle why MCOs that function physiologically as ferroxidases, e.g., the FET3 protein from yeast (15) and mammalian ceruloplasmin (19, 38), also bind and oxidize phenolic compounds by using the same reaction center. Protein structures that could impart sufficient substrate specificity to discriminate between such chemically different species as Fe(II) and diphenols are no doubt possible; thus, logic dictates that there may be a physiological rationale for having both activities catalyzed by a single catalytic center. With respect to bacteria, phenolate siderophores provide an obvious metabolic link between these two very different substrate species (11, 37). Excessive copper entering the cell through nonspecific divalent metal transporters, such as that encoded by the feo genes (28), may impede the uptake of other essential micronutrients, such as iron. It is possible that oxidation of phenolate siderophores releases chelated iron in such a way that levels of free iron are increased in the periplasm and, as a consequence, Fe(II) and Cu(II) uptake via the feo system becomes more appropriately balanced. Such a mechanism would confer a level of protection from elevated copper on the bacterium.

The strong enhancement of both the phenoloxidase and ferroxidase activities of yacK MCO upon copper addition was unexpected, and similar enhancement was not observed with the addition of any other metal ion. The stimulatory effect of copper addition was fully reversible and essentially disappeared at copper concentrations below 0.1 μM. As the average concentration of copper in the adult human body is approximately 20 μM (32), the response range of yacK MCO appears to be set at a level appropriate for the physiological conditions likely to be encountered by enteric bacteria. The stimulation of enzyme activity by copper addition did not appear to result from the replacement of intrinsic copper ions that might have been removed during purification, since copper addition enhanced both activities to roughly equivalent extents in both crude (heat treatment step) and highly purified (phenyl-Spherilose pool) preparations of the enzyme (data not shown). However, the appearance of two new EPR signals, increased A330, and decreased thermostability of the enzyme all suggest that the added copper interacts directly with the yacK-encoded protein in a specific and reversible manner. Thus, in addition to the intrinsic copper atoms in the enzyme, it appears that one or more labile metal-binding sites, upon specific binding of copper, can bring about a conformational change in the enzyme with a consequent increase in catalytic activity. Although we agree with Grass and Rensing (24) that a catalytic role for the yacK-encoded protein in protection against copper seems more likely, the possibility that the protein can also contribute to copper resistance by acting both as a metal-specific sink for covalently bound copper, as well as a sort of copper buffer by virtue of its reversible copper-binding sites, cannot be completely discounted.

That being said, the regulation of both yacK gene expression and enzyme activity of the encoded MCO in response to copper suggests that this could be a mechanism for careful tuning of the system to maintain a proper balance between iron and copper uptake. Although yacK MCO readily oxidized Fe(II), its Km for this substrate (70 μM) was significantly higher than those reported for the yeast FET3 protein (2 μM) (15) or ceruloplasmin (0.6 or 50 μM) (38), and the physiological relevance of this activity for E. coli is not obvious from our results. However, it may be significant that the dynamic ranges of the ferroxidase and phenoloxidase responses to added copper were so different—more than 200-fold versus ca. 15-fold, respectively (Table (Table3).3). Strong ferroxidase activity would tend to keep free iron in the ferric state, where it has low solubility and would be impossible to transport via divalent metal transporters. Thus, differential responses of the two enzyme activities to copper levels could allow the phenoloxidase activity to act on phenolate siderophores with minimal oxidation of Fe(II), thereby leaving the more soluble form of iron, Fe(II), available for uptake.

On a final note, oxidation of phenolic metabolites to form highly colored, insoluble polymers that protect organisms from damaging environmental conditions (e.g., UV light, dehydration, and enzymatic degradation) is an often-seen response in plants and fungi (14). Given the colored precipitate that formed when enterobactin was oxidized by yacK MCO, it is interesting to consider whether the oxidation of phenolate siderophores by bacterial MCOs might provide an adaptive physiological response whereby the bacteria use phenolic polymers to protect themselves until such time as growth conditions become more favorable.


This work was supported by U.S. Department of Energy grant DE-FG02-99ER20336.

We thank Jeremy Kaplan and Ken Rudd for providing stimulating and helpful advice during the initial stages of this project. Thanks also go to Carlos Gutierrez for providing the synthetic enterobactin used in these studies.


1. Alexandre G, Zhulin L B. Laccases are widespread in bacteria. Trends Biotechnol. 2000;18:41–42. [PubMed]
2. Askwith C, Kaplan J. Iron and copper transport in yeast and its relevance to human disease. Trends Biochem Sci. 1998;23:135–138. [PubMed]
3. Askwith C, Eide D, Vanho A, Bernard P S, Li L T, Davis-Kaplan S, Sipe D M, Kaplan J. The FET3 gene of Saccharomyces cerevisiae encodes a multicopper oxidase required for ferrous iron uptake. Cell. 1994;76:403–410. [PubMed]
4. Attieh Z K, Mukhopadhyay C K, Seshadri V, Tripoulas N A, Fox P L. Ceruloplasmin ferroxidase activity stimulates cellular iron uptake by a trivalent cation-specific transport mechanism. J Biol Chem. 1999;274:1116–1123. [PubMed]
5. Ausubel F M, Brent R, Kingston R E, Moore D D, Seidman J G, Smith J A, Struhl K, editors. Current protocols in molecular biology. 3rd ed. New York, N.Y: John Wiley & Sons, Inc.; 2000.
6. Berks B C, Sargent F, Palmer T. The TAT protein export pathway. Mol Microbiol. 2000;35:260–274. [PubMed]
7. Bertrand M G. Sur l'action oxydante des sels manganeux et sur la constitution chimique des oxydases. C R Acad Sci. 1897;124:1355–1358.
8. Blattner F R, Plunkett III G, Bloch C A, Perna N T, Burland V, Riley M, Collado-Vides J, Glasner J D, Rode C K, Mayhew G F, Gregor J, Davis N W, Kirkpatrick H A, Goeden M A, Rose D J, Mau B, Shao Y. The complete genome sequence of Escherichia coli K-12. Science. 1997;277:1453–1474. [PubMed]
9. Bligny R, Gaillard J, Douce R. Excretion of laccase by sycamore (Acer pseudoplatanus L.) cells: effects of a copper deficiency. Biochem J. 1986;237:583–588. [PubMed]
10. Bradford M M. A rapid and sensitive method for the quantitation of microgram quantities of protein utilizing the principle of protein-dye binding. Anal Biochem. 1976;72:248–254. [PubMed]
11. Braun V, Hantke K, Koster W. Bacterial iron transport: mechanisms, genetics, and regulation. In: Sigel A, Sigel H, editors. Iron transport and storage in microorganisms, plants, and animals. New York, N.Y: Marcel Dekker, Inc; 1998. pp. 67–145.
12. Brouwers G J, deVrind J P M, Corstjens P L A M, Cornelis P, Baysse C, DeJong E W D V. cumA, a gene encoding a multicopper oxidase, is involved in Mn2+ oxidation in Pseudomonas putida GB-1. Appl Environ Microbiol. 1999;65:1762–1768. [PMC free article] [PubMed]
13. Brown N L, Barrett S R, Camakaris J, Lee B T, Rouch D A. Molecular genetics and transport analysis of the copper-resistance determinant (pco) from Escherichia coli plasmid pRJ1004. Mol Microbiol. 1995;17:1153–1166. [PubMed]
14. Dean J F D, Eriksson K-E L. Laccase and the deposition of lignin in vascular plants. Holzforschung. 1994;48:21–33.
15. de Silva D, Davis-Kaplan S, Fergestad J, Kaplan J. Purification and characterization of Fet3 protein, a yeast homologue of ceruloplasmin. J Biol Chem. 1997;272:14208–14213. [PubMed]
16. de Silva D M, Askwith C C, Eide D, Kaplan J. The FET3 gene product required for high-affinity iron transport in yeast is a cell surface ferroxidase. J Biol Chem. 1995;270:1098–1101. [PubMed]
17. Diamantidis G, Effosse A, Potier P, Bally R. Purification and characterization of the first bacterial laccase in the rhizospheric bacterium Azospirillum lipoferum. Soil Biol Biochem. 2000;32:919–927.
18. Ducros V, Brzozowski A M, Wilson K S, Brown S H, Ostergaard P, Schneider P, Yaver D S, Pedersen A H, Davies G J. Crystal structure of the type-2 Cu depleted laccase from Coprinus cinereus at 2.2 Angstrom resolution. Nat Struct Biol. 1998;5:310–316. [PubMed]
19. Eggert C, Temp U, Dean J F D, Eriksson K-E L. Laccase-mediated formation of the phenoxazinone derivative, cinnabarinic acid. FEBS Lett. 1995;376:202–206. [PubMed]
20. Ezaz-Nikpay K, Uchino K, Lerner R E, Verdine G L. Construction of an overproduction vector containing the novel srp (sterically repressed) promoter. Protein Sci. 1994;3:132–138. [PubMed]
21. Fife R S, Moody S, Houser D, Proctor C. Studies of copper transport in cultured bovine chondrocytes. Biochim Biophys Acta. 1994;1201:19–22. [PubMed]
22. Freeman J C, Nayar P G, Begley T P, Villafranca J J. Stoichiometry and spectroscopic identity of copper centers in phenoxazinone synthase—a new addition to the blue copper oxidase family. Biochemistry. 1993;32:4826–4830. [PubMed]
23. Givaudan A, Effosse A, Faure D, Potier P, Bouillant M-L, Bally R. Polyphenol oxidase in Azospirillum lipoferum isolated from rice rhizosphere: evidence for laccase activity in non-motile strains of Azospirillum lipoferum. FEMS Microbiol Lett. 1993;108:205–210.
24. Grass G, Rensing C. Genes involved in copper homeostasis in Escherichia coli. J Bacteriol. 2001;183:2145–2147. [PMC free article] [PubMed]
25. Harris Z L, Durley A P, Man T K, Gitlin J D. Targeted gene disruption reveals an essential role for ceruloplasmin in cellular iron efflux. Proc Natl Acad Sci USA. 1999;96:10812–10817. [PubMed]
26. Hoopes J T, Dean J F D. Staining electrophoretic gels for laccase and peroxidase activity using 1,8-diaminonaphthalene. Anal Biochem. 2001;293:96–101. [PubMed]
27. Huang K X, Fujii I, Ebizuka Y, Gomi K, Sankawa U. Molecular cloning and heterologous expression of the gene encoding dihydrogeodin oxidase, a multicopper blue enzyme from Aspergillus terreus. J Biol Chem. 1995;270:21495–21502. [PubMed]
28. Kammler M, Schon C, Hantke K. Characterization of the ferrous iron uptake system of Escherichia coli. J Bacteriol. 1993;175:6212–6219. [PMC free article] [PubMed]
29. Keilin D, Mann T. Laccase, a blue copper-protein oxidase from the latex of Rhus succedanea. Nature. 1939;143:23–24.
30. Laemmli U K. Cleavage of structural proteins during the assembly of the head of bacteriophage T4. Nature. 1970;227:680–685. [PubMed]
31. Lesuisse E, Simon M, Klein R, Labbe P. Excretion of anthranilate and 3-hydroxyanthranilate by Saccharomyces cerevisiae—relationship to iron-metabolism. J Gen Microbiol. 1992;138:85–89. [PubMed]
32. Linder M C, Hazegh-Azam M. Copper biochemistry and molecular biology. Am J Clin Nutr. 1996;63:797S–811S. [PubMed]
33. Lindley P F, Card G, Zaitseva I, Zaitsev V, Reinhammer B, Selin-Lindgren E, Yoshida K. An X-ray structural study of human ceruloplasmin in relation to ferroxidase activity. J Biol Inorg Chem. 1997;2:454–463.
34. Link A J, Robison K, Church G M. Comparing the predicted and observed properties of proteins encoded in the genome of Escherichia coli K-12. Electrophoresis. 1997;18:1259–1313. [PubMed]
35. Mellano M A, Cooksey D A. Nucleotide sequence and organization of copper resistance genes from Pseudomonas syringae pv. tomato. J Bacteriol. 1988;170:2879–2883. [PMC free article] [PubMed]
36. Nielsen H, Engelbrecht J, Brunak S, von Heijne G. Identification of prokaryotic and eukaryotic signal peptides and prediction of their cleavage sites. Protein Eng. 1997;10:1–6. [PubMed]
37. O'Brien I G, Gibson F. Structure of enterochelin and related 2,3-dihydroxy-benzoylserine conjugates from Escherichia coli. Biochim Biophys Acta. 1970;215:393–402. [PubMed]
38. Osaki S. Kinetic studies of ferrous ion oxidation with crystalline human ferroxidase (ceruloplasmin) J Biol Chem. 1966;241:5053–5059. [PubMed]
39. Outten F W, Outten C E, Hale J, O'Halloran T V. Transcriptional activation of an Escherichia coli copper efflux regulon by the chromosomal MerR homologue, CueR. J Biol Chem. 2000;275:31024–31029. [PubMed]
40. Pitari G, Chichiricco G, Marcozzi G, Rossi A, Maccarrone M, Avigliano L. Expression of ascorbate oxidase in cultured green zucchini cells—effect of copper and abscisic-acid. Plant Physiol Biochem. 1993;31:593–598.
41. Ramirez R J A, Karamanukyan L, Ortiz S, Gutierrez C G. A much improved synthesis of the siderophore enterobactin. Tetrahedron Lett. 1997;38:749–752.
42. Shimizu A, Kwon J H, Sasaki T, Satoh T, Sakurai N, Sakurai T, Yamaguchi S, Samejima T. Myrothecium verrucaria bilirubin oxidase and its mutants for potential copper ligands. Biochemistry. 1999;38:3034–3042. [PubMed]
43. Solano F, Lucas-Elio P, Fernandez E, Sanchez-Amat A. Marinomonas mediterranea MMB-1 transposon mutagenesis: isolation of a multipotent polyphenol oxidase mutant. J Bacteriol. 2000;182:3754–3760. [PMC free article] [PubMed]
44. Solomon E I, Machonkin T E, Sundaram U M. Spectroscopy of multi-copper oxidases. In: Messerschmidt A, editor. Multicopper oxidases. Singapore: World Scientific Publishing Co.; 1997. pp. 103–127.
45. Solomon E I, Sundaram U M, Machonkin T E. Multicopper oxidases and oxygenases. Chem Rev. 1996;96:2563–2605. [PubMed]
46. Stanley N R, Palmer T, Berks B C. The twin arginine consensus motif of Tat signal peptides is involved in Sec-independent protein targeting in Escherichia coli. J Biol Chem. 2000;275:11591–11596. [PubMed]
47. Stearman R, Yuan D S, YamaguchiIwai Y, Klausner R D, Dancis A. A permease-oxidase complex involved in high-affinity iron uptake in yeast. Science. 1996;271:1552–1557. [PubMed]
48. Volk W A, Larsen J L. Bacterial ascorbic acid oxidase. Biochim Biophys Acta. 1963;67:576–580. [PubMed]
49. Vulpe C D, Kuo Y M, Murphy T L, Cowley L, Askwith C, Libina N, Gitschier J, Anderson G J. Hephaestin, a ceruloplasmin homologue implicated in intestinal iron transport, is defective in the sla mouse. Nat Genet. 1999;21:195–199. [PubMed]
50. Yamada T, Agui T, Suzuki Y, Sato M, Matsumoto K. Inhibition of the copper incorporation into ceruloplasmin leads to the deficiency in serum ceruloplasmin activity in Long-Evans cinnamon mutant rat. J Biol Chem. 1993;268:8965–8971. [PubMed]
51. Yuan D S, Stearman R, Dancis A, Dunn T, Beeler T, Klausner R D. The Menkes-Wilson disease gene homologue in yeast provides copper to a ceruloplasmin-like oxidase required for iron uptake. Proc Natl Acad Sci USA. 1995;92:2632–2636. [PubMed]

Articles from Journal of Bacteriology are provided here courtesy of American Society for Microbiology (ASM)