(Received for publication, July 23, 1996, and in revised form, December 12, 1996)
From the Department of Clinical Pharmacology, University of Berne, 3010 Berne, Switzerland
The cop operon of Enterococcus
hirae effects copper homeostasis in this organism. It encodes a
repressor, CopY, an activator, CopZ, and two P-type copper
ATPases, CopA and CopB. Expression of all four genes is
regulated by the ambient copper. In this regulation, CopY apparently
acts as a copper-inducible repressor. By DNase I footprinting, it was
shown that purified CopY protected two discrete sites in the region
encompassing nucleotides 71 to
11 relative to the translational
start site and containing hyphenated inverted repeats. Transcription is
initiated between these repeats at nucleotide
42, in a domain that
remained accessible to DNase I in the DNA-repressor complex. Chemical
cross-linking revealed that CopY exists as a dimer in solution. In DNA
band-shift assays, it was apparent that the CopY-DNA interaction
occurred in two discrete steps. Half-maximal binding of repressor to
the two operator sites was observed at 2 × 10
9
M and 5 × 10
9 M CopY,
respectively. Copper ions released CopY from the promoter/operator with
an apparent half-binding constant for Cu(I) of 20 µM. The site-directed mutations A
61T and A
30T essentially abolished the
binding of CopY to the respective binding sites, and the double mutation A
61T/A
30T inactivated both binding sites. Thus, CopY is a
copper-inducible repressor of the cop operon of E. hirae, exhibiting highly specific DNA-protein interactions with
two sites on the cop promoter/operator and playing a key
role in copper homeostasis in E. hirae.
Copper is an essential trace element, acting as a cofactor for enzymes as diverse as cytochrome c oxidases, lysyl oxidases, or tyrosinases. But copper can also cause serious cell damage through radical formation. Therefore, careful regulation of the intracellular copper concentration is required. Until a few years ago, copper homeostasis was believed to be effected by metallothioneins, small, cysteine-rich proteins that complex and are regulated by heavy metal ions (1). More recently, it was discovered that copper homeostasis in the Gram-positive bacterium Enterococcus hirae involves two P-type copper ATPases, CopA and CopB, of 727 and 745 amino acids, respectively (2, 3). The CopA ATPase was shown to be required for copper accumulation under conditions of copper limitation, and CopB is needed to extrude copper from the cells when ambient copper reaches toxic levels (4, 5).
The two copper ATPase genes, copA and copB, are arranged in the cop operon, together with two regulatory genes, copY and copZ, in the order copY, copZ, copA, copB. These four genes appear to make up the copper homeostatic system of E. hirae. copY and copZ encode hydrophilic proteins of 145 and 69 amino acids, respectively. Expression of the ATPases was shown to be regulated by the ambient copper concentration in a biphasic fashion: expression was minimal at 10 µM extracellular copper, and increased as well as decreased copper concentrations led to induction of the cop operon (4). Disruption of the copY gene caused massive overexpression of CopA and CopB ATPase, whereas null mutation of copZ suppressed expression to very low levels. Based on these findings, a model was proposed in which CopY acts as a repressor and CopZ as an activator of the cop operon (6). In partial support of this model, we here show the interaction of CopY with the operator/promoter region of the cop operon.
The transcriptional start site of the cop operon was mapped by primer extension, and the site of interaction of CopY with the operator was delineated by DNase I footprinting. Through site-directed mutagenesis, nucleotide residues that are critically involved in the CopY-operator interaction were identified. It was also shown that CopY is a homodimer that interacts with the operator in two steps in a copper-sensitive fashion.
E. hirae (ATCC 9790, formerly called Streptococcus faecalis or S. faecium) was obtained from the American Type Culture Collection. Escherichia coli strains XLmutS and XL1-Blue were obtained from Stratagene, Inc. The expression vector pQE12 was purchased from Qiagen. The construction of pOA1 has been described previously (6).
MaterialsPico green was from Molecular Probes, Inc. The
Chameleon mutagenesis kit (Stratagene, Inc.) was used for site-directed
mutagenesis. [-32P]ATP (3000 Ci/mmol) was obtained
from Amersham Corp., phenylmethylsulfonyl fluoride from Merck, and
growth media additives from BBL. Purification columns, column
materials, and the MluI
ScaI switch-toggle
primer were bought from Pharmacia Biotech Inc. Moloney murine leukemia virus reverse transcriptase was from Anawa Trading, Inc. (Wangen, Switzerland). Dynazyme DNA polymerase was provided by Finnzymes (Espoo,
Finland), and oligonucleotide primers were synthesized by Microsynth,
Inc. (Windisch, Switzerland). All other molecular biology reagents were
obtained from Boehringer Mannheim. DNase I and other chemicals were
from Sigma and were of the highest grade available.
Total cellular RNA of wild-type E. hirae was prepared either from uninduced cells or from cells that
were induced for 1 h with 1 mM CuSO4. RNA
was isolated as described previously (7). The primer
5-CCAATCCATCGAATCAGCTAATATC-3
was labeled with polynucleotide kinase
and [
-32P]ATP to a specific activity of 0.1 µCi/nmol. 20 µg of RNA in 30 µl of hybridization solution (40 mM PIPES,1 pH 6.4, 1 mM EDTA, 0.4 M NaCl, 80% formamide) were
heated for 5 min to 85 °C and annealed overnight at 30 °C with
0.5 pmol of the labeled primer. The nucleic acids were then ethanol
precipitated, washed, and dissolved in 100 µl of 10 mM
dithiothreitol, 0.5 mM dNTPs, 40 units of RNase inhibitor,
and reverse transcribed with 100 units of Moloney murine leukemia virus
reverse transcriptase in the buffer supplied by the manufacturer for 90 min at 42 °C. Following ethanol precipitation, the products were
analyzed by electrophoresis on 6% denaturing polyacrylamide sequencing
gels.
A PCR product of the copY
gene was generated with the two primers 5-CAAGAATGATCAAAGAATTT-3
and
5
-CCATTTTGGAATTCGTTGCC-3
and Dynazyme DNA polymerase. Following
cutting with BamHI and HindIII, the PCR product
was cloned into pQE12 and digested with the same enzymes. The resultant
clone encoded a CopY protein with a five-amino acid N-terminal
extension originating from the expression vector. These extra residues
were deleted by site-directed mutagenesis. The final clone, pWY145, was
verified by DNA sequencing.
For overexpression, 500-ml cultures of
XL1-Blue cells transformed with pWY145 were grown aerobically at
30 °C to an A550 of 0.5. Following induction
with 1 mM
isopropyl-1-thio--D-galactopyranoside for 2 h, the
cells were harvested by centrifugation for 10 min at 5000 × g. The cell pellet was washed once with 200 ml of TGD buffer
(50 mM Tris-SO4, pH 7.7, 5% (v/v) glycerol, 2 mM dithiothreitol) and resuspended in 100 ml of TGD buffer.
The cells were lysed by incubation with 25 µg of lysozyme and 1 mM phenylmethylsulfonyl fluoride for 10 min at room
temperature and broken open by one passage through a French pressure
cell at 40 MPa. The cell debris was collected by centrifugation for
1 h at 90,000 × g and the supernatant passed
through a Q-Sepharose column. CopY was eluted with a 0-100
mM K2SO4 gradient in TGD buffer.
Final purification was achieved by gel filtration on a Superose-12
column in TGD buffer plus 150 mM
K2SO4. Purified CopY was stored in 2-µg
aliquots at
70 °C. The concentration of CopY was determined by
amino acid analysis.
100 pmol of purified CopY were cross-linked at room temperature with the glutaraldehyde concentrations indicated under "Results" in a total volume of 40 µl of 10 mM Na-HEPES, pH 8.0, 5 mM magnesium acetate, 5 mM dithiothreitol, 50 mM sodium acetate, 12% (v/v) glycerol, for 10 min at room temperature, followed by quenching with 5 µl of 1 M ethanolamine. The products were resolved on a 15% SDS-polyacrylamide gel and visualized by silver staining.
Site-directed MutagenesisMutation A61T was generated in
pOA1 (6) with the Chameleon mutagenesis kit using the mutagenic primer
5
-CGATTTCAGTTGTAATCTAT-3
and a commercial
MluI
ScaI switch-toggle primer. The resultant plasmid pOM214 was used to generate the double mutant A
61T/A
30T by
PCR amplification with the primers 5
-CGAATCTTCCAACTGATCG-3
and
5
-GTTTTTTCACCTCCATCGATTACATTTGAAAACTTAAC-3
. Mutant A
30T was
generated similarly, but using wild-type DNA as a template. Mutants
T
23A and A
61T/T
23A were generated by PCR amplification of pOA1
and pOM214, respectively, with the primers
5
-GTTTTTCACCTCCATCGATTTCATTTGTAAAC-3
and
5
-CGAATCTTCCAACTGATCG-3
.
DNA band-shift assays were performed
essentially as described (8). 2-3 fmol (6-10 nCi) of labeled
wild-type cop promoter/operator DNA were incubated with a
10-300-fold molar excess of purified CopY for 30 min at room
temperature in a total volume of 20 µl of the buffer used for DNase I
footprinting. The reactions were analyzed on 6% polyacrylamide gels in
6.8 mM Tris acetate, pH 8.0, 3.3 mM sodium
acetate, 2.5% (v/v) glycerol. Gels were allowed to polymerize for at
least 12 h prior to use and were run at 5 W for 90 min.
Promoter/operator fragments with the A61T mutation were isolated by
PCR amplification of the mutated plasmid with the two primers
5
-CCAATCCATCGAATCAGCTAATATC-3
and 5
-CGAATCTTCCAACTGATCG-3
. The
promoter/operator fragments with the mutations A
30T and T
23A and
the double mutations A
30T/A
61T and T
23A/A
61T were similarly generated by PCR amplification of mutated plasmid DNA with the primers
used for mutagenesis.
Footprinting assays were performed in
200 µl of 20 mM Tris acetate, pH 8.0, 5 mM
dithiothreitol, 5 mM magnesium acetate, 50 mM
sodium acetate, 12% glycerol, 1 mM Ca(OH)2, 2 µg of poly(dI-dC), and 5 µg of bovine serum albumin. Assays
contained 4-6 fmol of cop promoter/operator fragment, that
was labeled with 32P at one end with polynucleotide kinase
as described (8) to a specific activity of 3 µCi/pmol. The amounts of
purified CopY specified under "Results" were added and the reaction
incubated for 30 min at room temperature. The DNA was then digested
with 0.012 unit of DNase I at 30 °C for 2 min. Digestion was
terminated by adding 700 µl of 92% ethanol, 0.75 M
ammonium acetate, 7 µg/ml tRNA. The products were precipitated at
70 °C for 15 min and centrifuged for 10 min at 10,000 × g. The pelleted material was washed once with 70% ethanol
and resolved by electrophoresis on an 8% polyacrylamide sequencing gel
together with a Sanger sequencing reaction of the same fragment with
the primers 5
-CGAATCTTCCAACTGATCG-3
for the sense strand and
5
-CCAATCCATCGAATCAGCTAATATC-3
for the antisense strand.
Standard molecular biology methods were used according to published procedures (8). DNA was sequenced according to the method of Sanger et al. (9). The concentrations of the DNA fragments used for band-shift assays were determined by fluorescent measurements with pico green as described by the manufacturer.
To characterize the function of the CopY repressor and to study
its interaction with the DNA target, CopY was overexpressed in E. coli and purified. When CopY was overexpressed as detailed under
"Experimental Procedures," approximately 60% of the protein was
trapped in inclusion bodies. However, CopY in the supernatant still
made up 10% of the total protein in this fraction. CopY could be
purified to approximately 90% purity by successive chromatography on
Q-Sepharose and gel filtration. Fig. 1A shows
the products of the different purification steps. Below the band of
apparent molecular mass 17 kDa, corresponding to CopY, there was always a minor band of 16 kDa in the preparation. This contaminant increased with increasing purification and could be a degradation product of
CopY. However, we could not observe variation in the results obtained
with preparations containing differing amounts of the 16-kDa band.
We assessed whether CopY was monomeric or multimeric in solution by cross-linking experiments. Fig. 1B shows that increasing concentrations of glutaraldehyde resulted in the formation of CopY homodimers with an apparent molecular mass of 34 kDa. No significant amounts of higher molecular mass species were generated even at very high concentrations of cross-linker, indicating that CopY homodimers are the principal species of CopY in solution. Inclusion of a 4-fold excess of lysozyme in the experiment did not result in significant cross-linking of this control protein to itself or to CopY (not shown). In analytical gel filtration, CopY migrated like a globular protein of apparent molecular mass 53 kDa, suggesting that the native CopY dimers are not globular but form an extended structure, as would be expected.2
Since CopY acts as a repressor of the cop operon, it was
expected to interact with a DNA target at or near the site of
transcription initiation. Primer extension analysis was thus carried
out to determine the transcription initiation site of the
cop operon. Total RNA from copper-induced and uninduced
wild-type E. hirae cells was reverse transcribed with a
primer as specified under "Experimental Procedures," and the
resultant DNA fragments were resolved on a sequencing gel (Fig.
2). A single product was generated from induced but not
from uninduced cells. This is in line with the previous observation
that in copper-induced cells expression of the cop operon
was increased 12-fold compared with uninduced cells (6).
Co-electrophoresis of the primer extension products with DNA sequencing
reactions conducted with the same primer allowed assignment of
the transcriptional start site to guanosine at position 42 with
respect to the start of translation. This assignment was further
verified by running gels of mixtures of sequencing reactions and
primer-extension product (not shown).
DNA binding properties of the CopY repressor were studied by band-shift
assays. When a radiolabeled 316-bp DNA fragment encompassing the
transcription start site was incubated with purified CopY, DNA-protein
complexes could be observed (Fig. 3). Increasing
concentrations of CopY changed the apparent mobility of the DNA
fragment in two steps: to an intermediate retarded form I, and to a
more retarded form II, indicating two binding events. Addition of a
10-fold molar excess of unlabeled cop promoter/operator DNA
to the binding mixture competed with the DNA-protein complex formation,
whereas added promoter DNA from the -lactamase gene of E. coli did not act as a competitor. Similarly, the promoter of the
NaH antiporter gene of E. hirae (10) did not compete with
the binding (data not shown), indicating the specificity of the
DNA-CopY interaction. By quantitative evaluation of the band-shift
data, the half-association concentrations of CopY under these
conditions were estimated to be 2 nM for retarded form I
and 5 nM for retarded form II.
In vivo, the expression of the E. hirae cop operon is induced by Cu2+, Cu+, Ag+ or Cd2+ but not by Ni2+ or Zn2+ (3, 6). It was thus of interest to test whether these heavy metal ions affected the in vitro interaction of CopY with the cop operator. Fig. 3B shows the effects of different heavy metal ions on the CopY-cop operator interaction as measured by band-shift assays. As little as 1 µM added copper in the binding reaction released some CopY from the cop operator. At 50 µM copper, formation of a CopY-DNA complex was completely inhibited. An approximately 50% displacement of CopY from the operator was observed with 20 µM Cu+, 50 µM Ag+, or 50 µM Cd2+. The same concentrations of Ni2+ and Zn2+ ions released approximately 5% of the bound repressor from the DNA. These results demonstrate that Cu+, Ag+, and Cd2+ are efficient inducers of the CopY repressor. This is in line with the induction properties of the cop operon observed in vivo. It is of course not possible to quantitatively compare in vitro band-shift experiments with whole cell studies, since it is not known how externally added metal ions change the cytoplasmic concentration in living cells. Also, the redox equilibrium between Cu(I) and Cu(II) makes it very difficult to determine the oxidation state of copper in vivo.
By site-directed mutagenesis of the cop promoter/operator
DNA, specific DNA-CopY interactions were identified. Inversion of the
adenosine at position 61 to a thymidine (A
61T) altered the binding
of CopY to the operator. Fig. 4A shows that
even a 200-fold molar excess of repressor to A
61T cop
promoter/operator resulted in the generation of only one retarded form.
The electrophoretic mobility of this complex was identical to the form
I intermediate observed with wild-type promoter DNA. This suggests that
adenosine
61 is an essential base for the interaction of CopY with
one site on the operator. The mutation A
30T revealed another
essential residue, inactivating the second site of interaction of CopY
with the operator (Fig. 4B), whereas the mutation T
23A in
the proximal half of the inverted repeat did not have an effect on the
repressor-DNA interaction (not shown). In the double mutant
A
30T/A
61T, the repressor-DNA interaction was practically abolished.
Clearly, single bases in the cop promoter/operator play key
roles in the interaction with the CopY repressor. Half-maximal binding
of repressor to the single remaining high affinity binding sites in the
mutants was estimated to occur at 7 nM CopY for the site
proximal to the translation start and at 12 nM CopY for the
distal binding site. These values are 2-3-fold higher than those
observed for the wild-type operators, which could indicate some
cooperativity between the binding sites. Since it is unlikely that the
relative binding affinities of the two sites are reversed in the
mutants, this would suggest that the proximal site exhibits the higher
affinity for CopY and interacts first with the repressor.
With DNase I footprinting experiments, the sites on the cop
promoter/operator DNA that are occupied by the bound CopY repressor were mapped. Two regions were protected from DNase I digestion by CopY.
They were slightly asymmetric on the two DNA strands and encompassed
nucleotides 13 to
38 and
44 to
69 on the sense strand (Fig.
5A) and nucleotides
11 to
37 and
44 to
71 on the complementary strand (Fig. 5C). Between these
protected sites, there is a region not protected from DNase I
digestion, namely from
39 to
43 on the sense strand and from
38
to
43 on the antisense strand. The start of transcription, mapped to
nucleotide
42, is located in this DNase I-sensitive domain (see Fig.
4C for a summary).
Band-shift experiments with mutant promoter/operator DNA had suggested
that CopY binding by one site was impaired in some of the mutants.
DNase I protection experiments with the mutated cop
promoter/operator fragment A61T showed indeed that only one site
could bind repressor under the conditions used. The distal half of the
inverted repeat that contained the A
61T mutation remained sensitive
to DNase I digestion (Fig. 5B). Similarly, the A
30T
mutated operator/promoter DNA was protected from DNase I by CopY at the
distal, nonmutated site but was protected only by high concentrations
of CopY at the mutated site (Fig. 5D). Finally, the double
mutant A
30T/A
61T remained essentially unprotected to DNase I
digestion by CopY at both binding sites. Only high concentrations of
CopY somewhat protected the proximal region from DNase I digestion
(Fig. 5E). These observations are in agreement with the gel
retardation experiments and show that mutations A
61T and A
30T both
strongly interfere with the binding of the CopY repressor to its DNA
target. Clearly, the nucleotides at these positions must fulfill
crucial roles in the DNA-protein interaction.
Taken together these results show that the cop promoter/operator possesses two independent binding sites for the CopY repressor. These sites exhibit somewhat different binding affinities for the repressor and appear well separated, since there are some DNase I-sensitive bases between the sites, including the site of transcriptional initiation.
We have characterized the interaction of the E. hirae copper-responsive repressor, CopY, with the cop operator in vitro. Three main features of this regulation system were revealed: (i) the CopY repressor interacts with the cop operator/promoter at two discrete sites at positions flanking the transcriptional start site; (ii) the repressor-DNA interaction is copper-sensitive; and (iii) single base changes can lead to half-of-the-site reactivity. Cross-linking and gel filtration experiments suggest that CopY is present as a homodimer in solution. Thus, it appears likely that the CopY homodimers interact with each of the two binding sites identified in the operator/promoter region.
The N terminus of the CopY repressor exhibits 27-32% sequence
identity to the -lactamase repressors MecI of Staphylococcus epidermis (11), PenI of Bacillus licheniformis (12),
and BlaI of Staphylococcus aureus (13). These repressors
possess two structurally and functionally different domains (14),
similar to the well characterized proteins of the large
Cro/LysR
family of repressors (15). The general concept of the function of these proteins is that the N-terminal part of the protein forms a
helix-turn-helix DNA binding motif, whereas the C-terminal part
interacts with the effector. This concept is supported by x-ray
crystallographic studies of the bacteriophage
Cro and 434 repressors
(16, 17), which revealed that the Q28-Q29 pair in the helix-turn-helix
motif of these proteins tightly interacts with the base pairs of an ACA
triplet of the operator DNA (18) and that the mutation Q28A abolishes
the interaction of the repressor with wild-type operator DNA (19). An
analogous Q30-Q31 pair is present in CopY, and the cop
promoter/operator contains four ACA triplets that would allow a similar
interaction. We have shown that mutation of two of these ACA triplets
to TCA almost completely abolish CopY-DNA interaction at the
corresponding sites. These findings are essentially complementary to
those obtained with the 434 repressors and suggest that the
CopY-operator interaction is very similar in nature.
CopY concentrations required for half-maximal interaction with the two
sites on the wild-type operator were determined in vitro as
2 × 109 M and 5 × 10
9 M, and thus are in the range of
10
9 to 10
10 M observed for
other repressor-operator interactions (20, 21). Although most
repressors are homodimeric proteins, only a few of them interact with
two discrete binding sites on the operator, such as NodD, IlvY, or CadC
(22, 23). In fact, the CadC regulatory system resembles the E. hirae CopY system in other ways. It also controls expression of
heavy metal ATPases: the cad operons of cadmium-resistant
bacteria encode a regulatory protein, CadC, and a cadmium efflux
ATPase, CadA. The expression of CadA is strongly induced by
Cd2+ but also by Pb2+, Bi3+, and
some other heavy metal ions (24). By gel retardation assays and DNase I
footprinting, it was recently shown that CadC binds to an inverted
repeat in the promoter/operator region of the cad operon and
is released by Cd2+, Pb2+, and, more
efficiently, Bi3+ and Hg2+ (25). Similar to the
CopY-DNA interaction, the binding of CadC to DNA took place in two
discrete steps. However, CadC does not exhibit significant sequence
identity to CopY but is related to ArsR from E. coli, which
regulates resistance to arsenic ions (26), and to SmtB, which controls
cyanobacterial metallothionein synthesis (27). So these proteins
apparently represent another class of metalloregulators.
The promoter of the cop operon features the consensus
sequences TGTAAT and TTGACA at the 10 and
35 positions relative to the transcriptional start site. These sequences are found in many bacterial promoters (28). The positioning of the two CopY binding sites
relative to these transcriptional signals lies within the range of
consensus positions for repressors and outside the range for activators
(29). The biological significance of two CopY binding sites with
similar affinities remains unclear at this time.
The control of the cop operon of E. hirae apparently also requires the function of CopZ (6). CopZ is a protein of 69 amino acids containing the presumed copper-binding motif GMXCXXC. In copZ-disrupted strains, the transcription of the cop operon was repressed and copZ function could be complemented in trans (6). A model for the observed biphasic regulation was proposed in which derepression of the cop operon can be achieved in two ways: by copper releasing CopY from the operator under conditions of excessive copper, or through activation of transcription by CopZ under copper-limiting conditions. Although our data support the activation of the cop operon by copper-induced release of the repressor, we could not obtain evidence of an involvement of CopZ in vitro, and the elucidation of the role of this protein requires further study.
We thank Thomas Weber for expert technical assistance, Hans-Ulrich Mösch for advice on DNase I footprinting, and Linda Thöni and Bernhard Erni for helpful suggestions.