Identification of Potential Cation-Interaction Sites in the N-domain of NKA
The binding of cations to the TM-domain of NKA switches the enzyme between the E1 and E2 states.[6, 24, 46–49] The structural effect of cation binding appears to extend over long distances, i.e., to all cytoplasmic domains [50]; including the N-domain [50]. However, the possibility of cation binding to any of the cytoplasmic domains with a consequent effect on function should not be ruled out. In this work, physicochemical analysis of the NKA N-domain amino acid sequence was performed and results showed an acidic isoelectric point (pI) of ∼5.05 [35]. This is due to the presence of 30 negatively (Asp and Glu) and 20 positively charged residues. In contrast, other P-type ATPases like Ca2+-ATPase contain an equal number of negatively and positively charged residues, thus resulting in neutral pI (∼7.0) [51], but cases occur where large variations are found in pI (5.2 to 8.9) [52]. In this regard, acidic proteins (those with low pI) are known to bind cations [33, 53]. The most acidic protein known in human tissues is dentin sialophosphoprotein (DSPP, pI = 3.3) [34], a protein involved in biomineralization processes. In this study, we hypothesized that the large number of negatively charged residues in the N-domain as a molecular structural-physicochemical property may play a role in NKA function/catalysis, that is, cation interaction with negatively charged residues may modulate nucleotide binding. In this regard, the identification of cation binding sites in proteins is complex, but recently there have been some advances and, software like metal ion binding site predictor (MIB) and IonCom now are available [54–56]. Hence, analysis of the N-domain amino acid sequence for cation interaction sites predicted their presence in the protein structure (Fig. 1). Notably, the Asp443 and Glu446 interaction with calcium was previously predicted in the Ca2+-ATPase [29]. Nonetheless, in addition to Asp and Glu, other amino acid residues appear to be important for interaction with cations (Fig. 1) [53]. Analysis for potassium interaction sites did not yield any results.
Intrinsic fluorescence of the NKA N-domain
In proteins, intrinsic fluorescence is a physicochemical property where the peak (λ) for maximum emission intensity depends on the presence and location of Trp residues in the three-dimensional (3D) structure [58–60]. That is, protein structural changes may be detected by variations in fluorescence intensity and by bato- and hypsochromic modifications of the spectrum [60–62]. In this regard, two Trp residues (Trp385 and Trp411) are present in the NKA N-domain hence useful for the study of the protein structure [63]. The 3D structure of the NKA N-domain was obtained using the IntFOLD online server (https://www.reading.ac.uk/bioinf/IntFOLD/) (Fig. 2) [64]; the N-domain model was essentially similar to that of the crystal structure of the entire NKA (Fig. 2A) [18, 49]. The location of both Trp385 and Trp411 was determined to be on the opposite side of the ATP binding site (Fig. 2A); 3D structure analysis showed that most of the molecular structure of these Trp residues is buried (Fig. 2B).
The N-domain synthetic gene was generated and cloned in the pET28a + expression vector, and the recombinant N-domain was expressed in Escherichia coli BL21 (DE3) and purified by affinity chromatography as described in Methods. Protein purity was determined to be ~ 95% by densitometry using ImageJ software (Fig. 3) [65, 66].
The solvent exposure of Trp residues in the NKA N-domain was determined by acrylamide-mediated quenching of intrinsic fluorescence. This assay has been used before in studies of the whole NKA, H+-ATPase, and SERCA [23, 67, 68], but also in the recombinant N-domains of SERCA and H+-ATPase [28, 29]. The intrinsic fluorescence spectrum of the purified N-domain was obtained upon excitation at a wavelength (λ) of 295 nm; as at this λ Trp residues are excited preferentially. The maximum peak for fluorescence emission of the NKA N-domain was determined to be at λ of 338 nm (Fig. 4A), which characterizes a folded protein with mostly buried Trp residues (Fig. 2); interestingly, similar results have been found in NKA (the whole protein) and the Ca2+-ATPase recombinant N-domain [29, 69].
The addition of acrylamide to the suspended NKA N-domain resulted in quenching of the protein fluorescence intensity (Fig. 4A), indicating exposure of Trp residues to the aqueous medium as in the 3D structural model (Fig. 2). The Stern-Volmer quenching constants calculated were: Ksv = 10.36 ± 0.84 M-1 and V = 0.24 ± 0.34 M-1 (Fig. 4B), the values indicated partially solvent-accessible Trp residues as described in Fig. 2 [36]. The upward slope of the curve in Fig. 4B occurs either when all Trp residues in a protein are equally solvent-accessible or when a single Trp residue contributes most to fluorescence [70].
Cation interaction mediated NKA N-domain intrinsic fluorescence quenching
In the NKA, Trp residues are mostly located in the TM-domain [71]. When in the presence of cations, NKA intrinsic fluorescence increases by ∼3% [69]. In this regard, the NKA N-domain contains only two Trp residues (Fig. 2). In the present study, the presence of cations Na+, K+, and Ca2+ resulted in intrinsic fluorescence quenching by up to ∼40% (Fig. 5A, 5C, and 5E). Sigmoid intrinsic fluorescence change (F0-F) was observed only in the presence of Na+ and K+ (Fig. 5B and 5D). The data was fitted to the Hill equation (Eq. 2) by nonlinear regression [72, 73]:
$${\text{F}}_{0}-\text{F}=\frac{\varDelta {\text{F}}_{\text{m}\text{a}\text{x}}\bullet {\text{L}}^{\text{n}}}{{\text{K}}_{0.5}^{\text{n}}+{\text{L}}^{\text{n}}}$$
2
where \({F}_{0}-F\) is the N-domain intrinsic fluorescence quenching when in the presence of the cation/ligand \(L\), \(n\) is the Hill number, \(\varDelta {F}_{max}\) is the maximum quenching of fluorescence, and \({K}_{0.5}\) is the cation concentration when ½ΔFmax is attained. Notably, K0.5 was similar for both Na+ and K+, but the Hill number was slightly different (Table 1). In contrast, in the presence of Ca2+, a hyperbolic saturation pattern was observed (Fig. 5F); i.e., suggesting classic Michaelis-Menten (n = 1) ligand site(s). In this regard, Mg2+ is known to generate structural changes at the N-domain alone and in complex with ATP [23, 74]. Notably, all cations generated similar ΔFmax in the N-domain. The conformational changes generated as a result of the binding of cations change the superficial electrostatics of the N-domain, as showed by intrinsic fluorescence quenching by iodide [23].
Table 1
N-domain interacting parameters with cations.a
Cationb
|
\({K}_{0.5}\) (μM)
|
\(\varDelta {F}_{max}\) (%)
|
Hill number (n)
|
Na+
|
1,131 ± 94.0
|
46.23 ± 3.1
|
2.3 ± 0.3
|
K+
|
1,263 ± 98.0
|
48.01 ± 3.6
|
2.9 ± 0.5
|
Ca2+
|
0.48 ± 0.01
|
48.75 ± 3.2
|
1.0 ± 0.02
|
aExperimental data from Fig. 5. The Hill equation (Eq. 2) was used to calculate interaction parameters with cations by nonlinear regression.
bThe chloride salts of the cations were used.
Nucleotide binding-mediated intrinsic fluorescence quenching of the NKA N-domain
In P-ATPases, ATP binding induces conformational changes in the N-domain structure [20, 24, 75, 76], that may also be observed in the isolated N-domain [20, 23, 28, 29]. In this study, ATP binding quenched the intrinsic fluorescence of the NKA N-domain (Fig. 6A); ATP affinity (212.7 ± 7.4 µM). ATP binding was slightly cooperative (n = 1.3) with high change in the fluorescence quenching amplitude: ΔFmax = 86% (Fig. 6B, Table 2); i.e., substantial structural changes seems to occur upon ATP binding [20, 23].
ATP binding experiment was then performed in the presence of cations Na+, K+, and Ca2+ (Fig. 7A, 7B, and 7C). Before ATP addition, the NKA N-domain was pre-incubated in a salt concentration quenching fluorescence only by 10%. Results showed that the presence of cations increased ATP affinity in the N-domain (Fig. 7D, Table 2); cooperativity for binding was also increased by the presence of Na+ and Ca2+ (Fig. 7D). In contrast, Ca2+ showed a slight decrease in the value of ΔFmax (Fig. 7D, Table 2). Fluorescence quenching results indicated therefore that the presence of cations does affect N-domain structure generating a more favorable interaction with the nucleotide (ATP). In the whole NKA enzyme, Na+ presence also increases ATP affinity [77]; in this regard, the N-domain seems to require the presence of additional structures to display a high nucleotide affinity [77]. Notably, a relatively low salt concentration was enough to observe a significant effect on ATP affinity (Fig. 7).
Table 2
ATP binding parameters to NKA N-domain in the presence of cations.a
Nucleotide + cationb
|
\({K}_{0.5}\) (μM)
|
\(\varDelta {F}_{max}\) (%)
|
Hill number \(\left(n\right)\)
|
ATP
|
212.7 ± 7.4
|
86.8 ± 1.5
|
1.32 ± 0.04
|
ATP + 0.75 mM Na+
|
133.7 ± 4.5
|
79.4 ± 5.3
|
1.49 ± 0.15
|
ATP + 0.75 mM K+
|
117.7 ± 5.6
|
77.1 ± 2.0
|
1.38 ± 0.07
|
ATP + 0.05 µM Ca2+
|
153.7 ± 9.3
|
70.6 ± 2.7
|
1.56 ± 0.09
|
aExperimental data from Fig. 7. The Hill equation (Eq. 2) was used to calculate binding parameters by nonlinear regression.
bThe chloride salts of the cations were used.
Analysis in silico of NKA N-domain in complex with ligands (ATP and FITC)
FITC (a fluorescent covalent label for P-ATPases) is known to react with either of two Lys residues located in the nucleotide-binding site [78, 79]. For Eosin Y, the molecular interactions with NKA are known [80]. However, molecular information on the FITC interaction with NKA is known only partially [81]. Thus, molecular docking of FITC to the 3D structure of the NKA N-domain was performed using the AutoDock Suite 1.5.6 toolkit (https://vina.scripps.edu/) [39]. First, ATP docking to the N-domain 3D model (Fig. 2) was performed as control: ATP molecular docking was similar to that already published for the NKA (Fig. 8A) [20].
The nucleotide-binding site is located between the boundaries of the A-, N-, and P-domains (Fig. 8A). The ATP::N-domain complex is stabilized by interactions with a relatively large number of amino acid residues; namely: Asp443, Ser445, Phe475, Lys501, Glu505, Tyr535, Arg544, and Val545 (Fig. 8B). The adenine group displays a π−π-stacking interaction with Phe475, and three hydrogen bonds with Asp443, Ser445, and Lys501 each. The ribose group interacts with Val545. Notably, the phosphate group shows a network of molecular interactions: 1) three H-bonds with Glu505; 2) two H-bonds with Val545; 3) one H-bond with Tyr535. Finally, a salt bridge links α-phosphate and Arg544; it is known to be important in the binding of both ATP and ADP through the stabilization of the phosphate moiety [82]. Importantly, published work has shown that mutations of Glu446, Phe475, Gln482 and Arg544 result in loss of ATP-binding capacity [21].
FITC chemical reaction occurs between the isothiocyanate group (-N = C = S) and the ε-amino group of a Lys residue [25]. In this regard, Lys480 and Lys501 have been identified as reactive to FITC [25, 78, 83], but only one occurs at a time [84], and importantly, Lys480 appears as non-essential for nucleotide binding [85]. Notably, FITC molecular docking to the N-domain showed two binding modes (I and II) where the isothiocyanate group was located close to Lys480 and Lys501 (Fig. 9A and 9B). In binding mode I (Fig. 9A), molecular interactions with amino acid residues were mainly hydrophobic with Asp443, Glu446, Phe475, and Leu546, and two H-bonds with Ala444 and Lys501. In binding mode II (Fig. 9B), FITC molecular interaction with amino acid residues was hydrophobic with Phe475 and Leu546, one H-bond with Ser445, and established a salt bridge with Lys501. Both FITC::ND-domain complexes were determined to be stable by molecular dynamics simulations (not shown).
FITC labeling of NKA N-domain in the presence of ATP
Labeling reaction of the NKA with FITC is known to be hampered by the presence of ATP [84]; as both bind to the same site (Fig. 8 and Fig. 9) [14, 25, 78]. The recombinant NKA N-domain was labeled with FITC and as in the whole NKA, the reaction was prevented by the presence of physiological concentrations of ATP (Fig. 10); clear gels showed a decrease in the FITC labeling of the N-domain as a function of ATP concentration; ∼80% in labeling was prevented by the presence of 4 mM ATP (Fig. 10). Similar results were obtained when in the presence of eosin Y (Not shown). Importantly, SDS denaturation of the N-domain lead to lost the reactivity toward FITC (Fig. 10). Therefore, an intact nucleotide binding site structure appears to be critical for FITC binding (Fig. 9). Hence, an apparent mechanism for FITC labeling of NKA N-domain is described by Eq. 3 (Eq. 3):
$$\text{F}\text{I}\text{T}\text{C}+\text{N}-\text{d}\text{o}\text{m}\text{a}\text{i}\text{n}\underleftrightarrow{{K}_{1}}\text{F}\text{I}\text{T}\text{C}:\text{N}-\text{d}\text{o}\text{m}\text{a}\text{i}\text{n}\underleftrightarrow{{K}_{2}}\text{F}\text{I}\text{T}\text{C}\colon\colon \text{N}-\text{d}\text{o}\text{m}\text{a}\text{i}\text{n} \underrightarrow{k}\text{F}\text{I}\text{T}\text{C}-\text{N}-\text{d}\text{o}\text{m}\text{a}\text{i}\text{n}$$
3
FITC labeling of NKA N-domain in the presence of cations.
The effect of the presence of cations (Na+, K+, and Ca2+) on the reactivity of FITC to the NKA N-domain was tested. A significant decrease (~ 30%) in FITC labeling was observed in the presence of 150 mM NaCl (Fig. 11A). Hence, Na+ presence seems to diminish FITC affinity in the nucleotide-binding site. The presence of K+ displayed similar results at the same concentration (Fig. 11B). There appears to be a structural effect of these cations in the N-domain that results in a decrease in the affinity of the nucleotide-binding site for FITC and ATP (Fig. 10 and Fig. 11). In contrast, the presence of Tris increases the affinity for eosin Y, thereby decreasing FITC labeling (not shown). Such an effect of cations (particularly Na+) on the affinity of the nucleotide-binding site could potentially affect function/catalysis. In the cell cytoplasm, Na+ and K+ are required in a specific ratio (120 and 12 mM for Na+ and K+, respectively) for NKA activity. In the absence of K+, Na+ can either inhibit or stimulate ATP-ADP exchange depending on its concentration [86, 87].
The effect of the presence of Ca2+ in FITC labeling of the NKA N-domain was non-significant (Fig. 12); only a slight decrease of ∼10% was observed in the presence of 200 nM CaCl2 (Fig. 12). However, it is known that calcium causes inhibition of ATP hydrolysis [14]. In the activity of para-nitrophenylphosphate phosphatase (pNPPase), K+ may be partially (20%) substituted by Ca2+ resulting in a decrease of the fluorescence signal of FITC bound to NKA [14].