CO binding improves the structural, functional, physical and anti-oxidation properties of the PEGylated hemoglobin

Abstract

Context: PEGylated hemoglobin (Hb) is a promising oxygen therapeutic agent for clinical application. However, it suffered from structural perturbation, functional instability and methemoglobin (metHb) formation. Objective: To improve the structural, functional, physical and anti-oxidation properties of the PEGylated Hb. Materials and methods: PEGylation of Hb with CO binding (HbCO) was conducted using maleimide and acylation chemistry, respectively. Physical and chemical parameters were measured for Hb samples. The circular dichroism spectra, dynamic light scattering and analytical ultracentrifugation were used to investigate the structure and conformation of PEGylated HbCO. Results: CO binding can inhibit the autoxidation of the PEGylated Hb, structurally stabilize its tetramer and improve its thermal and pH stability. Importantly, the circular dichroism spectra showed that CO binding can decrease the structural perturbation of Hb induced by PEGylation. The PEGylated HbCO with CO release showed slightly higher oxygen-delivery capacity than the PEGylated Hb. The PEGylated HbCO did not show metHb formation after 30-day storage at 4°C. Discussion and conclusion: CO binding structurally stabilized the PEGylated Hb, abolished its metHb formation, and significantly increased its physical stability. In particular, it also avoided the perturbation of PEG chains on the heme microenvironment. The functional property of the PEGylated HbCO can be maintained during its long-term storage, which is of great significance for field transfusion.

Keywords::

• autoxidation
• heme environment
• oxygen carrier
• PEGylation
• thermal stability

Introduction

PEGylation, chemical conjugation with polyethylene glycol (PEG), can reduce renal clearance, enzymatic degradation and immunogenicity of peptides and proteins, thereby improving their therapeutic potentials (Pasut and Veronese 2012, Xue et al. 2013, Mu et al. 2013). Hemoglobin (Hb) is a potential therapeutic agent as oxygen carrier and a protein responsible for storage and transport of oxygen in red blood cell (Rameez et al. 2008, Tsuchida et al. 2009). However, Hb suffers from the disadvantages of nephrotoxicity, short in vivo half-life and vasoactivity (Cole and Vandegriff 2011, Elmer et al. 2012). Animal tests and clinical trials suggest that PEGylation has emerged as a novel chemical approach to circumvent these disadvantages of Hb (Jahr et al. 2012, Hu et al. 2007). Therefore, the PEGylated Hb is a potent candidate for development of Hb based oxygen carriers (HBOC).

Hb in a ferrous state can be easily oxidized to methemoglobin (metHb) in a ferric state. MetHb may lose its oxygen-delivery capacity, release its cytotoxic heme in the circulation and generate harmful reactive oxygen species (ROS) (Alayash et al. 2001, Vandegriff et al. 2006, Dunne et al. 2006). Hb also undergoes in vivo rapid dissociation from its native tetrameric state (α2β2 tetramer) to dimeric state (αβ dimer), which leads to nephrotoxicity (Lunelli et al. 1994). Recent studies suggested that PEGylation can accelerate the autoxidation of Hb (Hu et al. 2008, 2011). On one hand, the heme environment of Hb may be structurally perturbed by PEGylation, particularly at the sites proximal to the heme pocket, such as at Cys-93(β) (Huang et al. 2012). On the other hand, the conjugated PEG chains can act as a reservoir of waters in the hydration layer of Hb (Hu et al. 2005) and enhance the oxidation of Hb by promoting nucleophilic attack of heme (Hu et al. 2008). Typically, PEGylation of Hb was conducted in an air-equilibrated environment and metHb formation may occur in the PEGylation process. PEGylation in anaerobic condition has been carried out in the presence of Cl⁻ and inositol hexaphosphate (IHP) to inhibit the metHb formation of Hb (Iafelice et al. 2007). However, achieving such an oxygen-free environment requires considerable effort to remove O₂ from Hb. PEGylation also induces extensive tetramer-dimer dissociation of Hb, thereby resulting in significant structural and functional perturbation of Hb (Hu et al. 2007, 2009). For example, tetramer-dimer dissociation of Hb unfavorably alters the heme environment and quaternary structure and destabilizes the α1β2 interface of Hb (Caccia et al. 2009).

It is well known that the bond between heme of Hb and CO is much more stable than that between heme and O₂ (Balakrishnan et al. 2009). This implies that CO binding with heme can inhibit the autoxidation of Hb and improve the physical stability of Hb. For example, MP4 (Sangart Inc., USA), a PEGylated Hb with ˜ 7 copies of 5 kDa PEG conjugated to human Hb, has displayed an increased stability during storage upon CO binding (Vandegriff et al. 2008). Since the PEGylated Hb with CO binding (HbCO) cannot be oxidized, its autoxidation may be avoided and the long-term storage of the PEGylated Hb can be achieved by CO binding. Thus, CO binding is a possible approach to decrease or circumvent the above side effects of the PEGylated Hb.

In the present study, we have investigated the effect of CO binding on the structural (heme environment and molecular shape, etc.), functional (oxygen affinity and Hill coefficient), physical (pH and thermal stability), and anti-oxidation properties of the PEGylated Hb. PEGylation of HbCO was conducted using acylation chemistry and maleimide chemistry, respectively. Acylation chemistry can stochastically modify the ε-NH₂ of Hb and has been used to prepare the Enzon's product (Enzon Inc., USA) (Conover et al. 1996). Maleimide chemistry can specifically modify Cys-93(β) of Hb and has been used to prepare MP4 (Sangart Inc., USA) (Winslow 2004). Our study is expected to decrease or circumvent the side effect of the PEGylated Hb by CO binding.

Materials and methods

Materials

Carbon monoxide (99.9%) and oxygen (99.9%) were purchased from Yanan Gas Technology Co., Ltd. (Beijing, China). PEG succinimidyl carbonate (PEG-sc, 5 kDa) and PEG maleimide (PEG-mal, 5 kDa) were provided by the National Research Center for Biotechnology (Beijing, China).

PEGylation of Hb and HbCO

Human Hb was purified from erythrocytes as described previously (Lu et al. 2004). The purified Hb in PBS buffer (pH 7.4) was bubbled with CO to obtain HbCO. HbCO (0.15 mM in PBS buffer, pH 7.4) and Hb (0.15 mM in PBS buffer, pH 7.4) were allowed to react with 20-fold molar excess of PEG-sc polymers with PEG mass of 5 kDa at 4°C overnight, respectively. The ε-NH₂ groups of Hb were stochastically modified by PEG-sc. Hb (0.15 mM in PBS buffer, pH 7.4) and HbCO (0.15 mM in PBS buffer, pH 7.4) were also allowed to react with a 10-fold molar excess of PEG5k-mal respectively, and the reaction mixtures were incubated at 4°C overnight. The thiol groups of Cys-93(β) of Hb were specifically modified by PEG-mal. The reaction mixture was loaded on a DEAE Sepharose Fast Flow column (1.6 × 10 cm, GE Healthcare, USA). The column was equilibrated and eluted with 50 mM Tris-Ac buffer (pH 8.5) to remove the free PEG reagent. The bound PEGylated Hb was eluted from the column with 50 mM Tris-Ac buffer (pH 7.0).

Analytical methods

Size exclusion chromatography (SEC) analysis was carried out on a Superdex 200 column (1 × 30 cm, GE Healthcare, USA). The column was equilibrated and eluted with PBS buffer (pH 7.4) at a flow rate of 0.5 ml/min. SDS-PAGE analysis was carried out on a 15% polyacrylamide gel and the gel was stained with Coomassie Blue R-250. Spectrometric analysis was performed on a double-beam spectrophotometer (Model U-3110, Hitachi, Japan).

Circular dichroism spectroscopy

Circular dichroism (CD) spectra were recorded on a J-810 spectropolarimeter (Jasco, Japan) using a 1-mm path length cuvette at 25°C. The spectra of buffer blanks were measured before the samples and subtracted from the spectra of the samples. The CD spectra (250–480 nm) were obtained using 26.0 μM Hb samples (in tetramer). Far UV CD spectra (200–250 nm) were obtained using 1.3 μM Hb samples (in tetramer). The molar ellipticity (θ) was expressed in deg cm²/dmol on a heme basis.

Autoxidation

Autoxidation experiments were carried out by incubation of the Hb samples (25 μM as tetramer) in PBS buffer (pH 7.4) containing 300 μM EDTA at 37°C. Aliquots of the samples were taken out at various time intervals. Total Hb concentration was calculated using ε₅₂₃ = 7.12 mM⁻¹ cm⁻¹ (Snell and Marini 1988). The methemoglobin (metHb) level was measured based on absorbance at 560, 576, 630 and 700 nm (Benesch et al. 1973). Spectrophotometric measurements were carried out in a double-beam spectrophotometer (Model U-3110, Hitachi, Japan). The first-order autoxidation rate constants (k_ox) were measured as Hu et al. (2008).

Oxygen affinity and Hill coefficient

The HbCO samples were converted to oxygenated state by exposure to an O₂ atmosphere under irradiation of a visible light (Sakai et al. 1993a). In addition, the Hb and HbCO samples were stored at 4°C for 30 days. The resultant HbCO samples were converted to oxygenated state. Oxygen equilibrium curves of the Hb and HbCO samples were carried out using a Hemox Analyzer (TCS Corp., PA, USA) in Hemox buffer (pH 7.4) at 37°C. The O₂ equilibrium curves for each sample were used to obtain the values of Hill coefficient (n) and the oxygen pressure at which Hb was half saturated (P₅₀).

The pH stability

The pH stability of the Hb and HbCO samples was investigated by measuring the far UV CD spectra of the samples under different pH conditions. The ellipticity at 222 nm of each sample was recorded, as the ellipticity at 222 nm can reflect the α-helix content of Hb (Kelly and Price 2000). Here, the α-helix content of Hb at pH 7.4 was set as 100%. The samples were at Hb concentration of 1.3 μM (in tetramer).

Thermal stability

The Hb and HbCO samples were heated at 65°C as a function of incubation time (Ruckpaul et al. 1971). Briefly, each Hb sample (9.0 ml, 2 mg/ml) in PBS buffer (pH 7.4) was divided into 9 aliquots of 1 ml. Each aliquot was kept in tightly-capped microcentrifuge tubes and heated at 65°C in a water bath. The tubes were removed from the water bath at 0, 5, 10, 15, 20, 30, 40, 50 and 60 min, respectively. The tubes were immediately incubated in an ice bath for 10 min and centrifuged at 10,000 g for 10 min, followed by removal of the precipitate. The absorbance at 523 nm of the supernatant was used to calculate the percent denaturation from the equation: Denaturation (%) = (A₀−A_t)/A₀ × 100, where A₀ is the absorbance of the unheated aliquot at 523 nm and A_t is the absorbance of the heated sample at 523 nm at time t.

The Hb and HbCO samples were heated for 30 min as a function of temperature and then immediately incubated in an ice bath for 10 min. The resultant samples were centrifuged at 10,000 g for 10 min, followed by removal of the precipitate. The supernatant was subjected to CD analysis in the far UV region (200–250 nm). The ellipticity at 222 nm of each sample was recorded. The retention of secondary structure was calculated from the equation: Retention (%) = ([θ]−[θ]₀)/[θ]₀ × 100%, where [θ]₀ and [θ] are the ellipticity values of the unheated and heated samples at 222 nm, respectively.

Dynamic light scattering

Dynamic light scattering analysis was performed using a Wyatt DynaPro Titan (Santa Barbara, CA) at 25°C. The samples were at a protein concentration of 1.0 mg/ml and centrifuged at 15,700 g for 10 min before analysis.

Sedimentation velocity

Sedimentation velocity was detected by analytical ultracentrifugation on a ProteomeLab XL-1 (Beckman, USA) equipped with an An-60Ti rotor. The samples at the nominal concentration (A₂₈₀ = 0.740) were centrifuged at 60,000 g in PBS buffer (pH 7.4) at 20°C. Boundary movement was detected at 280 nm using the absorption optics. The sedimentation coefficient (S) and the ratio of frictional coefficient (f/f₀) were calculated using v– of 0.744 ml/g for Hb and 0.806 ml/g for the PEGylated proteins (Kellett 1971). These values were normalized to standard conditions.

Results

Characterization of the PEGylated proteins

The PEGylated proteins were characterized by size exclusion chromatography (SEC). As shown in Figure 1, Hb was eluted as a single and symmetric peak. The elution peak of Hb was left-shifted upon PEGylation. This indicated that PEGylation can increase the hydrodynamic volume of Hb. Moreover, the PEGylated Hb prepared with PEG-sc (P-sc-Hb) was eluted earlier than that with PEG-mal (P-mal-Hb), due to the more conjugated PEG chains in P-sc-Hb. However, the elution peaks of Hb, P-sc-Hb and P-mal-Hb were not shifted upon CO binding. This indicated that CO binding did not alter the hydrodynamic volumes of the Hb samples.

Figure 1. Size exclusion chromatography analysis of the Hb and HbCO samples. The analysis was carried out on a Superdex 200 column (1 cm × 30 cm, GE Healthcare, USA) at room temperature. The column was eluted with PBS buffer (pH 7.4) at a flow rate of 0.5 ml/min. The effluent was monitored at 540 nm.

The molecular radii of the Hb and HbCO samples were determined by dynamic light scattering. As shown in Table I, PEGylation can significantly increase the molecular radii and volumes of Hb and HbCO. However, the molecular volume of HbCO is comparable to that of Hb. Similarly, the molecular volumes of P-sc-HbCO and P-mal-HbCO were comparable to those of P-sc-Hb and P-mal-Hb, respectively. This was consistent with the result of SEC analysis.

Table I. The radius and molecular volume of the PEGylated proteins.^a

Sample	Radius (nm)^b	Molecular volume (nm^3)^c
Hb	3.10 ± 0.05	124.72
HbCO	3.09 ± 0.05	123.52
P-sc-Hb	5.66 ± 0.07	759.13
P-sc-HbCO	5.61 ± 0.06	739.19
P-mal-Hb	4.43 ± 0.06	363.99
P-mal-HbCO	4.32 ± 0.05	337.54

^aEach sample was measured three times.

^bThe samples were at a protein concentration of 1.0 mg/ml and measured at 25°C.

^cThe molecular volume was calculated with the equation V = 4πr³/3, where r is the molecular radius.

As shown in Figure 2, P-mal-Hb (Lane 4) showed an additional band with respect to Hb (Lane 2). This was due to that Cys-93(β) in β globin of Hb was specifically conjugated with PEG-mal, and α globin of Hb was unmodified. P-sc-Hb (Lane 3) showed additionally one major band and one minor band with respect to Hb. This indicated that Hb was non-specifically conjugated with PEG-sc, and partial globins were unmodified. P-sc-HbCO (Lane 6) exhibited migration bands comparable to P-sc-Hb (Lane 3). Similarly, the migration bands of P-mal-HbCO (Lane 7) and HbCO (Lane 5) were comparable to those of P-mal-Hb (Lane 4) and Hb, respectively. This indicated that CO binding did not alter the PEGylation pattern of Hb.

Figure 2. SDS-PAGE analysis of the Hb and HbCO samples. The analysis was carried out on a 14% Tris-glycine gel. Lane 1, the standard markers; Lane 2, Hb; Lane 3, P-sc-Hb; Lane 4; P-mal-Hb; Lane 5, HbCO; Lane 6, P-sc-HbCO; Lane 7, P-mal-HbCO. Proteins were identified by Coomassie blue staining.

As shown by spectrometric analysis (Figure 3), Hb showed three peaks at 414, 540 and 575 nm. The three peaks of Hb were shifted to 419, 538 and 569 nm upon CO binding. The optical shift of Hb indicated the successful binding of CO with the heme of Hb. In contrast, the optical spectra of Hb and HbCO were not further altered upon PEGylation. This indicated that PEGylation did not alter the optical absorption of Hb and HbCO.

Figure 3. Spectrometric analysis of the Hb and HbCO samples. Analyses of P-mal-HbCO (a), P-sc-HbCO (b), HbCO (c), P-mal-Hb (d), P-sc-Hb (e) and Hb (f) were performed on a double-beam spectrophotometer.

Structural analysis

Circular dichroism spectroscopy

In the near UV and Soret band region (250–500 nm), CO binding and PEGylation both altered the CD spectra of Hb (Figure 4). L band (around 260 nm) is considered to be sensitive to the interactions between heme and the surrounding globin. PEG-sc mediated PEGylation apparently increased the L band of Hb (Figure 4a). This suggested that PEGylation unfavorably altered the heme environment of Hb. In contrast, the L band of P-sc-Hb was decreased upon CO binding (Figure 4b). This suggested that CO binding can decrease the structural perturbation of the heme environment induced by PEGylation.

Figure 4. CD spectra of the Hb and HbCO samples. The spectra of the Hb (a) and HbCO samples (b) were recorded on a J-810 spectropolarimeter at room temperature. The molar ellipticity, θ, is expressed in deg cm²/dmol on a heme basis.

The Soret band (around 420 nm) of Hb is informative regarding the interactions of the heme prosthetic group with the surrounding aromatic residues. The modifications of these residues in the spatial orientation around the heme may affect porphyrin transitions and π-π* transitions in the surrounding aromatic residues. As reflected by the apparent increase in the Soret bands of P-mal-Hb and P-sc-Hb, the conjugated PEG chains unfavorably altered the interactions between the heme and the surrounding globin chain of Hb (Figure 4a). In contrast, the Soret bands of P-mal-Hb and P-sc-Hb were essentially superimposed on that of HbCO (Figure 4b). This CO binding can decrease the structural perturbation of Hb induced by PEGylation.

Sedimentation velocity

Sedimentation velocity studies were conducted to gain the conformational information of the Hb and HbCO samples. Sedimentation coefficients (S_20,w) can be used to monitor the conformational change of proteins (Carvalho et al. 2011). As shown in Table II, the S_20,w of Hb decreased upon PEGylation, due to the hydrodynamic drag of the PEG chains (i.e., the parachute effect of PEG). P-sc-HbCO and P-mal-HbCO showed higher S_20,w values than P-sc-Hb and P-mal-Hb, respectively. This indicated that CO binding structurally altered P-sc-Hb and P-mal-Hb, thereby increasing their molecular density. In addition, P-sc-HbCO showed a lower S_20,W than P-mal-HbCO, due to more conjugated PEG chains in P-sc-Hb.

Table II. The sedimentation velocity and f/f₀ of the PEGylated proteins.

Sample	S20,W^a	f/f0^b
Hb	4.212 ± 0.208	1.11
HbCO	4.231 ± 0.293	1.09
P-sc-Hb	2.022 ± 0.235	2.20
P-sc-HbCO	2.273 ± 0.118	1.93
P-mal-Hb	3.259 ± 0.248	1.66
P-mal-HbCO	3.385 ± 0.147	1.56

^aThe sedimentation coefficient.

^bThe ratio of frictional coefficient.

The ratio of frictional coefficient (f/f₀) can be used to provide hydrodynamic shape information of macromolecules (Li et al. 2012). As shown in Table II, PEGylation can increase the f/f₀ of Hb, indicating the non-ideality and molecular asymmetry of P-sc-Hb and P-mal-Hb. P-sc-HbCO and P-mal-HbCO showed lower f/f₀ values than P-sc-Hb and P-mal-Hb, respectively. This indicated that CO binding can decrease the f/f₀ of P-sc-Hb and P-mal-Hb. Presumably, CO binding can tighten the tetrameric structure of P-sc-Hb and P-mal-Hb, thereby increasing their molecular symmetry and structurally stabilizing Hb.

Physical stability of the PEGylated proteins

The pH stability

The pH stability of the Hb and HbCO samples was evaluated by measuring their α-helix contents under different pH conditions. The α-helix content of Hb was stable at pH 5.0–8.5 (data not shown). All the Hb samples showed a decrease in the α-helix content at pH 3.0 (Figure 5), because Hb samples were unstable under acid condition. However, PEGylation and CO binding both can increase the retention in the α-helix contents of Hb at pH 3.0. Moreover, CO binding had higher ability than PEGylation to enhance the pH stability of Hb.

Figure 5. The pH stability of the Hb and HbCO samples. The pH stability was investigated by measuring the far UV CD spectra of the Hb and HbCO samples at pH 3.0. The ellipticity at 222 nm of each sample was recorded and the α-helix content of Hb at pH 7.4 was set as 100%. Bars represent mean ± S.D. from three repeated measurements.

Thermal stability

Thermal stability of the Hb and HbCO samples was investigated as a function of incubation time at 65°C. As shown in Figure 6, PEGylation can decrease the denaturation rate of Hb. Moreover, CO binding can significantly decrease the denaturation rate of P-sc-Hb and P-mal-Hb. In contrast, the denaturation rate of P-mal-Hb was higher than that of HbCO. Thus, the ability of CO binding to increase the thermal stability was higher than that of PEGylation.

Figure 6. Thermal stability of the Hb and HbCO samples as a function of incubation time. The samples were heated at 65°C in a water bath for different incubation time.

Thermal stability of the Hb and HbCO samples was also investigated as a function of incubation temperature for 30 min. As shown in Figure 7a, PEGylation and CO binding both can increase the retention in the α-helix of Hb upon heat treatment. As shown in Figure 7b, the T_m value of Hb can be increased by PEGylation and by CO binding. P-sc-HbCO showed the highest T_m in the samples, reflecting the synergistic effect of PEGylation and CO binding on the thermal stability of Hb. Since the T_m difference between P-sc-HbCO and P-sc-Hb is much higher than the one between P-sc-Hb and Hb, CO binding had higher ability than PEGylation to increase the thermal stability of Hb.

Figure 7. Thermal stability of the Hb and HbCO samples as a function of incubation temperature. The Hb and HbCO samples were heated at different temperature for 30 min. The T_m values of the samples (b) were measured according to the retention-temperature plot (a).

Anti-oxidation property

The metHb formation kinetics of the Hb and HbCO samples was shown in Figure 8. P-sc-Hb showed a metHb content higher than Hb and lower than P-mal-Hb during the incubation period at 37°C. Hb displayed a first-order autoxidation rate constant (k_ox) of 0.012 h⁻¹. In contrast, P-mal-Hb and P-sc-Hb displayed the k_ox of 0.020 h⁻¹ and 0.016 h⁻¹, respectively. This indicated that PEGylation can enhance the autoxidation of Hb. Moreover, PEGylation at Cys-93(β) that was proximal to the heme pocket can significantly increase the autoxidation of Hb. In contrast, CO binding of the Hb samples can abolish the occurrence of autoxidation, as reflected by the fact that metHb was essentially not observed in the HbCO samples. Presumably, CO binding can prevent ferrous (Fe²⁺) heme from the nucleophilic attack, thereby essentially inhibiting the metHb formation.

Figure 8. The metHb formation kinetics of the Hb and HbCO samples. The experiments were carried out by incubation of the Hb samples (25 μM as tetramer) in PBS buffer (pH 7.4) containing 300 μM EDTA at 37°C in sealed tubes.

Functional property

The functional property of the HbCO samples with CO release was measured and compared with that of the Hb samples. As shown in Table III, Hb showed a P₅₀ (oxygen affinity) of 13.0 mmHg with Hill coefficient (n) of 2.8. P-sc-Hb showed a P₅₀ of 7.0 mmHg with n of 2.0. P-mal-Hb showed a P₅₀ of 8.2 mmHg with n of 2.1. This suggested that PEGylation can increase the oxygen affinity of Hb. Interestingly, the HbCO samples with CO release showed slightly higher P₅₀ and n than the corresponding Hb samples. Presumably, CO binding can structurally stabilize Hb and improve the anti-oxidation property of Hb, which in turn improved the functional property of the Hb samples. The metHb contents in the Hb samples were over 15% after storage at 4°C for 30 days, which led to complete loss of oxygen-delivery capacity of the Hb samples. In contrast, the HbCO samples under this storage condition can maintain the oxygen-delivery capacity after CO release.

Table III. Oxygen-delivery properties of the PEGylated proteins.

Sample	P50 (mmHg)^a	n^b
Hb	13.0	2.8
HbCO^c	13.4	2.8
P-sc-Hb	7.0	2.0
P-sc-HbCO^c	7.9	2.0
P-mal-Hb	8.2	2.1
P-mal-HbCO^c	8.9	2.2

Oxygen equilibrium curves of the samples were measured at 37°C.

^aThe partial oxygen pressure at 50% saturation that is expressed in mmHg.

^bThe Hill coefficient.

^cThe HbCO samples were measured after CO release.

Discussion

Our present study was aimed to decrease or circumvent some side effects of the PEGylated Hb, such as structural perturbation, functional instability and autoxidation. Accordingly, CO binding was used to improve the structural, physical, functional and anti-oxidation properties of the PEGylated Hb.

Structurally, CO binding can alter the heme environment and molecular shape of the PEGylated Hb. Heme environment, the nonpolar environment around heme, can prevent heme from the contact with water or other polar molecules, thereby inhibiting oxidation of Fe²⁺ to Fe³⁺. Previous report suggested that the heme environment of Hb was perturbed by conjugation of the PEG chains, especially conjugation at the sites proximal to the heme pocket, such as at Cys-93(β) (Li et al. 2009). Our present study suggested that the heme environment of Hb can be structurally stabilized upon CO binding. Presumably, the relative position of porphyrin with the surrounding aromatic residues is changed upon CO binding and the side chain of Cys-93(β) points away from the surface of Hb, thereby stabilizing the Hb structure (Li et al. 2009). Our study also suggested that CO binding can slightly tighten the tetrameric structure of the PEGylated Hb and structurally stabilize the Hb tetramer, as reflected by its symmetric molecular shape and increased molecular density. Presumably, CO binding may interrupt the conformational transition of Hb from R state to T state.

Physically, CO binding can improve the thermal stability and pH stability of the PEGylated Hb in the present study. Consistently, Sakai et al. (1993b) reported that endothermic transition points of oxyHb and HbCO were observed at 62.2°C and at 77.7°C, respectively. Presumably, the energy barrier to break the heme-protein bond is greater in the PEGylated HbCO than that in the PEGylated Hb. Thus, higher temperature is needed to destabilize the PEGylated HbCO, as compared with the PEGylated Hb. In addition, the PEGylated HbCO is slightly more resistant to low pH than the PEGylated Hb. The higher stability of PEGylated HbCO may come from the tighter conformation of the HbCO protein core, which makes all the salt bonds and hydrogen bonds inside HbCO molecule more stable.

MetHb formation gives rise to a decrease in the oxygen-delivery capacity, generation of reactive oxygen species (ROS) and heme loss of Hb (Vandegriff et al. 2006) . Moreover, the PEG chains can increase the number of water molecules in the hydration layer of Hb and enhance the autoxidation by promoting the nucleophilic attack of heme. PEGylation may also alter the heme environment of oxyHb and increase the solvent accessibility to the heme pocket, thereby promoting the metHb formation (Hu et al. 2008). Our present study suggested that the autoxidation of the PEGylated Hb can be abolished upon CO binding. CO binding can strongly stabilize the bond between heme and CO and keep the iron atoms in the low-spin ferrous state (Baldwin 1980). This can prevent heme from the nucleophilic attack and avoid the autoxidation of the PEGylated Hb.

Functionally, the oxygen-delivery capacity of the PEGylated HbCO could be recovered after O₂ feeding under visible light irradiation (Sakai et al. 2011, Winslow 2000). Our present study suggested that the PEGylated HbCO with CO release even showed higher functional property than the PEGylated Hb. Interestingly, the PEGylated HbCO with CO release can maintain its oxygen-delivery capacity during long-term storage at 4°C. Moreover, metHb was not observed during this period. Thus, CO binding was useful for long-term storage of PEGylated Hb. Similarly, CO-MP4 did not oxidize and showed essentially no met-Hb formation over time (Vandegriff et al. 2008).

In summary, CO binding structurally stabilized the PEGylated Hb, abolished its metHb formation, and significantly increased its physical stability. In particular, the functional property of the PEGylated HbCO can be maintained during its long-term storage, which is of great significance for field transfusion.

Acknowledgments

The authors thank National Key Basic Research Program, No. 2013CB733604, Natural Science Foundation of China (Grant 21336010 and Grant 20906095) and 863 High-Tech Project, 2012AA02A406.

Declaration of interest

The authors report no declarations of interest. The authors alone are responsible for the content and writing of the paper.