Significant increase of oxidase activity through the genetic incorporation of a tyrosine-histidine cross-link in a myoglobin model of heme-copper oxidase

Abstract

Heme-copper oxidase (HCO) performs efficient four-electron reduction of oxygen to water without releasing toxic, reactive oxygen species (ROS). Essential for this function is a post-translationally modified histidine–tyrosine cross-link (Tyr-His) in its heme a3/Cu_B oxygen reduction center. Through the genetic incorporation of the Tyr-His ligand and Cu_B site into myoglobin, we recapitulated important features of HCO into this small soluble protein, which exhibits selective O₂ reduction activity while generating less than 6% ROS, at more than 1000 turnovers. These results support that Tyr-His crosslink is indeed important for HCO function, and creates the exciting opportunity to rapidly evolve better HCO model proteins to achieve higher activity and selectivity, which may be suitable as alternatives to precious metal catalyst in fuel cells.

In aerobic respiration, heme-copper oxidase^[1] (HCO) performs efficient four-electron reduction of oxygen to water without releasing toxic, reactive oxygen species (ROS) such as superoxide and peroxide.^[2] Due to its importance in biogenetics, fascinating chemistry and potential applications in fuel cells, HCO has been intensively investigated by structural biologists,^[3] enzymologists,^{[1, 4]} and synthetic chemists.^[5] While tremendous progress has been made from the above studies, several important questions about the HCO structural features and mechanism still remain. Of particular interests is the unique post-translationally modified covalent cross-link between C6 of tyrosine and Nε2 of a nearby histidine (Tyr-His crosslink, Figure 1A), revealed first by high-resolution x-ray structure^[3a] and then confirmed by biochemical and biophysical studies.^[6] Importantly, the Tyr-His crosslink is conserved in HCO from bacterial to mammal,^[5b] implicating a strong evolutionary connection and its functional significance.

Figure 1.

(A) Cu_B site of cytochrome c oxidase (bovine numbering). (B) Structure model overlay of imiTyrCu_BMb (cyan) and F33YCu_BMb (yellow). (C) Synthesis of imiTyr 1. a) Di-tert-butyl dicarbonate, THF, 1N NaOH rt, 4 h. b) imidazole, Cs₂CO₃, CuI, DMF, reflux, 16 h.

While the presence of covalent cross-linked Tyr-His is firmly established in HCOs, its exact role in HCO function is still not fully understood. A major limitation of our understanding is that such an post-translational modification cannot be directly probed through traditional site-directed mutagenesis methods. To overcome this limitation, synthetic model compounds have been made, whose studies have suggested that that the crosslink between Tyr and His significantly decreased pKa of phenol and imidazole sidechains. The decreased pKa of phenol is believed to facilitate proton delivery and tyrosyl radical formation.^{[5b, 7]} Despite these advances, questions about the roles of Tyr-His crosslink still remain, as no study has been able to compare the functions of His and Tyr residues at the same positions in the same protein with and without cross-linking.

To answer the above question and to gain new insights into the HCO catalytic process,^[7c] we report herein the genetic incorporation of an unnatural amino acid (UAA) imiTyr 1 (Figure 1C) that mimics the cross-linked Tyr-His ligand into a functional model of HCO (Figure 1B) in sperm whale myoglobin^[8](termed imiTyrCu_BMb). We then compared its activity with a mutant bearing His and Tyr at the same position in the same protein but lack such a crosslink (F33YCu_BMb). Our results show that imiTyrCu_BMb recapitulated important features of HCO, exhibiting selective O₂ reduction activity while generating less than 6% ROS, at more than 1000 turnovers. Importantly, incorporation of imiTyr in HCO model resulted in eight fold increase in selectivity and threefold increase in catalytic turnover, supporting that Tyr-His crosslink is indeed essential for HCO function,^[5b] and provides direct evidence for the first time that the covalent cross-linked Tyr-His moiety can dramatically enhance the oxidase activity.

We choose myoglobin for genetic incorporation of the unnatural amino acids and to carry out functional studies because, in comparison the large (~200 kDa) membrane HCO proteins consisting of multiple subunits that are difficult to produce homogeneously in large amounts, myoglobin is a much smaller (18 kD) water-soluble protein that can be expressed and purified easily in gram quantities.^[9] Previous biochemical and spectroscopic studies have indicated that the designed heme-copper center in myoglobin can be a functional model of HCOs.^[8]

We first synthesized (S)-2-amino-3-(4-hydroxy-3-(1H-imidazol-1-yl)phenyl)propanoic acid 1 (hereafter termed imiTyr), which contains the ortho-imidazole–phenol linkage (Tyr-His ligand), and then genetically encoded it into myoglobin. The unnatural amino acid imiTyr 1 was synthesized by first protecting the amine group of 3-iodotyrosine using Di-tert-butyl dicarbonate,^[10] affording Boc-L-3-iodotyrosine, which was then allowed to react with imidazole in the precence of Cs₂CO₃ and CuI in DMF under refluxing conditions to afford 1 in 50% yield^[7c] (Figure 1C).

To selectively incorporate imiTyr at defined sites in proteins in E. coli, a mutant Methanococcus jannaschii tyrosyl amber suppressor tRNA (MjtRNA^Tyr_CUA)/tyrosyl-tRNA synthetase (MjTyrRS) pair was evolved that uniquely specifies 1 in response to the TAG codon. An MjTyrRS library, pBK-lib-jw1, in which six residues (Tyr32, Leu65, Phe108, Gln109, Asp158, and Leu162) were randomized, and any one of the six residues (Ile63, Ala67, His70, Tyr114, Ile159, Val164) was either mutated to Gly or kept unchanged, was used to screen for active and selective mutants for 1, as previously reported.^[11] Nine MjTyrRS clones emerged after three rounds of positive selections and two rounds of negative selections (Table S1). The best clone grew at 120 μg/mL of chloramphenicol in the presence of 1 mM 1, but only at 20 μg/mL chloramphenicol in its absence, and was named imiTyrRS. Sequencing of this clone revealed five mutations: Tyr32Glu, Leu65Ser, His67Gly, Asp158Tyr and Leu162Asn. Among the five mutations, the His67Gly mutation was required to create additional space to accommodate the imidazole group. The other four mutations may make specific hydrogen bonding interactions to stablize the imiTyr sidechain. Work is underway to obtain a structure of the imiTyrRS in order to gain insight into these interactions.

To determine the efficiency and fidelity for the incorporation of 1 into proteins, an amber stop codon was substituted for Ser4 in sperm whale myoglobin containing a C-terminal His₆ tag. Protein expression was carried out in E. coli in the presence of the selected synthetase (imiTyrRS), MjtRNA^Tyr_CUA and 1 mM 1, or in the absence of 1 as a negative control. Analysis of the purified protein by SDS-PAGE showed that full-length myoglobin was expressed only in the presence of 1 (Figure 2A), indicating that imiTyrRS is specifically active for 1 but does not utilize any natural amino acid. The yield of the mutant myoglobin was 10 mg/L. For comparison, the yield of wild-type sperm whale myoglobin (wtMb) under similar conditions was 50 mg/L. ESI-MS spectrometric analysis of the mutant myoglobin gave an observed average mass of 18497 Da (H⁺), in close agreement with the calculated mass of 18496 Da for the Ser4→1 myoglobin mutant.

Figure 2.

(A) Coomassie-stained SDS-PAGE of TAG4 mutant (left) and mutant (right) myoglobin (indicated by black arrow) expression in the presence and absence of 1 mM UAA 1. (B) ESI-MS spectra of the TAG4 mutant. The insert shows the deconvoluted spectrum; expected mass: 18496 Da, found: 18497 Da.

Having demonstrated the efficient genetic incorporation of imiTyr 1 into wtMb, we then expressed myoglobin double mutant Leu29His/Phe33→1 (called imiTyrCu_BMb) as a functional HCO model,^[8] with a yield of 5 mg/L. The selective incorporation of imiTyr at position 33 was again confirmed by SDS-PAGE (Figure 2A) and ESI-MS (Figure S2). Together with distal His64 in sperm whale Mb, the newly introduced His29 and imiTyr33 constitute a Cu_B binding site according to our model, placing the copper ion in close proximity to heme iron (Figure 1B). The UV-vis spectrum of imiTyrCu_BMb at pH 7.4 is similar to that observed for wtMb and is typical of a six-coordinate, H₂O-bound, high-spin heme with a Soret band absorption maximum at 408 nm (Figure S3). The deoxy-imiTyrCu_BMb was obtained by reducing the purified met-imiTyrCu_BMb with excess dithionite. The UV-vis spectrum (Figure S3) of deoxy-imiTyrCu_BMb exhibits a Soret band at 430 nm and a visible absorption band at 562 nm. This spectrum is nearly identical to that of deoxy-wtMb, indicating that mutation in imiTyrCu_BMb did not cause major structural changes of the protein. Addition of Cu²⁺ at pH 7.4 to imiTyrCu_BMb resulted in spectrum difference in the Soret region (Figure S4), which was fitted to a double reciprocal plot and a Hill plot.^[8b] These results indicate that a single Cu²⁺ binding site were present in imiTyrCu_BMb, with a K_D of 1.6 μM.

Since the hallmark of native HCOs is the clean catalytic reduction of O₂ to water without releasing ROS, we then measured the rate imiTyrCu_BMb catalyzes oxygen reduction by using an O₂ electrode in the presence of 1000 equivalents of ascorbic acid as the reductant and 100 equivalents of TMPD (tetramethyl-p-phenylenediamine dihydrochloride) as a redox mediator. Similar method was previous used to measure enzymatic activity of native HCO.^[12] To differentiate between ROS and water products, we employed catalase which selectively reacts with H₂O₂ to produce O₂, and superoxide dismutase (SOD) which selectively reacts with superoxide to produce H₂O₂ and O₂. If O₂ consumption results in release of ROS but not formation of water, the apparent O₂ reduction rate will slow down in the presence of catalase and SOD, because they will convert ROS back to O₂. By comparing the rates of reduction in the absence and presence of catalase and SOD, the portion of O₂ reduction due to water formation (in blue) and due to ROS formation (in red) can be calculated (Figure 3A). Not surprisingly, most of the O₂ consumed by wtMb lead to ROS, both in the presence and absence of Cu²⁺. By contrast, imiTyrCu_BMb catalyzed O₂ reduction without generating more than 6% H₂O₂ in the presence of 1 eq Cu²⁺, with oxygen consumption rate of 12 μM/min (enzyme activity = 2 oxygen/min). In the absence of Cu²⁺, however, more than 30% of O₂ was converted to ROS, indicating that Cu_B is important for the catalytic turnover of O₂ into H₂O.

Figure 3.

(A) Rates of oxygen reduction to form either water (blue) or ROS (red) catalyzed by 6 μM wtMb, F33YCu_BMb, or imiTyrCu_BMb, in the presence or absence of 6 μM of Cu²⁺, with 0.6 mM TMPD and 6 mM ascorbic acid as reductants (B) O₂ reduction turnover number catalyzed by imiTyrCu_BMb or F33YCu_BMb.

The myoglobin system described here provides a unique capability to compare directly the effect of His and Tyr with and without cross-linking. To take advantage this unique feature, we also measured O₂ consumption for the myoglobin mutant Leu29His/Phe33Tyr/Phe43His (F33YCu_BMb). This mutant harbors a Cu_B site consisting of residues His29, His43 and His64, as we have shown previously,^[8b] and a tyrosine residue in position 33. While His43 and Tyr33 sidechains are in similar positions as the imidazole group and phenol group of imiTyr33, respectively (Figure 1B), no crosslink is present between them. Not only the overall activity of the F33YCu_BMb was less (~40%) than imiTyrCu_BMb, but also more than 50% of O₂ was converted to ROS with or without Cu²⁺ added, confirming that the cross-linked Tyr-His ligand plays a critical role for the selective reduction of O₂ into H₂O.

To test if imiTyrCu_BMb is a robust O₂ reduction catalyst, capable of performing many turnovers without self-destruction, we then measured O₂ consumption under multiple turnover conditions. To a solution containing 6 μM imiTyrCu_BMb, 6 μM CuSO₄ and 800 μM O₂, 0.6 mM TMPD and 6 mM ascorbic acid were added. O₂ reduction was then monitored by using an oxygen electrode until all the oxygen was consumed. Subsequently, a further addition of 800 μM O₂ was carried out by purging the solution with pure oxygen, and then oxygen reduction was again measured until all the oxygen was reduced. These stepwise additions resulted in the multiple plateaus observed in the traces (Figure 3B), and these processes were repeated until imiTyrCu_BMb achieved more than 1000 turnovers (Figure 3B). The turnover number reported here is an underestimate as the extra turnovers occurring during oxygen addition were not included in the calculation. Notably, little decrease in O₂ consumption rate was observed after the catalyst reached 1000 turnovers, indicating that at this point the imiTyrCu_BMb catalyst has remained intact. By contrast, F33YCu_BMb underwent less than 400 turnovers under similar conditions. These results again support that the cross-linked Tyr-His ligand is required for optimal O₂ reduction to H₂O. Here, the different catalytic activity between imiTyrCu_BMb and F33YCu_BMb can also be attributed to the different geometry of the copper center. To further investigate the mechanism, we are currently working to solve the crystal structure of imiTyrCu_BMb, and examine the enzymatic property of another mutant having a Phe-His type crosslinkage at the 33 position.

In conclusion, by directly incorporating the unnatural amino acid imiTyr 1 into myoglobins in E. coli in response to the amber codon TAG, we have successfully designed a functional HCO model imiTyrCu_BMb which catalyzes selective and efficient oxygen reduction to water. The imiTyrCu_BMb HCO model bearing Tyr-His crosslink is eight fold more selective with three fold more turnovers than F33YCu_BMb, which does not contain the crosslink but harbor His and Tyr at the same positions in the same protein. Since the synthesis of imiTyr is only two steps with 50% overall yield, and mutant proteins bearing imiTyr 1 at any site can be easily obtained and purified in milligram quantities through site-directed mutageneis and recombinant protein expression, further systematic investigation of Tyr-His crosslink function is now possible. While imiTyrCu_BMb exhibits lower enzymatic activity (2 oxygen/min)^[12] in comparison to native heme copper oxidase (~ 300 oxygen/sec), it is possible to rapidly improve our HCO model to achieve higher activity through directed evolution and unnatural amino acid incoporation. Our designed enzyme harbors unnatural amino acid imiTyr which is highly analogous to the post-translationally modified tyrosine-histidine ligand found in Cu_B site of HCO, serves as an ideal model for a more detailed understanding of HCOs, and allows for potential applications in synthetic biology and alternative energy.^{[5b, 9, 13]}

Experimental Section

To express mutant myoglobin protein imiTyrCu_BMb, plasmids pBAD-JYAMB-His29TAG33 was cotransformed with pBK-imiTyrRS into GeneHog®-Fis E. coli cells. Cells were amplified in LB media (5 mL) supplemented with kanamycin (50 μg/mL) and tetracycline (15 μg/mL). Starter culture (1 mL) was used to inoculate 100 mL of liquid LB supplemented with appropriate antibiotics and 1 mM 1. Cells were then grown at 37°C to OD600 of 0.5, and protein expression was induced by the addition of 0.2% arabinose. After 12 hours of growth at 37°C, cells were harvested by centrifugation. Wild type or mutant myoglobin was then purified by Ni-NTA affinity chromatography under denaturing conditions, and reconstituted with heme as previously described.^[9]

Oxygen consumption was measured by using an Oxygraph Clark-type oxygen electrode (Hansatech Instruments) at 25°C in 20 mM Tris buffer, pH 7.4. All experiments were performed three times to obtain standard deviation. The electrode was calibrated against air-saturated buffer and O₂-depleted buffer prior to use. Protein was exchanged to 20 mM Tris, pH7.4, which was treated with chelex beads overnight. The final concentration of protein was adjusted to 6 μM. For experiment with Cu²⁺, CuSO₄ was added to final concentration of 6 μM and stirred for 10 min. For experiment with catalase and SOD, catalase (Sigma-Aldrich) was added to final concentration of 7.3 U/L, SOD (Sigma-Aldrich) was added to final concentration of 0.5 U/μL. The reaction is initiated by adding TMPD and ascorbic acid to a final concentration of 0.6 mM and 6 mM, respectively. Oxygen concentration was read immediately after adding reductant. The turn-over experiment was performed similarly except after O₂ in the reaction chamber being exhausted, the chamber was vented with pure O₂ to boost O₂ concentration to 800 μM. Ascorbic acid was added to keep concentration at 6 mM.