2025 Volume 72 Issue 2 Article ID: 7202108
Enzymes and cofactors interactions play a significant role in enzymatic function. Particularly, the covalent bonds between proteins and flavin cofactors are important for enzymatic activity and redox potential in covalent flavoproteins. For example, in pyranose 2-oxidase from the basidiomycete Phanerochaete chrysosporium (PcPOx), the flavin adenine dinucleotide (FAD) cofactor forms a covalent bond with histidine (His158), while FAD in other flavoproteins can form a covalent bond with other amino acid residues, such as cysteine, tyrosine, and aspartic acid. Considering the mechanism of forming a covalent bond with FAD, new covalent FAD patterns in PcPOx were expected. Here, we explored the potential for amino acids other than histidine to covalently bind FAD in PcPOx by conducting comprehensive site-directed mutagenesis at His158, and evaluated 19 mutants for covalent-bond-forming ability with FAD, as well as for oxidase and dehydrogenase activities towards D-glucose. All the mutants failed to form a covalent bond with FAD, though they could bind FAD noncovalently to various extents, except for H158D and H158P, which lost not only the covalent bonds with FAD but also the whole of FAD cofactors. The His158 variants showed markedly reduced both the oxidase and dehydrogenase activity toward D-glucose compared with the wild-type enzyme. Moreover, the apo-enzymes H158D and H158P were inactive. Our findings are expected to be helpful in the design of artificial cofactors for flavoproteins.
DCPIP, dichlorophenolindophenol; FAD, flavin adenine dinucleotide; FMN, flavin mononucleotide; Pc, Phanerochaete chrysosporium; PDH, pyranose dehydrogenase; PMS, 1-methoxy-5-methylphenazinium methyl sulfate; POx, pyranose 2-oxidase; To, Trametes ochracea; VAO, vanillyl-alcohol oxidase; WT, wild type.
Prosthetic groups are important as electron acceptors and electron donors for biocatalytic oxidation. Some oxidoreductases form covalent bonds with cofactors, and such interactions play a significant role in the enzymatic function of heme- [1] and flavin-dependent proteins [2]. Flavin cofactors include flavin adenine dinucleotide (FAD) and flavin mononucleotide (FMN), both of which are derivatives of vitamin B2. Although a majority of flavoproteins bind noncovalently with FAD or FMN, some enzymes form covalent linkages with their flavin cofactors, and this significantly alters their properties [2]. Indeed, if the covalent bond with flavin is lost in covalently bound flavoproteins, redox potentials and the enzymatic activities decrease [2, 3]. Interestingly, a recent study succeeded in forming an artificial covalent bond between a flavoprotein and FMN [4], and this resulted in increased thermostability and catalytic performance. Therefore, it is important to understand the requirements for covalent flavin formation in order to engineer flavoproteins for applications in rare sugar synthesis [5], biosensors [6], and biofuel cells [7].
Flavoproteins in which various amino acid residues such as histidine [3, 8], cysteine [9], tyrosine [10], and aspartic acid [11] are linked to C8α of the isoalloxazine ring of FAD have been reported (Fig. 1). Cysteine can also be attached to C6 of the isoalloxazine ring of FAD [12]. The amino acid forming a covalent bond appears to be highly specific to each flavoprotein, because mutation of the amino acid linked with FAD is reported to cause loss of the covalent bond [3, 13]. However, in those studies, only a few amino acid mutations were examined [3, 13]. Thus, there is a need for a comprehensive analysis of 19 variants to determine whether covalent flavin patterns in flavoproteins are specific or variable.
FAD is attached to each amino acid colored in cyan. (A): VAO from Myceliophthora thermophila (MtVAO713) (PDB ID 6F74), (B): PcPOx (PDB ID 4MIF), (C): fumisoquin biosynthesis gene B (PDB ID 6GG2), (D): p-cresol methylhydroxylase (PDB ID 1WVE), (E): chloramphenicol halogenase (PDB ID 3I3L), and (F): glucooligosaccharide oxidase (PDB ID 1ZR6). The 8α-N1-histidyl bond in (F) is hidden for visualization purposes. The figures were prepared using PyMOL Molecular Graphics System, Version 1.20 (Schrödinger, LLC, New York, NY, USA).
Here, we focused on the covalent flavoprotein pyranose 2-oxidase (POx, EC 1.1.3.10) [14, 15], which catalyzes the oxidation of diverse sugars derived from lignocellulose, such as glucose, xylose, galactose, and melibiose, at the C2 position [14]. POx can also catalyze oxidase reactions at the C3 position when 2-deoxy-D-glucose and methyl-β-D-glucoside are provided [14]. POx exhibits both oxidase and dehydrogenase activities toward sugar substrates, suggesting that its natural roles include producing hydrogen peroxide for peroxidases and reducing quinones or manganese (Ⅲ) ions for lignin degradation in filamentous fungi [15, 16] and bacteria [17]. POx contains an 8α-N3-histidyl-FAD bond [18, 19], and mutation of the histidine to alanine results in loss of the covalent bond, accompanied with a decrease in the kcat value towards D-glucose from 50.5 s−1 for the recombinant wild-type (WT) POx from Trametes ochracea (ToPOx; previously named Trametes multicolor) to 8.83 s−1 without any additional flavin [20]. The redox potential of the ToPOx mutant was also decreased from −115 ± 3 to −147 ± 2 mV [20]. The covalent bond in ToPOx is formed by an autocatalytic reaction [21] after protein folding, and this process, known as the quinone methide mechanism, is thought to be a common pathway for covalent bond formation in flavoproteins [2]. Consequently, some covalent flavin patterns might also occur in POx.
In the current study, we focused on POx from the basidiomycete Phanerochaete chrysosporium (PcPOx) [22, 23], because PcPOx exhibits greater thermal stability than other fungal POxs [24], making PcPOx a promising candidate for engineering it for various applications. The crystal structure of PcPOx has been determined [25], including the covalent bond between the residue His158 and FAD. We considered that if amino acids other than histidine can form a covalent bond with FAD in PcPOx, it might be possible to tune the enzymatic activities by varying the amino acid linked with FAD. This might form the basis of a new strategy for flavoprotein engineering.
The aim of this work was to test this idea by conducting comprehensive site-directed mutagenesis at His158 of PcPOx to replace the histidine residue with 19 other amino acids. We then examined whether or not the mutants formed a covalent bond with FAD, and we also evaluated their oxidase and dehydrogenase activities towards D-glucose, the most preferred sugar substrate of PcPOx [22].
Preparation of WT PcPOx. The plasmid containing pET-27b(+) vector (Merck Biosciences Inc., Darmstadt, Germany) and the WT PcPOx gene [26] were transformed into ECOSTM SONIC Competent E. coli strain BL21 (DE3) (Nippon Gene Co., Ltd., Tokyo, Japan) by heat-shock.
For PcPOx expression, a selected single colony was inoculated into 5 mL liquid LB medium containing 50 μg/mL kanamycin and shaken at 37 °C and 300 rpm overnight. After precultivation, 1.2 mL of the culture solution was added to 60 mL liquid LB medium containing 0.02 mM riboflavin and shaken (200 rpm) at 37 °C for 2 h. Then, PcPOx expression was induced with 0.2 mM isopropyl-β-D-thiogalactopyranoside (IPTG) for 3.5 h at 26.5 °C and 200 rpm. Cells were harvested by centrifugation at 3,000 × G for 5 min at 4 °C. The pellet was re-suspended in 1.8 mL of binding buffer; 20 mM Tris-HCl (pH 7.4) containing 500 mM NaCl and 20 mM imidazole. The cells were disrupted in an ultrasonic homogenizer (VP-300N, TAITEC Corporation, Saitama, Japan). The homogenate solution was centrifuged at 12,000 × G for 15 min at 4 °C. The supernatant was applied to a His-spin column (GE HealthCare Technologies Inc., Chicago, IL, USA) pre-equilibrated with binding buffer, and the column was washed with binding buffer. PcPOx was eluted with elution buffer; 20 mM Tris-HCl (pH 7.4) containing 500 mM NaCl and 500 mM imidazole. The protein concentration was determined using Bio-Rad Protein Assay Dye Reagent Concentrate (Bio-Rad Laboratories, Inc., Hercules, CA, USA). The absorbance at 595 nm was measured with a Multiskan SkyHigh instrument (Thermo Fisher Scientific Inc., Waltham, MA, USA).
Preparation of PcPOx variants. Inverse PCR was employed for site-directed mutagenesis, using the template plasmid containing pET-27b(+) vector and the WT PcPOx. Primers are listed in Table S1 (see J. Appl. Glycosci. Web site). DNA polymerase PrimeSTAR Max (TaKaRa Bio Inc., Shiga, Japan) and KOD One® PCR Master Mix -Blue- (Toyobo Co., Ltd., Osaka, Japan) were used. The amplified gene was purified with the Wizard SV Gel PCR Clean-Up System (Promega Corporation, Madison, WI, USA). After ligation with T4 kinase and Ligation High (Toyobo Co., Ltd.), the mutated plasmids were transformed into SONIC BL21 (DE3) by heat-shock. The mutations were confirmed by sequencing. Expression and purification of PcPOx variants were conducted as described above.
SDS-PAGE and UV-visible spectra. SDS-PAGE was carried out as described by Laemmli [27] using 12 % polyacrylamide gel. The gel was stained with SimplyBlueTM SafeStain (Thermo Fisher Scientific Inc.). Covalent flavinylation was identified as described previously [28, 29]. Briefly, SDS-PAGE gel was incubated in ice-cold 10 % acetic acid for 15 min and fluorescence was detected under 365 nm illumination using a T12621D UV transilluminator (Pacific Image Electronics Co., Ltd., New Taipei City, Taiwan).
UV-visible spectra of 7.1 μM (0.5 mg/mL) PcPOx WT and His158 variants were recorded at room temperature using Jasco V660 spectrophotometer (Jasco Ltd., Tokyo, Japan). The concentration of FAD disodium salt was calculated by using an absorption coefficient of 11,300 M−1cm−1 at 450 nm [30]. The spectra were drawn, and curve fittings were performed and plotted using Igor Pro 9 (WaveMetrics, Inc., Portland, OR, USA). After measuring UV-visible spectra, the protein concentrations of WT, H158A, and H158P were re-determined using the molar extinction coefficient estimated by the Protein Calculator ver. 3.4 online tool [31].
Assay of oxidase activity toward D-glucose. Oxidase activity toward D-glucose was determined by measuring H2O2 production in a horseradish peroxidase assay. The reaction mixture contained 50 mM sodium phosphate buffer pH 7.0, 0.1 mM 4-aminoantipyrine, 1 mM N-ethyl-N-(2-hydroxy-3-sulfopropyl)-3-methylaniline, sodium salt (TOOS), 4 U/mL horseradish peroxidase, D-glucose, and 14 nM PcPOx WT or 140 nM PcPOx His158 mutant. The absorbance at 555 nm was monitored at 26.5 °C. Kinetic parameters were calculated by Kitaoka’s method [32] using Microsoft Excel (Microsoft Corporation, Redmond, WA, USA).
Assay of dehydrogenase activity toward D-glucose. Dehydrogenase activities were determined using dichlorophenolindophenol (DCPIP) assay as reported by Fujii et al. [26] and Ozawa et al. [33]. The concentration of DCPIP was calculated by using an absorption coefficient of 16.3 mM−1cm−1 at 600 nm [33]. The reaction mixture for DCPIP assay contained 50 mM sodium phosphate buffer pH 7.0, 0.1 mM DCPIP, glucose, and 36 nM PcPOx WT or 360 nM PcPOx His158 variant. The absorbance at 600 nm was monitored at 26.5 °C and pH 7.0. The absorbance at 520 nm was monitored in the case of DCPIP concentrations over 0.5 mM.
We conducted not only DCPIP assay alone, but also DCPIP assay with the addition of 1-methoxy-5-methylphenazinium methyl sulfate (PMS) as an electron mediator. The reaction mixture for DCPIP&PMS assay contained 50 mM sodium phosphate buffer pH 7.0, 0.1 mM DCPIP, 0.2 mM PMS, D-glucose, and 36 nM PcPOx WT or 360 nM PcPOx His158 variant. The absorbance at 600 nm was monitored at 26.5 °C and pH 7.0. Kinetic parameters were calculated using the same method as described for the oxidase activity. Specific activities were calculated based on the decrease in the absorbance at 600 nm over 30 min under 50 mM D-glucose. Deaeration was not conducted in the dehydrogenase activity assays.
Expression of PcPOx WT and His158 variants.
As shown in Fig. 2A, PcPOx WT and the His158 variants each exhibited a single band with a molecular mass of 65 kDa on SDS-PAGE gel [23]. The amounts of purified His158 variants were comparable to that of WT, indicating that the mutation of His158 did not alter the expression level of PcPOx.
(A) SDS-PAGE analysis of PcPOx WT and 19 H158 mutants. The single uppercase letters represent single-letter abbreviations for amino acid substitutions at His158. “(M)” denotes standard molecular weight marker proteins. (B) SDS-PAGE gel under 365 nm transillumination after incubation in ice-cold 10 % acetic acid for 15 min. The fluorescence of covalently bound FAD was detected. Protein samples were overloaded (80-100 μg/lane) to aid fluorescence detection. Lane nomenclature is to the same as in (A). The image was converted to monochrome, and black and white were inverted.
All protein samples were obtained from cultures in LB medium supplemented with 0.02 mM riboflavin, in line with previous work [26]. However, our experiments revealed that the expression level and oxidase activity toward D-glucose in PcPOx remained unaffected when riboflavin was omitted from the culture (see Fig. S1; see J. Appl. Glycosci. Web site).
Identification of covalent FAD.
Only PcPOx WT exhibited the fluorescence of covalent FAD in SDS-PAGE gel (Fig. 2B). During SDS-PAGE analysis, non-covalent FAD was released from protein samples upon boiling, and therefore only covalent FAD, which remained attached to the protein, exhibited fluorescence in the gel bands. To detect 8α-S-cysteinyl flavins, SDS-PAGE gel should be sprayed with performic acid before incubation in diluted acetic acid, thereby eliminating the internal quenching of fluorescence associated with flavin derivatives [28]. After the extra treatment and incubation in 10 % (v/v) acetic acid, PcPOx variants H158C and H158M showed no fluorescence due to 8α-S-cysteinyl flavin, while the released flavin showed fluorescence in the downstream SDS-PAGE gel (see Fig. S2; see J. Appl. Glycosci. Web site).
The UV-visible spectra of free FAD solution displayed maxima at 376 and 450 nm (Fig. 3A, B). These typical flavin peaks were observed in PcPOx WT and the His158 variants. PcPOx WT showed a blue shift of the former peak to 359 nm (Fig. 3C, D) compared with 386 nm in the case of H158A (Fig. 3E, F), which is consistent with covalent bond formation between protein and FAD, as reported for other covalent flavoproteins such as vanillyl-alcohol oxidase (VAO) [3] and choline oxidase [34]. The PcPOx H158P variant did not show any peaks derived from FAD (Fig. 3G, H), and neither did H158D (data not shown). We conducted a total of three cultivations for the preparations of H158P and H158D, but they always lacked flavin cofactors. The protein concentrations determined by the absorbance at 280 nm were 0.53 mg/mL for PcPOx WT, 0.33 mg/mL for H158A, and 0.31 mg/mL for H158P.
UV-visible spectra of (A, B) flavin adenine dinucleotide disodium salt, (C, D) PcPOx WT, (E, F) PcPOx H158A, and (G, H) PcPOx H158P. The region of the flavin absorption peak is magnified in (B, D, F, H). The protein concentrations determined by the absorbance at 280 nm were 0.53 mg/mL for PcPOx WT, 0.33 mg/mL for PcPOx H158A, and 0.31 mg/mL for PcPOx H158P. The elution buffer consisted of 20 mM Tris-HCl (pH 7.4), 500 mM NaCl, and 500 mM imidazole. Flavin adenine dinucleotide disodium salt was dissolved in Milli-Q water to a concentration of 140 μM. Each dotted line represents the peak at around 370 nm in the spectra of (B), (D) and (F). The blue dotted lines indicate the peak at around 359 nm of PcPOx WT. The green dotted lines indicate the peak of free FAD at 376 nm. The orange dotted lines indicate the peak of PcPOx H158A at 386 nm.
Thus, all of the PcPOx variants except for H158D and H158P exhibited typical flavin peaks at around 385 nm and 450 nm (Table S2; see J. Appl. Glycosci. Web site). The variants generally showed red shifts of the peak corresponding to the 376 nm peak of free FAD, though H158R showed a blue shift (368 nm). However, this blue shift was not derived from covalent FAD, since H158R did not exhibit the fluorescence of covalent FAD in SDS-PAGE gel. All POx solutions showed a yellow color, like that of free FAD, except for H158P and H158D. H158P and H158D solutions were colorless and transparent, suggesting that the two variants were apo proteins. Reinheitszahl (Rz)-value represented the ratio of the absorbance at the peak top derived from flavins (around 450 nm) to the absorbance at the peak top derived from proteins (around 280 nm), indicating the FAD / protein ratio. For comparison of Rz-values of PcPOxs, we used 1/Rz values as well as the previous work on VAO [3]. While the 1/Rz value of PcPOx WT was 10.8, those of the PcPOx variants, except for H158P and H158D, were below 13.4, indicating that more than 80 % of the protein retained FAD on any 17 variants. Although 1/Rz values of H158P and H158D could not be theoretically determined due to the absence of any peak top derived from flavin, the ratio of the absorbance at the peak top derived from proteins (around 280 nm) to the absorbance at the peak top at 450 nm (peak top of free FAD) could be calculated: 9.8 for H158P and 12.4 for H158D. We did not find evidence explaining why the 1/Rz values of the two variants H158P and H158D resembled those of flavin-containing PcPOx WT and other His158 variants. To elucidate the effect of background interference, we performed peak fitting of the spectra (see Fig. S3; see J. Appl. Glycosci. Web site) and recalculated as the modified Reinheitszahl (Rz′)-value using the two peaks at approximately 280 and 450 nm identified through UV-visible spectral analysis (see Table S2; see J. Appl. Glycosci. Web site). The peak fitting did not reveal any flavin-derived peak (around 450 nm) in the UV-visible spectra of the two apo variants. These results suggest that background interference affected to the 1/Rz values of the two apo variants.
Overall, these results indicate that only PcPOx WT contains covalently bound FAD, while other mutants such as H158A contain non-covalently bound FAD, and H158P and H158D do not contain FAD.
Oxidase activity towards D-glucose.
Although oxidase activity toward D-glucose of histidine-tagged POx from Trametes ochracea is enhanced by chelation treatment with EDTA to eliminate the inhibitory effect of Ni2＋ [20], that of PcPOx decreased after a similar chelation treatment (Fig. S4; see J. Appl. Glycosci. Web site). Therefore, all activity tests here were conducted with untreated protein samples. Oxidase activity toward D-glucose of PcPOx was not affected by the addition of free FAD to the mixture. This was confirmed using PcPOx WT and variant H158G, which is considered to be the mutant with the smallest steric hindrance (Fig. S5; see J. Appl. Glycosci. Web site).
Oxidase activity toward D-glucose of all PcPOx His158 variants was less than 10 % of that of WT in terms of kcat value. Even the most active variant H158Y exhibited only 8 % of WT’s kcat value (Table 1 & Fig. S6; see J. Appl. Glycosci. Web site). H158P and H158D, which lack the FAD cofactor, exhibited kcat values of less than 0.05 s−1. These two apo variants also remained inactive after the addition of 50 µM free FAD to the reaction mixture. While all the mutants exhibited higher Km values than WT, the aromatic amino acid mutants - H158W, H158Y, and H158F - had the highest values (Fig. 4). Notably, H158W and H158Y showed over six-fold higher Km values, suggesting a substantially reduced substrate affinity.
Table 1. Steady-state kinetic parameters of oxidase activity toward D-glucose in PcPOx and the mutants. The symbol ± represents the standard error, calculated from data collected in triplicate at each substrate concentration.
| WT | H158A | H158C | H158D | H158E | |
| Km (mM) | 1.73 ± 0.38 | 2.25 ± 0.25 | 2.89 ± 0.56 | N.D. | 2.65 ± 0.42 |
| kcat (s−1) | 10.5 ± 0.7 | 0.47 ± 0.02 | 0.40 ± 0.02 | N.D. | 0.24 ± 0.01 |
| kcat/Km (s−1mM−1) | 6.1 ± 1.1 | 0.21 ± 0.02 | 0.14 ± 0.02 | N.D. | 0.09 ± 0.01 |
| H158F | H158G | H158I | H158K | H158L | |
| Km (mM) | 6.33 ± 0.98 | 6.02 ± 1.66 | 2.14 ± 0.28 | 2.32 ± 0.51 | 3.48 ± 0.27 |
| kcat (s−1) | 0.41 ± 0.03 | 0.24 ± 0.03 | 0.21 ± 0.01 | 0.55 ± 0.04 | 0.49 ± 0.01 |
| kcat/Km (s−1mM−1) | 0.06 ± 0.01 | 0.04 ± 0.01 | 0.10 ± 0.01 | 0.24 ± 0.04 | 0.14 ± 0.01 |
| H158M | H158N | H158P | H158Q | H158R | |
| Km (mM) | 3.53 ± 0.40 | 2.56 ± 0.27 | N.D. | 3.79 ± 0.59 | 4.36 ± 1.42 |
| kcat (s−1) | 0.53 ± 0.02 | 0.16 ± 0.01 | N.D. | 0.33 ± 0.02 | 0.42 ± 0.05 |
| kcat/Km (s−1mM−1) | 0.15 ± 0.01 | 0.06 ± 0.01 | N.D. | 0.09 ± 0.01 | 0.10 ± 0.02 |
| H158S | H158T | H158V | H158W | H158Y | |
| Km (mM) | 2.54 ± 0.59 | 1.83 ± 0.48 | 2.79 ± 0.60 | 12.6 ± 2.6 | 11.3 ± 3.0 |
| kcat (s−1) | 0.22 ± 0.02 | 0.18 ± 0.01 | 0.30 ± 0.02 | 0.42 ± 0.04 | 0.85 ± 0.11 |
| kcat/Km (s−1mM−1) | 0.09 ± 0.02 | 0.10 ± 0.02 | 0.11 ± 0.02 | 0.03 ± 0.00 | 0.07 ± 0.01 |
n.d., not determined due to extremely low activity.
The vertical axis indicates Km values extracted from Table 1. The symbols used are consistent with those in Fig. 2. Error bars represent standard error of three repeats. H158D and H158P are not shown, as their affinity was too low.
Dehydrogenase activity towards D-glucose.
In PcPOx WT, the kcat value of dehydrogenase activity toward D-glucose was only 13 % of that of the oxidase activity when the concentration of DCPIP was 0.1 mM (Table 2 & Fig. S7; see J. Appl. Glycosci. Web site). Although the concentration of DCPIP was set at an equivalent level to that of dissolved oxygen (lower than 0.25 mM), the turnover of dehydrogenase activity became comparable of that of oxidase activity when the concentration of DCPIP was sufficiently high. We obtained only the specific dehydrogenase activity toward D-glucose of PcPOx His158 mutants since steady-state kinetic parameters were not obtained because of too low activity (kcat values < 0.1). As in the case of oxidase activity, the glucose dehydrogenase activity of all mutants was decreased compared to WT (Fig. 5). The ratios of dehydrogenase activity to oxidase activity in the presence of 50 mM D-glucose were shown in Table S3 (see J. Appl. Glycosci. Web site).
Table 2. Steady-state kinetic parameters of dehydrogenase activity toward D-glucose of PcPOx WT. The symbol ± represents the standard error, calculated from data collected in triplicate at each substrate concentration.
| Fixed substrate | Varied substrate | Km (mM) | kcat (s−1) | kcat/Km (s−1mM−1) |
| D-glucose (50 mM) as e- doner | DCPIP | 0.55 ± 0.07 | 11.2 ± 0.8 | 20 ± 1 |
| DCPIP (0.1 mM) as e- acceptor | D-glucose | 1.29 ± 0.12 | 1.41 ± 0.04 | 1.1 ± 0.1 |
| DCPIP (0.1 mM) as e- acceptor PMS (0.2 mM) as e- mediator | D-glucose | 1.13 ± 0.05 | 1.53 ± 0.02 | 1.4 ± 0.0 |
Specific dehydrogenase activity toward D-glucose of PcPOx WT and His158 variants. The symbols used are consistent with those in Fig. 2. Blue bars represent activity using DCPIP as an electron acceptor. Orange bars illustrate activity with DCPIP as an electron acceptor and PMS as an electron mediator. Error bars represent standard error of three repeats. H158D and H158P are not shown, as they were inactive.
All specific dehydrogenase activities toward D-glucose were measured under ambient oxygen conditions, with precautions taken to eliminate bubbles introduced during pipetting. H158K showed the highest activity among the mutants in DCPIP assay, as was the case for oxidase activity. However, H158A showed the highest activity upon addition of the electron mediator PMS. The impact of the mediator was different for each mutant. For example, in DCPIP&PMS assay, the activities of H158V and H158G were 230 % and 200 %, respectively, of those in DCPIP assay, while those H158K and H158R were slightly decreased to 98 % and 96 %.
Our results indicate that histidine is essential for covalent interaction with FAD in PcPOx. Although most of the His158 variants could bind FAD non-covalently, H158D and H158P completely lacked their FAD cofactors. This is a surprising result, since previous research on covalent flavoproteins, such as VAO [3] and sarcosine oxidase [35], did not find apo-mutants, though the number of mutants examined was limited.
The mutation of His158 decreased both oxidase and dehydrogenase activity toward D-glucose compared to WT. This decrease in activity is characterized by increased Km values and decreased kcat values. Increased Km values have also been observed in ToPOx variant H167A [20] and pyranose dehydrogenase (PDH) variant H103Y [13]. However, in VAO, His422 variants showed decreased Km values [3]. Therefore, there is no clear relationship between Km values and covalent bonding with FAD. The reason why His158 variants substituted with aromatic amino acids exhibited high Km values will be discussed based on the crystal structure of the H158W variant since we have recently succeeded in the crystallization of the H158W variant. In contrast to the increased Km values of PcPOx His158 variants, kcat values are consistently decreased for mutants of covalent flavoproteins [3, 13, 20]. The reductive half-reaction was significantly slowed and the oxidative half-reaction was slightly speeded up in VAO variant H422A [3] and PDH variant H103Y [36]. This is considered to be due to the lower redox potential of the mutants [3, 36]. The decreased kcat values in PcPOx His158 variants may be explained similarly. However, it should be noted that PcPOx variant H158Y showed a markedly decreased kcat value (10.5 ± 0.8 s−1 for WT and 0.85 ± 0.06 s−1 for H158Y), while PDH variant H103Y showed a slight decrease (37.8 ± 1.1 s−1 for WT, and 27.4 ± 0.5 s−1 for H103Y) [13]. The effects of a lower redox potential on turnover may differ for oxidase and dehydrogenase activities.
The blue shift in the UV-visible spectra in WT is due to 8α-N3-histidyl-FAD. A similar blue shift is also observed in 8α-aminoacyl riboflavin [28]. For example, peak wavelengths are 367 nm for 8α-S-cysteinyl riboflavins [37], 359 nm for 8α-O-tyrosyl riboflavins [38], 355 nm for 8α-N1-histidyl riboflavins [38, 39], and 355 nm for 8α-N3-histidyl riboflavins [38]; the corresponding value for riboflavin is 372 nm [40]. The blue shift may be due to the anionic flavin quinone of FAD. In previous studies using FMN (a derivative of riboflavin), the anionic state of oxidized FMN exhibited peaks at 355 and 445 nm in the UV-visible spectra [41, 42], while the neutral state of oxidized FMN showed peaks at 373 and 450 nm [41, 42]. Similar changes were observed in WT and the H158A variant of PcPOx.
Our study unveils critical insights into covalent protein-flavin interactions, highlighting the limitation of single mutations at His158 for forming new covalent flavins in PcPOx. A specific conformation of the amino acid residues may be needed for covalent bond formation in PcPOx, as considered on another covalent flavoprotein [43]. Nevertheless, chemical modification of FAD may enable covalent falvin formation. For example, a non-covalent flavoprotein, human D-amino oxidase G281C variant does not form a covalent bond with FAD but forms a covalent bond with 8-methylsulfonyl FAD [44]. Furthermore, lipoamide dehydrogenase WT, which does not form a covalent bond with FAD, requires 8-Cl-FAD for 8α-S-cysteinyl-FAD interaction [45]. Therefore, a detailed understanding of the interaction of chemically modified flavins and amino acid residues in proteins is needed to enable engineering of enzymes and cofactors with enhanced properties.
In conclusion, all the mutants of the FAD-linked residue His158 of PcPOx were unable to covalently bind FAD, and in particular, H158D and H158P were apo-enzymes that were completely unable to bind FAD, even noncovalently. We concluded that the two variants were apo proteins based on the colorless and transparent solutions, UV-visible spectra lacking any peaks derived from flavin, and no oxidase activity toward glucose. However, the reason why the 1/Rz values of the two apo variants were similar to those of the WT and flavin binding variants remained unclear. A more detailed analysis of these two variants could elucidate the mechanism underlying their release of flavin cofactors. Moreover, all of the PcPOx His158 variants showed significantly decreased D-glucose oxidase and dehydrogenase activities. Exploring the relationship between flavoprotein variants and chemically modified flavins is expected to be helpful to obtain new types of covalent flavins.
The authors declare that they have no competing interests.
This study received financial support from Grants-in-Aid for Scientific Research (A) from the Japan Society for the Promotion of Science (JSPS) (No. 23H00341 to KI) and a Grant-in-Aid for JSPS Fellows (No. 24KJ0866 to YY).
