The influence of pH and divalent/monovalent cations on the internal electron transfer (IET), enzymatic activity, and structure of fructose dehydrogenase

We report on the influence of pH and monovalent/divalent cations on the catalytic current response, internal electron transfer (IET), and structure of fructose dehydrogenase (FDH) by using amperometry, spectrophotometry, and circular dichroism (CD). Amperometric measurements were performed on graphite electrodes, onto which FDH was adsorbed and the effect on the response current to fructose was investigated when varying the pH and the concentrations of divalent/monovalent cations in the contacting buffer. In the presence of 10 mM CaCl2, a current increase of up to ≈ 240% was observed, probably due to an intra-complexation reaction between Ca2+ and the aspartate/glutamate residues found at the interface between the dehydrogenase domain and the cytochrome domain of FDH. Contrary to CaCl2, addition of MgCl2 did not show any particular influence, whereas addition of monovalent cations (Na+ or K+) led to a slight linear increase in the maximum response current. To complement the amperometric investigations, spectrophotometric assays were carried out under homogeneous conditions in the presence of a 1-electron non-proton-acceptor, cytochrome c, or a 2-electron-proton acceptor, 2,6-dichloroindophenol (DCIP), respectively. In the case of cytochrome c, it was possible to observe a remarkable increase in the absorbance up to 200% when 10 mM CaCl2 was added. However, by further increasing the concentration of CaCl2 up to 50 mM and 100 mM, a decrease in the absorbance with a slight inhibition effect was observed for the highest CaCl2 concentration. Addition of MgCl2 or of the monovalent cations shows, surprisingly, no effect on the electron transfer to the electron acceptor. Contrary to the case of cytochrome c, with DCIP none of the cations tested seem to affect the rate of catalysis. In order to correlate the results obtained by amperometric and spectrophotometric measurements, CD experiments have been performed showing a great structural change of FDH when increasing the concentration CaCl2 up to 50 mM, at which the enzyme molecules start to agglomerate, hindering the substrate access to the active site probably due to a chelation reaction occurring at the enzyme surface with the glutamate/aspartate residues. Graphical Abstract Fructose dehydrogenase (FDH) consists of three subunits, but only two are involved in the electron transfer process: (I) 2e−/2H+ fructose oxidation, (II) internal electron transfer (IET), (III) direct electron transfer (DET) through 2 heme c; FDH activity either in solution or when immobilized onto an electrode surface is enhanced about 2.5-fold by adding 10 mM CaCl2 to the buffer solution, whereas MgCl2 had an “inhibition” effect. Moreover, the additions of KCl or NaCl led to a slight current increase Electronic supplementary material The online version of this article (10.1007/s00216-018-0991-0) contains supplementary material, which is available to authorized users.

D-Fructose dehydrogenase (FDH; EC 1.1.99.11) from Gluconobacter japonicus NCBR 3260 is a heterotrimeric membrane-bound enzyme complex with a molecular mass of 146.4 kDa, consisting of three subunits/domains [45]: subunit I, which is the catalytic dehydrogenase domain (DH FDH ) with a covalently bound flavin adenine dinucleotide (FAD) cofactor, where D-(-)-fructose is involved in a 2H + /2eoxidation reaction to form 5-dehydro-D-(-)-fructose; subunit II, which is equivalent to the cytochrome domain (CYT FDH ) acts as electron acceptor to subunit I and contains three heme c moieties covalently bound to the enzyme scaffold and two of them involved in the one-by-one ET pathway [45]; subunit III, which is not involved in the ET but rather plays a key role for the stability of the enzyme complex [38,45,49,50]. FDH exhibits a strict substrate specificity to D-(-)-fructose and therefore has been used as bioelectrocatalyst for biosensor development both in direct and mediated ET (DET and MET) modes [51][52][53][54].
Unfortunately, the crystal structure of FDH is not yet available because of the difficulties involved with crystallization of a membrane bound protein with a high molecular weight (ca. 146 kDa). Despite these obvious difficulties, in the last few years much effort has been directed towards considering new crystallization methods [55]. The crystal structure would be of fundamental importance to clarify the ET mechanism of this enzyme with particular attention on the co-factor involved.
The suggested ET pathway for FDH when immobilised on the electrode surface and in the absence of any competing eacceptors [38], is assumed to occur according to Scheme 1: 1. Oxidation of D-(-)-fructose to 5-keto-D-(-)-fructose involving 2e -/2H + with the reduction of FAD to FADH 2 ; 2. FADH 2 is sequentially re-oxidized in two separate 1 ET steps. In the first FADH 2 is partially re-oxidized to FADH⋅ through the IET pathway between the DH FDH and CYT FDH domains, whereby one of the three heme c (heme c 1 ) is reduced. Next, the electron is transferred from heme c 1 to a second heme c (heme c 2 ) of the two hemes involved in the ET pathway and then to a final electron acceptor, which is the electrode when FDH is adsorbed onto the electrode surface;

NOT involved in the ET mechanism
Scheme 1 Suggested ET mechanism of the oxidation reaction of FDH. D-(-)-fructose is oxidized to 5-keto-D-(-)-fructose releasing two electrons, which are transferred one by one initially through the FAD, followed by two heme c working as monoelectronic acceptors 3. FADH⋅ is finally re-oxidized to FAD by heme c 1 and the electron is then transferred to heme c 2 (which gives the second internal electron transfer (IET) step), which in turn is re-oxidized by the electrode whereby FDH is returned to its fully oxidized state.
The influence of ionic strength as well as various cationic and anionic species on enzymes has already been proven to modulate enzymatic activity or stability in different ways [56,57]; e.g., CDH has two redox domains connected through a flexible linker [58][59][60], with an ET mechanism similar to FDH but with only one heme in its cytochrome domain as electron acceptor. Moreover, Sezer et al. showed an increase in terms of catalytic current of SOx by increasing the ionic strength of the buffer solution [61], while Feng et al. reported on the influence of the viscosity on the rate of the IET [62]. Nevertheless, each of these enzymes has its own and different dependency on pH and ionic strength because of the amino acids residues at the interface between the subunits connecting each other through linker regions.
The aim of this paper was to investigate whether there is any influence of pH and the concentration of various divalent/ monovalent cations on the rate of the IET and the activity and the structure of FDH. First, FDH was immobilized onto graphite electrodes by a drop-casting procedure and amperometric measurements were carried out when shifting the pH in presence of various concentrations of divalent cations (CaCl 2 and MgCl 2 ) or monovalent cations (KCl and NaCl). Furthermore, the influence of pH on the rate of the IET was investigated in the presence of the CaCl 2 concentration that yielded the highest current density for fructose. These results were compared with those obtained with the spectrophotometric assays carried out in the presence of 2,6-dichloroindophenol (DCIP, bielectronic acceptor) or of cytochrome c (cyt c, monoelectronic acceptor). Finally, circular dichroism (CD) measurements and homology modelling/docking studies were performed, confirming the hypothesis formulated on the IET of FDH.

Electrochemical measurements
Graphite rods (Alfa Aesar GmbH and Co. KG, AGKSP grade, ultra BF^purity, and 3.05 mm diameter, Karlsruhe, Germany) were polished on wet emery paper (Turfbak Durite, P1200) and then carefully rinsed with Milli-Q water [56]. The enzyme-modified electrodes were prepared by allowing 5 μL of a FDH solution (volumetric activity measured with potassium ferricyanide at pH 4.5 and found to be 420 ± 30 U mL −1 , specific activity = 250 ± 30 U mg -1 , protein concentration = 1.7 ± 0.2 mg mL −1 ) to physically adsorb on the top of the graphite rod electrodes, overnight at 4°C. The FDH modified graphite electrode, an Ag|AgCl (0.1 M KCl) reference electrode and a platinum wire counter electrode were fitted into a flow-through wall-jet electrochemical cell [63] connected to a flow injection analysis (FIA) system consisting of a peristaltic pump (Gilson, Villier-le-Bel, France) and a six-port valve electrical injector equipped with a 50 μL loop (injection volume 50 μL) (Rheodyne, Cotati, CA, USA) [63]. A potentiostat (Zäta Elektronik, Höör, Sweden) controlled the flow-through cell, and the response of the injected samples was shown on a strip chart recorder (Kipp and Zonen, Utrecht, The Netherlands) [64].

Spectrophotometric measurements
Two spectrophotometric assays were performed to measure the activity of FDH in solution as a function of both pH and the concentration of cations/ionic strength. In the first assay, the activity of the DH FDH domain was monitored measuring the time-dependent variation of the absorbance at λ = 520 nm (ε = 6.9 mM -1 cm -1 ) of a mixture containing 100 μL of 300 mM D-(-)-fructose, 780 μL of 50 mM NaAc buffer, 20 μL of enzyme solution, and 100 μL of 3 mM DCIP as twoelectrons/two protons acceptor [56]. In the second assay, the activity of FDH and the internal electron transfer (IET) between the DH FDH and CYT FDH domains was monitored by using the one-electron acceptor cyt c, which communicates only with the CYT FDH domain due to steric hindrance because of its dimensions. The absorption is followed at 550 nm (ε = 19.6 mM −1 cm −1 ) of a mixture of 20 μL of a 1 mM cyt c solution, 100 μL of 300 mM D-(-)-fructose, 860 μL of 50 mM NaAc buffer pH 4.5, and 20 μL of the enzyme solution [65]. One unit of DCIP and cyt c activity was defined as the amount of FDH that reduces 1 μmol of DCIP or cyt c, respectively, per min under the applied conditions. All measured enzyme activities are the average of three measurements at 25°C, whereas the average activity values in 50 mM NaAc buffer without additions of Bextra^KCl, NaCl, MgCl 2 , or CaCl 2 were set to 100% and the average activity values in the presence of divalent and monovalent cations in 50 mM NaAc buffer were related to those values. All measurements were carried out using a UV-Vis spectrophotometer 1800 (Shimadzu Europe GmbH, Duisburg, Germany).

Circular dichroism (CD) measurements
Changes in the secondary structure of FDH as a function of pH and addition of divalent cations to the solution were investigated through CD measurements using a CD spectrometer (J-815 Circular Dichroic Spectrometer, JASCO, Easton, MD, USA) [66]. These measurements were carried out in a 0.1 cm cuvette at 25°C using a final protein concentration of 0.15 mg mL -1 , diluted in 50 mM NaH 2 PO 4 at pH 4.5 (unfortunately we could not use NaAc because the -COOH groups adsorb the light so we needed to use a non-adsorbing medium to obtain reliable data [67]). Then, the concentration of CaCl 2 or MgCl 2 was increased in the range of 0-100 mM to receive information about any changes in the secondary structure. The obtained spectra were further fitted with a method called β-structure selection (BeStSel) that takes into account the twist of βstructures for estimation of the secondary structure. This method can reliably distinguish parallel and anti-parallel βsheets and accurately estimate the secondary structure for a broad range of proteins [68,69]. Moreover, the secondary structure components applied by the method are characteristic to the protein fold, which in turn can be predicted to the level of topology in the CATH classification from a single CD spectrum.

Results and discussion
Influence of pH and divalent/monovalent cations on the direct electron transfer (DET) reaction of FDH: electrochemical study In order to study the influence of pH on the DET reaction of FDH with graphite electrodes, we performed amperometric measurements with graphite electrodes simply modified with adsorbed FDH. The measurements were performed with 5 mM D-(-)-fructose as substrate and by varying the pH in the running buffer between 3 and 7 using NaAc buffer between pH 3 and 5.5, MOPS buffer between pH 5 and 7. Figure 1a shows the current density (J, μA cm -2 ) dependence on pH. The optimum pH was found at 4.5. The response current rapidly decreased when increasing the pH above 4.5, thus assessing no DET activity starting at pH 7. These results are in good agreement with previous values reported in the literature [76].
Next, the influence of divalent (CaCl 2 and MgCl 2 ) and monovalent (NaCl and KCl) cations was studied through registering calibration curves for D-(-)-fructose at pH 4.5 and varying the salt concentrations in the range of 0-100 mM. Figure 1b shows the dependence of maximum current, J max , on the concentration of CaCl 2 , MgCl 2 , NaCl, and KCl for FDH at pH 4.5. The J max values were obtained from the calibration curves and were set to 100% in the absence of extra added cations. All values obtained in the presence of the different cations are related to those values. Surprisingly, when increasing the concentration of CaCl 2 , a marked increase in J max was observed, with a maximum increase of 237%, corresponding to an average maximal current density of 467.4 ± 38.2 μA cm -2 for 10 mM CaCl 2 . However, at higher concentrations of CaCl 2 than 10 mM, an unexpected I max decrease was registered, yielding J max values of 140% and of 80%, at 50 mM 100 mM CaCl 2 , respectively. This particular trend for CaCl 2 is probably due to the presence of side-chain carboxyl functions of surface-exposed aspartic acid (pK a = 3.86) and glutamic acid (pK a = 4.07) residues at the interface of the DH FDH and CYT FDH domains. These -COOH groups arẽ 90% deprotonated at pH 4.5, creating a strong electrostatic repulsion between the two domains [77]. It can be supposed that a possible complexation of Ca 2+ with the carboxylic groups of the aspartic and glutamic acid residues takes place, resulting in a closer interaction and shortening the distance between the DH FDH and CYT FDH domains, thus increasing the rate of the IET [58]. On the other hand, the decrease in response signal when the concentration of Ca 2+ is higher than 10 mM can probably be ascribed to a possible saturation of the aspartate/glutamate residues available at the interface between the DH FDH and CYT FDH domains [78]. Therefore, at concentrations higher than 10 mM, Ca 2+ is now available for complexation of other aspartate/glutamate residues present on the surface of both subunits. Both the DH FDH and CYT FDH subunits show a partial positive charge at pH 4.5, which might now create a slight repulsion between the two subunits, which is, however, counter-balanced by the attractive interactions. A possible explanation for the inhibition effect observed at 100 mM CaCl 2 could be the interaction with other enzyme molecules through a Ca 2+ -bridge [79].
Contrary to the effect by addition of CaCl 2 , addition of MgCl 2 showed a 10% decrease in J max already at 10 mM MgCl 2 . This result could be ascribed to a lower affinity of Mg 2+ compared with Ca 2+ towards the carboxylic functions available at the interface of the DH FDH and CYT FDH domains. Mg 2+ is expected mainly to be involved in the complexation reaction occurring between different enzyme molecules (Mg 2+ -bridge), which will partially hinder the substrate access to the active site of FDH and will therefore show a slight inhibition behavior evident already at low MgCl 2 concentrations. Figure 1c shows that the enhanced effect on the catalytic current density by addition of CaCl 2 occurs only in the pH range between 3. enzyme activity, due to its presumed effect on the IET as described and discussed above. In a series of previous investigations, the effect of ionic strength and addition of Ca 2+ and Mg 2+ on the catalytic performance of various CDHs was investigated both in solution and when adsorbed on graphite [56,58,80,81]. The results obtained here for FDH are both similar and dissimilar with respect to those obtained for the various CDHs; e.g., for two ascomycete CDHs the addition of both Ca 2+ and Mg 2+ had a similar increasing effect on the catalytic efficiency and could even lead to a drastic shift in the pH optimum of the enzyme activity [80]. Thus, it seems as though further investigations on the background reasons for the effect by both ionic strength and by individual cations on the interactions between the various domains participating in the ET pathways in such multi-domain redox enzymes are necessary.
For the monovalent cations, further additions of both KCl and NaCl showed a slight linear increase in J max up to 100 mM, which is more evident in Fig. 1d, where the J max is plotted versus the ionic strength of the buffer. The concentration of the monovalent cations is directly related to the ionic strength, leading to a general increase in the catalytic current, as already reported in literature for many enzymes, whereas monovalent cations are usually not involved in any complexation reactions occurring at the protein surface [56].

Influence of pH and divalent/monovalent cations on the catalytic reaction of FDH: spectrophometric study
Two different spectrophotometric assays were performed to follow the catalytic reaction of FDH in solution; one with cyt c and one with DCIP, acting as mono-and bielectronic acceptors, respectively. DCIP may have a direct connection with subunit I (DH FDH ), where the catalytic oxidation of D-fructose to 5-keto-D-fructose occurs [59], whereas cyt c may interact with both subunits, namely subunits I and II. Figure 2a shows the dependence of the activity of FDH with pH using DCIP as electron acceptor, and one can see that the highest activity is found between 3 and 6 corresponding to an activity of 3086 U mL -1 at pH 5.5. For pH values higher than 6, the activity rapidly decreases, indicating a progressive inactivation of DH FDH (subunit I). The dependence of the activity of FDH with pH when using cyt c is shown in Fig. 2b. In this case the activity exhibits two pH optima: one at 4.5 and one at 7, in agreement with previous results already reported in the literature [82]. These results could be ascribed to the different mechanisms of the interaction between FDH and cyt c: (1) ET may either proceed from the active site (FAD containing subunit I, DH FDH ) via the heme groups (subunit II containing the three heme centers, CYT FDH ) toward cyt c or (2) directly from the DH FDH to cyt c. A similar conclusion was reached when investigating the ET between FDH and cyt c by cyclic voltammetry at different scan rates [83,84].
As far as for the influence of mono-and divalent cations, the results of the spectrophotometric activity assays are shown in Fig. 3a and b. The results obtained with cyt c (Fig. 3a) are comparable with those obtained with the amperometric measurements (Fig. 1b) for immobilized FDH on graphite electrodes, as Ca 2+ ions may affect the rate of the IET and therefore the rate of overall DET reaction. Figure 3b shows that there is no influence of either divalent or monovalent cations on the activity of FDH using DCIP as electron acceptor because DCIP is reduced directly by the FAD deeply buried Since the conversion rate of fructose at the FAD is not influenced much by the presence of different ions, the increase in current and cyt reduction found here may have arisen from an enhanced IET and/or a better interaction of the subunit II (heme subunit) with cyt c or the electrode.
The obtained results unequivocally confirm the influence of CaCl 2 on the rate of the IET reaction, whereas there is no influence observed on the catalytic oxidation reaction of D-(-)-fructose at the DH FDH domain by CaCl 2 or by the other cations tested.

Influence of CaCl 2 on the secondary enzyme structure
In order to confirm the results obtained with the amperometric and spectrophotometric assays, CD measurements of FDH were performed increasing the concentration CaCl 2 or MgCl 2 in the range of 0-100 mM. Figure 4a-d show the CD spectra recorded for FDH in the presence of CaCl 2 , MgCl 2 , NaCl, and KCl. It is possible to observe that the secondary structure only to some extent shows minor changes up to a concentration of 10 mM, whereas at higher concentrations of CaCl 2 the spectra reveal large changes that can be explained by agglomeration with other FDH molecules due to Ca 2+ -bridge interactions between enzyme molecules, as shown in Fig. 4a. Therefore, at low CaCl 2 concentrations, it is possible to observe an intra-complexation caused by Ca 2+ ions, which might increase the rate of the IET reaction, whereas at high CaCl 2 concentrations inter-complexation reactions occur due to aggregation of enzyme molecules caused by the high CaCl 2 concentration. Conversely, the CD spectra in the presence of low Mg 2+ concentration already showed large changes, probably due to intercomplexation reactions occurring between different enzyme molecules, as depicted in Fig. 4b. Moreover, Na + and K + did not show any influence on the secondary enzyme structure, as reported in Fig. 4c and d, because they are not involved in any chelation reactions normally occurring with divalent cations [85]. Finally, formation of FDH multimers or aggregates was clearly proven by using dynamic light scattering (DLS) measurements as reported in Figs. S1-S4 and Table S1 in the Electronic Supplementary Material (ESM).

Structural analysis and docking of FDH domains
The already characterized FAD-glucose dehydrogenase from Aspergillus flavus (4YNT) and thiosulfate dehydrogenase (TSDBA) from Marichromatium purpuratum Bas2 isolated^(5LO9) were used for modeling and docking studies of the DH FDH and CYT FDH domains, respectively. Figure 5a shows the homology models of the DH FDH and CYT FDH domains highlighting in the insets the position of the FAD cofactor (subunit I) and the three heme c (subunit II), respectively, while subunit III was not possible to model because of the absence of a similar structure. From previous studies [49] it was already stated that subunit III has no influence on the DET reaction but it is important for the overall stability of the enzymatic complex.
Moreover, the docking between the DH FDH and CYT FDH domains was needed to highlight those amino acids expected to be involved in the interaction between the two domains, viz., Asp, Glu, Lys, Arg, and Tyr. DH FDH exhibits five amino acid residues negatively charged at the interface, whereas CYT FDH contains three such amino acid residues. The occurrence of these residues could partially explain the reason for the highest rate of the IET reaction at pH 4.5, where the repulsion between the two domains is lowest, whereas at pH between 7 and 10 the repulsion between the two domains is high and will cause the low rate of the IET reaction. However, this model does not give any information about Ca 2+ -bridging interactions. Nevertheless, a concentration of 10 mM CaCl 2 results in the upper limit in terms of an enhanced IET, probably because only a few possible matches between the negatively charged amino acid residues and Ca 2+ ions are found at the interface between the DH FDH and the CYT FDH , while the number of negatively charged amino acids residues exposed on the other parts of the enzyme surface is much higher, confirming the possibility of other Ca 2+ -bridging interactions between different individual enzyme molecules allowing the aggregation process [86,87].

Conclusions
This paper demonstrates the possibility to enhance the measurable activity of FDH either in solution or when immobilized onto an electrode surface about 2.5-fold by adding 10 mM CaCl 2 to the buffer solution, whereas MgCl 2 had no such effect. Additions of KCl or NaCl led to a slight linear increase in J max of about 10%. However, CaCl 2 had an effect only at pH 4.5 because there was no shift in the optimum pH, in contrast to what was shown for ascomycete CDHs in previous papers [56,58,80,81]. Moreover, the amperometric and spectrophotometric results were confirmed through conformational changes observed in the secondary structure of FDH, revealing unfolding of the protein occurring at high CaCl 2 concentrations (50 and 100 mM) followed by aggregation.
By the homology models, it was assumed that Ca 2+ at lower concentrations was chelated by the few amino acid residues negatively charged, shortening the distance between the DH FDH and CYT FDH domains, leading to an enhanced IET, while at higher concentrations of Ca 2+ , they give Ca 2+ -bridging interactions with other enzyme molecules.
These results show that studies of the mechanism of the IET between the DH FDH and CYT FDH domains are of great interest to shed further light into the physiological functions of  Aspergillus niger FAD-glucose dehydrogenase (PDB ID: 4ynt, light blue structure); and between Subunit II of FDH (homology model) and thiosulfate dehydrogenase (tsdba) from Marichromatium purpuratum Bas2 isolated^form (PDB ID: 5lo9); (b) electrostatic map of Subunits I and II after docking obtained from the homology models