Zinc binding of a Cys2His2-type zinc finger protein is enhanced by the interaction with DNA

Abstract Zinc finger proteins specifically recognize DNA sequences and, therefore, play a crucial role in living organisms. In this study the Zn(II)-, and DNA-binding of 1MEY#, an artificial zinc finger protein consisting of three finger units was characterized by multiple methods. Fluorimetric, circular dichroism and isothermal calorimetric titrations were applied to determine the accurate stability constant of a zinc finger protein. Assuming that all three zinc finger subunits behave identically, the obtained thermodynamic data for the Zn(II) binding were ΔHbinding site =  − (23.5 − 28.0) kcal/mol (depending on the applied protonation state of the cysteines) and logβ’pH 7.4 = 12.2 ± 0.1, being similar to those of the CP1 consensus zinc finger peptide. The specific DNA binding of the protein can be characterized by logβ’pH 7.4 = 8.20 ± 0.08, which is comparable to the affinity of the natural zinc finger proteins (Sp1, WT1, TFIIIA) toward DNA. This value is ~ 1.9 logβ’ unit higher than those determined for semi- or nonspecific DNA binding. Competitive circular dichroism and electrophoretic mobility shift measurements revealed that the conditional stability constant characteristic for Zn(II) binding of 1MEY# protein increased by 3.4 orders of magnitude in the presence of its target DNA sequence. Graphical abstract Supplementary Information The online version contains supplementary material available at 10.1007/s00775-023-01988-1.


Section S2. Gene construction
The 1MEY# zinc finger protein (ZFP) described earlier in ref. [2] was expressed using a pETM11-SUMO3 vector, which was kindly provided by Dr. Milan Kožíšek (IOCB Prague, Proteases of Human Pathogens Research Group). The ZFP gene was recloned from the original pET-16b-P-1MEY# plasmid into the new vector by standard cloning procedures. The 303 bp long P-1MEY# gene was amplified in polymerase chain reaction (PCR) by DreamTaq polymerase (Thermo Scientific) using the 5'-aaaaggatcCGGCCATATCGAAGGTC-3' forward and 5'-ttttctcgagTCCTTAAGAGGTTTTTTTACCAG-3' reverse primers allowing for BamHI and XhoI recognition sites (underlined) in the amplified gene at its 5' and 3' termini, respectively (Fig. S3). The resulted DNA fragments were double digested by BamHI and XhoI restriction endonucleases (Thermo Scientific) and treated by FastAP thermosensitive alkaline phosphatase (Thermo Scientific). In parallel, the pETM11-SUMO3 plasmid was also digested by the same enzymes. The digested DNA mixtures were purified by extraction with phenolchloroform mixture and a subsequent precipitation by ethanol [3]. The insert and the vector were mixed in ~ 100:1 molar ratio and ligated by T4 ligase (Thermo Scientific). E. coli DH5α cells were transformed, plated and cultured in 5 mL LB medium. The resulting DNA was purified using EZ-10 Spin Column Plasmid DNA Miniprep Kit (BioBasic). The success of the mutagenesis was verified by standard DNA sequencing procedure.
The pETM11-SUMO3 system introduced an N-terminal hexa-His-SUMO3 tag instead of the previously applied deca-His tag [2]. This allowed the protein to be purified by Ni(II)affinity chromatography in the same way as before, but now the affinity tag could be removed not only by the previously described Ni(II)-promoted hydrolysis [2], but also by the ULP1 protease which selectively cleaves the sequence after two glycines following the SUMO3 domain [4]. The applied ULP1 protease itself also has a hexa-His affinity tag so it can be purified with Ni(II)-affinity chromatography.

Section S3. Protein expression and purification
The 1MEY# protein and the ULP1 protease was expressed in E. coli BL21 (DE3) cells, cultured to OD600 ~ 1.0 at 37 °C for 8 h, when the protein expression was induced by IPTG (isopropyl β-D-1-thiogalactopyranoside) at 0.2 mM final concentration, at 25 °C for 8 h. Cells were cultured in standard Lysogeny Broth (LB) medium [5] containing either ampicillin (100 μg/ml final concentration) or kanamycin (50 μg/ml final concentration). Cells were harvested by centrifugation at 4000× g for 15 min at 4 °C, yielding ~ 10 g wet cells / 1 L of the bacterial culture. 10 ml 1× binding buffer (500 mM NaCl; 100 mM HEPES (pH 8.2); 5 mM imidazole; 0.1 v/v % Triton X-100) was used to resuspend 1 g of the cell pellet. Cells were then lysed by ultrasonication at 50% amplitude (10 × 30 s) using a VCX 130 PB (130 W) ultrasonic processor equipped with a 129 mm long titanium probe with a 13 mm tip diameter. The extract was centrifuged at 4000× g for 35 min at 4 °C.
The soluble fraction of the lysate was mixed with 1/20× bed volume of Ni(II)-loaded His•Bind resin (Novagen) preequilibrated with 1× binding buffer and then equilibrated for additional 1 h at 4 °C. The resin was washed three times with 2× bed volume 1× wash buffer1 (500 mM NaCl; 100 mM HEPES (pH 8.2); 50 mM imidazole), and 3 times with 2× bed volume 1× wash buffer2 (150 mM NaCl; 100 mM HEPES (pH 8.2); 60 mM imidazole). At this point the resin portions loaded with 1MEY# and ULP1 were mixed together. Zn(ClO4)2 was added to the suspension at 50 µM final concentration followed by rotation at 16 °C for 12 h to cleave the hexa-His-SUMO3 affinity tag. Zn(II)-excess was applied to make it sure that the protein remained in its holo-form during the whole procedure and thus, to protect it from the oxidation at cysteine residues. After this incubation, 100 mM HEPES (pH 7.4) was added to the resin to reach final imidazole concentration of ~ 20 mM. The rotation was continued for 1 h at 16 °C and then the supernatant containing pure 1MEY# protein was separated from the resin together with the resin-bound hexa-His-SUMO3 affinity tag and ULP1 protease. The above purification steps were monitored by SDS PAGE (Fig. S4). The buffer of the purified 1MEY# sample was exchanged to 10 mM Cl⁻-free HEPES (pH 7.4) using Amicon 3K 15 ml filters (Merck) at 4000× g, 8×30 min at 15 °C followed by filtration through a 0.22 µm, Ø = 13 mm PES filter (Merck).

. Evaluation of the Zn(II) -EDTA reference ITC titrations
As it was indicated in the manuscript, the enthalpy of the Zn(II)-EDTA binding process was determined experimentally under the same conditions (10 mM HEPES (pH 7.4)) applied in the competition reaction. ZnCl2 was titrated with EDTA in the same buffer mentioned above.
EDTA to buffer titrations were performed to obtain the dilution heat in HEPES buffer and this was subtracted from the Zn(II)-EDTA titration to obtain the baseline corrected enthalpy (ΔHITC = −16.2 kJ/mol).

Section S4.2. Evaluation of the holo 1MEY# -EDTA competition ITC titration
The refined Δ ZnEDTA enthalpy, was used to fit the competition titration data. EDTA to 1MEY# flow through 1 titrations were carried out to obtain the dilution heat, which was subtracted from the heat of the competition titration. The following equation was used for data fitting: where Δ ZnEDTA referred to the previously determined enthalpy for the ZnEDTA complex formation and Δ ITC to all other processes in the system. From the fitted Δ ITC the reaction enthalpy of Zn(II) binding to the ZF subunits of 1MEY# was calculated based on the following equation: where Δ Zn1MEY#' is the enthalpy of the Zn(II)-binding to a single subunit of 1MEY# and the enthalpy of cysteine protonation is: Δ H 1 Cys = −29.7 kJ/mol [10]. In this process, 1.064 equivalents of protons dissociate from EDTA and beside the buffer, the free cysteine thiolates of 1MEY# can be protonated as well. In theory at pH 7.4 it would be expected, that the two cysteines takes up two protons, but Blasie et al., found experimentally that for the CP1 model peptide only ~0.5 cysteine per ZF subunit was protonated probably due to the positively charged sidechains close to cysteines [11]. Due to the high sequence similarity between CP1 and the ZF units of 1MEY# (Fig. 1 a) this result was accepted in the fitting procedure. The remaining protons were supposed to interact with HEPES. Thus, the following equation was solved:

Section S5. Effect of EDTA treatment on cysteine oxidation
The ellipticity decrease cannot be addressed to cysteine oxidation, because as it was proved in earlier research, cysteine oxidation cannot occur while Zn(II) is coordinated, and after it is removed, it does not matter whether the free cysteines get oxidized. It is also demonstrated by other researchers, that the chelation reactions with ZFs can be slow [12].
Holo-ZFPs with high Zn(II)-affinity are not sensitive to Cys oxidation due to the protecting effect of Zn(II) [13,14]. under these conditions during the incubation time based on DTNB assay (Fig. S10). J. Bjerrum introduced the following equation to estimate the stepwise stability constants of a system:

Section S6. Simulation of zinc finger Zn(II)-affinity in the presence of DNA
where is the "spreading factor" and is the value of the quotient of the constants to be expected on a statistical basis [11,12]. In this case the connection between the average determined stability constant ̅ and the stepwise stability constants can be written as: where is the number of occupied binding sites, is the total number of identical binding sites.
If the binding sites are identical it can be assumed, that = 1, thus in case of 1MEY# the following equations can be written: 2 In this chapter it will be labeled as ̅ for simplicity.  [14].  ᵁ indicates the uncertain fit of the fragments due to overlapping peaks. The assignment was performed manually in FreeStyle 1.6 (Thermo Scientific)    Based on the SDS-PAGE analysis ULP1 is extremely poorly soluble in the reaction buffers, but that small amounthardly visible on the gel (lanes e, f, and g)could cleave ~50% of the hexa-His-SUMO3 affinity tag from 1MEY# during 12 h incubation (lane i). ~50% of the cleaved 1MEY# stuck to the resin as it was observed previously [29]. In general ~25% of the initial protein was purified in the supernatant as pure native 1MEY# ZFP (lane h). 60 mM imidazole concentration was optimal for batch affinity tag cleavage. This allowed for a small soluble fraction of hexa-His-SUMO3-1MEY#. Lane f shows that a small portion of affinitytagged 1MEY# can be found in the wash fraction under these conditions.   Fig. S7. a,) with the theoretical maximum intensity values (yellow dots Fig. S7. a,)). Comparison with simulated data was performed assuming that logβ'pH=7.4 of the Zn1MEY# complex is either 8 (green dots), 9 (light blue dots) or 10 (red dots). The simulations were performed by the PSEQUAD program [7].   In panel a, all peaks with higher than 5% relative abundance were assigned. These peaks were all related to apo-1MEY# in multiple charge states. In panel b, all peaks with higher than 10% relative abundance were assigned. These peaks were all related to Ni(II)1Zn(II)21MEY#

TABLES
and Ni(II)1Zn(II)11MEY#. "z" indicates the charge. Multiple peaks could be seen because these MS spectra were not deconvoluted.