Nickel and GTP Modulate Helicobacter pylori UreG Structural Flexibility

UreG is a P-loop GTP hydrolase involved in the maturation of nickel-containing urease, an essential enzyme found in plants, fungi, bacteria, and archaea. This protein couples the hydrolysis of GTP to the delivery of Ni(II) into the active site of apo-urease, interacting with other urease chaperones in a multi-protein complex necessary for enzyme activation. Whereas the conformation of Helicobacter pylori (Hp) UreG was solved by crystallography when it is in complex with two other chaperones, in solution the protein was found in a disordered and flexible form, defining it as an intrinsically disordered enzyme and indicating that the well-folded structure found in the crystal state does not fully reflect the behavior of the protein in solution. Here, isothermal titration calorimetry and site-directed spin labeling coupled to electron paramagnetic spectroscopy were successfully combined to investigate HpUreG structural dynamics in solution and the effect of Ni(II) and GTP on protein mobility. The results demonstrate that, although the protein maintains a flexible behavior in the metal and nucleotide bound forms, concomitant addition of Ni(II) and GTP exerts a structural change through the crosstalk of different protein regions.


28
The discovery of antimicrobials to defeat bacterial pathogens is among the most important 29 medical advances of the last century. However, antimicrobial resistance (AMR) has impaired the 30 efficacy of antibiotics against infections in the last decades, and is considered by the World Health oligomerization with a [(αβ)3]4 quaternary structure. 16 Despite the different oligomeric organization, 58 the structure of the known urease enzymes is fully conserved, and they present a substantially 59 identical active site found in the α subunit. This site contains two Ni(II) ions bridged by the 60 carboxylate group of a carbamylated lysine, essential to maintain the ions at the correct distance for 61 catalysis, and by a hydroxide ion, the nucleophile in the hydrolysis reaction. 4,[14][15] 62 Although previous studies identified several molecules that bind urease and inhibit it 63 competitively or uncompetitively, none of them is generally used in therapy, due to their severe side 64 effects or limited ability to pass the bacterial membrane. [14][15] Recently, an alternative strategy to design 65 urease inhibitors has been proposed by targeting, instead of the enzyme, the process that delivers 66 nickel ions into the enzyme active site, precluding enzyme maturation to the active Ni(II)-loaded 67 urease. 17 This activation process is governed by the interplay of at least four accessory proteins, 68 named UreD, UreE, UreF and UreG, coded by genes belonging to a single operon together with the 69 structural genes. 4 UreE acts as the metallo-chaperone of the system that delivers Ni(II) into urease, 18 70 through tunnels that pass across a complex formed by UreD, UreF and UreG, the last acting as a 71 molecular chaperone that prepares urease to incorporate the metal ion. 19 Precluding urease 72 maturation by blocking delivery of Ni(II) into its active site could thus represent a novel approach to 73 enzyme inhibition.

74
The central player of the urease chaperone activation network is UreG, a GTPase that couples 75 the energy obtained from GTP hydrolysis to urease maturation. 20 HpUreG interacts either with 76 HpUreE, forming a heterodimeric HpUreG2E2 complex, 18 or with HpUreF and HpUreD, forming a 77 ternary HpUreG2F2D2 complex. 21 The multiplicity of partners of UreG is reflected in its folding      Table 1). The fit of the obtained data, performed using the AFFINImeter software 46 and a    Table 1. Thermodynamic parameters of Ni(II) titrations over HpUreG and its variants, in the absence 214 and in the presence of GTPS.

217
The obtained results confirmed that HpUreG undergoes dimerization when both Ni(II) and GTPS, a

233
µM. In this case, Ni(II) binding occurs with thermodynamic parameters similar to the ones observed 239 for the apo-protein, with favorable enthalpy and a negative entropic contribution (Table 1). On the 240 other hand, dimerization is an entropy-driven endothermic process as expected ( Table 1)      The EPR spectrum of WT HpUreG labeled with MA-Proxyl nitroxide (WT Prox , Figure 1C) arises dissect the conformational flexibility of different regions of HpUreG, the EPR spectrum of the 276 nitroxide-labeled HpUreG variants that contain a single labeled cysteine (C7 prox , C48 prox , and C66 prox ) C7 prox to C48 prox and then to C66 prox , reflecting an increased mobility of the nitroxide moiety and, 281 consequently, of the protein structural motif to which the label is attached ( Figure 1C)

304
Note that, in the following paragraphs, we named "fast" all spectral components characterized 305 by τc values included between 0.3-0.7 ns, "slow" those characterized by τc values included between 306 0.7-4.0 ns and "rigid" components characterized by τc values included between 4.0-6.2 ns. The 307 sharpest EPR line shape is observed for the nitroxide grafted to Cys66. This spectrum is constituted 308 by two components having almost the same proportion, one with τc = 0.6 ns ("fast") and the other fragments 27, 41, 53 demonstrates that this region is highly flexible. This dynamic behaviour is similar to 312 that observed for the SpUreG orthologue containing the nitroxide label grafted onto the 313 corresponding cysteine residue, which features two conformers with similar correlation times (τc Similarly, two components with different degrees of flexibility and comparable relative 316 abundance are observed for C48 prox : one features a "fast" behavior (τc = 0.6 ns), while the other 317 ("rigid") is consistent with a less flexible dynamic (τc = 4.9 ns) (Table 2 and Figure 3). On the other 318 hand, a "rigid" component (τc = 6.1 ns) is dominant (80%) in the case of C7 prox , for which a less 319 abundant component (20 %) shows a "fast" behavior (τc = 0.3 ns) (Table 2 and Figure 3). The latter 320 case can be explained by considering that the nitroxide moiety resides in a well-structured region or 321 in a buried site. The high yield of labeling reached for this site (80-100%, Figure S1) suggests that 322 Cys7 is accessible, so the observed rigid behavior is indicative of the presence of a highly rigid protein 323 segment.

324
All these data experimentally confirm previously reported molecular dynamic simulations on   double-Cys variants were constructed and labeled (C7 prox /C48 prox , C7 prox /C66 prox , C48 prox /C66 prox ) and 337 their CW EPR spectra are reported in Figure S7 in the SI. For all the DEER data shown in this section, 338 the error on distance distribution results was calculated with the validation tool of DeerAnalysis (see 339 Figure S8). 55 DEER data of the C7 prox /C48 prox variant ( Figure 4A) showed well-resolved distance 340 distribution with 2 peaks centered at 2.4 and 3.5 nm, while the one obtained for C7 proxyl /C66 proxyl ( Figure 4B) and for C48 proxyl /C66 proxyl ( Figure 4C) displayed broad distance distribution. These results 342 confirm the presence of considerable conformational heterogeneity in the protein sample, supporting 343 the highly flexible behavior of HpUreG, as already observed by previous NMR studies 20 and by the 344 CW EPR data described above.

345
In all cases, the average distance distributions measured by DEER span a broader range than 346 those predicted by an MMM 56 analysis based on the HpUreG crystal structure (PDB: 4HI10) 30

381
Differently, no significant spectral changes were detected on all protein variants after the 382 addition of GTP or GDP alone ( Figure S9 and S10). We also tested the effect of Mg(II) to protein   Table 2). These results were confirmed for the HpUreG C48 prox /C66 prox variant, which 399 showed a similar behavior as found for C66 prox in glycerol ( Figure S12).

473
Variants containing two labeling sites and thus needing only one cysteine mutated were 474 designed to perform distances measurements by DEER-EPR (SI Table) were named as following:   incubated for 2 hours at 37 °C. Every 30 minutes, 40 µL from the RM were incubated with 40 µL of absorbance at 600 nm was recorded. All the experiments were reproduced two times before estimate Elexsys500 Bruker spectrometer equipped with a Super High Q sensitivity resonator operating at X 501 band (9.9 GHz). The microwaves power was 10 mW, the magnetic field modulation amplitude was 502 0.1 mT, the field sweep was 15 mT, the receiver gain was 60dB. All the samples were analyzed in 503 quartz capillaries whose sensible volume was 40 µL.

504
The spin concentration was obtained by double integration of the EPR signal obtained under 505 non-saturating conditions and the labeling yield was evaluated comparing the spin concentration 506 with that one of a standard solution. For all variants, high labeling yields were obtained ranging from 507 80% to 100% for mono labeled samples and 150-170% for double-labeled samples. 508 X-band cw EPR spectra at room temperature were recorded at 50 µM of protein concentration 509 in Tris 20mM, pH=8, NaCl=150mM. When present, Ni(II) was 2.5mM (NiSO4), GTP/GDP 3 mM 510 (Sigma-Aldrich).

511
The EPR spectra were simulated using SimLabel program, 48 a Matlab graphical user interface 512 using the Easyspin toolbox. 49 Inter-label distance distributions were obtained using the four-pulse DEER sequence. 59 515 Experiments were performed on a Bruker ELEXSYS E580 spectrometer at Q-band using the standard 516 EN 5107D2 resonator. The system was equipped with an Oxford helium temperature regulation unit 517 and the data were acquired at 60K. This temperature has been optimized according to the relaxation 518 times measured at variable temperatures in the range of 20-100 K with 10 K steps. All the