Effect of histidine protonation state on ligand binding at the ATP-binding site of human protein kinase CK2

Histidine residues contribute to numerous molecular interactions, owing to their structure with the ionizable aromatic side chain with pKa close to the physiological pH. Herein, we studied how the two histidine residues, His115 and His160 of the catalytic subunit of human protein kinase CK2, affect the binding of the halogenated heterocyclic ligands at the ATP-binding site. Thermodynamic studies on the interaction between five variants of hCK2α (WT protein and four histidine mutants) and three ionizable bromo-benzotriazoles and their conditionally non-ionizable benzimidazole counterparts were performed with nanoDSF, MST, and ITC. The results allowed us to identify the contribution of interactions involving the particular histidine residues to ligand binding. We showed that despite the well-documented hydrogen bonding/salt bridge formation dragging the anionic ligands towards Lys68, the protonated His160 also contributes to the binding of such ligands by long-range electrostatic interactions. Simultaneously, His 115 indirectly affects ligand binding, placing the hinge region in open/closed conformations.

www.nature.com/scientificreports/We have recently shown that a balance of hydrophobic and electrostatic interactions contribute predominantly to the binding of halogenated benzotriazoles at the ATP-binding site of hCK2α 19,20,39 .Five residues neighboring the CK2 ATP-binding site may contribute to the effective electrostatic potential.Among them, there is positively charged Lys68 from the N-lobe involved in the binding of α-and β-phosphate in the productive complex with the ATP.The geometry of ligands allows the anionic ones to form hydrogen bonds or salt bridges with the Lys68, while neutral ones preferably bind closer to the hinge region 40 .Our recent studies demonstrated that electrostatic interactions predominate over possible halogen/hydrogen bonding with the hinge region of human CK2α 41 .We showed that the apparent binding affinity of titratable 5,6-dibromobenzotriazole (5,6-Br 2 Bt; pK a = 6.93) 42 to hCK2α varies non-linearly with pH, and the inflection point identifies the dissociation of the triazole proton upon ligand binding 18 .However, a slight inverse trend of pH-dependent changes was observed for 4,5,6,7-tetrabromobenzotriazole (TBBt), in which anionic form predominates in the tested pH range.Such observation indicates an alternative source of variation in the apparent binding affinity, presumably associated with the ionization states of neighboring histidine residues.Two of the sixteen histidine residues are within the 5 Å distance from the bound ligand: His115 from the hinge region and His160 of the catalytic loop (see Fig. 1).
While several studies showed that hydrogen bonding/salt bridge formation drags the ligands towards Lys68 41 , nothing was reported on how His115 and His160 (or their equivalents in other PK) may affect the binding of charged ligands and how such interactions can be used to optimize Type I inhibitors (i.e., those that bind at the ATP-binding site of a protein kinase).
To address this problem, we studied the interaction between five variants of hCK2α (WT protein and four histidine mutants) and three reference ionizable bromo-benzotriazoles and their conditionally non-ionizable benzimidazole counterparts (see Fig. 2) to identify the contribution of long-range electrostatic interactions involving the particular histidine residue to ligand binding.

Figure 1.
The interaction of TBBt within the binding pocket of hCK2α (PDB:6tll).His115 and His160 are identified in two conformations, and TBBt adopts four alternative poses, of which two proximal to His115 are denoted in magenta/blue, and two closer to His160 are in orange/blue.The green arrows identify the shortest histidine-ligand contacts.The hinge region and the catalytic loop are denoted in green and red, respectively.

CK variants
Besides the WT protein, four variants of hCK2α were analyzed-two in which a single histidine residue was substituted with alanine (H115A and H160A) and two other ones in which phenylalanine was introduced to preserve the existence of the aromatic ring (H115F and H160F).These four replacements were designed based on the in silico analysis of the hCK2α structure (see Fig. 1).All these variants fold correctly, as confirmed with Circular Dichroism spectroscopy (see Fig. 3A).
We also assessed the enzymatic activity of the tested protein variants (Fig. 3B and Table 1).The H115F variant retains the activity of the WT protein, and both H160A and H160F are minutely less active, while H115A displays only vestigial activity.However, the observed decrease in the activity of the H115A variant is likely due to its reduced stability rather than weaker or non-productive substrate binding.We applied differential scanning fluorimetry (DSF) to determine the thermal stability of these four hCK2α variants and, indeed, replacing His115 with alanine shifts the thermal denaturation middle-point temperature from 46.2 down to 38.5 °C with the protein aggregation starting at 35.5 ± 1.0 °C.The replacement effects were additionally assessed in silico with the aid of FoldX.We also checked with the aid of MST that ATP binding affinities for the H160A variant and the WT protein are the same within the estimated experimental uncertainties (K diss = 1.7 ± 0.5 and 2.2 ± 0.4 μM, respectively).

Ligand binding by His160 and His115 hCK2α mutants
We used differential scanning fluorescence (nanoDSF) to study the binding of model ligands by four variants of hCK2α relative to the WT protein.We choose three benzotriazoles (5,6-Br 2 Bt,4,5,6-Br 3 Bt, and 4,5,6,7-Br 4 Bt) that differ in their physicochemical properties together with their three conditionally indissociable benzimidazole analogs (5,6-Br 2 Bz, 4,5,6-Br 3 Bz, 4,5,6,7-Br 4 Bz).Such a ligand series enables sensing protein-dependent changes in the contribution of electrostatic and hydrophobic interactions to ligand binding.The binding poses at the ATP-binding site of hCK2α are known for all three tested benzotriazoles, while for benzimidazoles, the only available crystal structure is of TBBz bound to maize CK2α (see Fig. 4).
According to the analysis of available structures of halogenated ligands with CK2, neutral ligands, irrespective of the organism (here maize or human), bind closer to the hinge, possibly forming halogen bonds with the backbone carbonyl oxygen atoms of residues from the hinge region 18,40 .However, for anionic ligands, there are some slight differences depending on the origin of CK2α.TBBt is bound in a single pose to maize CK2α, forming a salt bridge with Lys68, not halogen bonds with the hinge.By contrast, TBBt in human CK2α is bound in multiple poses, some of which allow for halogen bonding with the hinge region (Fig. 4B).Multiple poses are also observed for 4,5,6-Br 3 Bt bound to hCK2α (Fig. 4D).Interestingly, the opposite relation was identified for 5,6-Br 2 Bt.Thus, the single pose in hCK2α (close to Lys68) is confronted with the two poses for mCK2α, one with halogen bonding to the hinge and another with salt bridge formed to Lys68 in mCK2α.
The binding affinities were initially estimated from the shift of the thermal unfolding profile, ∆T m , registered for the protein in apo form and complex with a ligand (see Table 2).The observed ∆T m values for H160A are generally lower by 0.5-1 °C than those for WT protein, while subtle effects are observed for other hCK2α variants.It must be, however, mentioned that the nanoDSF method is regarded as semi-quantitative since binding affinities are temperature-dependent, and differences in the thermal stability of each variant might also contribute to ∆T m change.A balance of enthalpic and entropic contributions to the binding may also affect the thermal shift of the complex.Ligands whose binding is entropically driven usually exhibit higher ∆T m than those of the same affinity but with enthalpically driven binding.That is why we only considered the ∆T m variation exceeding 1 °C as an indicator of change in ligand binding affinity.
For both H115 variants, there is no significant difference in ΔT m for benzotriazole derivatives at pH 8.However, for the H115A, but not for H115F, an evident change in TBBz binding affinity is observed.Also, for H160 variants, DSF experiments are minutely contradictory-as for 4,5,6-Br 3 Bt, which stabilizes the H160A by ~ 1.5 °C less than WT protein, when no such difference is observed when the same ligand binds to H160F variant.These discrepancies might stem from different melting temperatures of the apo form of hCK2α variants reflecting the temperature dependence of the binding affinity.However, they can also result directly from different physicochemical properties of alanine, phenylalanine, and histidine residues.
These observations were further confirmed with Microscale Thermophoresis (MST) and, additionally, for WT protein and alanine variants, with Isothermal Titration Calorimetry (ITC).Both methods allowed the direct Table 1.The stability of WT hCK2α and its four variants, determined with DSF (middle-point denaturation temperature, T m ) and assessed in silico with FoldX change in free energy of unfolding (ΔΔG unf ), and relative enzymatic activity.3).

Protein variant T m [°C] T aggregation [°C] ΔΔG unf [kJ•mol
As reported previously 19,39 , MST-derived data displayed the second inflection point at a higher ligand concentration, indicating the second weaker binding site.This second transition should be assigned to the binding at the site located at the interface between α and β subunits of CK2, identified in crystal structures of variously brominated benzotriazoles bound by hCK2α 43 .Consequently, all MST data were analyzed according to the model of two independent binding sites.ITC analysis differs depending on ligand-protein systems.The model assuming a single binding site was found sufficient for most ligands, and results obtained with the MST and ITC methods were consistent.However, significant discrepancies between the two methods occurred for the binding of TBBt and 4,5,6-Br 3 Bt by WT, and TBBt binding by the H160A variant.For these interactions studied by ITC, the model  Table 3.The thermodynamic data for binding events monitored with MST and ITC at pH 8. a Data from Refs. 41,49recalculated as described in "Methods".
of two independent binding sites was scored better according to the F-test.Supplementary Table S2 summarizes the justification for the fitting model selection and thermodynamic parameters for the second binding.Intense aggregation of the H115A variant during the required labeling precluded the application of MST, so for this hCK2α variant, we analyzed ligand binding with ITC.Measured binding affinities for a particular ligand fall for both mutants within the limits of experimental error except for a slight difference in binding 4,5,6-Br 3 Bt.Interestingly, both H115A and H115F bind TBBt, 4,5,6-Br 3 Bt, and TBBz weaker than the WT protein, while their affinity towards 5,6-Br 2 Bt, 5,6-Br 2 Bz and 4,5,6-Br 3 Bz remains virtually unaffected (Fig. 5A).So, the binding affinity of H115 variants decreases for ligands that putatively interact with the hinge region 40,41 .Moreover, the analysis of the enthalpic and entropic components (Fig. 5C) proves that binding of all benzimidazoles to the H115A mutant is much more entropy-driven, while for two benzotriazoles possibly interacting with the hinge region, this trend is reversed.Interestingly, 5,6-Br 2 Bt, which binds distantly to the hinge, follows the trend observed for neutral benzimidazoles.
Upon His160 replacements, the binding affinities of the studied ligands do not significantly change at pH 8 (Fig. 5B), and no significant change is also observed for the enthalpy-entropy balance (Fig. 5D).

pH-dependent binding
Our recent study demonstrated that pH-dependent binding of 5,6-Br 2 Bt is due to forced ligand deprotonation upon the binding event 18 .However, a slight opposite trend was observed for TBBt binding, for which anionic form predominates at all tested pH.We had assigned this apparent variation in TBBt binding by WT protein to the pH-induced (de)protonation of the neighboring histidine residue (putatively His160, with a theoretical pK a of ~ 7) 18 .To confirm this hypothesis, we analyzed a series of DSF data collected at different pH for both histidine mutants against the values previously reported for the WT protein (Table 4 and Fig. 6).We verified whether the pH-dependent change in ΔT m can be described by simple sigmoidal relation (1): where T m0 is the estimate of melting temperature unaffected by protonation/deprotonation event, T m is a difference between T m between two extreme states, pX is the pH of the inflection point, and p is a measure of (de)protonation cooperativity (Hill coefficient).
As expected, for complexes formed with H160A and H160F, the ligand-induced changes in ΔT m remain pHindependent for TBBt at the significance level of 0.05.Contrary to the latter, for H115A complexes with TBBt, a similar trend as in WT is preserved.For H115A, the changes in ΔT m in the presence of TBBt at low pH are even more significant, reaching 12 °C and pH-dependence of ΔT m in the presence of 5,6-Br 2 Bt also minutely differs from WT.Such a discrepancy results putatively from temperature-dependent effects since the H115A variant displays a significantly lower mid-point temperature of the thermal denaturation than the WT protein.At the same time, no such differences are observed for the H115F variant.We further applied MST to analyze quantitatively the pH dependence of 5,6-Br 2 Bt and TBBt binding affinities identified in the nanoDSF.Such an approach enables precise determination of K diss values for protein-ligand systems not affected by the slight heat effect associated with TBBt-hCK2α binding that precluded the application of ITC.Results for 5,6-Br 2 Bt binding to WT protein agree with those determined previously with ITC 18 .The pH-dependent change of TBBt binding affinity is confirmed for WT protein with MST, while no such effect is observed for the H160A variant.Based on these observations, the pH-dependent change of the ligand-binding affinity is most likely associated with the binding-induced histidine protonation that can be described according to Eq. ( 2) 44 where K intd is the intrinsic dissociation constant for the neutral His160, pK a f , and pK a c are pK a values for His160 protonation in the apo protein and complex with TBBt, respectively.The same equation may be applied (as reported previously 18 ) to describe the binding-induced 5,6-Br 2 Bt deprotonation where pK a values refer to the ligand properties, and K intd is the intrinsic dissociation constant for the deprotonated ligand.The intrinsic dissociation constant for the cationic form of His160 in the presence of the neutral ligand (K intp ) may be further calculated as We tested two alternative models for each protein-ligand system, assuming either pH-dependent (according to Eq. ( 2)) or pH-independent affinities for the strongly-binding site.The model that assumed pH-dependent K d1 values (Eq.( 2)) was scored better than the alternative for both 5,6-Br 2 Bt and TBBt binding to WT protein (p < 10 -8 ).For 5,6-Br 2 Bt, we assumed pK a values determined by us recently 18 .For TBBt binding, we used the PlayMolecule ProteinPrepare application 45 to estimate His160 pK a in the apo hCK2α based on the available protein structures (1JWH and 1NA7).According to these calculations, the pK a of His160 in the apo hCK2α was estimated to be 4.7, which was used in the analysis of MST data (pK a f in Eq. ( 2)).Interestingly, the estimated from the MST data value of His160 pK a in the complex (6.9 ± 0.7; see Table 5) corresponds to the values estimated in silico with ProteinPrepare for the hCK2α complexes with TBBt (6TLL) and 5,6-Br 2 Bt (6TLP): 5.69 and 6.62, respectively.Analogous analysis performed for CK2-TBBz complex (2OXY) confirms that the pK a of His160 is unaffected by the binding of that neutral ligand.
In contrast to WT protein, TBBt binding by H160A was found pH-independent, while the alternative hypothesis (i.e., pH-dependent K diss ) has to be rejected at the significance level of 0.05.Consistently, the binding of the titratable 5,6-Br 2 Bt by the H160A varies with pH (Table 5); however, the applied model assuming binding-induced shift of ligand protonation equilibrium does not satisfactorily describe the observed effect.The (2) Table 5. pH-dependence of ligand binding by the WT CK2 and H160A mutant monitored by MST (see Eq. ( 2)). a Value calculated using PlayMolecule ProteinPrepare application 45 based on the 1JWH and 1NA7 structures.b pK values reported previously 18 .*K intp and K intd refer to protonation state: protonated (cation for His160 and neutral for 5,6Br 2 Bt) and deprotonated (neutral for His160 and anion for 5,6Br 2 Bt), respectively.

Discussion
Recent studies demonstrated that besides shape and hydrophobicity, the affinity of halogenated benzimidazoles and benzotriazoles to hCK2α (and similarly to other EPKs) is mainly driven by a balance of halogen/hydrogen bonding with the hinge region and hydrogen bonding/salt bridge formation with Lys68 18,40,46 .However, our recent studies 18 pointed out that other factors might also contribute to the binding at the ATP-binding site of hCK2α, turning attention to the proximal histidine residues.
According to the available structures of hCK2α complexes, the His115 side chain is not directly involved in any interactions with ligands bound at the ATP-binding site.However, it is located just between the two residues that may form halogen/hydrogen bonds with the ligand (Glu114 and Val116 in hCK2α).Nevertheless, substituting His115 with alanine or phenylalanine affects the binding affinities of ligands interacting with the hinge region (TBBt and TBBz among the tested ones).Inspection of the hCK2α structures shows that His115 is partially solvent exposed but interacts with proximal residues (Glu114 and Val116) to place the hinge in the desired conformation.The conformation of the hinge region may directly interfere with ligand binding as it is for human and maize ATP-binding sites of CK2.In maize protein, the hinge displays solely a single, so-called open conformation, while in human protein, both open and closed conformations exist 47 .Since the only residue difference in the hinge region is His115 (human) vs. Tyr115 (maize), this residue may be directly involved in placing the hinge in open/closed conformations.The decrease of ligand-binding affinities is mutual for H115A and H115F variants.In maize CK2, there is tyrosine at this position, the direct electrostatic or hydrogen-bonding interactions involving histidine imidazole ring, rather than those with an aromatic system, seem crucial for the hinge positioning.Change in hinge conformation for the H115A variant explains the observed variations in enthalpy-entropy balance.
His160, the second of the studied histidine residues, was even more noteworthy.It may interact with numerous neighboring residues, including electrostatic interactions with Arg47, Asp120, Lys122, and Lys158, and π-π interactions formed with proximal Phe121and Asn161, and less possibly with Phe197 (Supp Fig. S46).Depending on the protonation state of His160, the thermodynamic contribution of these interactions may differ.However, the protonation state of histidine primarily affects the binding of charged ligands.Its compulsory protonation at neutral and basic conditions decreases the apparent binding affinity towards anionic ligands.We previously showed that for WT hCK2α, the pH-dependency of apparent binding affinity of 5,6-Br 2 Bt results from bindinginduced perturbation of the pK a of this ligand, which in the tested pH range of 6.5-8.7 predominantly exists in the anionic form 42 .Interestingly, for H160A and H160F hCK2α variants, the decrease of 5,6-Br 2 Bt binding affinity at acidic conditions is even more significant, which is reasonable since in WT, the two opposite pH-dependent effects may compensate for each other.Thus, at low pH, the estimated difference in free energy of binding caused by histidine protonation is ~ 3 kJ•mol −1 (deduced from TBBt binding by WT protein), while caused by 5,6-Br 2 Bt deprotonation is from 3 to 5 kJ•mol −1 (based on 5,6-Br 2 Bt binding affinities to WT).So, at moderately acidic conditions, the unfavorable contribution of forced 5,6-Br 2 Bt deprotonation is partly compensated by accompanying histidine protonation.One can thus assume that at pH of 6.5, the net change of the free energy of binding caused by 5,6-Br 2 Bt deprotonation is up to 8 kJ•mol −1 , and this is indeed the difference in free energy of binding for H160A-5,6-Br 2 Bt (i.e., ΔG 1 (pH 6.5)-ΔG 1 (pH 8.0)).
However, for the H160A variant, the observed change of 5,6-Br 2 Bt binding affinity with pH cannot be explained, assuming binding-induced deprotonation of the ligand model (Eq.( 3)).At pH 6.5, the maximal penalty for deprotonation of 5,6-Br 2 Bt (with pK f a 6.93) is ~ 3 kJ•mol −1 , as calculated according to relation: The binding affinity of non-dissociable analog 1-CH 3 -5,6-Br 2 Bt was previously estimated with ITC to 1.56 ± 0.52 μM (− 33.1 ± 0.8 kJ•mol −1 ) 18 .K d for the binding of the neutral form of 5,6-Br 2 Bt was estimated previously to ~ 1 µM 18 , and based on results presented here, ΔG intp is − 34.9 ± 1.1 kJ•mol −1 .All these values agree with the value estimated for the H160A variant at pH 6.5 (− 34.0 ± 0.6 kJ•mol −1 ).Altogether indicates that at moderately acidic conditions, the binding of an anionic ligand is thermodynamically less favored for hCK2α variants in which a neutral residue replaces His160.It could be that this replacement makes the binding pose of 5,6-Br 2 Bt at the ATP-binding site of hCK2α pH-controllable.
To summarize, the binding of the anionic ligand by WT hCK2α is preferred not only due to hydrogen bond/ salt-bridge formation with Lys68 but also by long-range electrostatic interactions with positively charged His160.While His115 residue probably is directly involved in placing the hinge in open/closed conformations in hCK2α.

Methods
This manuscript does not involve the use of any animal or human sample/data.

Site-directed mutagenesis of hCK2α
hCK2α amino acid-substituted mutants (H115A, H115F, H160A, and H160F) were obtained via site-directed mutagenesis using the QuikChange Lightning Site-Directed Mutagenesis Kit (Agilent Technologies, Santa Clara, USA).The DNA primers used for mutagenesis are listed in Table S1.The starting DNA template pET28-hCK2α plasmid was used 39 .Following the manufacturer's guidelines, mutagenesis reactions were performed in a 50-μL

Expression and purification of WT hCK2α and mutants
The catalytic subunit of human CK2 (hCK2α) and all its variants were expressed and purified as described previously 39 .Protein sample homogeneity was routinely confirmed by gel electrophoresis.-benzimidazole (0.5 mmol, 138 mg) was dissolved in 1.5 mL of conc.nitric acid, and Bromine (0.5 mmol, 40 mg) was added dropwise.The reaction mixture was heated at 60 °C overnight.Afterward, the reaction was cooled to room temperature, and a saturated NaHCO 3 solution (10 mL) was added.Precipitate was filtered, then purified by column chromatography on silica gel using hexane/AcOEt (8:2; v/v) as eluent to give 4,5,6-tribromo-1H-benzimidazole (95 mg); yield = 53%; 1

UV-monitored titration
pK a values of the newly synthesized compounds were determined at 298 K by spectrophotometric titration in the pH range of 13-3.5.Measurements were carried out on Perkin Elmer Lambda 25 UV-VIS spectrometer in the 200-500 nm range with a thermostated cell holder.The spectra corrected for the scattering caused by solute aggregates were further globally analyzed according to the Henderson-Hasselbach formula: where ε (λ, pH) is the spectrum recorded at a given pH, and ε n (λ), ε a (λ) are the reference spectra for the neutral and dissociated forms.

Circular Dichroism spectroscopy (CD)
Circular dichroism spectra were recorded on a Jasco J-815 spectropolarimeter.Each CD spectrum was measured thrice at 25 °C in the 195-270 nm range.All the spectra were recorded using a 1-mm path-length quartz cell for the ~ 2 μM protein solution in 25 mM Tris-HCl pH 8 buffer containing 0.5 M NaClO 4 .The exact protein concentration was determined for each sample based on UV absorbance at 280 nm, assuming a molar extinction coefficient ε = 61,895 M −1 cm −1 .

Activity assay
The enzymatic activity was monitored based on the luminescence measured with the SpectraMax iD3 Multi-Mode Microplate Reader (Molecular Devices, San Jose, CA, USA), with the aid of ADP-Glo kinase assay (Promega, Walldorf, Germany).The measurements were carried in a 96-well plate in a volume of 25 μl in 20 mM Tris-HCl buffer (pH 7.5), containing 100 ng of tested hCK2α variant (1 μl), 10 μM CK2 substrate peptide RRRD-DDSDDD (Biaffin GmbH & Co KG), 10 μM ATP and 20 mM MgCl 2 .The reaction was initiated by adding the enzyme and continued for 20 min at 30 °C.For each protein variant, at least three independent experiments were performed.The activity was calculated as average with standard deviation and presented as a percent of WT protein activity.

Low-volume differential scanning fluorimetry (nanoDSF)
All experiments were carried out with constant protein and ligand concentrations of 2.5 μM and 25 μM, respectively.The effect of ligand was measured in 25 mM Tris-HCl pH 8, 0.5 NaCl buffer, and pH-dependent binding was monitored for a pH range of 6.5-8.7 in 25 mM Bis-Tris Propane, 0.5 M NaCl buffer.The samples were loaded into nanoDSF Grade Standard Capillaries (NanoTemper Technologies) and analyzed using the Prometheus NT.48 nanoDSF device (NanoTemper Technologies).Detailed experimental setup and data analysis procedure was described previously 41 .

Microscale thermophoresis (MST)
Experiments were carried out in two different buffers depending on the pH (MES for pH < 7 and HEPES for pH ≥ 7) with 0.5 M NaCl.The protein sample was initially labeled with the His-Tag Labeling Kit RED-tris-NTA ε , pH = ε n ( )•10 pH + ε a ( )•10 pK a 10 pH +10 pK a

Figure 2 .
Figure 2. Schematic diagrams of the compounds used in the study.

Figure 3 .
Figure 3. CD Spectra of the studied hCK2α variants (A) and their enzymatic activity (B).

Figure 5 .
Figure 5. Thermodynamics of the interaction of hCK2α variants with the tested ligands.Binding by the catalytic subunit of human protein kinase CK2 vs. H115 (A) and H160 mutants (B) was assessed with ITC and MST methods.Additionally, the entropy-enthalpy compensation is compared for WT vs. H115A (C) and H160A (D).

Figure 6 .
Figure 6.pH-dependence of ΔT m for the complexes of different variants of hCK2α with TBBt (A) and 5,6-Br 2 Bt (B) determined using nanoDSF.

Table 2 .
Thermal stabilization of four variants of hCK2α by the presence of tenfold excess of tested ligands, assessed by nanoDSF at pH 8. Values that changed significantly are in bold.

Table 4 .
18ermal stabilization of four variants of hCK2α by the presence of tenfold excess of tested ligands at different pH, assessed by nanoDSF.cdatafrom Ref.18.