Oxygen binding and nitric oxide dioxygenase activity of cytoglobin are altered to different extents by cysteine modification

Cytoglobin (Cygb), like other members of the globin family, is a nitric oxide (NO) dioxygenase, metabolizing NO in an oxygen (O2)‐dependent manner. We examined the effect of modification of cysteine sulfhydryl groups of Cygb on its O2 binding and NO dioxygenase activity. The two cysteine sulfhydryls of Cygb were modified to form either an intramolecular disulfide bond (Cygb_SS), thioether bonds to N‐ethylmaleimide (NEM; Cygb_SC), or were maintained as free SH groups (Cygb_SH). It was observed that the NO dioxygenase activity of Cygb only slightly changed (~ 25%) while the P50 of O2 binding to Cygb changed over four‐fold with these modifications. Our results suggest that it is possible to separately regulate one Cygb function (such as O2 binding) without largely affecting the other Cygb functions (such as its NO dioxygenase activity).

Cytoglobin (Cygb), like other members of the globin family, is a nitric oxide (NO) dioxygenase, metabolizing NO in an oxygen (O 2 )-dependent manner. We examined the effect of modification of cysteine sulfhydryl groups of Cygb on its O 2 binding and NO dioxygenase activity. The two cysteine sulfhydryls of Cygb were modified to form either an intramolecular disulfide bond (Cygb_SS), thioether bonds to N-ethylmaleimide (NEM; Cygb_SC), or were maintained as free SH groups (Cygb_SH). It was observed that the NO dioxygenase activity of Cygb only slightly changed (~25%) while the P 50 of O 2 binding to Cygb changed over four-fold with these modifications. Our results suggest that it is possible to separately regulate one Cygb function (such as O 2 binding) without largely affecting the other Cygb functions (such as its NO dioxygenase activity).
The Cygb monomer contains two exposed cysteine residues (Cys 38 and Cys 83) that enable Cygb to form an intramolecular disulfide bond [20][21][22]. The formation of this intramolecular disulfide bond greatly increases the dissociation rate constant of a bound intrinsic histidine, resulting in a greater apparent binding constant of extrinsic ligands [21,22]. As a NO dioxygenase, O 2 must bind to Cygb before it metabolizes NO. Under hypoxic conditions, Cygb with intramolecular disulfide bond (Cygb-SS) holds more O 2 in the form of Cygb(Fe 2+ O 2 ) than Cygb with free sulfhydryl group (Cygb-SH) and Cygb with thioether bonds between a cysteine residue and N-ethylmaleimide (NEM; Cygb-SC). However, the NO dioxygenase activity of Cygb is limited and effectively controlled by the rate of Cygb reduction [23]. Thus, if the rate of Cygb reduction is not altered by modification of the sulfhydryl groups in the Cygb, the NO dioxygenase activity of Cygb may not be greatly affected even under hypoxic conditions.
It was suggested that Cygb in vivo is monomeric [20,24]. However, Cygb dimer and other oligomers have been reported to exist in Cygb preparations [22,25]. Although both Cygb monomer and oligomers are of interest to study, the monomer is reported to be the predominant form both in vivo and in vitro [18,20,24]. Therefore, in this study, we focus on examination of the effect of modifying the sulfhydryl groups of the cysteine residues in monomeric Cygb on the affinity of O 2 binding and the rate of Cygb-mediated O 2 -dependent NO metabolism in the presence of cellular reductants.

Materials and methods
Modification of the sulfhydryl groups of the cysteine residues (RSH) of Cygb The expression and purification process of Cygb was described in our prior reports [5]. As isolated Cygb was first treated with dithiothreitol (DTT, C 4 H 10 O 2 S 2 ) in phosphatebuffered saline (PBS) buffer (pH = 7.4 with 0.1 mM EDTA) on ice for about 30 min to reduce any remaining disulfide bonds in the Cygb preparation, forming a nearly homogeneous solution of Cygb in the sulfhydryl form (Cygb-SH). This protein was run through a G-25 column to remove the excess DTT then aliquoted for treatment to modify the free thiol groups as follows: reaction with 2 mM NEM (C 6 H 7 NO 2 MW: 125.13) to alkylate reactive thiol groups (Cygb-SC), or 4 mM 4,4 0 -dithiodipyridine (4-PDS, C 10 H 8 N 2 S 2 MW: 220.3) to drive disulfide bond formation (Cygb-SS) at room temperature for 1 h. After the incubation period, the samples were concentrated and again run through a G-25 column to remove excess reactants. The purity of the three products Cygb-SS, Cygb-SH, and Cygb-SC was determined by comparing the concentration of a given sample of Cygb (measured by the pyridine hemochromagen assay [26]) with the concentration of free thiols of the same Cygb sample (measured by the 4-PDS method [27]). The final average percentage of thiol modification is > 90%.

PAGE and Coomassie staining of pure Cygb
Pure Cygb (10 ng) was loaded to 4-20% gradient gels under nonreducing conditions. Protein samples were separated at 150 V for~90 min. At the end of separation, gels were incubated with 1% Coomassie stain in 40% methanol and 10% acetic acid. After 2 h of staining, gels were destained with repeated 15 min incubations with 20 mL of destaining solution (40% methanol and 10% acetic acid) until bands were clear. Gels were then imaged with a Bio-Rad (Hercules, CA, USA) Versadoc imaging system using QUANTITYONE imaging software.

Measurements of O 2 binding and redox state of Cygb(Fe 2+ )
A Cary 50 UV/Vis spectrophotometer was used to measure the changes in absorbance at 428 nm (A 428 ; peak absorbance of Cygb(Fe 2+ )) with time in the process of reduction of Cygb(Fe 3+ ) and O 2 binding to Cygb(Fe 2+ ). Simultaneously, a Clark O 2 electrode was placed in the solution to monitor the changes in O 2 concentration ([O 2 ], 37°C). Both A 428 and [O 2 ] were sampled at the same rate (10 readings per second). After 1.5 mL of buffer solution and Cygb(Fe 3+ ) were added into the cuvette, the cuvette was covered with a Parafilm membrane. An argon gas tube with an outer diameter of 1 mm was inserted into the cuvette that was covered by a parafilm membrane and positioned above the solution surface to keep an argon flow in the cuvette to gradually remove O 2 from the solution. When O 2 in the solution decreased by about half,~1 mM dithionite (DT) was injected into the solution to rapidly scavenge all O 2 and reduce Cygb(Fe 3+ ) to Cygb(Fe 2+ ) in the cuvette. Although the recordings of O 2 concentration and the absorbance of Cygb(Fe 2+ ) were started before injection of DT, time 0 was assigned to the time point at which DT was added into the solution. At this moment, both [O 2 ] and A 428 suddenly changed. This time point was used as the starting point for synchronizing the readings of A 428 and [O 2 ], which is important for matching the percent of O 2 bound to Cygb(Fe 2+ ) with the corresponding [O 2 ] in the solution. After Cygb was reduced and O 2 was scavenged from both the solution and the gas phase, the gas tube was taken out of the cuvette, and the hole in the Parafilm membrane for the gas tube was sealed by a piece of Parafilm. A syringe needle was used to make a tiny hole on the Parafilm so that O 2 in the air could enter the gas phase of the cuvette slowly. The entered O 2 was gradually dissolved into the test solution to raise the O 2 concentration in the solution at a rate of~0.04-0.06 lMÁs À1 to maintain the O 2 binding process near the equilibrium. The dissolved O 2 can bind on Cygb(Fe 2+ ) to form Cygb(Fe 2+ O 2 ).

Electrochemical measurements of the rate of O 2 -dependent NO metabolism by Cygb
The measurements were performed in a four-port waterjacketed electrochemical chamber (NOCHM-4 from WPI, Sarasota, FL, USA) containing 2 mL of Dulbecco's PBS (Thermo Scientific, South Logan, UT, USA) as previously described [5,19]. The NO solution was also prepared as previously described [28,29]. After the NO and O 2 electrodes had stabilized, NO (0.5 lM) was injected into the aerated buffer solution (under room air) in the absence of Cygb and reductants. When the NO concentration decreased to baseline, 0.4 lM Cygb, 400 UÁmL À1 superoxide dismutase (SOD) from horseradish (Sigma-Aldrich, St. Louis, MO, USA) or 50 lM SOD mimetic (GC4419, Galera Therapeutics Inc., Malvern, PA, USA), and either ascorbate (Asc; Sigma-Aldrich) or cytochrome b5 reductase (b5R) with NADH (Sigma-Aldrich) and cytochrome b5 (b5) were added to the chamber. Then 0. 5  Spectrophotometric measurements of Cygb(Fe 3+ ) reduction by Asc or by the b5R/b5/NADH enzyme-reducing system Measurements were performed on a Cary 50 UV/Vis spectrophotometer. After 1.5 mL of buffer solution was added into the cuvette, the cuvette was covered with a Parafilm membrane. A Clark O 2 electrode was placed in the solution to monitor O 2 concentration. The measured O 2 concentrations were recorded by an Apollo 4000 Free Radical Analyzer (WPI) using a Clark electrode. The solution was stirred using a magnetic stirring bar that was placed on the bottom of the cuvette. An argon gas tube was inserted in the cuvette to bubble argon into the solution for 15 min to quickly remove O 2 . Before injecting Cygb (for reduction by Asc) or Cygb + b5 + NADH (for reduction by the reducing system b5R/b5/NADH) into the solution, the argon gas tube was placed just above the solution surface to keep an argon flow in the cuvette. About 20 min after injection of Cygb or Cygb/b5/NADH, Asc or b5R was added to the solution to initiate reduction of Cygb(Fe 3+ ), respectively. The reaction process was monitored by measuring the changes in absorbance at 416 nm (peak absorbance of Cygb(Fe 3+ )) with time [30].

Equations for determining rate constants of Cygb reduction
The reduction scheme of Cygb(Fe 3+ ) reduction by Asc has been proposed in our previous paper [30]: where A is reductant, B is Cygb(Fe 3+ ), C is the complex Cygb(Fe 3+ A), and D is Cygb(Fe 2+ A + ). In the following derivation process, concentrations of Thus, we can obtain: Under the steady-state approximation (dC/dt = 0), the rate of B consumption is equal to the rate of D formation. From the above equations we can obtain (see Data S1): where g 1 can be considered as the rate constant of reduction of B by A. In Eqn (2) where g 0 is an integration constant. In experiments for measuring the reduction of Cygb(Fe 3+ ), we used a UV/Vis spectrophotometer to monitor the changes in absorbance (Abs) at wavelength 416 nm. According to the Beer-Lambert Law, we have: where e B , e C and e D are the molar extinction coefficients of B, C and D, respectively. Abs is the absorbance at time t. If time t is large (theoretically t approaches infinity), Eqn (4) can be written in the following form: where Abs b is the aborbance as t approaches infinity.  (4) and (5) gives: Then we can get: Under steady-state approximation, C is a constant during the measurements of absorbance. Thus, we can obtain: Substituting Eqn (9) into Eqn (3), we have: Equation (10) indicates that the plot of ln(Abs À Abs b ) vs. t from experimental data will generate a straight line with a slope of Àg 1 .

PAGE and Coomassie staining of Cygb
PAGE followed by Coomassie staining was performed to characterize each of the Cygb preparations including the reduced protein Cygb-SH (SH), the oxidized disulfide Cygb-SS (SS), and NEM-modified Cygb-SC (SC). The resolved gel of the three types of Cygb preparations is shown in Fig. 1. The pure Cygb bands appear with a molecular weight of~21 kDa. The bands in the right column are protein markers with molecular weights indicated. In all three of the Cygb preparations, only monomeric Cygb is present. Any possible dimer form may have been converted into monomer in the preparation of Cygb samples with DTT.

O 2 -binding curves of modified Cygbs
To examine the effect of modifying the sulfhydryl groups of cysteine on the affinity of O 2 binding to Cygbs, we measured O 2 -binding curves for each of the three Cygb preparations at 37°C. Figure 2 shows typical experimental O 2 -binding curves. It can be seen that the P 50 of Cygb has the following order: Cygb-SS < Cygb-SH < Cygb-SC. The mean and standard errors of the P 50 values (n = 3) for the three Cygbs are listed in Table 1 (unit in mmHg or Torr).

Reduction of modified Cygbs
During NO dioxygenation, O 2 binds to Cygb(Fe 2+ ) to form Cygb(Fe 2+ O 2 ) which can be oxidized to Cygb (Fe 3+ ) by NO. To maintain continuity of NO dioxygenation, a reductant (such as Asc) is required for reducing Cygb(Fe 3+ ) back to Cygb(Fe 2+ ). The reduction of the three Cygb(Fe 3+ ) preparations by Asc (1-20 mM) and by b5R (5-120 nM)/b5 (0.5 lM)/NADH (100 lM) was measured with a spectrophotometric assay. It was observed that either Asc or b5R in the presence of b5 (0.5 lM) and NADH (100 lM) cannot fully reduce each of the Cygbs including Cygb-SH, Cygb-SC, and Cygb-SS in the concentration range   listed above, but full reduction could be accomplished by 1 mM DT. In Fig. 3 we demonstrated the typical experimental curves for reduction of Cygb-SH by Asc (Fig. 3A) and b5R/b5/NADH (Fig. 3B). From the experimental data, we plotted ln(Abs À Abs b ) vs. time t based on Eqn (10) (Fig. 3C,D). All plots with Asc as the reductant are linear with time after 2-5 s from the initial injection of Asc (Fig. 3C). From the slopes of these lines, we obtained the pseudo first-order rate constant g 1 , which (shown as k Asc or k b5R in Fig. 3E-F) is linear with reductant concentration [Asc] or [b5R]. In Fig 4A,B, we demonstrated the plots of ln (Abs À Abs b ) vs. time t for Cygb-SH, Cygb-SS, and Cygb-SC reduction by 10 mM Asc or 30 nM b5R in the presence of 0.5 lM b5 and 100 lM NADH. All plots are nearly linear. From the slopes of the lines, we can determine the pseudo first-order rate constants of reduction of the three Cygbs by Asc (10 mM) or by b5R (30 nM) in the presence of 0.5 lM b5 and 100 lM NADH. The pseudo first-order rate constants of Cygb-SH, Cygb-SS, and Cygb-SC at other Asc and b5R concentrations are shown in Fig 4C (for Asc) and Fig. 4D (for b5R).  largely change the equilibrium binding constant of the intrinsic distal histidine [21], which can significantly change the apparent O 2 -binding equilibrium constant. The disulfide bond, if formed, can be reduced by DTT to form two free SH groups. It was reported that the P 50 of O 2 binding to Cygb is 1.8 Torr after Cygb is treated with DTT [20,31], while the P 50 of O 2 binding to Cygb without DTT treatment was reported in the range between 0.2 and 1 Torr [3,20,32]. We tested O 2 binding behaviors for Cygb-SS (P 50 = 0.7 Torr), Cygb-SH (P 50 = 1.4 Torr) and Cygb-SC (P 50 = 2.9 Torr). The measured P 50 values for Cygb-SS and Cygb-SH ( Fig. 2 and Table 1) are in the range of reported values in the literature. Furthermore, we observed that P 50 for Cygb-SC is greater than that for Cygb-SH.
We then examined the effect of modification of the sulfhydryl group of the cysteines of Cygb on the rate of Cygb reduction. Chemical kinetics with mathematical modeling is a powerful tool to study complicated reaction processes [5,[33][34][35]. Using this tool, we obtained Eqn (10) for determining the rate constant of Cygb reduction from experimental data of spectrophotometric measurements. It was observed that the rate of Cygb-SH, Cygb-SC, and Cygb-SS reduction by Asc and by b5R/b5/NADH increases with Asc and b5R concentration, respectively. A group of typical experimental curves for the reduction of Cygb-SH with varying concentrations of Asc (1-20 mM) and b5R (5-120 nM) were shown in Fig. 3A,B. Using Eqn (10), we determined the pseudo first-order rate constants from the experimental curves in Fig. 3A,B. The data process was demonstrated in Fig. 3C-F. From Fig. 3E,F, we can see that the determined pseudo first-order rate constants are nearly linear with Asc or b5R concentrations, and the intercepts of the fitted lines are greater than zero. From the slopes of the plots shown in Fig. 4A-D, we can see that the pseudo first-order rate constants of Cygb-SH, Cygb-SC, and Cygb-SS reduction by Asc and b5R/b5/NADH are close to each other with the rate constant of Cygb-SS reductioñ 25% lower than those of Cygb-SH and Cygb-SC. Because the rate of Cygb reduction is the rate-limiting step for NO dioxygenase activity of Cygb [23], the small differences (~25%) in rate constants of Cygb reduction after modification of sulfhydryl groups imply that the modification of sulfhydryl groups on the Cygb only cause a small effect on the NO dioxygenase activity of Cygb although this modification more largely shifts the apparent O 2 -binding constant. Correspondingly, the differences in the O 2 -dependent NO dioxygenase activity for the three Cygbs (Fig. 5) parallel the differences in their reduction rates with~25% decrease seen in Cygb-SS compared to Cygb-SH and Cygb-SC (Fig. 4A).
The process of NO dioxygenation requires O 2 binding to Cygb to form Cygb(Fe 2+ O 2 ) which can rapidly react with NO. If we gradually remove O 2 from the test solution, O 2 concentration in the solution will gradually decrease. Since the modification of sulfhydryl groups of Cygb can shift the P 50 , differences in the Cygb(Fe 2+ O 2 ) concentration for the three Cygb preparations may be seen when pO 2 drops below 10 Torr (Fig. 2). From Fig. 2 we see that Cygb-SS will have the highest Cygb(Fe 2+ O 2 ) concentration and Cygb-SC will have the lowest. Thus, under low pO 2 conditions, one could speculate that the NO dioxygenase activity by Cygb-SS might be the highest among the three Cygbs. However, only a small change in the NO dioxygenase activity was observed for Cygb-SS compared to Cygb-SH or Cygb-SC at O 2 concentrations below 120 lM (Fig. 5). Thus, although the modification of the sulfhydryl groups on the Cygb can largely increase the apparent O 2 -binding constant of Cygbs in the absence of NO, the bound O 2 on ferrous Cygb or Cygb (Fe 2+ O 2 ) can be immediately removed by NO if NO is present in the solution because NO rapidly reacts with Cygb(Fe 2+ O 2 ) to form NO À 3 and Cygb(Fe 3+ ). As a result, almost all of Cygb(Fe 2+ O 2 ) will react with NO in a short time. The NO dioxygenase activity is then limited by the rate-limiting step, reduction of Cygb. The modification of the sulfhydryl groups on Cygb does not greatly change the rate of Cygb reduction; therefore, the NO dioxygenase activity is only modestly affected by the modification of the sulfhydryl groups of Cygb.
In summary, the modification of the sulfhydryl groups on Cygb can largely change (~4-fold) the apparent constant of O 2 binding to Cygb. However, this modification only causes a small difference (~25%) in the rate constant of Cygb reduction and NO consumption by Cygb. These results indicate that it is possible to largely change one property of Cygb (such as O 2 binding ability) and keep another property of Cygb (such as NO dioxygenase activity) almost unchanged or only modest change.