Structure and metal binding properties of ZnuA, a periplasmic zinc transporter from Escherichia coli

Abstract

ZnuA is the periplasmic Zn(2+)-binding protein associated with the high-affinity ATP-binding cassette ZnuABC transporter from Escherichia coli. Although several structures of ZnuA and its homologs have been determined, details regarding metal ion stoichiometry, affinity, and specificity as well as the mechanism of metal uptake and transfer remain unclear. The crystal structures of E. coli ZnuA (Eco-ZnuA) in the apo, Zn(2+)-bound, and Co(2+)-bound forms have been determined. ZnZnuA binds at least two metal ions. The first, observed previously in other structures, is coordinated tetrahedrally by Glu59, His60, His143, and His207. Replacement of Zn(2+) with Co(2+) results in almost identical coordination geometry at this site. The second metal binding site involves His224 and several yet to be identified residues from the His-rich loop that is unique to Zn(2+) periplasmic metal binding receptors. Electron paramagnetic resonance and X-ray absorption spectroscopic data on CoZnuA provide additional insight into possible residues involved in this second site. The second site is also detected by metal analysis and circular dichroism (CD) titrations. Eco-ZnuA binds Zn(2+) (estimated K (d) < 20 nM), Co(2+), Ni(2+), Cu(2+), Cu(+), and Cd(2+), but not Mn(2+). Finally, conformational changes upon metal binding observed in the crystal structures together with fluorescence and CD data indicate that only Zn(2+) substantially stabilizes ZnuA and might facilitate recognition of ZnuB and subsequent metal transfer.

Fig. 1

Structure of ZnZnuA. a Ribbon diagram of crystal form I showing the N-terminal domain in pink, the C-terminal domain in blue, and the connecting α-helix (α_c) in green. The zinc ions are represented as gray spheres. The ends of the disordered His-rich loop are indicated by asterisks. b Viewed approximately 90° around the vertical axis from the orientation in a

Fig. 2

Structural comparison between ZnuA protein from Escherichia coli (Eco-ZnuA) with both metal binding sites occupied (ZnZnuA, molecule A, green) and Eco-ZnuA with one metal binding site occupied (CoZnuA, molecule E, magenta). a Stereo superposition of the Cα atoms. The β6α6 loop is labeled with arrows. b Closeup view of the conformational changes in the C-terminal domain. Residues His224 and Arg237 are shown as stick representations, and the Zn²⁺ ions are shown as gray spheres. The position of the Co²⁺ site is identical to that of Zn1

Fig. 3

The primary metal binding site in Eco-ZnuA. a ZnZnuA molecule A, final 2.0-Å resolution 2F_o – F_c electron density map (blue, contoured at 1.2σ) superimposed on an anomalous difference Fourier map calculated using data collected near the zinc absorption edge (yellow, contoured at 5σ). b Superposition of the metal binding sites in ZnZnuA molecules A, B, C, and D (shown in green, light green, red, and dark salmon, respectively) and CoZnuA molecules E and F (shown in blue and light blue, respectively). c CoZnuA molecule A, final 2.15-Å resolution 2F_o – F_c electron density map (blue, contoured at 1.2σ). d ApoZnuA molecule H, final 2.45-Å resolution 2F_o – F_c electron density map (contoured at 1.2σ). The water molecule is not observed in molecule G. e Superposition of the metal binding sites in apoZnuA, molecules G and H (shown in gray and black, respectively) and in ZnZnuA, with one (molecule D, dark salmon) and two (molecule A, green) occupied metal binding sites

Fig. 4

The second metal binding site in Eco-ZnuA. a Final 2.0-Å resolution 2F_o – F_c electron density map (blue, contoured at 0.7σ) superimposed on the anomalous difference Fourier map calculated using data collected near the zinc absorption edge (yellow, contoured at 5σ). The unidentified ligands are labeled ligand 2, ligand 3, and ligand 4. b Superposition of all the His224 residues with ZnZnuA molecules A, B, C, and D shown in green, light green, red, and dark salmon, respectively, CoZnuA molecules E and F shown in blue and light blue, respectively, and apoZnuA molecules G and H shown in gray and black, respectively. The two different positions of the β6α6 loop are clear

Fig. 5

Far-UV circular dichroism (CD) spectra of Eco-ZnuA. The spectra were acquired for 5.4 μM apoZnuA (bold solid line), with 1 equiv of Zn²⁺ (dashed line), with 2 equiv of Zn²⁺ (dotted line), and with 5 equiv of Zn²⁺ (thin solid line) in 20 mM potassium phosphate, pH 7.5, 20 mM NaF

Fig. 6

Thermal denaturation of Eco-ZnuA monitored by CD spectroscopy at 222 nm. Data for 1.4 μM apoZnuA (filled squares), and ZnuA in the presence of 10 μM Zn²⁺ (filled circles), Cu²⁺(up triangles), Cu⁺(inverted triangles), Ni²⁺ (open squares), Co²⁺(open circles), and Mn²⁺ (asterisks) were fit to the two-state model (equilibrium between the native and the unfolded state) assuming zero ΔC_P (solid line) or nonzero ΔC_P (not shown). Buffer, 20 mM potassium phosphate, pH 7.5, 20 mM NaF

Fig. 7

Titration of Eco-ZnuA/Mag-Fura-2 (MF) with Zn²⁺. a Representative UV/vis spectra obtained during titration of 15.5 μM MF and 21.9 μM Eco-ZnuA with 4-μL aliquots of 1 mM Zn²⁺ in 50 mM N-(2-hydroxyethyl)piperazine-N′-ethanesulfonic acid (HEPES), pH 7.5, 200 mM NaCl at room temperature. Arrows indicate the direction of the absorbance changes as increasing concentrations of Zn²⁺ are added. b Absorbance change at 366 nm of Eco-ZnuA/MF mixture as a function of added Zn²⁺

Fig. 8

UV/vis spectrum of CoZnuA loaded with 1 equiv of Co²⁺

Fig. 9

Electron paramagnetic resonance (EPR) spectra of CoZnuA. Traces A, B, F, and G are experimental EPR spectra of ZnuA with 1 equiv of Co²⁺ (A, B) or 2 equiv of Co²⁺ (F, G). Spectra A and F were recorded at 5 K, 50-mW microwave power, and spectra B and G were recorded at 13 K, 2-mW microwave power. Spectra F and G are shown with twofold reduced amplitude compared to spectra A and B, respectively. Trace D is the difference spectrum obtained by subtraction of trace B from trace G (prior to amplitude reduction of trace G). Trace C is a simulation of trace B assuming a spin Hamiltonian H = βg·B·S + S·D·S, where S = 3/2, and parameters D = 50 cm⁻¹ (i.e., D ≫ gβBS, M_S = |±1/2〉), E/D = 0.04, g_⊥ = 2.25, and g_‖ = 2.38. Trace E is a simulation of trace D with parameters D = 50 cm⁻¹ (i.e., D ≫ gβBS, M_S = |±1/2〉), E/D = 0.11, g_⊥ = 2.25, and g_‖ = 2.36. Trace H is the sum of traces C and E, shown with twofold reduced amplitude

Fig. 10

Extended X-ray absorption fine structure (EXAFS) data. Fourier transforms (FT) of k³-weighted EXAFS for a 1-ZnZnuA and b Co²⁺ in the second site in ZnCoZnuA. Solid lines represent raw data, while diamonds represent the best fits [fit 1–2 (Zn) and fit 2–4 (Co), Table 3]

Fig. 11

8-Anilino-1-naphthalenesulfonic acid (ANS) fluorescence data. Changes in ANS fluorescence intensity at 510 nm of 5 μM apoZnuA, 0.5 mM ANS solution upon addition of a various metal ions to 5 and 10 μM final concentration, and b to 0.1, 0.5, and 1.0 mM final concentration (note the different scales for relative intensity). The columns labeled buffer correspond to the fluorescence of 5 μM apoZnuA, 0.5 mM ANS solution diluted with metal-free buffer H (50 mM HEPES, pH 7.5, 200 mM NaCl) containing 0.6% HNO₃. The volume of buffer H, 0.6 % HNO₃ corresponds to the volume of metal aliquots added to the protein/ANS solution. All the spectra were corrected for the fluorescence of ANS alone by subtracting the fluorescence of ANS solution of equivalent volume and concentration from the fluorescence of all the spectra. Experimental conditions as follows: λ_ex = 380 nm, λ_em = 510 nm, 296 K, buffer 50 mM HEPES, pH 7.5, 200 mM NaCl