Solid Phase Pegylation of Hemoglobin

Abstract

A solid phase conjugation process was developed for attachment of polyethylene glycol to hemoglobin molecule. Bovine hemoglobin was loaded onto an ion exchange chromatography column and adsorbed by the solid medium. Succinimidyl carbonate mPEG was introduced in the mobile phase after the adsorption. Pegylation took place between the hemoglobin on the solid phase, and the pegylation reagent in the liquid phase. A further elution was carried out to separate the pegylated and the unpegylated protein. Analysis by HPSEC, SDS-PAGE, and MALLS demonstrated that the fractions eluted from the solid phase contained well-defined components. Pegylated hemoglobin with one PEG chain was obtained with the yield of 75%, in comparison to the yield of 30% in the liquid phase pegylation. The P₅₀ values of the mono-pegylated hemoglobin, prepared with SC-mPEG 5kDa, 10kDa and 20kDa, were 19.97, 20.23 and 20.54 mmHg, which were much closer to the value of red blood cells than that of pegylated hemoglobin prepared with the conventional method.

• solid phase
• polyethylene glycol
• hemoglobin
• red blood cell substitute

INTRODUCTION

Modified hemoglobin has been used as various oxygen carriers for different medical purposes [4]. Among the modification methods, polyethylene glycol (PEG) modification, known as pegylation, is very effective to suppress the immunogenicity and prolong the circulation half life of the protein [1], [9]. Studies have demonstrated that pegylation can increase the molecular weight, decrease the renal excretion, reduce the proteolytic digestion, and stabilize the protein structure [3]. Clinical trials have been conducting on the effectiveness of pegylated hemoglobin as a red blood cell substitute and other purposes [7], [20].

Hemoglobin is a complex molecule having four subunits. There are many possible sites for the attachment of PEG molecules. For example, bovine hemoglobin has 46 lysine residue amino groups and 4 N-terminal amino groups [14]. These groups can react with common pegylation reagents in a random way, creating a mixture of products with varied number of PEG molecules attached to one protein at different sites. The random reaction in attachment number and site also resulted in the potential loss of bioactivity and the difficult reproducibility from batch-to-batch. In addition, the separation of reaction mixture needs extra steps of chromatography operation.

The concept of solid phase pegylation is to adsorb the protein onto a solid surface and to pegylate it in the adsorption mode. In this way, multiple site reaction could be minimized due to the hindrance of the solid medium. By selection of the adsorption site, it is possible to avoid the attachment of PEG to an unwanted site of a protein such as the catalytic pocket of an enzyme. Solid-phase bio-reaction has been used in peptide synthesis for a number of years [19]. However, the concept of solid-phase pegylation was proposed just 10 years ago in a conference [18]. A publication appeared recently on conjugation of interferon alpha using a cation exchange resin [16]. Although it is not highly selective, solid phase pegylation offers an alternative to the present liquid pegylation strategy which focuses on the reaction in a solution environment. Solid phase pegylation is still in its early development stage. Until now there has been no report on pegylation of hemoglobin. The author's laboratory reported conjugation of albumin and hemoglobin with ion exchange chromatography [11]. The concept could be extended to pegylation reaction.

In this paper, we investigated the pegylation of hemoglobin with ion exchange chromatography. Bovine hemoglobin was adsorbed onto the ion exchange column. Pegylation took place between the hemoglobin on the solid phase and the pegylation reagent in the mobile liquid phase. The reagent was succinimidyl carbonate monomethoxy polyethylene glycol (SC-mPEG), with molecular weights of 5000, 10000, and 20000, denoted as SC-mPEG5kDa, SC-mPEG10kDa, and SC-mPEG20kDa. By a linear gradient elution, the pegylated hemoglobin was removed from the column while unpegylated bHb was still adsorbed. Therefore a simultaneous reaction and separation procedure was realized through one chromatography process. Analysis was conducted with size exclusion high performance liquid chromatography (HPSEC), multi-angle laser light-scattering (MALLS), and polyacrylamide gel electrophoresis (SDS-PAGE). The adsorption of the protein on the chromatography medium surface could prevent or eliminate the multiple site attachment of PEG molecules due to possible steric hindrance.

MATERIALS AND METHODS

Materials

Bis (3, 5-dibromosalicyl) fumarate (DBBF) and N-hydroxysuccinimide (NHS) were purchased from Sigma. CM Sepharose Fast Flow, Superdex 200 (HR 10/30) was purchased from GE Healthcare (USA). Monomethoxy polyethylene glycols (mPEG5kDa, mPEG10kDa, and mPEG20kDa) were purchased from Union Carbide (Danbury, CT, USA). Conversion of mPEG5kDa, mPEG10kDa, mPEG 20kDa to their active derivatives of succinimidyl carbonate monomethoxy polyethylene glycol (SC-mPEG5kDa, SC-mPEG10kDa, and SC-mPEG20kDa) was carried out as described previously [21]. Bovine hemoglobin was obtained freshly from a local slaughterhouse. Stroma-free hemoglobin solution was prepared by following the procedures of Lu et al. [17].

All other chemicals obtained were of analytical reagent grade. Water used in the experiments was from Milli-Q ultrapure water purifications system (Millipore, Bedford, MA, USA).

Pegylation of Hemoglobin Using Conventional Liquid Phase Method

Hemoglobin was first intramolecularly crosslinked by bis (3, 5-dibromosalicyl) fumarate (DBBF), and then used as the starting material to prepare pegylated Hb. The process involved mixing 5 milliliters of 0.1 mM hemoglobin solution with 0.15mM DBBF in 0.05 M PBS buffer, pH 6.8, and allowed the reaction going on for 2 hours at 4°C. SC-mPEG was dissolved in the DBBF-Hb solution (pH6.8, 10 mM PBS), and the reaction was performed for 2 hours at 4°C (Figure 1).

Figure 1.  Simplified scheme of the reaction between succinimidyl carbonate mPEG (SC-mPEG) and hemoglobin.

Pegylation of Hemoglobin Using Solid Phase Adsorption Method

DBBF-Hb was adsorbed on to CM Sepharose Fast Flow at pH6.8 in 10mM PBS The feed solution was 5 milliliters of 2 mg /ml DBBF-Hb. The chromatography column was 16mm*150mm (GE Healthcare, USA) containing 5 milliliters of the ion exchange medium. SC-mPEG solution of 1 mg /ml was added one hour later at a flow rate of 1ml/min until 5 milliliters of solution had been loaded. The column was incubated for 2 hours at 4°C and then was eluted with 5-6 column volume (CV) of 10 mM PBS buffer at pH 6.8. Elution was initiated by applying a linear salt gradient from 0.1 M to 1.0 M NaCl in 10 mM PBS buffer at a rate of 1.0 ml/min. All chromatographic operations were performed with AKTA Purifier 10 Chromatography Station (GE Healthcare, USA). Elution fractions were detected using UV-VIS monitor at 280 nm and 405nm. Fractions were collected and analyzed by HPSEC, SDS-PAGE, and MALLS.

Qualitative Analysis of Pegylated Hemoglobin

High performance size exclusion chromatography (HPSEC) was performed on Agilent 1100 chromatography system (Agilent Co., USA) with a Superdex 200 column (10mm x 300mm), and 50 mM sodium phosphate containing 0.15 M sodium chloride (pH 7.4) as the mobile phase. The flow rate was set to 0.5 ml/min.

Multi-angle laser light scattering (MALLS) was conducted with DAWN EOS-OPTILAB DSP system (Wyatt Technology Corp., USA). The DAWN EOS detector contained a K10 cell and a He-Ne laser, λ = 690 nm and the OPTILAB DSP was a refractive index detector.

HPSEC was connected with MALLS in series. The liquid flow out of the column went through the UV detector of the Agilent 1100, and then through the DAWN EOS detector and the OPTILAB DSP detector. The mobile phase was 550 mM sodium phosphate containing 0.15 M sodium chloride (pH 7.4). The analysis was carried out at a flow rate of 0.5 ml/min at room temperature. The UV detector and the OPTILAB DSP detector were calibrated using sodium chloride and toluene, respectively. The DAWN EOS detector was normalized with bovine serum albumin as the sample according to the protocols recommended by Wyatt Technology. The signal from the DAWN EOS detector was denoted as MALLS signal while the signal from the OPTILAB DSP detector was named RI signal in this study. The molecular weight was calculated based on the MALLS and RI signals by using the ASTRA® software (Wyatt Technology, USA).

SDS-PAGE analysis was performed under standard Laemmli conditions [15]. The polyacrylamide concentration was 5% in the stacking gel and 12% in the separation gel using Mini-protein III cell (Bio-Rad, USA). Each sample (20 ml, about 10 mg) was mixed with the five-folds of the reducing buffer (4 ml), boiled for 5 min, and loaded into each well. Silver staining was used to visualize the proteins.

Biological Activity of Hemoglobin

The O₂ affinity of the pegylated hemoglobin was evaluated for P₅₀ (the O₂ pressure at which hemoglobin is half saturated) on a Hemox Analyzer (TCS Corp, PA, USA). The Hill coefficient was obtained from the O₂ equilibrium curves measured with Hemox Analyzer [5]. The concentrations of oxy-, deoxy-, and methemoglobin in pegylation products were calculated from their absorbance at three wavelengths (630, 576, and 560 nm) and the molar extinction coefficients were used for the calculation [8].

RESULTS

Pegylation of Hemoglobin Using Conventional Method

To compare with the new investigation of the solid phase, we first performed conventional liquid phase pegylation and analyzed the results. Figure 2 contained four HPSEC analytical chromatograms which related to the liquid pegylation products. Figure 2A showed natural hemoglobin before pegylation reaction. A single peak (28.1 min)was observed. Figure 2B, 2C, and 2D were the analytical results of pegylation with three reagents: SC-mPEG5kDa, SC-mPEG10kDa, and SC-mPEG20kDa. For each reagent, there were more than three peaks of HPSEC pattern, indicating the existence of multi-components of the reaction. Unmodified hemoglobin was found in 2B, 2C, and 2D as Peak 3 by its elution time. Pegylated products were in Peak 1 and Peak 2. By comparison of 2B, 2C, and 2D, one can find that the Peak 1 of Figure 2B appeared at similar elution time to the Peak 2 of Figure 2C (24.1 min and 24.3 min), indicating the possibility of similar molecular weights of the products. Because SC-mPEG10kDa (Figure 2C) is about twice SC-mPEG5kDa (Figure 2B) in molecular weight, if Peak 2 in Figure 2C is one SC-mPEG10kDa attached to one hemoglobin, Peak 1 in Figure 2B should be at least two SC-mPEG5kDa attached to one hemoglobin molecule. Furthermore, we can also find that Peak 2 of Figure 2D and Peak 1 of Figure 2C had a similar elution time (20.2 min and 20.1 min). Therefore, we could assume that Peak 1 of Figure 2C is at least di-pegylated hemoglobin with SC-mPEG10kDa if Peak 2 of Figure 2D is mono-pegylated protein with SC-mPEG20kDa.

Figure 2.  HPSEC analysis of pegylated hemoglobin in liquid phase. Samples were loaded on a Superdex 200 gel filtration column with 50 mM sodium phosphate containing 0.15 M sodium chloride (pH 7.4) as the mobile phase at a flow rate of 0.5 ml/min. A was the nature hemoglobin, and B to D were the results of pegylated-Hb with PEG5kDa, PEG10kDa, and PEG20kDa, respectively.

Due to its linear structure and its ability to associate a large amount of water molecules, it is possible that the attachment of one PEG molecule would hinder a further reaction of hemoglobin with another PEG molecule, especially for large PEG such as SC-mPEG20kDa. Therefore, we could speculate that Peak 2 of Figure 2D contained mono-pegylated hemoglobin. Furthermore, we could also assume that Peak 2 of Figure 2C and 2B all contained mono-pegylated hemoglobin with SC-mPEG 10kDa and SC-mPEG 5kDa, respectively. The Peak 1 in Figure 2B, 2C, and 2D contained di-pegylated and other multi-pegylated products. The experimental results revealed that conventional liquid phase pegylation produced a mixture of products with mono-, di-, or multi-attachment of PEG to a hemoglobin molecule.

Pegylation of Hemoglobin Using Solid Phase Adsorption Method

As described in the Materials and Methods section, the solid phase pegylation was also carried out for the same reaction time to the liquid pegylation. We could record the elution profile easily through the UV-VIS monitor change at the exit of the column during elution process. Figure 3 is such a profile. The first peak was produced during elution with 10 mM PBS buffer, pH 6.8. The second peak appeared when we increased the salt concentration up to 1.0 M NaCl in the same buffer of 10 mM PBS, pH 6.8. The pegylation reagent in Figure 3 was SC-mPEG5kDa. The left vertical axis is absorbance at 405nm, which reflects the concentration of hemoglobin, showing that both peaks might contain hemoglobin. The fractions corresponding to the first peak and the second peak, named F1 and F2, were collected and analyzed by HPSEC, MALLS, and SDS-PAGE. The same solid phase pegylation strategy was applied for SC-mPEG10kDa and SC-mPEG20kDa. The elution profiles of these two reactions were similar to Figure 3.

Figure 3.  Elution process of the pegylated hemoglobin using ion-exchange chromatography on CM Sepharose Fast Flow. The column was equilibrated with 10 mM PBS at pH6.8, and the column was eluted with 1M NaCl in 10 mM PBS at pH6.8. The flow rate was 1.0 ml min/ml and the elution profiles were monitored by absorbance at 280 nm and 405nm.

Compositional Analysis

High performance size exclusion chromatography (HPSEC) and multi-angle laser light scattering (MALLS) were used to determine the products of solid phase pegylation. The MALLS was connected with HPSEC in series as described in section 2.4. The analytical results showed that the first fraction F1 in Figure 3 was pegylated product; there was only one sharp peak during HPSEC analysis, while F2 fraction was a mixture of mainly unpegylated hemoglobin and a little pegylated hemoglobin (Figure 4). From the MALLS software calculation, the apparent molecular weight of the pegylated hemoglobin (PEG5kDa) was 99.7kDa, compared with 67.8 kDa of unpegylated hemoglobin. The difference is 32 kDa, about 6 times the pegylation reagent SC-PEG5kDa. The results of SC-PEG10kDa and SC-PEG20kDa are shown in Figure 5 and Figure 6 in which A is the HPSEC and B is the MALLS. Near uniform composition was found in HPSEC. The apparent molecular weights of the two reactions, analyzed with MALLS, were 119.1kDa and 156.kDa, respectively. The difference to native hemoglobin was 52kDa (about 5 times) and 89kDa (about 5 times).

Figure 4.  HPSEC-MALLS chromatograms analysis of PEG5kDa-modified hemoglobin by solid adsorption method. The samples were loaded on a Superdex 200 gel filtration column with 50 mM sodium phosphate containing 0.15 M sodium chloride (pH 7.4) as the mobile phase at a flow rate of 0.5 ml/min. A was the result of the first peak (F1) by HPSEC; B was the result of the first peak (F1) by HPSEC-MALLS; C was the result of the second peak (F2) by HPSEC; D was the result of the second peak (F2) by HPSEC-MALLS.

Figure 5.  HPSEC-MALLS chromatograms analysis of PEG10kDa-modified hemoglobin by solid adsorption method. The samples were loaded on a Superdex 200 gel filtration column with 50 mM sodium phosphate containing 0.15 M sodium chloride (pH 7.4) as the mobile phase at a flow rate of 0.5 ml/min. A was the result of the first peak (F1) by HPSEC; B was the result of the first peak (F1) by HPSEC-MALLS.

Figure 6.  HPSEC-MALLS chromatograms analysis of PEG20kDa-modified hemoglobin by solid adsorption method. The samples were loaded on a Superdex 200 gel filtration column with 50 mM sodium phosphate containing 0.15 M sodium chloride (pH 7.4) as the mobile phase at a flow rate of 0.5 ml/min. A was the result of the first peak (F1) by HPSEC; B was the result of the first peak (F1) by HPSEC-MALLS.

Figure 7 and Figure 8 demonstrate the results of SDS-PAGE for natural Hb, DBBF-Hb and pegylated Hb prepared using conventional method and the solid phase adsorption method. The main band of MW 16kDa is the subunit of hemoglobin. An extra band at 31kDa, existing in the lane of DBBF cross-linked Hb and the covalent bonds formed between subunits, could not be broken by mercaptoethanol. The apparent molecular weight of the pegylated hemoglobin (PEG5kDa) was about 30kDa in SDS-PAGE, more than the true molecular weight (16kDa + 5kDa). PEG10kDa and PEG20kDa also had larger apparent molecular weights than theoretical ones.

Figure 7.  SDS-polyacrylamide gel electrophoresis of Pegylated-Hb with PEG5kDa. The gel concentration is 13.5% and was stained by silver. Lane 1, natural hemoglobin; Lane 2, the result of intramolecularly cross-linked hemoglobin; Lane 3, standard protein markers; Lane 4, the result of modified-Hb by liquid phase; Lane 5, the result of the first elution peak by solid phase; Lane 6, the result of elution peak by solid phase.

Figure 8.  SDS-polyacrylamide gel electrophoresis of Pegylated-Hb with PEG10kDa and PEG20kDa. The gel concentration is 13.5% and was stained by silver. Lane 1, natural hemoglobin; Lane 2, the result of intramolecularly cross-linked hemoglobin; Lane 3, the result of PEG10kDa-modified hemoglobin by liquid phase; Lane 4, the result of PEG20kDa-modified hemoglobin by liquid phase; Lane 5, standard protein markers; Lane 6, the result of the first elution peak of PEG10kDa-modified hemoglobin by solid phase; Lane 7, the result of elution peak of PEG10kDa-modified hemoglobin by solid phase; Lane 8, the result of the first elution peak of PEG20kDa-modified hemoglobin by solid phase; Lane 9, the result of elution peak of PEG20kDa-modified hemoglobin by solid phase.

By comparison of Figure 7 and Figure 8, it can be found that the second PEG5kDa-hemoglobin band (the fourth band (from bottom to top) of lane 4 in Figure 7) and the first PEG10kDa-hemoglobin band (the third band (from bottom to top) of lane 6 in Figure 8) appeared at almost the same location; the second PEG10kDa-hemoglobin band (the fourth band (from bottom to top) of lane 3 in Figure 8) and the first PEG20kDa-hemoglobin band (the third band (from bottom to top) of lane 8 in Figure 8) were at almost the same location. Because SC-mPEG10kDa is about twice SC-mPEG5kDa in molecular weight, if the second PEG5kDa-hemoglobin band is two SC-mPEG5kDa attached to one hemoglobin, the first PEG10kDa-hemoglobin band should be one SC-mPEG10kDa attached to one hemoglobin molecule. Therefore, we could assume that the first pegylated hemoglobin band (Figure 7, lane 5; Figure 8, lane 6 and lane 8) were mono-pegylated hemoglobin with SC-mPEG5kDa, SC-mPEG10kDa, and SC-mPEG20kDa.

Bioactivity Examination

Figure 9 indicated the results of P₅₀. For natural bovine hemoglobin, the P₅₀ value showed about 28.52 mm Hg. The P₅₀ values for DBBF-Hb and PEG-Hb were decreased from the native value. For Hill coefficient, natural bovine Hb was 2.25, whereas those for DBBF-Hb, PEG-Hb in the liquid phase and in the solid phase were reduced correspondingly (Table 1). The content of MetHb for natural bovine Hb was 3.31, whereas those for DBBF-Hb, PEG-Hb by liquid phase, and solid phase were increased. However, compared with the PEG-Hb conjugates prepared using conventional method, the products prepared using solid phase adsorption method possessed better bioactivities, with better Hill coefficient and P₅₀.

Figure 9.  O₂ equilibrium curves of pegylated hemoglobin. A was the result of natural hemoglobin; B was the result of PEG5kDa-modified hemoglobin by solid phase; C was the result of PEG5kDa -modified hemoglobin by liquid phase.

Table 1. The biological activity of the pegylated hemoglobin and the yield of mono-pegylated-Hb

	P50(mmHg)	Hill coefficient	MetHb(%)	Mono-PEGylated-Hb(%)
Natural Hb	28.52	2.25	3.31	—
DBBF-Hb	24.09	2.04	7.27	—
Solid phase (PEG5kDa)	19.97	1.98	15.41	68.5
Liquid phase (PEG5kDa)	17.78	1.34	18.25	34.2
Solid phase (PEG10kDa)	20.23	1.85	14.23	70.4
Liquid phase (PEG10kDa)	18.35	1.35	16.83	38.1
Solid phase (PEG20kDa)	20.54	1.92	13.75	75.2
Liquid phase (PEG20kDa)	18.66	1.56	16.54	31.7

The biological activity parameters of the hemoglobin include the P₅₀ values, Hill coefficient, and the content of MetHb.

DISCUSSIONS

Succinimidyl carbonate monomethoxy polyethylene glycol (SC-mPEG) is known to react with ε-amino group of lysine residues, which account for approximately 10% of amino acids in a typical protein. Due to the presence of these multiple sites, conventional liquid conjugation between PEG and protein cannot be precisely controlled. The resulting products are often heterogeneous as illustrated by Figure 10A. This may raise a concern regarding product quality during the manufacturing process. Moreover, coupling of the PEG group could occur at or near the bioactive site of the protein, which may reduce its activity. In this method, hemoglobin was first adsorbed onto solid medium and then reacted with the PEG reagent. Due to the solid phase hindrance, the PEG could only react with the accessible sites of the adsorbed protein, hence minimizing random reaction. By proper selection of the adsorption site, it is possible that solid phase pegylation brings about better selectivity, as indicated in Figure 10B where pegylation took place at the opposite site of a protein to its adsorption site. The adsorption medium in this study was CM-Sepharose Fast Flow, a non-selective adsorbent. Therefore, we could not say that the pegylation is selective, but can say that it improved the selectivity compared with conventional liquid phase pegylation.

Figure 10.  Brief illustration of the pegylation of hemoglobin. A was common liquid phase method for preparation of pegylated hemoglobin; B was solid phase adsorption method for preparation of pegylated hemoglobin.

Because we used an ion exchange medium, the selection of pH during the process was very important. Hemoglobin has a pI about 7.0. We chose pH 6.8 as adsorption condition so that the protein was adsorbed by the medium. Upon pegylation, the pI of the product was reduced due to the reaction of amino group. The pegylated protein was then eluted easily by the washing buffer while unreacted protein was still on the medium. In this way, simultaneous pegylation and product separation were realized. This is certainly the advantage of the solid phase method.

The molecular weight value of SC-mPEG5kD conjugate, determined by MALLS and SDS-PAGE, was larger than its true values (64kDa + 5kDa). This behavior could be due to PEG ability to coordinate many water molecules per each ethylene unit and to its highly chain flexibility [10], [12], thus giving to PEG an apparent molecular weight about 5-10 times higher than that of a globular protein of a comparable mass, as verified by gel permeation chromatography [2], [13]. Furthermore, due to that attachment of polymer, the protein concentration was difficult to precisely determine by Bradford, Lowery, and other methods [6]. This might also be a major cause for the error of molecular weight estimation. Therefore the apparent molecular weights of the pegylated hemoglobin (5kDa, 10kDa, and 20kDa) were about 99.7kDa, 119.1kDa, and 156.kDa, respectively.

O₂ affinity of the hemoglobin is an important factor to determine O₂ delivery and unloading of tissues, and the decrease of P₅₀ is indicative of the increase of O₂ affinity. Hill coefficient is another important parameter, representing the cooperativity of hemoglobin. Some reaction might change the molecule conformation of Hb, or affect the conformation changing from tense (T) to relax (R) state. Solid phase pegylation gave a much better P50 and Hill coefficient, probably because of its better selectivity of the pegylation site and more uniform product than the liquid method. The methemoglobin (MetHb) content was also lower in the solid phase than in the liquid phase, probably due to its avoidance of high oxygen affinity.

Acknowledgements

The authors are thankful for financial support from Natural Science Foundation of China (Grant NO.20636010 and 20536050) and National Basic Research Development Program of China (Grant NO.2007CB714305) Declaration of interest: The authors report no conflicts of interest. The authors alone are responsible for the content and writing of the paper.