A zinc-binding site by negative selection induces metallodrug susceptibility in an essential chaperonin

Abstract

GroES is an indispensable chaperonin virtually found throughout all life forms. Consequently, mutations of this protein must be critically scrutinized by natural selection. Nevertheless, the homolog from a potentially virulent gastric pathogen, Helicobacter pylori, strikingly features a histidine/cysteine-rich C terminus that shares no significant homology with other family members. Additionally, three more (H45, C51, and C53) are uniquely present in its apical domain. The statistical analyses show that these residues may have originated from negative selection, presumably driven by either dependent or independent amino acid mutations. In the absence of the C-terminal metal-binding domain, the mutant protein still exhibits a substantial capacity for zinc binding in vivo. The biochemical properties of site-directed mutants indicate that H45, C51, and C53 make up an oxidation-sensitive zinc-binding site that may donate the bound metal to a zinc acceptor. Of interest, bismuth antiulcer drugs strongly bind at this site (K(d) of approximately 7 x 10(-26) M), replacing the bound zinc and consequently inducing the disruption of the quaternary structure. Because biological features by negative selection are usually inert to change during evolution, this study sheds light on a promising field whereby medicines can be designed or improved to specifically target the residues that uniquely evolved in pathogenic proteins so as to retard the emergence of drug resistance.

Fig. 1.

(A) Position-specific polymorphic proportions related to the three metal-binding residues. A fraction of the GroES from H. pylori (HpGroES) is shown as the reference of residue position. The substitution modes are identified from the alignment of multiple GroES homologs, and the corresponding proportions of each substitute are represented by the height of the letters. The numbers in brackets are the substitution scores for relevant substitution modes, individually given by GSM474 score times polymorphic proportion. (B) Data plot of neighbor preference based on the UniProtKB database. Filled circles are used to represent positive values and open circles indicate negative values. The circular area equals the absolute value of corresponding data, and the thresholds are given by the statistical analysis. The six amino acid pairs discussed in this work are marked by arrows, of which the total score is shown below the thresholds. The y axis symbols are appointed as the leader residue of an amino acid pair, and the x axis as the follower.

Fig. 2.

Determination of in vivo zinc binding for GroES variants. The wild-type and point mutants of HpGroESΔMBD as well as the wild-type EcGroES were isolated from E. coli cells grown in media consisting of different nutrients, and then desalted into zinc-free buffer containing 50 mM Tris-HCl (pH 7.5) and 100 mM NaCl. The protein concentration was estimated by Bradford assay, and the zinc content was quantitated by ICP-MS with reference to the standard solution of zinc. Each column represents the mean ± the standard deviation in triplicate.

Fig. 3.

Zinc transfer conducted by the two domains of the chaperonin, with or without oxidative sensitivity. (A) The bound Zn²⁺ of the H-X₅-C-X-C motif and the MBD was separately determined with increasing molar ratios of H₂O₂ versus the protein. (B) The hydrolysis of p-NPP by active AP was monitored by the changes of the absorption at 400 nm (A₄₀₀) with or without the presence of several GroES variants, which were incubated with zinc and H₂O₂ (or not) before desalting. The examined samples are holo-AP (up-pointing triangle), apo-AP with HpGroESΔMBD (square), apo-AP with 2 molar equivalents H₂O₂-treated EcGroES-MBD (circle), apo-AP with EcGroES-MBD (plus), apo-AP with 0.5 molar equivalents H₂O₂-treated HpGroESΔMBD (diamond), apo-AP with C51A/C53A (cross), and apo-AP (down-pointing triangle). The half of maximal A₄₀₀ is shown by a dashed line. All tests were performed in 50 mM Tris-HCl (pH 7.5) and 100 mM NaCl.

Fig. 4.

Competition between Zn²⁺ and Bi³⁺ to the chaperone domain of HpGroES (represented by HpGroESAΔMBD) (A) and the histidine/cysteine-rich MBD (represented by EcGroES-MBD) (B). The apoproteins were presaturated by ZnSO₄ or Bi(NO₃)₃ and then titrated by Bi(NO₃)₃ or ZnSO₄, accordingly. The absorption at 360 nm (A₃₆₀) was monitored for the formation of Bi-S bond. The experiments were performed at ambient temperature and in a buffer containing 50 mM Tris-HCl (pH 7.5) and 100 mM NaCl.

Fig. 5.

Oligomeric states of apo- and metal-bound HpGroESΔMBD variants. (A) One molar equivalent of ZnSO₄ (solid line) or Bi(NO₃)₃ (shaded curve) was loaded into proteins, and the samples were run on a Superose 12 column. The molecular masses were estimated by standard globular proteins, and A₂₁₅ was monitored to indicate the protein elution. (B) Apo-, Zn-, and Bi-bound proteins were visualized by Coomassie Blue staining in 15% polyacrylamide gel under nondenaturing conditions. Both were carried out at room temperature.

Fig. 6.

Growth rates of E. coli BL21 cells hosting HpGroESΔMBD variants in M9 minimal medium supplemented with 50 μM bismuth(III) citrate. The bacteria were induced by IPTG in nutrient-rich medium in advance, and appropriately diluted into M9 medium (with A₆₀₀ = 0.1–0.2) as the starter culture. At each selected time point, A₆₀₀ of an aliquot of bacterial culture was determined to indicate the cell number, and the A₆₀₀ values were normalized to that at the beginning. The data were collected in triplicate and represented as the mean ± the standard deviation.