Expression, Purification, and Characterization of Mammalian and Earthworm Hemoglobins

Dissertation

Presented in Partial Fulfillment of the Requirements for the Degree Doctor of Philosophy in the Graduate School of The Ohio State University

Jacob James Elmer, B.S.

Chemical and Biomolecular Engineering Graduate Program

The Ohio State University

2011

Dissertation Committee:

Andre F. Palmer, Advisor

David Wood

Jessica Winter

Jacob James Elmer

2011

Abstract

The frequent shortages, risks of disease transmission, and storage issues associated with donated blood illustrate a significant demand for a red blood cell (RBC) substitute. Such a substitute should be able to effectively transport oxygen throughout the body with minimal side effects. Several hemoglobin-based oxygen carriers (HBOCs) have been developed and clinically tested, but they have all caused severe side effects. The problems associated with these HBOCs may all be attributed to removing hemoglobin from the RBC. Therefore, this work focuses on the use of the extracellular hemoglobin of the earthworm Lumbricus terrestris (LtEc) as a new class of HBOC. Since earthworms lack RBCs, their hemoglobin is freely dissolved in the bloodstream and has already adapted to solve many of the challenges facing modern synthetic

HBOCs. It has a lower rate of oxidation, avoids harmful side reactions with nitric oxide (NO), and it is extremely stable. We have developed a novel purification technique to highly purify large amounts of LtEc at costs that are comparable to donated blood. The LtEc product also transports oxygen similarly to human blood. Transfusion of LtEc into hamsters does not elicit the harmful side effects observed with other HBOCs and preliminary studies have not revealed any immune or allergic reactions in vivo. Therefore, this work shows that LtEc might be an effective and safe oxygen carrier that warrants further study and suggests the need for a paradigm shift in the HBOC field from cellular to extracellular hemoglobins.

Dedication

I dedicate this work to my Mother, for her dedication to me.

iii

Acknowledgements

This work is the product of collaborations between myself and many other individuals. I would like to thank Dr. Andre Palmer for giving me one of the best graduate experiences possible. I would also like to thank my lab mates for their continuous support and friendship:

 David Harris

 Guoyong Sun

 Ning Zhang

 Guo Chen

 Sharon Gunderson

 Alex Roth

 Jorge Fontes

 Shahid Rameez

 Yipin Zhou

 Uddyalok Banerjee

I would also like to thank the faculty of the chemical engineering department for being excessively generous with their equipment and their advice. Specifically, I would like to thank

Dr. David Wood for mentoring me towards the end of my graduate tenure. Much of this work would not have been possible without him.

Last but not least, I would like to specifically thank all of my undergraduate assistants.

Besides conducting experiments, my students also inspired me to work harder and perform some experiments which I wouldn’t have done otherwise. I wish them all the best (in chronological order):

 Henry White

 Mark Politz

 Katie Zorc

 Ilse Fernandez

 Zeinab Mohammed

 Parth Patel

Vita

1999-2003 ...... Seckman Senior High School

2003-2007 ...... B.S. Chemical Engineering

B.S. Biological Sciences

Missouri University of Science & Technology

Summer, 2005 ...... Trainee, NASA Space Flight and Life Sciences Training Program

Summer, 2006 ...... Intern, Donald Danforth Plant Sciences Center

2004-2007 ...... Undergraduate Researcher, Missouri University

of Science and Technology

2007-present ...... Graduate Research Associate, Department of

Chemical Engineering, The Ohio State University

PUBLICATIONS

Elmer J, Harris DR, Sun G, Palmer AF. Purification of hemoglobin by tangential flow filtration with diafiltration. Biotechnol Prog. 2009 25(5):1402-10.

Elmer J, Buehler PW, Jia Y, Wood F, Harris DR, Alayash AI, Palmer AF. Functional comparison of hemoglobin purified by different methods and their biophysical implications. Biotechnol Bioeng 2010 106(1):76-85.

Elmer J, Cabrales P, Wang Q, Zhang N, Palmer AF. Synthesis and biophysical properties of polymerized human serum albumin. Biotechnol Prog. 2011 27(1):290-6.

Elmer J, Harris D, Palmer AF. Purification of hemoglobin from red blood cells using tangential flow filtration and immobilized metal ion affinity chromatography. J Chromatogr B Analyt Technol Biomed Life Sci. 2011 879(2):131-8.

FIELDS OF STUDY

Major Field: Chemical and Biological Engineering Specialization: Protein expression and purification

vii

TABLE OF CONTENTS

Abstract ...... ii Dedication ...... iii Acknowledgements...... iv Vita……...... vi List of Tables ...... xi List of Figures ...... xiii CHAPTER 1: INTRODUCTION TO TRANSFUSION, HEMOGLOBIN, AND BLOOD SUBSTITUTES ..... 1 1.1- The Past, Present, and Future of Transfusion ...... 1 1.2 - Hemoglobin...... 2 1.3 – Hb Dissociation, Oxidation, and Interactions with Nitric Oxide ...... 7 1.4 - Previous Generations of HBOCs...... 13 1.5 - References ...... 14 CHAPTER 2: PURIFICATION OF HEMOGLOBIN BY TANGENTIAL FLOW FILTRATION ...... 22 2.1 – Introduction ...... 22 2.2 - Materials and Methods ...... 25 2.3 – Advantages and Disadvantages of Diafiltration ...... 32 2.4 – Comparison of TFF-Purified Hbs and Commercially Prepared Hbs ...... 36 2.5 - Conclusion ...... 45 2.6 – References ...... 45 CHAPTER 3: PURIFICATION OF HEMOGLOBIN BY IMMOBILIZED METAL AFFINITY CHROMATOGRAPHY ...... 48 3.1 – Introduction ...... 48 3.2 – Materials and Methods ...... 51

viii

3.4 - Conclusion ...... 60 3.5 - References ...... 60 CHAPTER 4: INTRODUCTION TO ERYTHROCRUORIN ...... 62 4.1 - Extracellular Hbs: A New Paradigm ...... 62 4.2 - Structure and Stability of LtEc ...... 63

4.3 - O2 Transport by LtEc ...... 66 4.4 - Autoxidation of LtEc ...... 67 4.5 - Interactions between LtEc and other Ligands ...... 69 4.6 - Preliminary Animal Studies with LtEc ...... 70 4.7 - Conclusion ...... 70 4.8 - References ...... 71 CHAPTER 5: PURIFICATION AND IN VITRO CHARACTERIZATION OF ERYTHROCRUORIN ...... 76 5.1 – Introduction ...... 76 5.2 – Materials and Methods ...... 78 5.3 – Results and Discussion ...... 80 5.4 - Economic Considerations...... 91 5.5 – Conclusion ...... 92 5.6 – References ...... 93 CHAPTER 6: TRANSFUSION OF ERYTHROCRUORIN IN AN EXTREME HEMODILUTION MODEL . 95 6.1 - Introduction ...... 95 6.2 - Materials and Methods ...... 96 6.3 - Results ...... 101 6.4 - Conclusion ...... 110 6.5 - References ...... 111 CHAPTER 7: EXPRESSION OF HEME PROTEINS IN E. COLI ...... 113 7.1 – Introduction: A Recombinant Hb Review ...... 113 7.2 - Materials and Methods ...... 115 ix

7.3 – Results and Discussion ...... 118 7.4 – Conclusion ...... 126 7.5 – References ...... 127 CHAPTER 8: FUTURE WORK ...... 130 8.1 – Mammalian Hemoglobin Purification ...... 130 8.2 – LtEc Purification ...... 131 8.3 – LtEc Characterization (in vitro)...... 132 8.4 – Animal Studies with LtEc ...... 134 8.5 – Expression and Purification of Recombinant Globin Proteins ...... 134 Comprehensive Bibliography ...... 137 Appendix A: Vector Sequences used for Gene Expression ...... 156 Appendix B: Protein Amino Acid Sequences and Molecular Weights ...... 163

List of Tables

Table 2.1 – Technical specifications of the TFF cartridges ...... 26

Table 2.2 – Hb recovery after each stage ...... 32

Table 2.3 – Initial and Final Purity of bhb with and without diafiltration ...... 34

Table 2.4 – Endotoxin level of stage III retentate samples ...... 34

Table 2.5 – Oxygen affinity and cooperativity of bHb samples compared to bovine RBCs ...... 36

Table 2.6 – Oxygen affinities and cooperativities of Hbs ...... 42

Table 2.7 – Ligand binding kinetics of Hbs ...... 43

Table 2.8 – Autoxidation rate constants of Hbs with and without catalase & SOD ...... 44

Table 3.1 – MetHb level of purified Hb products ...... 56

Table 3.2 – Oxygen affinity and cooperativity of purified Hbs ...... 59

Table 4.1 – Size, molecular weight, oxygen affinity, and cooperativity of HbA, AmEc, LtEc, and human RBCs...... 66

Table 4.2 – Autoxidation rates and redox potentials of HbA, LtEc, and AmEc ...... 68

Table 5.1 – Average yield and percent of oxidized heme in purified LtEc samples ...... 82

Table 5.2 – Oxygen affinity and cooperativity of LtEc compared to HbA and bHb ...... 83

Table 5.3 – Ligand binding kinetics for purified LtEc with HbA and bHb as controls ...... 85

Table 5.4 – Viscosity and COP values of LtEc at 5 & 10 g/dL ...... 86

Table 5.5 – Maxima (nm)/Extinction Coefficients (, AU/((mM heme)(cm)) of LtEc and HbA

spectra ...... 88

Table 5.6 – Maxima (nm)/Extinction Coefficients (, AU/((mM heme)(cm)) of metLtEc and

metHbA spectra ...... 89

Table 6.1 – Properties of infused HBOCs ...... 102

Table 6.2 – Systemic parameters of all animals before and after exchange transfusion ...... 103

Table 7.1 – Reaction rate constants of recombinant globins with O2, CO, and NO...... 121

xii

List of Figures

Figure 1.1 – Hb and the heme structure. Hb and the heme structure ...... 3

Figure 1.2 – Differential O2 affinity of Hb ...... 4

Figure 1.3 – Oxygen equilibrium curves for blood & PFCs and purified bovine and human Hbs .... 5

Figure 1.4 – Formation of 2,3 BPG from glucose through glycolysis and the enzyme 2,3 BPG mutase ...... 6

Figure 1.5 – Oxidation of HbA heme (Fe2+ to Fe3+) over a period of 3 days at 37oC ...... 10

Figure 1.6 – Reaction pathways involving Hb, metHb, NO, nitrite, O2, and H2O2 ...... 12

Figure 2.1 – TFF purification scheme for Hbs...... 24

Figure 2.2 – PAGE analysis of Hb samples after every stage with diafiltration and without diafiltration ...... 33

Figure 2.3 – Oxygen equilibrium curves for bovine RBCs and TFF-purified bHb ...... 35

Figure 2.4 – Comparison of AEX and TFF Hb purification strategies ...... 38

Figure 2.5 – SDS-PAGE analysis of Hb samples ...... 39

Figure 2.6 – SEC profiles of Hb samples ...... 40

Figure 2.7 – Direct infusion ESI-MS analysis of Hb preparations ...... 41

Figure 2.8 – Catalase activity of Hb samples (47) ...... 42

Figure 2.9 – Effects of IHP, NaCl, pH, and temperature on the O2 affinity (P50) of Hbs ...... 44

Figure 3.1 – Schematic of different IMAC purification strategies...... 50

Figure 3.2 – PAGE analysis of samples from IMAC purification of bHb, HbA, and cHb ...... 55

Figure 3.3 – MALDI analysis of the IMAC-purified Hbs ...... 56

Figure 3.4 – OECs of HbA and bHb samples before and after IMAC purification ...... 58 xiii

Figure 3.5 – OECs of cHbA and cHbD samples before and after IMAC purification ...... 58

Figure 4.1 – Assembly of LtEc ...... 65

Figure 5.1 – LtEc purification scheme ...... 77

Figure 5.2 – SDS-PAGE analysis of samples from the LtEc purification process ...... 82

Figure 5.3 – MALDI-TOF analysis of TFF-purified LtEc ...... 84

Figure 5.4 – OECs for LtEc, bHb, and HbA ...... 84

Figure 5.5 – UV-Vis spectra of purified LtEc and HbA with various ligands ...... 87

- Figure 5.6 – UV-Vis spectra of purified metLtEc and metHbA with and without bound NO2 ...... 88

Figure 5.7 – Interaction of HbA and LtEc with NO ...... 90

Figure 5.8 – Effects of KCN and K3Fe(CN)6 on HbA and LtEc ...... 91

Figure 6.1 – Hemodilution protocol ...... 98

Figure 6.2 – Partial pressure of O2 (pO2) in arterioles, tissues, and venules after transfusion of test solutions ...... 104

Figure 6.3 – Effects of test solutions on functional capillary density ...... 105

Figure 6.4 – Effects of test solutions on mean arterial pressure and heart rate ...... 106

Figure 6.5 – Effects of test solutions on arteriolar diameter and blood flow ...... 107

Figure 6.6 – Pharmacokinetics and in vivo oxidation of each test solution...... 109

Figure 6.7 – Effects of LtEc on mean arterial pressure, heart rate, arteriolar diameter, and functional capillary density after consecutive injections over a period of 5 days ... 110

Figure 7.1 – Expression vectors for recombinant globins ...... 116

Figure 7.2 – Expression and purification scheme for recombinant globins...... 117

Figure 7.3 - PAGE analysis of recombinant , and Mb samples ...... 120

xiv

Figure 7.4 – MALDI analysis of purified , and Mb ...... 122

Figure 7.5 – O2 release by recombinant globins ...... 123

Figure 7.6 – CO binding concentration dependence of recombinant globins ...... 124

Figure 7.7 – CO binding time courses for recombinant globins ...... 125

Figure 7.5 – NO dioxygenation concentration dependence by recombinant globins ...... 125

Figure 7.9 – NO dioxygenation time courses for recombinant globins ...... 126

CHAPTER 1: INTRODUCTION TO TRANSFUSION, HEMOGLOBIN, AND BLOOD SUBSTITUTES

1.1- The Past, Present, and Future of Transfusion

Even our earliest ancestors understood the vital importance of blood. However, it has taken hundreds of years for us to understand the true nature of blood and how it works. The

Greek philosopher Galen (ca. 180 AD) was the first to show that arteries and veins contain blood1, but over 1,400 years would pass before William Harvey proved that blood circulates from the lungs and heart throughout the rest of the body.2 Shortly thereafter, the invention of the microscope allowed Swammerdam and Malpighi to observe red blood cells (RBCs) and capillaries.3, 4 Oxygen and its role in respiration was then discovered by Priestley and Lavoisier in the late 1700’s.5, 6 The science of transfusion also began to grow in the 1700’s, culminating in the first successful transfusion of human blood by James Blundell in 1818.7 Transfusions became increasingly popular over the next century, but patients were plagued with severe or fatal allergic reactions until the A, B, O, and Rh blood types were discovered and blood type matching began in the early 1900s.8 The introduction of nontoxic anticoagulants (i.e. sodium citrate) also made short term storage of donated blood possible, allowing the Red Cross to begin the first national blood bank in 1941.8 Since then, blood transfusions have become commonplace and save millions of lives each year.

While transfusions are usually considered safe and free of side-effects, they are still burdened with a few problems. For example, transfusion of donated RBCs poses risks of disease

1 transmission (HIV, hepatitis, variant Creutzfeldt-Jakob disease, & West Nile virus)9-15, immune system suppression16-21, acute lung injury22-24 , and allergic reactions. It is important to mention that most of these risks are extremely low in the developed world, but they are much higher in third world countries that lack proper disease screening techniques or processing facilities.25

Donated RBCs also possess a short shelf life (42 days)26 and must be refrigerated, making them inaccessible in some emergency or battlefield situations. The supply of RBCs is also frequently limited by seasonal shortages (usually during the summer months) and shortages of rare blood types (A-, B+, B-, and O-).27 The frequency of these shortages may soon be dangerously high, since the demand for transfusions is increasing twice as fast as the rate of blood donations (6-

8%/year vs. 2-3%/year).28 If this trend continues, it is estimated that the U.S. will suffer an annual shortage of 4 million units of blood per year by 2030.29

All of these problems clearly illustrate the need for an alternative to donated RBCs.

Such an alternative must be an effective O2 transporter, available at low cost in large amounts, and stable at ambient temperature for long periods of time. Several attempts have been made to create such an alternative O2 carrier (AOC), leading to the formation of two classes of AOCs – perfluorocarbons (PFCs) and hemoglobin (Hb)-based oxygen (O2) carriers (HBOCs). PFCs

(Fluosol & Oxygent) showed promise in early clinical trials, but their development was terminated after clinical trials revealed short circulation times and increased risk of stroke.27, 30-36

1.2 - Hemoglobin

The other class of AOCs, HBOCs, uses Hb as the precursor since it has many unique

properties that make it an excellent O2 carrier. It is a tetrameric protein consisting of two  and two  subunits. Each subunit contains a heme with an iron center that is able to bind a single

37 molecule of O2. Therefore, each tetramer is able to bind 4 molecules of O2 or roughly 1.34 mL of O2 per gram Hb. Hemoglobin is found in high concentrations (30-40 mg/mL) within RBCs. As

38 a result, the O2 capacity of human blood (140 mg/mL Hb) is 70 times higher than plasma.

Figure 1.1– Hb and the heme structure. The porphyrin ring structure of the heme is shown on the left, with the central iron atom coordinated by four nitrogen atoms and highlighted in red. Within the heme binding pocket of Hb, a histidine residue coordinates one side of the heme, while the other side is free to bind O2 (middle). Each subunit of the Hb tetramer (right) contains a heme group.

A good O2 carrier must do much more than simply increase the O2 capacity of blood.

39 After all, PFCs increased the dissolved O2 concentration in plasma (10-20 times) , but they were ultimately unsuccessful. The key difference between PFCs and Hbs is that the Hb subunits are able to carefully control O2 release as they circulate through the body. Max Perutz was the first to reveal the structure of Hb and explain this phenomenon in detail (for which he was later

40 awarded the Nobel Prize). He showed that O2 binding induces conformational shifts in the heme pockets of the remaining subunits which increase their O2 affinity. Hb in the 3 deoxygenated state has the lowest affinity for O2, but when one of the Hb subunits binds an O2 molecule, the O2 affinity of the remaining deoxy-Hb subunits increase (see Figure 1.2). A similar effect can be observed during O2 release. Fully saturated Hb has a high O2 affinity that rapidly

41 decreases as O2 molecules are released. As a result, Hb tends to quickly release and bind all of its O2 at a specific low O2 concentration (pO2 = 10-40 mm Hg), rather than slowly releasing it.

This unique trait allows Hbs to quickly take up O2 in the lungs and hold on to the O2 until it reaches the capillaries, which have a low partial pressure of O2 that favors O2 release.

Figure 1.2 – Differential O2 affinity of Hb. Globin subunits are shown in red, hemes are black, and O2 molecules are light blue. Deoxygenated Hb has a low O2 affinity, but every time an O2 molecule binds to one of the hemes the O2 affinity of the remaining subunits increases.

The effects of O2 affinity are best illustrated by the O2 equilibrium curves shown in

Figure 1.3. PFCs are unable to regulate O2 release, so their rate of O2 release follows Henry’s

27 Law and is linearly proportional to the pO2 of the solution. In contrast, RBCs (whole blood in

Figure 1.3) retain most of their O2 reserve until the pO2 reaches a critical point around 4-40 mm

Hg (pO2 within muscle tissue). The most important consequence of this difference is that PFCs

4 will release much of their dissolved O2 before reaching the extremities, while Hbs are able to delay O2 release until they reach the capillaries.

The right panel of Figure 1.3 shows Hill plots of purified bovine Hb (bHb) and human Hb

(HbA). This data can be used to calculate the P50 and cooperativity coefficient. The P50 value, a measure of O2 affinity, is the partial pressure of O2 at which half of the hemes in the Hb solution are saturated with O2 (Y = 0.5). By drawing vertical lines down from each curve at Y = 0.5, we can see that bHb has a P50 around 25 mm Hg, while HbA has a P50 around 13 mm Hg. Practically speaking, a lower P50 value indicates that the Hb solution holds on to its O2 tighter and has a higher O2 affinity. For example, since the P50 of HbA is less than bHb, HbA has a higher O2 affinity. We can also use these plots to estimate cooperativity (i.e. the Hill coefficient or “n”), which is a way of measuring how much the subunits influence each other after they bind consecutive O2 molecules. In general, a steeper slope indicates a higher cooperativity. For example, both bHb and HbA, and blood have high cooperativity (n = 2.9) and a sigmoidal O2 release curve. PFCs are not cooperative (n=1.0) and exhibit a linear release curve.

Figure 1.3 – Oxygen equilibrium curves for blood & PFC’s (left graph, modified from Reiss27) and purified bovine and human Hbs (right graph, my data). In both plots, the x-axis represents the partial pressure of O2 (pO2) within the solution while the y-axis reflects the O2 content of the 5 solution (left plot) or O2 saturation of the Hb (right plot). Dashed lines represent raw data, while solid lines have been fit to the Adair equation.

The O2 affinity of Hb is also controlled by two allosteric effectors within the RBC – protons (H+) and 2,3 biphosphoglycerate (2,3 BPG). 2,3 BPG may be formed as an offshoot of glycolysis (instead of ATP and pyruvate) by the enzyme 2,3 BPG mutase. 2,3 BPG has a high negative charge (-5) that tightly binds to a site between the beta subunits of HbA that has several positively charged amino acids (2 amino termini, 2 lysines, and 4 histidines). By neutralizing these otherwise repulsive charges, 2,3 BPG brings the subunits closer together and

42 causes conformational changes throughout all of the subunits that decrease O2 affinity.

However, the O2 affinity of HbA increases again when 2,3 BPG is removed during purification, causing the left shift in its O2 equilibrium curve shown in Figure 1.3. 2,3 BPG does not bind to bHb, so the O2 affinity of pure bHb is similar to whole bovine blood and there is no shift in its O2 equilibrium curve.

Figure 1.4 – Formation of 2,3 BPG from glucose through glycolysis and the enzyme 2,3 BPG mutase (left). 2,3 BPG tightly binds to positively charged amino acids between the 1 and 2 subunits, shown in blue (right). 6

Protons and pH also influence the O2 affinity in a phenomenon called the “Bohr Effect.”

As the pH of an Hb solution drops, amino acid residues become protonated by the excess protons in solution. The changes in the size and charge of the residues cause conformational changes in both HbA and bHb which decrease O2 affinity. This phenomenon is crucial for proper

O2 release within muscle tissue. During exercise, cellular respiration consumes O2 and produces

CO2 that accumulates in the capillaries and forms carbonic acid, which may then generate two protons:

+ - + -2 H2O + CO2  H2CO3  H + HCO3  H + CO3

+ The increase in proton concentration and drop in pH (pH = -log[H ]) then causes the O2 affinity of Hb to drop. As a result, more O2 is released where it is needed most – in active tissues with low pO2 and high pCO2. Hb also transports CO2 away from the capillaries and prevent a severe drop in pH (a harmful condition known as acidosis).43

1.3 – Hb Dissociation, Oxidation, and Interactions with Nitric Oxide

Despite all of its advantages, Hb still has three major flaws – tetramer dissociation, susceptibility to oxidation, and undesired side reactions with the physiological signaling molecule nitric oxide (NO). Hb is protected from all of these hazards within the RBC, but special care must be taken to avoid them after purifying Hb. For example, consider the stability of the

HbA or bHb tetramers. The subunits are held together solely by electrostatic and hydrophobic interactions, no covalent linkages are formed during tetramer assembly. Therefore, the subunits dissociate and associate freely, creating different equilibria (tetramer to dimer and/or dimer to monomer) at various concentrations of Hb. At the high concentrations of Hb inside the

RBC (300-400 mg/mL), Hb exists solely as a tetramer.44, 45 However, at lower concentrations

near the dissociation constant (KD) of Hb, the 22 tetramer begins to dissociate into  dimers.

The dissociation constant of Hb is highly variable and depends on many different factors, including the presence of O2, pH, temperature, and even chlorine ions (Cl-). For example, the dissociation constant of oxy-HbA is 3.2 M at pH 6.0, falls sharply to 32 nM around pH 8.5, and increases again at higher pH values. This trend indicates that the oxy-HbA tetramer is most stable at pH 8.5 (at the maximum KD value) and should be mostly in the tetrameric form at concentrations above 32 nM at that pH. In contrast, deoxy-HbA dissociation constants are several magnitudes lower than oxy-HbA and tend to linearly increase with pH (from KD = 1 pM at pH 6.5 to KD = 10 M at pH 11.0). Since most Hb solutions are actually a mixture of oxy- and deoxy-Hb, the best way to prevent dimerization is to keep Hb concentrations above 50 M (Hb concentration inside RBCs = 4-6 mM).46

Dimerization may not occur within the RBC, but HbA is known to dimerize quickly in the plasma where the HbA concentration is much lower. Once dissociated in the plasma, the dimers are usually safely sequestered and processed by haptoglobin47, a serum protein which tightly binds the HbA dimer. However, if large amounts of Hb dimer accumulate and haptoglobin is depleted, the dimers are small enough to extravasate into the tissues or accumulate in the kidneys. The dimers then degrade over time, releasing toxic amounts of heme into the surrounding cells.48-50

The heme iron within each Hb subunit is also highly susceptible oxidation. With respect to Hb, the term oxidation is used to describe the transition of iron from the Fe2+ state to Fe3+ and even Fe4+ states. Oxidation is physiologically significant because Hb with hemes in the Fe3+ state (metHb) is unable to bind or transport O2. Instead, metHb is unstable and generates free

8 radical species. Oxidation may be caused by any free radical species or oxidizing agent, but most

- Hb oxidation in vivo is caused by either hydrogen peroxide (H2O2) or superoxide (O2 ) generation.

H2O2 is released by white blood cells during infection or injury and released into the blood

2+ 3+ 4+ stream, where it may react with Hb. H2O2 quickly reacts with Fe to form Fe and even Fe at

51 high concentrations. A special type of oxidation, autoxidation, may also occur while O2 is bound to Hb. The bound O2 can strip an electron away from the heme iron and leave the heme

- 3+ - pocket, creating O2 and Fe . The O2 may then react with water to produce H2O2 which oxidizes another heme iron, making this an autocatalytic reaction (hence auto-oxidation).

Within the RBC, all of these reactions are kept in check by a set of antioxidant enzymes,

- superoxide dismutase and catalase. Superoxide dismutase reacts with O2 to form molecular O2,

52-54 while catalase catalyzes the transformation of H2O2 into H2O. The RBC also has other antioxidant enzymes, like metHb reductase, which are able to reduce metHb to functional Hb

(an otherwise irreversible reaction). Unfortunately, purifying Hb from the RBC removes all of these beneficial enzymes, making pure HbA and bHb more susceptible to oxidation. The oxidation rate of Hb may be calculated by measuring the concentration of Fe2+ spectroscopically over time and fitting the data to a double exponential decay function:

Figure 1.5 – Oxidation of HbA heme (Fe2+ to Fe3+) over a period of 3 days at 37oC.

Where b is the initial (fast) rate of oxidation and d is the overall (slightly slower) rate of

-1 oxidation, Kox, with units of hr . The overall rate of oxidation is most commonly reported and is approximately twice as much for bHb (0.26 hr-1) as HbA (0.13 hr-1). Increasing temperature,

[H2O2], and even [Cl-] can all increase the oxidation rates of bHb and HbA. The oxidation rates of dimers are also higher than tetramers. At any rate, oxidation of the heme can create several problems. First of all, metHb is less stable and prone to precipitation, making long term storage of Hbs a problem. In vivo, metHb can create free radicals which increase oxidative stress on tissues, causing lipid peroxidation in cell membranes and other problems.51 Fortunately,

o oxidation may be avoided by storing Hb at -80 C and limiting exposure to O2 and other oxidizing agents.

Hb O2 transport has been studied for decades, but recent studies have shown that Hb may have other physiologically important side reactions as well. The most important side reactions relate to NO scavenging and nitrite reduction. In the bloodstream, NO is produced by 10 the endothelial cells which line the blood vessels. The NO then diffuses to nearby smooth muscles cells and signals them to relax, thereby increasing blood vessel diameter (vasodilation) and decreasing blood pressure.55 In contrast, if NO levels are lower than normal, blood vessel diameter decreases (vasoconstriction) and blood pressure increases (hypertension). When Hb is released into the blood stream, it can scavenge NO in a reaction known as NO dioxygenation:56

The two most important consequences of this reaction are that NO is consumed and the heme is oxidized. The disappearance of NO induces vasoconstriction and hypertension, while the oxidized heme loses its ability to transport O2. This undesired reaction is usually kept to a minimum within the RBC, since the RBC membrane creates a barrier between Hb and NO.57

However, in cases of hemolysis or during clinical trials of some HBOCs (see next section), the hypertensive effect is quite severe.

In contrast to its harmful NO scavenging activity, Hb is also able to generate NO by reducing nitrite. The complex network of reactions between Hb and nitrite is shown in Figure

1.6. Both oxy-Hb58 and deoxy-Hb59 react undergo redox reactions with nitrite to form metHb and NO. It has been suggested that this reaction may be a significant source of NO in vivo.60

The mechanism of NO escape from the RBC after nitrite reduction is still under debate, since it seems likely that any NO generated should be consumed by oxy-Hb through NO dioxygenation.

However, both deoxy-Hb and metHb are able to bind NO, forming a stable intermediate that might avoid dioxygenation by oxy-Hb. NO may also reversibly nitrosylate free thiols (SH groups, specifically 93Cys) on the surface of Hb, yielding Hb-SNO.61 Others have suggested another possibility in which nitrite reacts with NO inside the heme pocket to form N2O3, a relatively more

11 stable molecule that does not interact with oxy-Hb and may quickly diffuse out of the RBC. The

62, 63 N2O3 may then spontaneously degrade into NO and nitrite in the blood stream.

- Both NO and NO2 have also been observed to protect against the harmful effects of

4+ - 4+ Fe =O. As shown in the bottom panel of Figure 1.6, NO and NO2 can react with Fe =O to form

- - 4+ 3+ NO2 and NO3 , respectively, while reducing Fe =O to Fe .

Figure 1.6 – Reaction pathways involving Hb (red), metHb (brown), NO (green), nitrite, O2, and

H2O2. Species with unique UV-Vis spectra are marked with red stars, while species with unique electron paramagnetic resonance (EPR) spectra are marked with green stars.

1.4 - Previous Generations of HBOCs

Since Hb appears to be an ideal and natural O2 carrier, researchers have been trying to use it as an alternative O2 carrier for decades. Early attempts to use purified HbA as an AOC caused severe side-effects, such as kidney failure and increased mortality.48-50 Renal failure was later determined to be caused by the dissociation of HbA tetramers and toxic accumulation of heme in the kidney. To prevent dimerization and its side-effects, Baxter Healthcare developed two new HBOCs (HemAssist64-69 and recombinant HbA 1.170-75) with cross-linked subunits. The kidney toxicity of these products was much lower, however, they were discontinued after they were found to cause severe hypertension and increased the risk of death in clinical trials. Other companies tried to improve on Baxter’s methods by polymerizing human (Hemolink76-78 &

PolyHeme45, 79) and bovine (Hemopure80-83) Hb to yield high molecular weight (MW) HBOCs

(128-500 kDa). Unfortunately, these products were discontinued after they were also shown to cause hypertension, cardiac toxicity, and an increased risk of myocardial infarction (MI) and death.84

The adverse effects associated with these HBOCs were likely due to heme oxidation85, extravasation, and side reactions with NO.55, 86 In vivo metHb levels as high as 40% were observed in some clinical trials71, suggesting that oxidative stress may have been responsible for the cardiac toxicity and MIs observed in vivo. It is important to mention that another HBOC,

MP487-89 (a PEGylated HbA developed by Sangart Inc., currently in phase III clinical trials) also scavenges NO, but does not cause vasoconstriction. It is able to avoid vasoconstriction by reacting with endogenous nitrite to form NO, thereby maintaining NO levels.90 Unfortunately, this reaction oxidizes the Hb and increases oxidative stress. MP4 has also been shown to have increased rates of heme loss91, oxidation91, and dimerization92 in vitro.

Other HBOCs with novel properties are currently in development, including encapsulated Hb93, higher MW Hb polymers (ZL-BvHb, MW > 10MDa)94, 95, and Hbs which are conjugated to antioxidant enzymes.96 Some of these HBOCs show reduced NO activity and oxidation rates, but no clinical data are available for them yet. Therefore, decades of HBOC research and development have still not yielded a viable AOC. However, we now know that a safe HBOC must be highly stable (for easy storage and to prevent dimerization), minimize reactions with NO (to prevent hypertension), and have a low rate of oxidation (to limit oxidative stress).

1.5 - References

1. Hankinson RJ. The Cambridge Companion to Galen: Cambridge University Press, 2008.

2. Harvey W. An Anatomical Study of the Motion of the Heart and of the Blood in Animals, 1628.

3. Swammerdam J. Bybel der natuur (The book of nature). London: C. G. Seyffert, 1758.

4. Mezzogiorno A, Mezzogiorno V. Marcello Malpighi (1628-1694). American Journal of Nephrology 1997;17(3-4):269-273.

5. Priestley J. Experiments and Observations on Different Kinds of Air. 1774.

6. Lavoisier A. Traite Elementaire de Chemie. 1789.

7. Blundell J. Researches Physiological and Pathological. 1824.

8. Giangrande PLF. The history of blood transfusion. British Journal of Haematology 2000;110(4):758-767.

9. Dodd RY. Current risk for transfusion transmitted infections. Curr Opin Hematol 2007 Nov;14(6):671-676.

10. Fiebig E. Safety of the Blood Supply. Clinical Orthopaedics and Related Research 1998;357:6-18.

11. Goodnough LT. Risks of blood transfusion. Anesthesiol Clin North America 2005 Jun;23(2):241-252, v.

12. Llewelyn CA, Hewitt PE, Knight RSG, Amar K, Cousens S, Mackenzie J, et al. Possible transmission of variant Creutzfeldt-Jakob disease by blood transfusion. The Lancet 2004;363(9407):417-421.

13. Houston F, Foster JD, Angela C, Hunter N, Bostock CJ. Transmission of BSE by blood transfusion in sheep. The Lancet 2000 9/16;356(9234):999-1000.

14. Hunter N, Foster J, Chong A, McCutcheon S, Parnham D, Eaton S, et al. Transmission of prion diseases by blood transfusion. J Gen Virol 2002 Nov;83(Pt 11):2897-2905.

15. Aguzzi A, Glatzel M. vCJD tissue distribution and transmission by transfusion--a worst- case scenario coming true? Lancet 2004 Feb 7;363(9407):411-412.

16. Blumberg N. Deleterious clinical effects of transfusion immunomodulation: proven beyond a reasonable doubt. Transfusion 2005 Aug;45(2 Suppl):33S-39S; discussion 39S-40S.

17. Blumberg N, Heal JM. Blood transfusion immunomodulation: the silent epidemic. Arch Pathol Lab Med 1998 Feb;122(2):117-119.

18. Vamvakas EC. Transfusion-associated cancer recurrence and postoperative infection: meta-analysis of randomized, controlled clinical trials. Transfusion 1996;36(2):175-186.

19. Klein HG. Immunomodulatory Aspects of Transfusion: A Once and Future Risk? Anesthesiology 1999;91(3):861.

20. Amato A, Pescatori M. Effect of perioperative blood transfusions on recurrence of colorectal cancer. Diseases of the Colon & Rectum 1998;41(5):570-585.

21. Taylor RW, Brien J, Trottier SJ, Manganaro L, Cytron M, Lesko MF, et al. Red blood cell transfusions and nosocomial infections in critically ill patients *. Critical Care Medicine 2006;34(9):2302-2308 2310.1097/2301.CCM.0000234034.0000251040.0000234037F.

22. Kopko PM, Holland PV. Transfusion-related acute lung injury. British Journal of Haematology 1999;105(2):322-329.

23. Wallis JP. Transfusion-related acute lung injury (TRALI)--under-diagnosed and under- reported. Br J Anaesth 2003 May;90(5):573-576.

24. Rana R, Fernandez-Perez ER, Khan SA, Rana S, Winters JL, Lesnick TG, et al. Transfusion- related acute lung injury and pulmonary edema in critically ill patients: a retrospective study. Transfusion 2006 Sep;46(9):1478-1483.

25. Moore A, Herrera G, Nyamongo J, Lackritz E, Granade T, Nahlen B, et al. Estimated risk of HIV transmission by blood transfusion in Kenya. The Lancet 2001;358(9282):657-660.

26. Mozzarelli A, Ronda L, Faggiano S, Bettati S, Bruno S. Haemoglobin-based oxygen carriers: research and reality towards an alternative to blood transfusions. Blood Transfus 2010 Jun;8 Suppl 3:s59-68.

27. Riess JG. Overview of progress in the fluorocarbon approach to in vivo oxygen delivery. Biomater Artif Cells Immobilization Biotechnol 1992;20(2-4):183-202.

28. O'Donnel E, Filipak S, Yanovsky A. The push for blood substitutes: is it all in vain? (Having my say). Healthcare Purchasing News 2008;32(1):56.

29. Vamvakas EC, Taswell HF. Epidemiology of blood transfusion. Transfusion 1994 Jun;34(6):464-470.

30. Cleman M, Jaffee CC, Wohlgelernter D. Prevention of ischemia during percutaneous transluminal coronary angioplasty by transcatheter infusion of oxygenated Fluosol DA 20%. Circulation 1986 Sep;74(3):555-562.

31. Young LH, Jaffe CC, Revkin JH, McNulty PH, Cleman M. Metabolic and functional effects of perfluorocarbon distal perfusion during coronary angioplasty. Am J Cardiol 1990 Apr 15;65(15):986-990.

32. Kent KM, Cleman MW, Cowley MJ, Forman MB, Jaffe CC, Kaplan M, et al. Reduction of myocardial ischemia during percutaneous transluminal coronary angioplasty with oxygenated Fluosol. Am J Cardiol 1990 Aug 1;66(3):279-284.

33. Bajaj AK, Cobb MA, Virmani R, Gay JC, Light RT, Forman MB. Limitation of myocardial reperfusion injury by intravenous perfluorochemicals. Role of neutrophil activation. Circulation 1989 Mar;79(3):645-656.

34. Forman MB, Perry JM, Wilson BH, Verani MS, Kaplan PR, Shawl FA, et al. Demonstration of myocardial reperfusion injury in humans: results of a pilot study utilizing acute coronary angioplasty with perfluorochemical in anterior myocardial infarction. J Am Coll Cardiol 1991 Oct;18(4):911-918.

35. Spence RK, McCoy S, Costabile J, Norcross ED, Pello MJ, Alexander JB, et al. Fluosol DA- 20 in the treatment of severe anemia: randomized, controlled study of 46 patients. Crit Care Med 1990 Nov;18(11):1227-1230.

36. Gould SA, Rosen AL, Sehgal LR, Sehgal HL, Langdale LA, Krause LM, et al. Fluosol-DA as a red-cell substitute in acute anemia. N Engl J Med 1986 Jun 26;314(26):1653-1656.

37. Dominguez de Villota ED, Ruiz Carmona MT, Rubio JJ, de Andres S. Equality of the in vivo and in vitro oxygen-binding capacity of haemoglobin in patients with severe respiratory disease. Br J Anaesth 1981 Dec;53(12):1325-1328.

38. Costanzo LS. Physiology. Hagertown: Lippincott, Williams, & Wilkins, 2007.

39. Tremper KK, Anderson ST. Perfluorochemical emulsion oxygen transport fluids: a clinical review. Annu Rev Med 1985;36:309-313.

40. Perutz MF, Rossmann MG, Cullis AF, Muirhead H, Will G, North AC. Structure of haemoglobin: a three-dimensional Fourier synthesis at 5.5-A. resolution, obtained by X-ray analysis. Nature 1960 Feb 13;185(4711):416-422.

41. Perutz MF. Regulation of oxygen affinity of hemoglobin: influence of structure of the globin on the heme iron. Annu Rev Biochem 1979;48:327-386.

42. Bunn HF, Briehl RW. The interaction of 2,3-diphosphoglycerate with various human hemoglobins. J Clin Invest 1970 Jun;49(6):1088-1095.

43. Hsia CCW. Respiratory Function of Hemoglobin. New England Journal of Medicine 1998;338:239-248.

44. Chiancone E. Dissociation of hemoglobin into subunits. II. Human oxyhemoglobin: gel filtration studies. J Biol Chem 1968 Mar 25;243(6):1212-1219.

45. Gould SA, Moore EE, Hoyt DB, Ness PM, Norris EJ, Carson JL, et al. The life-sustaining capacity of human polymerized hemoglobin when red cells might be unavailable. J Am Coll Surg 2002 Oct;195(4):445-452; discussion 452-445.

46. Atha DH, Riggs A. Tetramer-dimer dissociation in homoglobin and the Bohr effect. J Biol Chem 1976 Sep 25;251(18):5537-5543.

47. Graversen JH, Madsen M, Moestrup SK. CD163: a signal receptor scavenging haptoglobin-hemoglobin complexes from plasma. Int J Biochem Cell Biol 2002 Apr;34(4):309- 314.

48. Brandt JL, Frank NR, Lichtman HC. The effects of hemoglobin solutions on renal functions in man. Blood 1951 Nov;6(11):1152-1158.

49. Miller JH, McDonald RK. THE EFFECT OF HEMOGLOBIN ON RENAL FUNCTION IN THE HUMAN. The Journal of Clinical Investigation 1951 10/01;30(10):1033-1040.

50. Savitsky JP, Doczi J, Black J, Arnold JD. A clinical safety trial of stroma-free hemoglobin. Clin Pharmacol Ther 1978 Jan;23(1):73-80.

51. Herold S, Rehmann FJ. Kinetics of the reactions of nitrogen monoxide and nitrite with ferryl hemoglobin. Free Radic Biol Med 2003 Mar 1;34(5):531-545.

52. Yubisui T, Matsuki T, Tanishima K, Takeshita M, Yoneyama Y. NADPH-flavin reductase in human erythrocytes and the reduction of methemoglobin through flavin by the enzyme. Biochem Biophys Res Commun 1977 May 9;76(1):174-182.

53. Kuma F. Properties of methemoglobin reductase and kinetic study of methemoglobin reduction. J Biol Chem 1981 Jun 10;256(11):5518-5523.

54. Scott MD, Lubin BH, Zuo L, Kuypers FA. Erythrocyte defense against hydrogen peroxide: preeminent importance of catalase. J Lab Clin Med 1991 Jul;118(1):7-16.

55. Schechter AN, Gladwin MT. Hemoglobin and the paracrine and endocrine functions of nitric oxide. N Engl J Med 2003 Apr 10;348(15):1483-1485.

56. Eich R, Li T, Lemon DD, Doherty DH, Curry S, Aitken JF, et al. Mechanism of NO-induced oxidation of myoglobin and hemoglobin. Biochemistry 1996;35:6976-6983.

57. Liu X, Miller MJ, Joshi MS, Sadowska-Krowicka H, Clark DA, Lancaster JR, Jr. Diffusion- limited reaction of free nitric oxide with erythrocytes. J Biol Chem 1998 Jul 24;273(30):18709- 18713.

58. Keszler A, Piknova B, Schechter AN, Hogg N. The reaction between nitrite and oxyhemoglobin: a mechanistic study. J Biol Chem 2008 Apr 11;283(15):9615-9622.

59. Gladwin MT, Grubina R, Doyle MP. The new chemical biology of nitrite reactions with hemoglobin: R-state catalysis, oxidative denitrosylation, and nitrite reductase/anhydrase. Acc Chem Res 2009 Jan 20;42(1):157-167.

60. Huang Z, Shiva S, Kim-Shapiro DB, Patel RP, Ringwood LA, Irby CE, et al. Enzymatic function of hemoglobin as a nitrite reductase that produces NO under allosteric control. J Clin Invest 2005 Aug;115(8):2099-2107.

61. Gladwin MT, Ognibene FP, Pannell LK, Nichols JS, Pease-Fye ME, Shelhamer JH, et al. Relative role of heme nitrosylation and beta-cysteine 93 nitrosation in the transport and metabolism of nitric oxide by hemoglobin in the human circulation. Proc Natl Acad Sci U S A 2000 Aug 29;97(18):9943-9948.

62. Jeffers A, Xu X, Huang KT, Cho M, Hogg N, Patel RP, et al. Hemoglobin mediated nitrite activation of soluble guanylyl cyclase. Comp Biochem Physiol A Mol Integr Physiol 2005 Oct;142(2):130-135.

63. Robinson JM, Lancaster JR, Jr. Hemoglobin-mediated, hypoxia-induced vasodilation via nitric oxide: mechanism(s) and physiologic versus pathophysiologic relevance. Am J Respir Cell Mol Biol 2005 Apr;32(4):257-261. 18

64. Przybelski RJ, Daily EK, Kisicki JC, Mattia-Goldberg C, Bounds MJ, Colburn WA. Phase I study of the safety and pharmacologic effects of diaspirin cross-linked hemoglobin solution. Crit Care Med 1996 Dec;24(12):1993-2000.

65. Xu L, Sun L, Rollwagen FM, Li Y, Pacheco ND, Pikoulis E, et al. Cellular responses to surgical trauma, hemorrhage, and resuscitation with diaspirin cross-linked hemoglobin in rats. J Trauma 1997 Jan;42(1):32-41.

66. Schubert A, O'Hara JF, Jr., Przybelski RJ, Tetzlaff JE, Marks KE, Mascha E, et al. Effect of diaspirin crosslinked hemoglobin (DCLHb HemAssist) during high blood loss surgery on selected indices of organ function. Artif Cells Blood Substit Immobil Biotechnol 2002 Jul;30(4):259-283.

67. Lamy ML, Daily EK, Brichant JF, Larbuisson RP, Demeyere RH, Vandermeersch EA, et al. Randomized trial of diaspirin cross-linked hemoglobin solution as an alternative to blood transfusion after cardiac surgery. The DCLHb Cardiac Surgery Trial Collaborative Group. Anesthesiology 2000 Mar;92(3):646-656.

68. Sloan EP, Koenigsberg M, Gens D, Cipolle M, Runge J, Mallory MN, et al. Diaspirin cross- linked hemoglobin (DCLHb) in the treatment of severe traumatic hemorrhagic shock: a randomized controlled efficacy trial. Jama 1999 Nov 17;282(19):1857-1864.

69. Saxena R, Wijnhoud AD, Carton H, Hacke W, Kaste M, Przybelski RJ, et al. Controlled safety study of a hemoglobin-based oxygen carrier, DCLHb, in acute ischemic stroke. Stroke 1999 May;30(5):993-996.

70. Hayes JK, Stanley TH, Lind GH, East K, Smith B, Kessler K. A double-blind study to evaluate the safety of recombinant human hemoglobin in surgical patients during general anesthesia. J Cardiothorac Vasc Anesth 2001 Oct;15(5):593-602.

71. Siegel JH, Fabian M, Smith JA, Costantino D. Use of recombinant hemoglobin solution in reversing lethal hemorrhagic hypovolemic oxygen debt shock. J Trauma 1997 Feb;42(2):199-212.

72. Loeb AL, McIntosh LJ, Raj NR, Longnecker DE. Resuscitation after hemorrhage using recombinant human hemoglobin (rHb1.1) in rats: effects on nitric oxide and prostanoid systems. Crit Care Med 1998 Jun;26(6):1071-1080.

73. Sillerud LO, Caprihan A, Berton N, Rosenthal GJ. Efficacy of recombinant human Hb by 31P-NMR during isovolemic total exchange transfusion. J Appl Physiol 1999 Mar;86(3):887-894.

74. Raat NJ. Effects of recombinant-hemoglobin solutions rHb2.0 and rHb1.1 on blood pressure, intestinal blood flow, and gut oxygenation in a rat model of hemorrhagic shock. J Lab Clin Med 2005 Nov;146(5):304-305.

75. Lowe KC. Blood substitutes: from chemistry to clinic. Journal of Materials Chemistry 2006;16(43):4189-4196. 19

76. Greenburg AG, Kim HW. Hemoglobin-based oxygen carriers. Crit Care 2004;8 Suppl 2:S61-64.

77. Cheng DC, Mazer CD, Martineau R, Ralph-Edwards A, Karski J, Robblee J, et al. A phase II dose-response study of hemoglobin raffimer (Hemolink) in elective coronary artery bypass surgery. J Thorac Cardiovasc Surg 2004 Jan;127(1):79-86.

78. Winslow RM. Red cell substitutes. Semin Hematol 2007 Jan;44(1):51-59.

79. Caron A, Menu P, Faivre-Fiorina B, Labrude P, Alayash AI, Vigneron C. Cardiovascular and hemorheological effects of three modified human hemoglobin solutions in hemodiluted rabbits. J Appl Physiol 1999 Feb;86(2):541-548.

80. Kasper SM, Grune F, Walter M, Amr N, Erasmi H, Buzello W. The effects of increased doses of bovine hemoglobin on hemodynamics and oxygen transport in patients undergoing preoperative hemodilution for elective abdominal aortic surgery. Anesth Analg 1998 Aug;87(2):284-291.

81. Levy JH, Goodnough LT, Greilich PE, Parr GV, Stewart RW, Gratz I, et al. Polymerized bovine hemoglobin solution as a replacement for allogeneic red blood cell transfusion after cardiac surgery: results of a randomized, double-blind trial. J Thorac Cardiovasc Surg 2002 Jul;124(1):35-42.

82. LaMuraglia GM, O'Hara PJ, Baker WH, Naslund TC, Norris EJ, Li J, et al. The reduction of the allogenic transfusion requirement in aortic surgery with a hemoglobin-based solution. J Vasc Surg 2000 Feb;31(2):299-308.

83. Sprung J, Kindscher JD, Wahr JA, Levy JH, Monk TG, Moritz MW, et al. The use of bovine hemoglobin glutamer-250 (Hemopure) in surgical patients: results of a multicenter, randomized, single-blinded trial. Anesth Analg 2002 Apr;94(4):799-808.

84. Natanson C, Kern SJ, Lurie P, Banks SM, Wolfe SM. Cell-free hemoglobin-based blood substitutes and risk of myocardial infarction and death: a meta-analysis. Jama 2008 May 21;299(19):2304-2312.

85. Buehler PW, Alayash AI. Toxicities of hemoglobin solutions: in search of in-vitro and in- vivo model systems. Transfusion 2004 Oct;44(10):1516-1530.

86. Winslow RM. alphaalpha-crosslinked hemoglobin: was failure predicted by preclinical testing? Vox Sang 2000;79(1):1-20.

87. Vandegriff KD, Winslow RM. Hemospan: design principles for a new class of oxygen therapeutic. Artif Organs 2009 Feb;33(2):133-138.

88. Vandegriff KD, Young MA, Keipert PE, Winslow RM. The safety profile of Hemospan®: a new oxygen therapeutic designed using maleimide poly(ethylene) glycol conjugation to human hemoglobin. Transfusion Alternatives in Transfusion Medicine 2007;9(4):213-225.

89. Vandegriff KD, Malavalli A, Wooldridge J, Lohman J, Winslow RM. MP4, a new nonvasoactive PEG-Hb conjugate. Transfusion 2003 Apr;43(4):509-516.

90. Lui FE, Dong P, Kluger R. Polyethylene glycol conjugation enhances the nitrite reductase activity of native and cross-linked hemoglobin. Biochemistry 2008 Oct 7;47(40):10773-10780.

91. Vandegriff KD, Malavalli A, Minn C, Jiang E, Lohman J, Young MA, et al. Oxidation and haem loss kinetics of poly(ethylene glycol)-conjugated haemoglobin (MP4): dissociation between in vitro and in vivo oxidation rates. Biochem J 2006 Nov 1;399(3):463-471.

92. Caccia D, Ronda L, Frassi R, Perrella M, Del Favero E, Bruno S, et al. PEGylation promotes hemoglobin tetramer dissociation. Bioconjug Chem 2009 Jul;20(7):1356-1366.

93. Arifin DR, Palmer AF. Polymersome encapsulated hemoglobin: a novel type of oxygen carrier. Biomacromolecules 2005 Jul-Aug;6(4):2172-2181.

94. Bucci E, Kwansa H, Koehler RC, Matheson B. Development of zero-link polymers of hemoglobin, which do not extravasate and do not induce pressure increases upon infusion. Artif Cells Blood Substit Immobil Biotechnol 2007;35(1):11-18.

95. Matheson B, Kwansa HE, Bucci E, Rebel A, Koehler RC. Vascular response to infusions of a nonextravasating hemoglobin polymer. J Appl Physiol 2002 Oct;93(4):1479-1486.

96. Nemoto M, Mito T, Brinigar WS, Fronticelli C, Koehler RC. Salvage of focal cerebral ischemic damage by transfusion of high O2-affinity recombinant hemoglobin polymers in mouse. J Appl Physiol 2006 May;100(5):1688-1691.

CHAPTER 2: PURIFICATION OF HEMOGLOBIN BY TANGENTIAL FLOW FILTRATION

2.1 – Introduction

If a viable HBOC is ever developed, large amounts of Hb would be needed to meet the global demand. The Hb would also need to be readily available and highly pure. Two major sources of Hb are human and bovine RBCs. Donated human RBCs expire after 42 days (mostly due to damage to the RBC membrane)1, but the Hb inside the RBCs may still be used for HBOC synthesis. A small amount of metHb (1-5% of total Hb) may be present in the expired RBCs, but the metHb can easily be reduced using standard reducing agents.2 Alternatively, bovine RBCs are readily available from the cattle industry or companies like Quad Five (Ryegate, Montana,

USA). Directly transfusing bovine RBCs into humans would cause severe or fatal allergic responses, but the amino acid sequence of pure bHb is almost identical to HbA (88% sequence similarity)3-5 and does not elicit an immune response even after multiple administrations.

Therefore, both HbA and bHb are readily available starting materials for HBOC synthesis. In fact, two major HBOC companies Hemosol Inc. and OPK Biotech LLC have previously purified Hb from human and bovine RBCs, respectively, to synthesize their HBOCs.6, 7

The first step in any Hb purification process involves RBC lysis. Lysing the RBC releases

Hb, along with hundreds of other protein impurities and RBC membrane debris.8, 9 Some of these impurities may be beneficial (i.e. antioxidant enzymes like catalase and superoxide dismutase), but the remaining impurities can interfere with HBOC synthesis or cause severe 22 immune reactions in vivo and must be removed.10-12 Many different purification techniques have been developed to isolate Hb, including heat treatment13, aqueous phase extraction14, and anion exchange chromatography.6, 7, 15 Both Hemosol and OPK Biotech use anion exchange chromatography (AEX), since it produces the highest purity Hb. However, AEX also requires expensive equipment, resin, and a long process time.

The Palmer lab has previously developed an alternative Hb purification technique that utilizes hollow fiber-based tangential flow filtration (TFF) to purify Hb from RBC lysate.16 TFF is a relatively simple procedure that has been used to effectively purify target proteins from mixtures of impurities based on protein size and net charge.17, 18 The TFF process previously established by Palmer et al. to purify Hb from RBC lysate consists of three stages (see Figure

2.1). In the first stage, the lysate is clarified by passing it through glass wool and a TFF cartridge with a pore size of 50 nm. This filter also retains most viruses, including HIV (diameter = 100 nm).19 Large RBC proteins are then removed by passing the Hb solution through another TFF cartridge with a molecular weight cutoff (MWCO) of 500 kDa. Finally, small protein impurities are removed while the purified Hb is concentrated by a third TFF cartridge with a MWCO of 50 kDa (Note: These TFF cartridges are advertised as having an estimated MWCO of 100 kDa, but the actual MWCO is ~50 kDa, therefore HbA with a MW of ~64 kDa is retained by these filters).

The flow rates are kept high throughout the process (480 mL/min) to maintain a high wall shear rate and to minimize fouling of the filters. The product obtained is highly pure, concentrated

(100-500 mg/mL), and sterile Hb that may be used for HBOC synthesis.

Figure 2. 1 - TFF purification scheme for Hbs. The TFF process consists of 3 stages with hollow fiber filters of decreasing pore size (50 nm, 500 kDa, and 50 kDa). During each stage, the retentate is continuously pumped through the cartridge and the filtrate is collected for the next stage. During the third stage, Hb is collected in the retentate.

My initial focus in the lab was on improving the existing TFF process and comparing the biophysical properties of TFF-purified bHb to HbA and bHb purified by AEX. The final stage of the TFF process was modified by adding diafiltration steps to increase the final purity of the sample. Diafiltration is a simple modification of normal filtration in which the volume of the retentate solution is maintained by adding fresh buffer as volume is lost in the filtrate stream.20,

21 For our new diafiltration strategy, the volume of the stage 3 (MWCO 50 kDa) retentate was maintained at 1 L until 10 L of buffer had been added. The purity, yield, and oxygen binding equilibria of the diafiltrated and undiafiltrated samples were then compared to identify any possible advantages or disadvantages of the diafiltration process.

Hb purified with TFF (without diafiltration) was also extensively compared to Hbs purified by AEX (which were generously provided by Hemosol and OPK Biotech) to identify any

24 significant differences between the Hb products. Hb purity, ligand binding kinetics, and autoxidation rates were measured for each Hb sample.

2.2 - Materials and Methods

RBC Lysis

Each round of Hb purification started with 1 L of bovine or human RBCs in a 0.4 % citrate solution. RBCs were initially harvested by centrifugation (3716 g, 30 min at 4°C). Plasma proteins and acellular Hb were washed away by resuspending the cells in a 0.9 % w/v saline solution, centrifuging (3716 g, 30 min at 4°C) the mixture, and discarding the supernatant. This wash step was repeated twice. After the final wash step, bRBCs were resuspended in phosphate buffer (PB; 3.75 mM, pH 7.2) to a final total volume of 2 L and lysed on ice for one hour.

TFF of Hb

A 3-stage filtration process was used to purify bHb from bRBC lysate (Figure 2.1). The

RBC lysate was first passed through a glass wool column to remove large cellular debris. The solution was then sequentially passed through 50 nm, 500 kDa, and 50 kDa MWCO TFF membranes. Table 2.1 shows the technical specifications of each TFF membrane. The retentate was continuously recycled during each stage, while the filtrate was collected and directly applied to the next stage. When diafiltration was applied at stage III, the recycled retentate was kept at a constant volume of approximately 1 L until a total of 10 L of phosphate buffer (PB, pH

7.4) had been added to the retentate bottle. 10 mL samples of filtrate and retentate were collected at each stage for further analysis. TFF cartridges were sterilized immediately after each use with 0.5 M NaOH for 1 hour and stored in a 0.01 % SDS/0.02 % sodium azide solution at 4°C. Before each use, the cartridges were rinsed thoroughly with deionized H2O and tested to

25 ensure that the physical integrity of the hollow fibers was maintained. The entire process was performed on ice to minimize metHb formation. Each purification process was repeated three times.

Pore Size Material Surface Area (cm2) Part Number

50 nm Polysulfone 1050 M10S-360-01S

500 kDa Polysulfone 1050 M1-500S-260-01S

100 kDa Polysulfone 1050 M1AB-360-01S Table 2.1 – Technical specifications of the TFF cartridges

Size Exclusion Chromatography (SEC)

100 L of Hb was injected onto a BioSep-SEC-S 3000 (600 mm x 7.5 mm) SEC column

(Phenomenex, Torrance, CA) attached to a Waters 600 pump, Waters 2489 multi-wavelength detector, and controlled by a Waters 600s controller using EmpowerTM2 software (Waters

Corp., Milford, MA). The running buffer consisted of 0.1 M phosphate buffer, pH 7.0, pumped at a rate of 1.0 mL/min. Absorbance was monitored at 280 and 405 nm. SEC runs were terminated at 40 minutes, followed by a column equilibration of approximately 20 minutes between chromatography runs.

Hb, MetHb, and Total Protein Concentration

UV-visible spectroscopy was used to determine the Hb and metHb concentration at each stage of the purification process using the cyanomethemoglobin method.22, 23 The absorbance of the sample was first measured at 630 nm (L1). More specifically, the sample was diluted (by a factor of D1) to yield an absorbance reading (L1) between 0.1–1.0. Three drops of

10% KCN solution were added to the 3 mL sample and the absorbance was measured again at

630 nm (L2). The metHb concentration was calculated using Eq. 2.1:

(2.1)

−1 23 Where E1 = 3.7 (cm mM) is the extinction coefficient of metHb at 630 nm and λ is the path length of the cuvette, λ = 1 cm. To measure total Hb concentration, 3 drops of K3Fe(CN)6 were added to the diluted (D2) 3 mL sample. After 2 minutes of incubation at room temperature, 3 drops of 10% KCN were added to the sample. The final absorbance at 540 nm was then measured to obtain the value of L3. The total Hb concentration was calculated using

Eq. 2.2:

(2.2)

−1 23 Where E2 = 11.0 (cm mM) is the extinction coefficient of Hb at 540 nm. Finally, the metHb level was calculated using Eq. 2.3:

(2.3)

Total protein concentration was measured by the Bradford method using the Coomassie

Plus protein assay kit (Pierce Biotechnology, Rockford, IL). The mass of total protein at each stage was then calculated by multiplying the protein concentration by the total volume of the sample.

Endotoxin Assay

The endotoxin level of purified Hb was measured with an endotoxin test kit (Pyrogent

Plus, Lonza, Walkersville, MD).

Sodium Dodecyl Sulfate-PolyAcrylamide Gel Electrophoresis (SDS-PAGE)

Undiluted samples from each stage of the purification process were mixed 1:1 with 5%

β-mercaptoethanol Laemmli sample buffer and incubated at 95°C for 5 minutes. Each lane of the gel (12% acrylamide resolving, 4% stacking) was loaded with 30 g of total protein. Gels were then run for 1 hour at 130 V, stained with Coomassie blue R250 (stain buffer, Bio-Rad) for one hour, and finally destained with destaining buffer (70% H2O, 20% ethanol, 10% glacial acetic acid). All gels were scanned on a Gel Doc XR system (Bio-Rad) and analyzed with Quantity One software (Bio-Rad laboratories, Hercules, CA).

Hb-O2 Equilibria

A Hemox Analyzer (TCS Scientific Corporation, New Hope, PA) was used to measure O2 equilibrium curves. This instrument measures the O2 tension with a Clark O2 electrode (Model

5331 Oxygen Probe; Yellow Springs Instruments, Yellow Springs, OH) and simultaneously measures the Hb saturation using a dual-wavelength photometer. The concentration of Hb sample was between 60-75 M heme and the temperature was maintained at 37°C. Samples were prepared by mixing 50–100 L of purified Hb or RBCs with 5 mL of Hemox Buffer (135 mM

NaCl, 5 mM KCl, and 30 mM N-tris (hydroxymethyl)methyl-2-aminoethane sulfonic acid, pH 7.4),

20 L of Additive A, 10 L of Additive B (not added to RBC samples), and 10 L of anti-foaming reagent (TCS Scientific Corporation). To minimize Hb autoxidation, 4 L of the Hayashi enzymatic reduction system (208 μM glucose-6-phosphate, 6.8 units/mL glucose-6-phosphate dehydrogenase, 50 μM NADP, 6.3 μg/mL ferredoxin, 8 mU/mL ferredoxin-NADP reductase, and

246 U/mL catalase)24 was added as well. The samples were then allowed to equilibrate with air at a pO2 of 145 ± 2 mm Hg. After the sample equilibrated, the gas stream was switched to pure nitrogen to deoxygenate the sample. The absorbance of oxy- and deoxy-Hb was then recorded

28 as a function of pO2. The resulting Hb-O2 equilibrium curve was then fit to the Hill equation (Eq.

2.4) to regress Ao, A∞, P50 and n.

(2.4)

Where Y is the fractional saturation of Hb, A0 is the absorbance at 0 mm Hg O2 and A∞ is the absorbance at full saturation. Oxygen binding equilibria of each sample were also measured as a function of chloride ion concentration [Cl-], temperature, pH, and inositol hexaphosphate

(IHP) concentration according to established methods25. IHP is an analog of 2,3 BPG that has similar effects on Hb O2 affinity, but is much less expensive. NaCl concentrations ranging from 0 to 1000 mM were used in 0.01 M phosphate buffer, pH 7.4, at 37oC. Temperature effects were evaluated from 15 to 37oC in 10 mM potassium phosphate with 0.1 M NaCl, pH 7.4, at 37oC. The

Bohr effect was measured from pH = 5.5-9.0 in 0.1 M Bis-Tris buffer and the effect of IHP molar ratios of Hb:IHP (1:0.5 to 1:10) were also evaluated.

Liquid Chromatography Mass Spectrometry (LC-MS)

Mass spectroscopy was used to determine the molecular weight of the α and β globin chains of purified HbA and bHb. Purified Hb was sent to the Mass Spectrometry and Proteomics

Facility of the Campus Chemical Instrument Center at The Ohio State University (Columbus, OH) for analysis. Separation of the α and β chains was achieved using a Dionex U300 HPLC system

(Dionex Corporation, Sunnyvale, CA) connected in series to a Micromass A-TOF II Mass

Spectrometry system (Waters Corporation, Milford, MA). The mobile phase was composed of

0.1 % trifluoroacetic acid (TFA) in deionized water as Buffer A and 0.1 % TFA in acetonitrile (ACN) as Buffer B. The flow rate was maintained at 25 L/min throughout all experiments. The separation was performed on a Discovery BIO Wide Pore C18 HPLC column (1.0 mm × 10 mm, 3

m, Sigma Aldrich Inc., Atlanta, GA). The gradient was initiated with 30% Buffer B and was 29 increased to 70% Buffer B over 60 min, then reduced to 30% Buffer B at 60.1 min and kept constant at 30% Buffer B until 90 min. Twenty L of sample (~10 g) was injected for each run.

Direct Infusion Mass Spectrometry Analysis

A linear ion trap Fourier transform ion cyclotron mass spectrometer (Thermo Electron,

San Jose, CA) was used to analyze Hb by the direct infusion method. Before introduction into the mass spectrometer, 5 L of each Hb was desalted using C18 ziptips (Millipore), dried, and reconstituted in 1 mL water:acetonitrile:formic acid (50:50:0.1). Reconstituted Hb was injected into the mass spectrometer at a flow rate of 3 L/min over a 500 ms injection time. The following instrument settings were used: source voltage = 3.05 kV, capillary voltage = 40.00 V, tube lens voltage = 80.00 V, and capillary temperature = 200°C. Direct infusion ESI-MS spectra were deconvoluted using Xtract for Qual Browser software (Thermo Electron, San Jose, CA).

Rapid Kinetic Measurements

Oxygen dissociation rate constants were measured by rapidly mixing Hb (30 μM heme) with 1.5 mg/mL sodium dithionite in an Applied Photophysics SF-17 microvolume stopped-flow spectrophotometer as previously described26 by monitoring deoxygenation at 437.5 nm in 0.1 M

Tris buffer, pH 7.4 at 25°C. The kinetics of carbon monoxide (CO) association with deoxygenated

Hb was measured in a stopped flow apparatus and the process was monitored at 437.5 nm in 50 mM Tris buffer, pH 7.4, at 25°C after mixing Hb and CO solutions in the presence of 1.5 mg/mL sodium dithionite.27 The kinetics of NO dioxygenation by oxyHb was also measured as previously described.26 NO stock solutions (~ 2 mM) were prepared by bubbling NO gas through a deoxygenated solution of 1 M NaOH before using it to saturate a deoxygenated 0.05

M Bis-Tris buffer, pH 7.0, in a gas-tight serum bottle at room temperature. This stock solution was then transferred to a gastight syringe for appropriate dilutions with deoxygenated Bis-Tris

30 buffer. Solutions of air-equilibrated Hb were mixed with anaerobic solutions of NO and the conversion of oxy-Hb to ferric Hb was monitored by absorbance changes at 420 nm. The value of the bimolecular rate constant for this reaction is known to be extremely large, on the order of

1 × 107 M-1 s-1, so the concentration of NO after mixing was kept low (≤25 M) to minimize loss of the reaction in the dead time of the instrument. Under these conditions, the rate of the autoxidation reaction of NO with dissolved O2 is negligible compared with the rate of reaction with oxy-Hb.

Autoxidation Experiments

All Hb samples were converted to ferrous (Fe2+) oxy-Hb immediately prior to autoxidation experiments.28 Experiments were carried out with 20-25 M heme in sealed cuvettes with room air equilibrated 50 mM Chelex-treated potassium phosphate buffer at 37 °C.

Absorbance changes in the range of 450-700 nm due to spontaneous oxidation of Hb were recorded in a temperature-controlled photodiode array spectrophotometer (Hewlet Packard

8453). Similar oxidation assays were also performed in the presence of superoxide dismutase

(4.6 U/mL) and catalase (414 U/mL) for all Hbs. Oxidation in the presence of IHP (20 M) was measured for human Hb and oxidation in the presence of NaCl (0.1 M) was measured for bovine

Hb. Autoxidation reactions were tracked to near completion (~24 h), at which time 22 M

2+ potassium ferricyanide (K3Fe(CN)6) was added to completely oxidize the remaining Fe heme. A multicomponent analysis was performed to calculate the oxy, met, and hemichrome species based on their individual extinction coefficients.29 Autoxidation rates were obtained from plots of HbFe2+ versus time using nonlinear least-squares curve fitting (single-exponential, two parameter decay) techniques using Sigma-Plot (SPSS, Chicago IL) and equation 2.5:

(2.5)

Catalase Activity

H2O2 (15 M) and catalase (3.36 nM) were added to each Hb (1 M) in a total volume of

1 mL in sealed cuvettes with room air-equilibrated 50 mM Chelex-treated potassium phosphate buffer. Absorbance changes were monitored for 2 minutes at 240 nm and H2O2 concentrations were determined using a molar extinction coefficient of 43.6 cm-1 M-1. All experiments were performed at 37oC in a temperature-controlled photodiode array spectrophotometer (Hewlet

Packard 8453).

Statistical Analysis

All comparisons were performed using a Student paired t-test with ANOVA. In all analyses, α = 0.05 was taken as the level of statistical significance using JMP software (Cary, NC).

2.3 – Advantages and Disadvantages of Diafiltration

The effects of diafiltration on Hb recovery and purity are shown in Tables 2.2 & 2.3, along with SDS-PAGE analysis in Figure 2.2. There seems to be no significant differences in Hb recovery at any stage of the process, with or without filtration. However, approximately 35% of the initial Hb is lost in the retentate of the first two stages or the filtrate of the final stage.

Hb Recovery (%) Filtration Stage No Diafiltration 10 L Diafiltration Post wool 99.30±4.15 96.93±2.16 Stage I Filtrate 87.73±3.97 80.90±2.16 Stage II Filtrate 81.07±3.84 78.07±2.78 Stage III Retentate 66.38±3.82 63.20±11.75 Table 2.2 - Hb recovery (%) after each stage.

The PAGE gels in Figure 2.2 show dark bands in each lane around 15 kDa as expected for bHb ( subunit = 15,053 g/mol &  subunit = 15,954 g/mol). The bands consistently appear lower than their expected MWs because the gels were overloaded with Hb to reveal impurities.

In most lanes, an Hb dimer band is also seen around 32 kDa. Concentrated RBC impurities (most likely insoluble or membrane-bound proteins) are highlighted by the many bands seen in the stage 1 retentate lane. For the most part, the final bHb samples (Lane 3R) appear to be equally pure. However, a faint band around 21 kDa that is present in the retentate lanes and the undiafiltrated sample is absent in the diafiltrated sample. Therefore, it seems that diafiltration may have succeeded in removing at least one major impurity.

Figure 2.2 – PAGE analysis of Hb samples after every stage with diafiltration (left) and without diafiltration (right).

The results in Table 2.3 suggest that there is no significant change in purity throughout the entire TFF process, with or without diafiltration. There does seem to be an increase in the average purity of the diafiltered samples (which would agree with the PAGE results), but the standard deviations are very large. The large amount of error may be due to the inherent inconsistency of the total protein concentration measurements, which were measured using the

Bradford Assay. Therefore, the results shown in Table 3 are inconclusive, but the PAGE gels suggest that diafiltration does increase final sample purity.

Purity Filtration Stage No Diafiltration 10 L Diafiltration Lysate 82.2 + 15% 82.4 + 35% Stage III Retentate 83.4 + 7% 93.7 + 13%

Table 2.3 – Initial and Final Purity of bHb with and without diafiltration

Since the purified bHb product may be used in HBOC synthesis and used in vivo, the levels of harmful contaminants (endotoxin and metHb) must be measured as well. As shown in

Table 2.4, both processes contain acceptable levels of endotoxin (< 5 EU/mL), but there appears to be slightly more endotoxin in the diafiltered samples. This additional endotoxin may have been introduced in the extra PB buffer that was used during diafiltration, but should not pose any significant risk. The metHb level of the diafiltered bHb is also significantly higher (2X) than the undiafiltered bHb. This increase in metHb is likely due to the longer process time introduced by the diafiltration step. However, all metHb levels are below acceptable levels (< 5%).

Endotoxin Level metHb Level

(EU/mL) (%) No Diafiltration 1.0-2.0 0.28 + 0.1 10 L Diafiltration 3.0-4.0 0.55 + 0.3

Table 2.4 - Endotoxin level of stage III retentate samples.

The O2 equilibrium curves for the purified bHb samples are shown in Figure 2.3, while the O2 affinity and cooperativity coefficient of each bHb sample is shown in Table 2.5.

Qualitatively, there appears to be no significant differences between the OECs of bovine RBCs and the purified bHb samples. The O2 affinity and cooperativity of each sample are also similar.

There may be a slight decrease in P50 and cooperativity for the diafiltered samples compared to bovine RBCs, but it is not statistically significant. The O2 binding characteristics of bHb appear to be unaffected by diafiltration.

1.0

0.8 bRBCs No Diafiltration 0.6 10 L Diafiltration Solid line: measured curve Y Dashed line: fitted curve 0.4

0.2

0.0 0 20 40 60 80 100 120 140

pO2 (mm Hg)

Figure 2.3 – Oxygen equilibrium curves for bovine RBCs and TFF-purified bHb (with and without diafiltration). The y-axis represents the fractional O2 saturation of Hb with O2 (Y), while the x- axis represents the partial pressure of O2 (pO2) within the Hb sample.

Oxygen Affinity Cooperativity Sample (P50, mm Hg) (Hill coefficient, n) Bovine RBCs 27.2 + 0.5 2.7 + 0.29 No Diafiltration 27.1 + 1.2 2.5 + 0.08 10 L Diafiltration 24 + 2.9 2.5 + 0.15

Table2. 5 – Oxygen affinity and cooperativity of purified bHb samples compared to bovine RBCs

Overall, it appears that diafiltration does not significantly affect the purity, recovery, or

O2 affinity of TFF-purified bHb. PAGE analysis suggests that diafiltration removes at least one small MW (21 kDa) impurity. The other impurities may simply be too large (>50 kDa) to be filtered out of the 50 kDa filter. In addition, the diafiltration step does appear to slightly increase metHb and endotoxin levels. Therefore, it appears that the diafiltration step is unnecessary.

2.4 – Comparison of TFF-Purified Hbs and Commercially Prepared Hbs

The risks of oxidative and structural damage to Hb are often overlooked or underestimated during purification. An ideal Hb purification process should take steps to prevent damage to Hb by minimizing processing time and exposure to harsh or oxidative conditions. Most Hb purification processes consist of three general steps: 1) RBC fractionation;

2) RBC lysis, and 3) a final purification step(s). In the first step, whole blood is centrifuged to yield a packed layer of RBCs underneath a thin white blood cell layer and a serum supernatant.

After washing the RBCs with an isotonic saline solution to remove any serum proteins and antibodies, fractionated RBCs are lysed in a hypotonic buffer solution to release Hb and other

RBC proteins. The RBC lysate may then be centrifuged or filtered to remove any RBC debris. The

RBC lysate is finally purified using TFF or AEX to remove RBC protein impurities and yield highly 36 pure Hb. This section compares the biophysical properties of bHb and HbA purified by TFF or

AEX (Figure 2.4). The bovine and human TFF Hb samples were purified on a small scale, as described previously.16 AEX-purified bovine Hb provided by OPK Biotech (Cambridge, MA) was prepared in a similar fashion6, with two main differences. Both processes use three filtration steps to process the RBC lysate, however, the pore size of the filters are different. In the TFF process, the Hb product is concentrated with a 100 kDa filter, while the Biopure Hb is concentrated with a 30 kDa filter. Since some RBC proteins have a molecular weight between

30-50 kDa, some protein impurities are retained with the Biopure bHb that are filtered out in the TFF process. The remaining impurities in Biopure bHb are removed with an additional anion exchange chromatography step. The bHb in the lysate is bound to the AEX resin at high pH and then eluted with a decreasing pH gradient to obtain the final bHb product.

Commercially prepared human Hb (HbA) was provided by Hemosol Inc. (Toronto,

Canada). The Hemosol purification process7 is distinctly different from both the TFF and Biopure methods. Initially, RBCs are fractionated and washed by filtration instead of centrifugation.

Filtering the RBCs prevents premature lysis of the RBC that is normally associated with centrifugation and may increase overall yield. Secondly, the RBC lysate is treated with CO to stabilize the Hb before it is pasteurized at 62oC for 10 hours to sterilize any viral contaminants.

The lysate is finally purified using a specific AEX technique called displacement chromatography.

RBC lysate is loaded onto the AEX resin, allowing both HbA and some other impurities to bind.

More and more lysate is added to the resin until impurities with stronger affinities for the resin begin to displace the bound Hb. Lysate is added to the column until the bound HbA is completely eluted from the column. This technique allows more HbA to be purified on a smaller

37 volume of resin compared to conventional IEC techniques, substantially lowering the cost of

HbA purification.

Figure 2.4 – Comparison of AEX and TFF Hb purification strategies.

PAGE analysis of the TFF and commercially purified Hbs is shown in Figure 2.5. Overall,

4 major bands are seen in the Hb samples. The bands close to 16, 32, and 64 kDa correspond to

Hb monomers, dimers, and tetramers, respectively. The band around 55 kDa appears to be catalase (catalase controls are included in lanes 1 and 12 of each gel which also have dark bands around 55 kDa). The TFF and Biopure bHb samples appear to be equally pure, with the same amount of catalase contamination in each. However, while the TFF-purified HbA sample also appears to have catalase, the catalase band in the Hemosol HbA samples is much weaker or not present at all. Therefore, there seems to be at least less catalase in the Hemosol HbA than TFF

HbA.

Figure 2.5 - SDS-PAGE analysis of Hb samples. [A] – Comparison of bHb samples. (Lanes 1 & 12) Catalase (indicated by arrow, tetramer = 240 kDa, monomer = 60 kDa) and its degradation products, (3 & 10) Blue MW standard, (2, 5, 7, 9) OPK Biotech bHb (1:10) and (4, 6, 8, 11) TFF bHb (1:20) [B] – Comparison of HbA samples. (1 & 12) Catalase and its degradation products, (3 & 10) Blue MW standard, (2, 5, 7) TFF HbA (1:10) and (4, 6, 8, 9, 11) Hemosol HbA.

The SEC elution profiles for each Hb are shown in Figure 2.6. The profiles are all essentially identical, revealing no major impurities. In each case, the HbA or bHb band elutes 39 around 19.8 minutes with strong absorbances at 280 and 405 nm. There is a single major peak in the TFF HbA profile, but it disappeared when the experiment was repeated.

Figure 2.6 - SEC profiles of Hb samples. Separation was performed on a BioSep 3000 (500 x 7.5 mm) analytical SEC column (Phenomenex) using 0.1 mM phosphate buffer at a flow rate of 1.0 mL/min.

MW verification of each Hb is shown by the direct infusion ESI-MS spectra in Figure 2.7.

The observed MWs of TFF HbA ( = 15,125 g/mol,  = 15,865 g/mol) and Hemosol HbA ( =

15,125 g/mol,  = 15,865 g/mol) are identical and agree with the expected MWs of HbA subunits

( = 15,126 g/mol,  = 15,867 g/mol). The observed MWs of TFF bHb ( = 15,053 g/mol,  =

15,954 g/mol) and OPK Biotech bHb ( = 15,053 g/mol,  = 15,951 g/mol) are also similar and

40 agree with the expected MWs of bHb subunits ( = 15,053 g/mol,  = 15,954 g/mol). The intensity of the  peaks is consistently higher than the intensity of the  peaks in each sample because the  subunit has a greater net charge than the  subunit, thereby affecting the m/z ratio even though the subunits are in a 1:1 ratio. Therefore, the different purification techniques seem to have no significant effects on the chemical composition of the Hbs.

Figure 2.7 – Direct infusion ESI-MS analysis of Hb preparations

The PAGE results in Figure 2.5 suggested that some catalase contamination was present in the Hb samples, so a catalase activity assay was performed on each sample. The control, 3.36 nM catalase, is shown in green in Figure 2.8. Both commercial Hb samples show no change in

H2O2 concentration over time, indicating that catalase is either absent or inactive. In comparison with the PAGE results it seems likely that catalase may be absent in the Hemosol

HbA sample and rather inactivated in the OPK Biotech bHb sample. On the other hand, both TFF bHb and HbA clearly show trace catalase activity. Therefore, catalase and other impurities still must be present in the TFF Hb samples, although at very low (<3.35 nM) concentrations.

Figure 2.8 - Catalase activity in Hb samples. Catalase control (solid green), TFF bHb (solid blue), TFF HbA (dashed red), OPK Biotech bHb (solid blue), Hemosol HbA (solid red)

The O2 binding characteristics of the Hb samples are shown in Table 2.6. There appear to be no significant differences in the O2 affinity or cooperativity of the Hemosol or TFF HbA samples. The O2 affinities of the OPK Biotech and TFF bHb samples also appear to be similar, but the Biopure bHb sample has significantly reduced cooperativity relative to TFF bHb.

Oxygen Affinity Cooperativity Sample (P50, mm Hg) (Hill coefficient, n) TFF-bHb 28.58 + 0.45 2.5 + 0.04 Biopure bHb 29.22 + 0.54 2.1 + 0.09 TFF-HbA 14.07 + 0.47 2.2 + 0.05 Hemosol HbA 14.50 + 0.37 2.3 + 0.03

Table 2.6 – Oxygen affinities and cooperativities of Hbs

The ligand binding kinetics of the Hb samples are shown in Table 2.7. The oxygen release rate of the bHb samples are similar (~37 s-1), as are the HbA samples (~41 s-1) The rate of

CO binding for all the Hb samples is almost identical as well (0.2 M-1s-1). Likewise, the rate of

NO-induced oxidation of all the Hb samples appear to be unaffected by the method of purification (~18.5 M-1s-1).

Oxygen CO Binding NO oxidation -1 -1 -1 - Sample Release kon,CO, M s k’ox,NO, M s -1 1 koff, s TFF-bHb 36.1 0.21 18.3 Biopure bHb 37.8 0.20 19.0 TFF-HbA 40.4 0.21 18.5 Hemosol HbA 42.9 0.20 18.8

Table 2.7 – Ligand binding kinetics of Hbs.

The effects of IHP, NaCl, pH, and temperature are shown in Figure 2.9. Overall, there appears to be no significant differences in any of these effects for all of the Hbs. IHP, as an analog of 2,3 BPG, increases the P50 and decreases the O2 affinity of the HbA samples as expected. NaCl has a similar effect, increasing P50 as [NaCl] increases and reaching a maximum around 500 mM NaCl. The negatively charged Cl- ions neutralize the positively charged amino acid residues between the  subunits just like 2,3 BPG or IHP does (see section 1.2 for a detailed explanation). All samples exhibit a similar Bohr effect, with a maximum P50 around pH 6.5.

Finally, increasing the temperature of the Hb solution also increases the P50, as expected.

Figure 2.9 – Effects of IHP, NaCl, pH, and temperature on the O2 affinity (P50) of Hbs

Autoxidation rates of the various samples are shown in Table 2.8, with and without SOD and catalase. As expected, the addition of SOD and catalase decreases the autoxidation rates

(with the exception of Hemosol HbA). The oxidation rates of the bHb samples are identical, but the oxidation rate of TFF HbA appears to be double that of Hemosol HbA. This effect may be due to a higher initial concentration of metHb in the TFF HbA sample.

Control +SOD & Catalase Sample -1 -1 Kox, min Kox, min TFF-bHb 6x10-4 4 x10-4 Biopure bHb 6 x10-4 3 x10-4 TFF-HbA 9 x10-4 6 x10-4 Hemosol HbA 4 x10-4 4 x10-4

Table 2.8 – Autoxidation rate constants of Hbs with and without catalase & SOD. 44

2.5 - Conclusion

Taken together, these results suggest that there are no significant differences between

TFF and AEX purified Hbs and all of the Hbs appear to be excellent starting materials for HBOCs.

The TFF Hbs appear to have some trace catalase contamination, but the HbA and bHb samples have similar ligand binding rates, allosteric effects, and O2 equilibria. The catalase contamination in the TFF samples may even be beneficial from an antioxidant perspective, but the autoxidation results indicate that the amount of catalase in the TFF Hb samples is not enough to prevent a significant amount of oxidation. Most importantly, the TFF process is much quicker and less expensive than the AEX processes. Therefore, if low levels of impurities are acceptable, TFF should be the preferred process for Hb purification since it saves both time and money.

2.6 – References

1. Mozzarelli A, Ronda L, Faggiano S, Bettati S, Bruno S. Haemoglobin-based oxygen carriers: research and reality towards an alternative to blood transfusions. Blood Transfus 2010 Jun;8 Suppl 3:s59-68.

2. Dalziel K, O'Brien JR. Side reactions in the deoxygenation of dilute oxyhaemoglobin solutions by sodium dithionite. Biochem J 1957 Sep;67(1):119-124.

3. Marotta CA, Forget BG, Cohen-Solal M, Weissman SM. Nucleotide sequence analysis of coding and noncoding regions of human beta-globin mRNA. Prog Nucleic Acid Res Mol Biol 1976;19:165-175.

4. Michelson AM, Orkin SH. The 3' untranslated regions of the duplicated human alpha- globin genes are unexpectedly divergent. Cell 1980 Nov;22(2 Pt 2):371-377.

5. Aranda Rt, Cai H, Worley CE, Levin EJ, Li R, Olson JS, et al. Structural analysis of fish versus mammalian hemoglobins: effect of the heme pocket environment on autooxidation and hemin loss. Proteins 2009 Apr;75(1):217-230.

6. Houtchens RA, Rausch CW, inventors. Method for Producing a Purified Hemoglobin Product. U.S., 2000.

7. Pliura DH, Wiffen DE, Ashraf S, Magnin AA, inventors. Purification of Hemoglobin by Displacement Chromatography. U.S., 1996.

8. Pasini EM, Kirkegaard M, Mortensen P, Lutz HU, Thomas AW, Mann M. In-depth analysis of the membrane and cytosolic proteome of red blood cells. Blood 2006 Aug 1;108(3):791-801.

9. Kakhniashvili DG, Bulla LA, Jr., Goodman SR. The human erythrocyte proteome: analysis by ion trap mass spectrometry. Mol Cell Proteomics 2004 May;3(5):501-509.

10. Brandt JL, Frank NR, Lichtman HC. The effects of hemoglobin solutions on renal functions in man. Blood 1951 Nov;6(11):1152-1158.

11. Miller JH, McDonald RK. THE EFFECT OF HEMOGLOBIN ON RENAL FUNCTION IN THE HUMAN. The Journal of Clinical Investigation 1951 10/01;30(10):1033-1040.

12. Savitsky JP, Doczi J, Black J, Arnold JD. A clinical safety trial of stroma-free hemoglobin. Clin Pharmacol Ther 1978 Jan;23(1):73-80.

13. Sakai H, Takeoka S, Nishide H, Tsuchida E. Convenient method to purify hemoglobin. Artif Cells Blood Substit Immobil Biotechnol 1994;22(3):651-656.

14. Lee CJ, Kan P, inventors. Hemoglobin Purification. U.S., 1993.

15. Sun G, Palmer AF. Preparation of ultrapure bovine and human hemoglobin by anion exchange chromatography. J Chromatogr B Analyt Technol Biomed Life Sci 2008 May 1;867(1):1- 7.

16. Palmer AF, Sun G, Harris DR. Tangential flow filtration of hemoglobin. Biotechnol Prog 2009 Jan-Feb;25(1):189-199.

17. Reis Rv, J. M. Brake, Charkoudian J, Burns DB, Zydney AL. High-performance tangential flow filtration using charged membranes. Journal of Membrane Science 1999;159(1-2):133-142.

18. Riess JG. Oxygen carriers ("blood substitutes")--raison d'etre, chemistry, and some physiology. Chem Rev 2001 Sep;101(9):2797-2920.

19. Gelderblom HR, Hausmann EH, Ozel M, Pauli G, Koch MA. Fine structure of human immunodeficiency virus (HIV) and immunolocalization of structural proteins. Virology 1987 Jan;156(1):171-176.

20. Henderson LW, Besarab A, Michaels A, Bluemle LW, Jr. Blood purification by ultrafiltration and fluid replacement (diafiltration). Hemodial Int 2004 Jan 1;8(1):10-18.

21. Lipnizki F, Boelsmand J, Madsen RF. Concepts of industrial-scale diafiltration systems. Desalination 2002;144(1-3):179-184.

22. Crosby WH, Munn JI, Furth FW. Standardizing a method for clinical hemoglobinometry. U S Armed Forces Med J 1954 May;5(5):693-703.

23. Zijlstra WG, Kampen EJv. Standardization of hemoglobinometry: I. The extinction coefficient of hemiglobincyanide at 540 nm. Clinica Chimica Acta 1960;5(5):719-726.

24. Hayashi A, Suzuki T, Shin M. An enzymic reduction system for metmyoglobin and methemoglobin, and its application to functional studies of oxygen carriers. Biochim Biophys Acta 1973 Jun 15;310(2):309-316.

25. Fronticelli C, Bucci E, Orth C. Solvent regulation of oxygen affinity in hemoglobin. Sensitivity of bovine hemoglobin to chloride ions. J Biol Chem 1984 Sep 10;259(17):10841- 10844.

26. Alayash AI, Summers AG, Wood F, Jia Y. Effects of glutaraldehyde polymerization on oxygen transport and redox properties of bovine hemoglobin. Arch Biochem Biophys 2001 Jul 15;391(2):225-234.

27. Antonini E, Brunori M. Frontiers of biology. Amsterdam, 1971.

28. Alayash AI. Effects of intra- and intermolecular crosslinking on the free radical reactions of bovine hemoglobins. Free Radic Biol Med 1995 Feb;18(2):295-301.

29. Winterbourn CC. CRC Handbook of Methods of Oxygen Radical Research. Boca Raton, 1985.

CHAPTER 3: PURIFICATION OF HEMOGLOBIN BY IMMOBILIZED METAL AFFINITY

CHROMATOGRAPHY

3.1 – Introduction

TFF can be used to quickly purify large amounts of Hb, however, the final sample is only

80-90% pure (see Table 2.3).1 This is not surprising, since the TFF filters used can only remove impurities with MWs above 500 kDa and below 50 kDa. Besides Hb (MW = 64.5 kDa), many different RBC proteins also have MWs within the range of 50-500 kDa.2, 3 Even AEX purified Hbs show some impurities after PAGE analysis (Figure 2.5). Therefore, if 100% pure Hb is required for HBOC synthesis or other studies, an alternative purification technique must be used.

The best way to purify a specific protein from a mixture of impurities is to take advantage of any unique characteristics (charge, stability, etc.) or ligand interactions it may have. For example, Hbs from many different species have a very high affinity for zinc

7 -1 4, 5 2+ (association constant Ka = 1.3x10 M ). The zinc ion (Zn ) tightly binds to 143His and

139Asp in the negatively charged 2,3-BPG binding pocket of HbA, while potentially interacting with 93Cys as well.4, 6 There are also several histidines on the surface of HbA (6 on the  subunit and 7 on the  subunit)7 which may also bind Zn2+. At low concentrations, Zn2+ binds to

4 Hb and significantly increases its O2 affinity. At high zinc concentrations (> 10:1 molar ratio of

Zn:Hb), Lehman et al. have shown that Hb may be selectively precipitated from RBC lysate and resuspended upon addition of EDTA, thereby purifying Hb.8-10

Plomer et al. improved the zinc precipitation method by using immobilized metal ion

(Zn2+) affinity chromatography (IMAC) to purify recombinant HbA (rHbA) from bacterial lysate.11

In this technique, a chelating resin is used to sequester the Zn2+ ion which then binds Hb. The resin may then be washed to remove weakly bound impurities and purify the Hb. The Hb is finally eluted by stripping the Zn2+ from the resin with the powerful chelating agent ethylenediaminetetraacetic acid (EDTA). The Hb may also be eluted from the column without stripping the zinc by using high concentrations of imidazole or low pH (3-4.0), but imidazole can oxidize the heme iron and low pH can damage the Hb. Therefore, IMAC with EDTA elution is commonly used to purify rHbA from bacterial lysates.12-14 However, IMAC purification of Hbs from RBC lysate has not yet been fully examined.

In this study, Hb was purified from RBCs using IMAC and TFF to produce ultrapure human (HbA), bovine (bHb), and chicken Hb (cHb). Two types of processes were examined (see

Scheme 1) in which (1) a single 50 nm TFF filter is used to prepare the lysate for IMAC or (2) a three-stage TFF process (using 50 nm, 500 kDa, and 50/100 kDa TFF cartridges) is used to filter and concentrate the Hb samples prior to IMAC. The samples were also diafiltrated using a 10 kDa TFF cartridge to remove any Zn2+ or EDTA in the final sample. The relative purities and equilibrium O2 binding characteristics of the Hb products obtained with these processes were compared to determine the efficacy of IMAC purification of Hb and the necessity of the 500 kDa and 50/100 kDa TFF stages.

Figure 3.1 – Schematic of different IMAC purification strategies. In process 1, RBC lysate is filtered through a single 50 nm TFF cartridge and applied to IMAC resin. In process 2, RBC lysate is filtered through 50 nm and 500 kDa TFF cartridges, then concentrated on a 50 kDa TFF cartridge. All samples were diafiltered using a 10 kDa TFF cartridge to remove zinc and EDTA after IMAC purification.

3.2 – Materials and Methods

RBC Lysis and TFF of Hb

RBC lysis and TFF were performed according to the method described in Section 2.2, with the following modifications. Chicken RBCs were acquired from King & Sons in Bradford,

OH. During cHb purification, a TFF cartridge with a 25 kDa MWCO was used instead of the usual

50 kDa MWCO cartridge. Five mL aliquots were taken at the end of stage 1 (50 nm) and stage 3

(50 or 25 kDa) and frozen at -80oC until needed for IMAC.

IMAC of Hb

A XK 50/30 column was packed with 200 mL of Chelating Sepharose Fast Flow resin (GE

Healthcare, Piscataway, NJ). The resin was initially washed with ultrapure H2O until the conductivity of the effluent decreased below 0.5 mS/cm. The resin was then charged with 300 mL of 15 mM zinc acetate until the conductivity of the effluent was stable at ~2.5 mS/cm.

Excess zinc acetate was removed from the column by flushing it with 0.2 M NaCl until the conductivity leveled off at 20.4 mS/cm. At this point, the column was cooled with ice water to maintain a resin temperature of ~2oC. Hb samples were diluted 10 with Wash Buffer 1 (20 mM

Tris, 500 mM NaCl, pH = 8.3) and loaded onto the column. During sample loading, the effluent was collected as soon as the absorbance at 280 nm (A280) began to spike to collect the first impurity fraction. After all of the sample was loaded onto the column, the column was washed further with Wash Buffer 1 until the A280 off the column effluent returned to zero. The column was then washed with Wash Buffer 2 (200 mM Tris for HbA and bHb, 500 mM Tris + 500 mM

NaCl for cHb, pH = 8.3) and Wash Buffer 3 (20 mM Tris, pH = 8.3). During both washes, the column effluent was collected and concentrated with 10 kDa centrifugal filters (Millipore,

Billerica, MA). Finally, the column was flushed with 15 mM EDTA in 20 mM Tris (pH = 8.3) and

51 the effluent was collected as soon as the absorbance at 540 nm (A540) began to increase. When all the Hb had been collected (effluent A540 < 1 mAU), the column was washed with 0.2 M NaOH and stored at room temperature. The Hb product was subsequently concentrated on a 10 kDa

TFF cartridge (Spectrum Labs) until the sample volume was reduced to 20 mL, then it was diluted 10 with 20 mM Tris (pH 8.3) and concentrated again. This process was repeated once more to remove any remaining EDTA-Zn complex from the sample. The final Hb product was then stored at -80oC until it could be analyzed.

Matrix-Assisted Laser Desorption/Ionization (MALDI)

Hb samples were diluted to 1 mg/mL, then 1 L of the diluted Hb was mixed with 1 L

HCl and 1 L acetonitrile saturated with sinapic acid. A calibration sample containing cytochrome C (12360.97 Da) and apomyoglobin (16952.31 Da) was prepared in a similar fashion.

The samples were then analyzed using a Bruker Daltonics FLEX instrument (Billerica, MA). The data from the Hb samples was adjusted using a linear model based on the observed and predicted MWs in the calibration sample.

SDS-PAGE and O2 Equilibrium Curves

SDS-PAGE and O2 equilibrium measurements were performed as described in Section

2.2.

3.3 – Results of IMAC Purification

The impurities removed from the resin during each step of the purification and the purity of the initial and final samples are clearly illustrated by the PAGE analysis in Figure 3.2. In all cases, it is clear that some impurities are present in the initial TFF-purified Hb samples, albeit at low concentrations relative to the dark Hb monomer band at ~15 kDa. In fact, the same impurities are observed in both of the samples taken from the stage 1 filtrate or stage 3

52 retentate of the TFF process. Therefore, it seems that the 500 kDa and 50 kDa TFF steps do not significantly increase the purity of the Hb samples.

Even more impurities in the TFF-purified samples are revealed in lanes 3-5 of the gels, which show the impurities removed during the three IMAC washes. Most of the impurities were removed during the 500 mM NaCl wash (lane 3). The high concentration of [Cl-] in this wash step removes impurities which are electrostatically bound to the positively charged IMAC resin.

A trace of each Hb was also eluted during this wash.

The second wash (200 mM Tris, lane 4) also removed many impurities in each sample.

The high concentration of Tris, which is a weak chelating agent, binds Zn2+ and displaces weakly bound impurities and even broadens the red Hb band on the resin. The cHb band split into two red bands when it was washed with 200 mM Tris. One of the red bands eluted from the column with the other impurities and seems to be a type of cHb since it gives a dark band around 15 kDa. The resin beneath the red band which stayed on the column during the 200 mM Tris wash turned a slight pink color instead of the usual white color. Adding 500 mM NaCl to the 200 mM

Tris wash buffer lightened the pink color and accelerated the elution of the second red band. It has been previously shown that chickens contain two major types of Hb, an A-type (cHbA, with

15 A and  subunits) and a D-type (cHbD, with D and  subunits). cHbD has fewer histidines

(D= 6,  = 5) than cHbA (A = 9,  = 5) and it lacks the 143His residue implicated in zinc binding. Therefore, the Hb that eluted during the second wash of IMAC is most likely cHbD, since it should have a much lower affinity for zinc than cHbA.

Only a few trace impurities were removed with the final 20 mM Tris wash (shown in lane 5). During cHb purification, a trace amount of the pink color beneath the dark red band also elutes, giving another dark band around 15 kDa. All Hbs began to elute as soon as the

53 column was flushed with 20 mM EDTA. The final Hb products are shown in Lane 6 of each gel.

In each sample, the expected Hb monomer (15 kDa), dimer (32 kDa), and a trace amount of tetramer (64.5 kDa) bands are present. Two unidentified light bands are also observed between

20-25 kDa in each sample. These bands were still present even when the resin was washed with high concentrations of imidazole or low pH (4.0-6.0) buffers. Therefore, these two bands are either a degradation product of Hb formed during PAGE sample preparation or unidentified impurities which have a stronger affinity for the IMAC resin than the Hbs.

MALDI analysis was used to verify the MWs of the IMAC-purified Hbs and the identity of the red species which eluted when cHb was washed with 200 mM Tris (Figure 3.3). The observed MWs of the HbA subunits,  = 15127.2 g/mol &  = 15867.8 g/mol, closely agree with their expected MWs ( = 15126.4 g/mol,  = 15867.2 g/mol). The observed MWs of the bHb subunits,  = 15053.9 g/mol &  = 15954.6 g/mol, also match their expected MWs ( = 15053.2 g/mol,  = 15954.4 g/mol). The two impurities observed around 20-25 kDa in the PAGE gels were not detected by MALDI.

Analysis of the two Hb fractions obtained during IMAC purification of cHb verifies that the Hb removed during the second wash step is cHbD. The MWs observed in this fraction

(Figure 3.3, upper right panel), 15694.8 g/mol & 16333.3 g/mol, are almost identical to the MWs

expected for the D subunit (15694.9 g/mol) and the  subunit (16334.9 g/mol) of cHbD.

Meanwhile, the final Hb IMAC product obtained after EDTA elution (Figure 3.3, lower right panel) has three peaks at 15297.5, 15694.2, and 16333.3 g/mol. The first peak corresponds to

the A subunit (15297.7 g/mol), while the latter two peaks match the D and  subunits, respectively. While the cHbD sample appears to contain no cHbA whatsoever, a small amount of cHbD remains in the cHbA fraction. The column was extensively washed with high

54 concentrations of Tris and NaCl, however, we were unable to completely remove the trace amount of cHbD from the cHbA fraction.

Figure 3.2 – PAGE analysis of samples from IMAC purification of bHb, HbA, and cHb. The following descriptions apply to lanes in all gels: Lanes 1 & 7 = Protein MW Ladder, Lane 2 = Initial Hb sample loaded onto IMAC resin, either Stage 1 TFF Filtrate (S1F) or Stage 3 TFF Retentate (S3R), Lane 3 – Impurities removed with 500 mM NaCl wash, Lane 4 – Impurities

55 removed with 200 mM Tris wash, Lane 5 – Impurities removed with 20 mM Tris wash, Lane 6 – Final EDTA Hb elution.

Figure 3.3 – MALDI analysis of the IMAC-purified Hbs. Both the long range (10-120 kDa) and short range (15-18 kDa) scans are shown for each sample.

The metHb levels of the Hb samples before and after IMAC purification are shown in

Table 3.1. All of the metHb percentages are below acceptable levels (<5%) for HBOC synthesis.

In most cases, the IMAC process had little or no significant effect on metHb levels.

Source Material S1F HbA S3R HbA S1F bHb S3R bHb S1F cHbA S3R cHbA Before IMAC 3.3% 0.9% 1.3% 0.9% 2.9% 5.0% After IMAC 2.8 + 0.6% 1.4 + 0.5% 0.8 + 0.4% 1.1 + 0.1% 2.1 + 2.0% 3.4 + 1.6%

Table 3.1 – MetHb level of purified Hb products

Characteristic O2 equilibrium curves (OECs) for HbA and bHb are shown in Figure 3.4 and values for P50 and n are given in Table 3.2. The data for HbA and bHb fit the Hill model extremely well. The IMAC purified bHb and HbA OECs overlapped the control OECs, which were taken from Stage 1 and Stage 3 of the TFF process. The O2 affinity at half saturation (P50) and cooperativity coefficient (n) show no significant difference between the starting Hb material used for IMAC and the final purified Hb samples. These values also agree with the expected values for HbA (P50 = 12-14 mm Hg, n = 2.5-2.9) and bHb (P50 = 24-28 mm Hg, n = 2.5 – 3.0).

O2 equilibrium analysis of the IMAC purified A-type and D-type cHb samples was somewhat unpredictable (Figure 7). While the control cHb samples (from both S1F and S3R of

TFF) displayed consistent OECs and values for P50 (28-30 mm Hg) and n (2.2-2.4), the cHbA and cHbD samples were quite inconsistent. The irregular shape of the A-type OECs made them impossible to fit to the Hill model, giving them extremely low cooperativities (n < 1.5). Their P50 values were also quite varied and much lower than the control samples. On the other hand, the cHbD samples produced sigmoidal curves with good cooperativity (n = 2.5-2.6). However, their

P50 values were as random as the cHbA samples, ranging from 8-25 mm Hg. Measurements of the cHb samples were repeated several times and consistently produced the same results.

Figure 3.4 – OECs of HbA and bHb samples before and after IMAC purification. Raw and fitted data are represented by dashed and solid lines, respectively.

Figure 3.5 - OECs of cHbA and cHbD samples before and after IMAC purification. Raw and fitted data are represented by dashed and solid lines, respectively.

HbA bHb cHbA cHbD S1F S3R S1F S3R S1F S3R S1F S3R

Initial P50 12.3 12.6 24.8 26.4 28.8 30.2 28.8 30.2 Final P50 12.2 + 0.2 13.0 + 0.9 25.5 + 0.7 26.1 + 0.5 17.0 + 2.3 25.1 + 5.7 13.2 + 8.5 19.8 + 8.9 Initial n 2.5 2.8 3.0 2.7 2.2 2.4 2.2 2.4 Final n 2.7 + 0.2 2.7 + 0.4 2.8 + 0.1 2.7 + 0.1 1.1 + 0.1 1.6 + 0.4 2.6 + 0.4 2.5 + 0.3

Table 3.2 – O2 affinity and cooperativity of purified Hbs. “Initial” represents samples before IMAC purification, “Final” represents samples after IMAC purification.

These results show that IMAC is a superior purification technique for isolating Hb from

RBCs. PAGE and MALDI suggest that IMAC-purified HbA and bHb are absolutely pure. PAGE suggests that the IMAC purified samples may have two impurities with MWs around 20-25 kDa, but these impurities were not detected by MALDI. Therefore, they may just be degradation products of Hb that form during PAGE sample preparation. LC-MS analysis of the PAGE bands would need to be performed to positively identify these bands. IMAC is also able to separate the A and D types of chicken Hb. The cHbD product is contaminated with some RBC impurities and the cHbA product is contaminated with small amounts of cHbA, but they are each >80% pure.

There is also no significant oxidation during the IMAC process. The OECs of HbA and bHb show that the purified Hbs are still able to effectively bind O2oxygen, so no zinc contamination is present in the final products. In contrast, the OECs of the IMAC purified cHbA and cHbD products are unusual and unpredictable. The same batch of TFF-purified cHb was used for all of the cHb IMAC purifications and all of the conditions were constant. More work will need to be done to determine the cause of their erratic behavior.

3.4 - Conclusion

Overall, it seems that IMAC is an excellent purification technique for Hbs from many different species. The resin is extremely stable and has a high binding capacity around 500 mg

Hb/100 mL resin. Considering the ease of scaling up IMAC columns, this technique could easily be adapted to an industrial setting with only minor adjustments to produce large amounts of highly pure Hbs for HBOC synthesis. Similarly, this technique could be scaled down into a high- throughput assay for isolating Hb in which a small amount of resin is packed into a spin column

(much like a plasmid prep). The buffers and sample could be quickly applied, providing pure Hb samples within 10 minutes.

3.5 - References

1. Elmer J, Buehler PW, Jia Y, Wood F, Harris DR, Alayash AI, et al. Functional comparison of hemoglobin purified by different methods and their biophysical implications. Biotechnol Bioeng 2010 May 1;106(1):76-85.

2. Pasini EM, Kirkegaard M, Mortensen P, Lutz HU, Thomas AW, Mann M. In-depth analysis of the membrane and cytosolic proteome of red blood cells. Blood 2006 Aug 1;108(3):791-801.

3. Kakhniashvili DG, Bulla LA, Jr., Goodman SR. The human erythrocyte proteome: analysis by ion trap mass spectrometry. Mol Cell Proteomics 2004 May;3(5):501-509.

4. Rifkind JM, Heim JM. Interaction of zinc and hemoglobin: binding of zinc and the oxygen affinity. Biochemistry 1977 Oct 4;16(20):4438-4443.

5. Gilman JG, Oelschlegel FJJ, Brewer GJ. Erythrocyte Structure and Function. New York, 1975.

6. Gilman JG, Brewer GJ. The oxygen-linked zinc-binding site of human haemoglobin. Biochem J 1978 Mar 1;169(3):625-632.

7. Xu Y, Zheng Y, Fan J-S, Yang D. A new strategy for structure determination of large proteins in solution without deuteration. Nature Methods 2006;3:931.

8. Carrell RW, Lehmann H. Zinc acetate as a precipitant of unstable haemoglobins. J Clin Pathol 1981 Jul;34(7):796-799. 60

9. Lehmann H, Williamson D, Carrell RW, Lucas JE. The precipitation of hemoglobin by zinc: its application to the isolation of a minor hemoglobin fraction (HbB2 delta 16 Gly replaced by Arg) from lysed whole blood. Hemoglobin 1982;6(2):183-186.

10. Tye RW, inventor. Preparation of stroma-free, non-heme protein-free hemoglobin. U.S., 1983.

11. Plomer JJ, Ryland JR, Matthews M-AH, Traylor DW, Milne EE, Durfee SL, et al., inventors. Purification of hemoglobin. U.S., 1998.

12. Hartman JC, Argoudelis G, Doherty D, Lemon D, Gorczynski R. Reduced nitric oxide reactivity of a new recombinant human hemoglobin attenuates gastric dysmotility. Eur J Pharmacol 1998 Dec 18;363(2-3):175-178.

13. Asmundson AL, Taber AM, van der Walde A, Lin DH, Olson JS, Anthony-Cahill SJ. Coexpression of human alpha- and circularly permuted beta-globins yields a hemoglobin with normal R state but modified T state properties. Biochemistry 2009 Jun 16;48(23):5456-5465.

14. Villarreal DM, Phillips CL, Kelley AM, Villarreal S, Villaloboz A, Hernandez P, et al. Enhancement of recombinant hemoglobin production in Escherichia coli BL21(DE3) containing the Plesiomonas shigelloides heme transport system. Appl Environ Microbiol 2008 Sep;74(18):5854-5856.

15. Richards RI, Wells JR. Chicken globin genes. Nucleotide sequence of cDNA clones coding for the alpha-globin expressed during hemolytic anemia. J Biol Chem 1980 Oct 10;255(19):9306- 9311.

CHAPTER 4: INTRODUCTION TO ERYTHROCRUORIN

4.1 - Extracellular Hbs: A New Paradigm

As of 2011, the only hemoglobin-based oxygen carriers (HBOCs) that have entered phase III clinical trials are polymerized human1-4 and bovine5-8 hemoglobins (Hbs).

Unfortunately, these products have been discontinued due to increased risks of death and other complications9. The major problems associated with these HBOCs (instability, oxidative stress, and nitric oxide (NO) scavenging) can be directly attributed to removing Hb from the protective internal environment of the red blood cell (RBC). The RBC has enzymes to prevent oxidation10-12,

13 14 a cell membrane to reduce interactions with NO , allosteric effectors to modulate O2 delivery , and high Hb concentrations that minimize dimerization of the Hb tetramer 15.

Since mammalian Hbs purified from RBCs are burdened with so many problems, extracellular Hbs from other organisms may be better suited for use in HBOC development. A special class of Hbs, known as erythrocruorins (Ecs), are found in organisms which lack RBCs

(annelids16, mollusks17, and some insects18). Consequently, Ecs have already adapted to the harsh conditions in the bloodstream with unique structural and functional modifications that make them attractive natural HBOCs. This review will focus on the unique properties of Ec from the Earthworm Lumbricus terrestris (LtEc).

4.2 - Structure and Stability of LtEc

Ecs come in a wide variety of shapes and sizes, including the spherical Ec of Riftia pachyptila (~400 kDa)19, the hexagonal bilayer (HBL) Ecs of L. terrestris20 or Arenicola marina21, and the huge cylindrical Ec of the clam Cardita borealis (12 MDa)22. These Ecs are all held together by covalent disulfide bonds and strong electrostatic or hydrophobic forces within large subunit interfaces. Therefore, they are not susceptible to dissociation at low concentrations like mammalian Hbs, which lack intermolecular disulfide bonds23.

LtEc consists of a macromolecular assembly of 144 globin subunits and 36 linker

20, 24, 25 chains . There are 5 types of globins (A, B, C, and D1’or D2) and 4 types of linkers (L1, L2,

L3, and L4)26, 27. Each of the globin subunits has a single intramolecular disulfide bond and a structure that is more similar to myoglobin than individual Hb subunits28. Each subunit also contains a heme group, which binds oxygen (O2) and even contributes to subunit association by forming hydrogen bonds with adjacent subunits through propionate groups25. The A, B, and C subunits also have intermolecular disulfide bonds which form an ABC trimer. The ABC trimer and D monomer self-associate through electrostatic and hydrophobic interactions to form the

29 ABCD tetramer . Next, the A3B3C3D3 dodecamer spontaneously forms from three ABCD tetramers through disulfide bonds. The linker chains are not required for dodecamer formation29. The dodecamer is hemi-spherical and has a structure that is reminiscent of the spherical Ec of R. pachyptila (RpEc), suggesting the LtEc may have also been spherical at some point30.

Meanwhile, three linker chains self-assemble to form a linker trimer. The linker chains are degenerate, meaning that several combinations of L1, L2, L3 or L4 can create the trimer. In

63 fact, the minimum requirement for linker trimer formation is only a binary mixture of L1 or L2

30 with L3 or L4 . The linker trimer is also held together by numerous disulfide bonds and strong hydrophobic interactions within a coiled coil domain31. The linker trimer also has large low density lipoprotein (LDL) domains which strongly bind the dodecamer to form the protomer.

Finally, 12 protomers assemble through interactions between the coiled coil domains of the linker trimers to form the hexagonal bilayer structure of LtEc, which has a molecular weight

(MW) of approximately 3.6 MDa and a diameter of 30 nm20. To put these numbers into context, human Hb (HbA) has a MW of 0.064 MDa and a diameter of 5 nm32.

Several other elements also contribute to the structure of LtEc. Approximately 50 calcium ions (Ca2+) are bound at various sites throughout LtEc. Copper and zinc atoms are also bound to LtEc35. The Ca2+ increases the stability of LtEc and helps it resist unfolding at high temperatures29, 34. Barium (Ba2+) has a similar effect and addition of EDTA (which chelates Ca2+) decreases the thermal stability of LtEc34. LtEc is also extremely stable in the presence of chemical denaturants, exhibiting a half-life of 28 hours in 1.75 M urea29. However, LtEc is prone to dissociation at alkaline pH (>8.0)36. It is important to mention that other Ecs are not as stable as LtEc. For example, the marine worm A. marina expresses an Ec (AmEc) which is adapted to a high ionic strength. Since the ionic strength of human plasma is relatively low, AmEc dissociates into dodecamers when injected into mice36. In contrast, LtEc comes from the terrestrial

Earthworm and is stable at the ionic strength of human blood37.

Figure 4.1 - Assembly of LtEc. HbA33 and myoglobin (Mb)34 are shown in the bottom right to provide a sense of scale. 65

4.3 - O2 Transport by LtEc

Human blood and LtEc bind and release O2 in a similar fashion. The O2 affinity or P50

(pO2 at which half of the hemes are saturated with O2) of human blood (26 mm Hg) is almost identical to LtEc (28 mm Hg)37. This is in contrast to pure HbA and AmEc, which both have significantly lower P50 values (higher O2 affinities) than human blood. The O2 affinity of HbA decreases when it is purified from human blood due to the removal of its allosteric effector 2,3-

14 2+ BPG. The allosteric effector of LtEc is Ca , which increases the O2 affinity of LtEc and is freely available in the bloodstream29. Other divalent cations, like Ba2+, Sr2+, and Mg2+, have a similar

38, 39 effect on the O2 affinity of LtEc . The relatively high O2 affinity (low P50) of AmEc is probably another effect of its exposure to low ionic strength buffers or an adaptation to the low O2 environment in which A. marina is found 36.

MW Diameter P50 n (kDa) (nm) (mm Hg) (---) HbA 64 5 11 2.7 AmEc 3,60036 3036 2.6 2.536 LtEc 3,600 30 2837 3.7 RBC --- 8,000 2640 2.7540

Table 4.1 - Size, molecular weight (MW), O2 affinity (P50), and cooperativity (n) of HbA, AmEc, LtEc, and human RBCs.

Cooperativity is unique trait of Hbs in which small changes in one subunit (i.e. ligand binding) affect the conformations and ligand affinities of adjacent subunits. This phenomenon allows Hbs to bind O2 quickly in the lungs, hold onto it in the arteries, then release it in large

66 amounts in the arterioles and capillaries. The cooperativities of HbA, AmEc, and blood are all around 2.5-2.7 under physiological conditions. The cooperativity of LtEc is relatively higher under physiological conditions (3.7), due to the increased number of subunit interactions within the LtEc dodecamer. In fact, the maximum cooperativity of LtEc is 7.9 at 25oC and pH 7.7 with

41 25 mM CaCl2 . The effects of cooperativity also appear to be mostly within the dodecamers and only slightly (if at all) transmitted between dodecamers42.

As previously mentioned, the LtEc dodecamer spontaneously forms in the absence of the linker chains and resembles the structure of the spherical Ec of R. pachyptila, suggesting that

LtEc may have also been a spherical Ec before the advent of the linker chains. Interestingly, isolated dodecamers and ABCD tetramers have O2 affinities and cooperativities similar to LtEc in its full form. The isolated ABC trimer and D monomer, however, have significantly higher O2 affinities and lower cooperativities. Therefore, the linker chains are not required for O2 transport and appear to simply increase the stability of LtEc 42.

4.4 - Autoxidation of LtEc

2+ 3+ Oxidation of the heme iron (Fe  Fe ) is an inevitable side-effect of O2 transport for all Hbs. After O2 binds to the heme iron, it can strip away an electron and escape the heme

- 3+ pocket, forming the pro-oxidant superoxide (O2 ) and oxidized Hb (metHb, Fe ). MetHb can be further oxidized to the ferryl form (Fe4+) and/or generate toxic hemichrome and other free radicals which greatly increase lipid oxidation in cell membranes and overall oxidative stress 43.

The size, structure, and amino acid composition of the heme pocket all have significant effects

- 44, 45 on the rate of Hb autoxidation. Large heme pockets allow O2 to easily escape , while

- aromatic amino acids (i.e. tyrosine or phenylalanine) within the heme pocket stabilize O2 and reduce oxidation rates 45.

The heme pockets of LtEc are much smaller than HbA20, 24. Each subunit of LtEc also has additional phenylalanine or tryptophan residues which are not present in the heme pockets of

HbA subunits.20 These differences are clearly expressed in the redox potentials of HbA and LtEc

(see Table 2). The redox potential of a species is a measure of how likely it is to accept or donate electrons. Species with positive redox potentials are more likely to accept electrons

(reduction), while negative redox potentials indicate that a species is more likely to donate electrons (oxidation). The redox potential of HbA is negative (-50 mV), whereas LtEc has a highly positive redox potential (+112 mV). Therefore, LtEc is much less likely to undergo autoxidation than HbA 23, 46, 47. In fact, experiments have shown that the autoxidation rate of LtEc (<0.010 hr-

1) is much lower than HbA (0.014 hr-1)46. LtEc is also easily reduced by reducing agents that are found in the bloodstream (ascorbic acid or glutathione), while HbA is not as easily reduced 23, 46.

-1 kox (hr ) Eo (mV) HbA 0.01436 -5046 LtEc < 0.010 +11246 AmEc 0.00548 +6346

Table 4.2. Autoxidation rates and redox potentials of HbA, LtEc, and AmEc

Divalent cations also influence oxidation of LtEc. For example, Ba2+ and Ca2+ both reduce the rate of LtEc autoxidation. Sr2+ and Mg2+ have a similar, yet less significant effect34.

The Cu and Zn atoms which are bound to LtEc also appear to have some superoxide dismutase 68

- (SOD) activity. SODs are a family of enzymes which react with O2 to form water, thereby

- preventing formation of harmful H2O2 from O2 . The SOD activity of LtEc is approximately 10% of the human SOD enzyme, but any anti-oxidant activity is beneficial from an HBOC development perspective49.

4.5 - Interactions between LtEc and other Ligands

Hbs are known to bind, transport, and or react with several other ligands besides O2.

For example, Hbs bind both O2 and carbon monoxide (CO), but release CO much slower than O2.

This competitive inhibition of O2 binding is the reason CO is a poisonous gas. LtEc can also bind

CO. Unlike HbA, however, the LtEc subunits appear to have varying affinities (either high or low) for CO50, 51.

The interactions between Hb and NO have recently become crucially important with respect to HBOC development. Mammalian Hbs have been shown to catalyze a NO

- dioxygenation reaction in which O2 and NO react to form NO3 and metHb. The metHb increases oxidative stress, but the elimination of NO can have much more significant effects in vivo52. NO is a signaling molecule which regulates blood vessel diameter, relaxing them at high concentrations (vasodilation) and constricting them at low concentrations (vasoconstriction).

Vasoconstriction also increases blood pressure and causes harmful systemic hypertension.

Therefore, reducing NO dioxygenation is a primary focus in HBOC design.

Mutagenesis studies in HbA have shown that mutations that reduce autoxidation also reduce the rate of NO dioxygenation. For example, substituting large apolar or aromatic residues (leucine, tryptophan, or phenylalanine) within the heme pocket or charged amino acids

(glutamine) near the heme pocket entrance greatly reduce the rate of NO dioxygenation in oxy-

HbA52, 53. Both LtEc and AmEc have naturally occurring additional phenylalanine and tryptophan residues within their heme pockets25. Interestingly, oxy-AmEc has been shown to bind NO instead of undergoing NO dioxygenation. Experiments with oxy-LtEc and NO must be conducted to determine if LtEc also avoids NO dioxygenation.

4.6 - Preliminary Animal Studies with LtEc

Preliminary experiments have been conducted in which small amounts of LtEc37 and

AmEc36 have been injected into healthy mice and rats. AmEc quickly dissociated after injection, but caused no other severe side-effects and the animals were healthy 18 weeks after injection.

In vitro experiments also indicated that AmEc is not scavenged by haptoglobin. Haptoglobin is a serum protein which strongly binds to free HbA and clears it from the bloodstream.

Injection of LtEc into mice and rats also lacked any side-effects. Most importantly, no immune response was observed even after repeated injections of LtEc. These are only preliminary results, yet they indicate that LtEc may have great potential as an HBOC. Further studies with larger doses of LtEc will have to be done to accurately determine the efficacy and safety of LtEc as an HBOC.

4.7 - Conclusion

Altogether, these results suggest that LtEc may be a superior HBOC, since it successfully avoids many of the problems associated with other HBOCs. It is highly stable, resistant to oxidation, and may avoid the NO dioxygenation reaction altogether. It also has beneficial antioxidant properties which should minimize oxidative stress in vivo.

Further studies must be done to determine the safety of LtEc. Larger exchange transfusion studies need to be done to determine if larger doses of LtEc pose any significant risks or side-effects. Models of hemorrhagic shock must also be used to simulate the effects of

LtEc in animals subjected to shock. In vitro, more work must be done to determine the exact nature of the interactions between LtEc and NO. For example, does LtEc possess NO dioxygenase activity or just a reduced rate of NO dioxygenation? The effects of H2O2 on LtEc oxidation and stability must also be determined to predict the effects of transfusing LtEc into patients with sepsis. Once all of these questions are answered, human clinical trials with LtEc may begin.

4.8 - References

1. Greenburg AG, Kim HW. Hemoglobin-based oxygen carriers. Crit Care 2004;8 Suppl 2:S61-64.

2. Cheng DC, Mazer CD, Martineau R, Ralph-Edwards A, Karski J, Robblee J, et al. A phase II dose-response study of hemoglobin raffimer (Hemolink) in elective coronary artery bypass surgery. J Thorac Cardiovasc Surg 2004 Jan;127(1):79-86.

3. Winslow RM. Red cell substitutes. Semin Hematol 2007 Jan;44(1):51-59.

4. Caron A, Menu P, Faivre-Fiorina B, Labrude P, Alayash AI, Vigneron C. Cardiovascular and hemorheological effects of three modified human hemoglobin solutions in hemodiluted rabbits. J Appl Physiol 1999 Feb;86(2):541-548.

5. Kasper SM, Grune F, Walter M, Amr N, Erasmi H, Buzello W. The effects of increased doses of bovine hemoglobin on hemodynamics and oxygen transport in patients undergoing preoperative hemodilution for elective abdominal aortic surgery. Anesth Analg 1998 Aug;87(2):284-291.

6. Levy JH, Goodnough LT, Greilich PE, Parr GV, Stewart RW, Gratz I, et al. Polymerized bovine hemoglobin solution as a replacement for allogeneic red blood cell transfusion after cardiac surgery: results of a randomized, double-blind trial. J Thorac Cardiovasc Surg 2002 Jul;124(1):35-42.

7. LaMuraglia GM, O'Hara PJ, Baker WH, Naslund TC, Norris EJ, Li J, et al. The reduction of the allogenic transfusion requirement in aortic surgery with a hemoglobin-based solution. J Vasc Surg 2000 Feb;31(2):299-308.

8. Sprung J, Kindscher JD, Wahr JA, Levy JH, Monk TG, Moritz MW, et al. The use of bovine hemoglobin glutamer-250 (Hemopure) in surgical patients: results of a multicenter, randomized, single-blinded trial. Anesth Analg 2002 Apr;94(4):799-808.

9. Natanson C, Kern SJ, Lurie P, Banks SM, Wolfe SM. Cell-free hemoglobin-based blood substitutes and risk of myocardial infarction and death: a meta-analysis. Jama 2008 May 21;299(19):2304-2312.

10. Yubisui T, Matsuki T, Tanishima K, Takeshita M, Yoneyama Y. NADPH-flavin reductase in human erythrocytes and the reduction of methemoglobin through flavin by the enzyme. Biochem Biophys Res Commun 1977 May 9;76(1):174-182.

11. Kuma F. Properties of methemoglobin reductase and kinetic study of methemoglobin reduction. J Biol Chem 1981 Jun 10;256(11):5518-5523.

12. Scott MD, Lubin BH, Zuo L, Kuypers FA. Erythrocyte defense against hydrogen peroxide: preeminent importance of catalase. J Lab Clin Med 1991 Jul;118(1):7-16.

13. Liu X, Miller MJ, Joshi MS, Sadowska-Krowicka H, Clark DA, Lancaster JR, Jr. Diffusion- limited reaction of free nitric oxide with erythrocytes. J Biol Chem 1998 Jul 24;273(30):18709- 18713.

14. Bunn HF, Briehl RW. The interaction of 2,3-diphosphoglycerate with various human hemoglobins. J Clin Invest 1970 Jun;49(6):1088-1095.

15. Chiancone E. Dissociation of hemoglobin into subunits. II. Human oxyhemoglobin: gel filtration studies. J Biol Chem 1968 Mar 25;243(6):1212-1219.

16. Terwilliger RC. Structures of Invertebrate Hemoglobins. American Zoologist 1980 January 1, 1980;20(1):53-67.

17. Boffi A, Verzili D, Chiancone E, Leone M, Cupane A, Militello V, et al. Stereodynamic properties of the cooperative homodimeric Scapharca inaequivalvis hemoglobin studied through optical absorption spectroscopy and ligand rebinding kinetics. Biophys J 1994;67(4):1713-1723.

18. Di Iorio E, Tavernelli I, Yu W. Dynamic properties of monomeric insect erythrocruorin III from Chironomus thummi-thummi: relationships between structural flexibility and functional complexity. Biophys J 1997;73(5):2742-2751.

19. Zal F, Lallier FH, Wall JS, Vinogradov SN, Toulmond A. The multi-hemoglobin system of the hydrothermal vent tube worm Riftia pachyptila. I. Reexamination of the number and masses of its constituents. Journal of Biological Chemistry 1996;271:8869.

20. Royer WE, Jr., Sharma H, Strand K, Knapp JE, Bhyravbhatla B. Lumbricus erythrocruorin at 3.5 A resolution: architecture of a megadalton respiratory complex. Structure 2006 Jul;14(7):1167-1177.

21. Royer WE, Omartian MN, Knapp JE. Low resolution crystal structure of Arenicola erythrocruorin: influence of coiled coils on the architecture of a megadalton respiratory protein. Journal of Molecular Biology 2007;365(1):226-236.

22. Terwilliger N, Terwilliger RC. Oxygen binding domains of a clam (Cardita borealis) extracellular hemoglobin. Biochimica et Biophysica Acta 1978;537(1):77-85.

23. Dorman SC, Kenny CF, Miller L, Hirsch RE, Harrington JP. Role of redox potential of hemoglobin-based oxygen carriers on methemoglobin reduction by plasma components. Artif Cells Blood Substit Immobil Biotechnol 2002 Jan;30(1):39-51.

24. Fushitani K, Matsuura MS, Riggs AF. The amino acid sequences of chains a, b, and c that form the trimer subunit of the extracellular hemoglobin from Lumbricus terrestris. J Biol Chem 1988 May 15;263(14):6502-6517.

25. Strand K, Knapp JE, Bhyravbhatla B, Royer WE, Jr. Crystal structure of the hemoglobin dodecamer from Lumbricus erythrocruorin: allosteric core of giant annelid respiratory complexes. J Mol Biol 2004 Nov 12;344(1):119-134.

26. Suzuki T, Riggs AF. Linker chain L1 of earthworm hemoglobin. Structure of gene and protein: homology with low density lipoprotein receptor. Journal of Biological Chemistry 1993;268(18):13548-13555.

27. Kao WY, Qin J, Fushitani K, Smith SS, Gorr TA, Riggs CK, et al. Linker chains of the gigantic hemoglobin of the earthworm Lumbricus terrestris: primary structures of linkers L2, L3, and L4 and analysis of the connectivity of the disulfide bonds in linker L1. Proteins: Structure, Function, and Bioinformatics 2006;63(174-187):174.

28. Royer WEJ, Strand K, Van Heel M, Hendrickson WA. Structural hierarchy in erythrocruorin, the giant respiratory assemblage of annelids. PNAS 2000;97(13):7101-7111.

29. Sharma PK, Kuchumov AR, Chottard G, Martin PD, Wall JS, Vinogradov SN. The role of the dodecamer subunit in the dissociation and reassembly of the hexagonal bilayer structure of Lumbricus terrestris hemoglobin. J Biol Chem 1996 Apr 12;271(15):8754-8762.

30. Lamy ML, Daily EK, Brichant JF, Larbuisson RP, Demeyere RH, Vandermeersch EA, et al. Randomized trial of diaspirin cross-linked hemoglobin solution as an alternative to blood transfusion after cardiac surgery. The DCLHb Cardiac Surgery Trial Collaborative Group. Anesthesiology 2000 Mar;92(3):646-656.

31. Xu Y, Zheng Y, Fan J-S, Yang D. A new strategy for structure determination of large proteins in solution without deuteration. Nature Methods 2006;3:931.

32. Birukou I, Soman J, Olson JS. Blocking the gate to ligand entry in human hemoglobin. Journal of Biological Chemistry 2011;286:10515.

33. Standley P, Mainwaring MG, Gotoh T, Vinogradov SN. The calcium, copper and zinc content of some annelid extracellular haemoglobins. Biochemistry Journal 1988;249(3):915-916.

34. Harrington JP. Multimeric Lumbricus hemoglobin stabilization by alkali and alkaline earth cations. Comparative Biochemistry and Physiology Part A: Physiology 1994;109(3):799- 803.

35. Chiancone E, Vecchini P, Rossi Fanelli MR, Antonini E. Studies on erythrocruorin. II. Dissociation of earthworm erythrocruorin. Journal of Molecular Biology 1972;70(1):73-76.

36. Rousselot M, Delpy E, Drieu La Rochelle C, Lagente V, Pirow R, Rees JF, et al. Arenicola marina extracellular hemoglobin: a new promising blood substitute. Biotechnol J 2006 Mar;1(3):333-345.

37. Hirsch RE, Jelicks LA, Wittenberg BA, Kaul DK, Shear HL, Harrington JP. A first evaluation of the natural high molecular weight polymeric Lumbricus terrestris hemoglobin as an oxygen carrier. Artif Cells Blood Substit Immobil Biotechnol 1997 Sep;25(5):429-444.

38. Fushitani K, Imai K, Riggs AF. Oxygenation properties of hemoglobin from the earthworm, Lumbricus terrestris. Effects of pH, salts, and temperature. Journal of Biological Chemistry 1986;261(18):8414.

39. Ochiai T, Weber RE. Effects of magnesium and calcium on the oxygenation reaction of erythrocruorin from the marine polychaete Arenicola marina and the terrestrial oligochaete Lumbricus terrestris. Zoological Science 2002;19(9):999-1000.

40. Zapletal C, Bode A, Lorenz MW, Gebhard MM, Golling M. Effects of hemodilution with a hemoglobin-based oxygen carrier (HBOC-201) on ischemia/reperfusion injury in a model of partial warm liver ischemia of the rat. Microvasc Res 2009 Dec;78(3):386-392.

41. Vidugiris G, Harrington JP, Friedman JM, Hirsch RE. Absence of ligand binding-induced tertiary changes in the multimeric earthworm Lumbricus terrestris hemoglobin. A resonance Raman study. Journal of Biological Chemistry 1993;268:26190-26192.

42. Fushitani K, Riggs AF. The extracellular hemoglobin of the earthworm, Lumbricus terrestris. Oxygenation properties of isolated chains, trimer, and a reassociated product. J Biol Chem 1991 Jun 5;266(16):10275-10281.

43. Alayash AI. Hemoglobin-based blood substitutes and the hazards of blood radicals. Free Radical Research 2000;33:341-348.

44. Stellwagen E. Haem exposure as the determinate of oxidation-reduction potential of haem proteins. Nature 1978 Sep 7;275(5675):73-74.

45. Gow AJ, Payson AP, Bonaventura J. Invertebrate hemoglobins and nitric oxide: how heme pocket structure controls reactivity. Journal of INorganic Biochemistry 2005;99(4):903- 911.

46. Harrington JP, Kobayashi S, Dorman SC, Zito SL, Hirsch RE. Acellular invertebrate hemoglobins as model therapeutic oxygen carriers: unique redox potentials. Artif Cells Blood Substit Immobil Biotechnol 2007;35(1):53-67.

47. Dorman SC, Harrington JP, Martin MS, Johnson TV. Determination of the formal reduction potential of Lumbricus terrestris hemoglobin using thin layer spectroelectrochemistry. Journal of Inorganic Biochemistry 2004;98(1):185-188.

48. Harnois T, Rousselot M, Rogniaux H, Zal F. High-level production of recombinant Arenicola marina globin chains in Escherichia coli: a new generation of blood substitute. Artif Cells Blood Substit Immobil Biotechnol 2009;37(3):106-116.

49. Liochev SI, Kuchumov AR, Vinogradov SN, Fridovich I. Superoxide dismutase activity in the giant hemoglobin of the earthworm, Lumbricus terrestris. Arch Biochem Biophys 1996 Jun 15;330(2):281-284.

50. Giacometti GM, Focesi A, Jr., Brunori M, Wyman J. Effect of light on carbon monoxide binding by erythrocruorin. Journal of Biological Chemistry 1975;98(2):333-339.

51. Giacometti GM, Focesi A, Giardina B, Brunori M, Wyman J. Kinetics of binding of carbon monoxide to lumbricus erythrocruorin: a possible model. Proceedings of the National Academy of Sciences 1975;72(11):4313-4316.

52. Eich R, Li T, Lemon DD, Doherty DH, Curry S, Aitken JF, et al. Mechanism of NO-induced oxidation of myoglobin and hemoglobin. Biochemistry 1996;35:6976-6983.

53. Olson JS, Eich RF, Smith LP, Warren JJ, Knowles BC. Protein engineering strategies for designing more stable hemoglobin-based blood substitutes. Artificial Cells, Blood Substitutes, and Immobilization Biotechnology 1997;25:227-241.

CHAPTER 5: PURIFICATION AND IN VITRO CHARACTERIZATION OF ERYTHROCRUORIN

5.1 – Introduction

HBOCs must be effective, safe, and available in large quantities at low cost. Preliminary results indicate that LtEc can effectively transport O2 without any known side-effects, but existing purification techniques (ultracentrifugation1 and size exclusion chromatography2) are limited by high equipment costs and low product yields. Therefore, a new purification process must be developed which produces large quantities of LtEc at low cost. The LtEc product must be highly pure, since Earthworm blood contains hemolysins which can lyse RBCs3 and clotting proteins which might aggregate and block capillaries.4

Since LtEc is such a relatively large protein, it may be easily purified from Earthworm blood with TFF. TFF is a relatively simple technique that is easily scaled up. It is also inexpensive, since the only equipment required are a pump and the TFF hollow fiber cartridges.

In our process (see Figure 5.1), Earthworms are homogenized and centrifuged to produce a clear red solution of crude LtEc. The homogenate is then further clarified and sterilized by passing it through a 0.22 m TFF cartridge. The LtEc is then purified by extensive diafiltration with a 500 kDa TFF cartridge. LtEc is retained by the filter, while impurities are removed in the filtrate. The purity and composition of the LtEc product was confirmed by SDS-PAGE, MALDI, and LC-MS. O2 binding equilibria and ligand binding kinetics were also measured for LtEc. Finally, the interactions of LtEc with NO were investigated. A novel cyanide (CN-) reductase activity was 76 also discovered for LtEc. Our results show that TFF produces highly pure LtEc which effectively binds O2 and other ligands, but it appears to lack NO dioxygenase activity and has a significantly lower rate of autoxidation than HbA or bHb.

Figure 5.1 – LtEc purification scheme. Earthworm homogenate is centrifuged twice at progressively higher speeds to remove debris and other insoluble aggregates. The crude LtEc is then sterilized with a 0.22 m TFF cartridge and purified by diafiltration with a 500 kDa TFF cartridge.

5.2 – Materials and Methods

Earthworm Blood Extraction and Clarification

For each round of purification, 1,000 Canadian nightcrawlers (Lumbricus terrestris) were purchased from Wholesale Bait Company (Hamilton, OH). Worms were initially rinsed with tap water to remove excess dirt, then 2 L batches of worms were extensively washed with 20-30 L of tap water to remove as much mucus as possible. A blender was used to homogenize the worms

(puree mode for ~10 seconds) and the homogenate was immediately centrifuged at 3,716 g for

40 min at 4oC. Solid debris were discarded and the cloudy red supernatant was centrifuged again at 18,000 g for 20 min at 10-15oC. The clear red supernatant (~2.0 L per 1,000 worms) was then put through filter paper to remove any remaining large particles.

Purification of LtEc

The Earthworm homogenate was pumped through two 0.22 m TFF cartridges in parallel (1050 cm2 surface area, Spectrum Labs, Rancho Dominguez, CA) at 480 mL/min until the majority of the sample volume was transmitted through the filter. The filter pores clogged several times during the filtration process (indicated by a clear filtrate) and were cleaned with deionized water. The red retentate of the 0.22 m filter was stored at -80oC for future analysis.

The 0.22 m filtrate (1.8-2.0 L) was then concentrated on two 500 kDa TFF cartridges (1050 cm2 surface area, Spectrum Labs) at 480 mL/min down to a volume of approximately 200 mL. The

500 kDa retentate was then diafiltrated by diluting it to 2.0 liters with buffer and concentrating it down to 200 mL for a total of ten cycles. The retentate was diluted with 20 mM Tris buffer

(pH 7.0) during the first eight rounds of diafiltration and modified lactated Ringer’s buffer (115 mM NaCl, 4 mM KCl, 1.4 mM CaCl2, 13 mM NaOH, 12.25 mM N-acetyl cysteine, 0.3% sodium lactate, pH 7.0) during the last two rounds of diafiltration. During the final round of diafiltration,

78 the retentate was concentrated to 20-50 mL and sterilized by passing it through a 0.45 m syringe filter. The purified LtEc was then stored at -80oC until needed. After each round of purification, all filters were rinsed and soaked in 0.2 M NaOH for 1 hour, then rinsed with distilled water and stored at 4oC.

It is important to mention that other purification techniques were also evaluated. LtEc bound to both Zn2+ and Ca2+ charged IMAC resin, but with terribly low binding capacities. While large amounts of HbA bind to IMAC resin and turn it a dark red color, a very small amount of

LtEc can bind to IMAC resin, coloring it a very light shade of pink. The binding capacity of AEX resin for LtEc is also very low. However, when the pink resin is washed with low concentrations of salt (100 mM NaCl in 20 mM Tris, pH 7.0), only the LtEc dodecamers elute while the linker proteins remain bound until the resin is washed with much higher salt concentrations (> 250 mM NaCl). Therefore, AEX cannot be used to purify LtEc, but it may be a useful technique for rapidly isolating small amounts of dodecamer.

Viscosity and Colloid Osmotic Pressure

The viscosity of each solution was measured in a cone/plate viscometer DV-II Plus with a cone spindle CPE-40 (Brookfield Engineering Laboratories, Middleboro, MA) at a shear rate of

160 s-1. The colloid osmotic pressure (COP) of each sample was measured using a Wescor 4420

Colloid Osmometer 47 (Wescor, Logan, UT). Measurements for LtEc solution viscosity and COP were made at a protein concentration of either 50 or 100 mg/mL.

UV-Vis Spectroscopy

Samples were diluted with 20 mM phosphate buffer saline (PBS, pH 7.4) to a concentration of 50-100 M heme. Sodium dithionite (1 mg/mL) was added to the diluted samples to completely reduce all of the heme iron (Fe3+Fe2+). Excess sodium dithionite was

79 immediately removed using G-20 desalting resin. The protein concentration of the reduced and desalted Hbs was measured using the Bradford method. All spectra were recorded using 4 mL

2+ of Hb sample in quartz cuvettes. Fe -O2 spectra were measured immediately after reduction/desalting. Sodium dithionite (1 mg/mL) was then added to the same sample to obtain the deoxy spectra. Finally, 0.5 mg/mL MAHMA NONOate (an NO donor) was added to the same sample to obtain the Fe2+-NO spectra. Separate sealed cuvettes were flushed with humid CO gas for 30 minutes to prepare the Fe2+-CO spectra. Soret spectra (250-475 nm) were obtained by diluting the samples 10, except for the CO samples which were diluted 20.

Samples were oxidized by adding 500 L of 20% KFe(CN)6 to 40 mL samples of the diluted Hbs. Excess KFe(CN)6 was then removed using G-20 desalting resin and the Bradford method was used to measure the protein concentration of the desalted sample. MetHb spectra

- were recorded immediately after desalting. MetHb-NO2 spectra were obtained by adding 100

L of a 10% sodium nitrite solution to the same cuvette. Soret spectra (250-475 nm) were obtained by diluting the samples 20 with PBS buffer.

Other Methods

SDS-PAGE, metHb estimation, MALDI, O2 binding equilibria, and stopped flow kinetics were all performed as previously described in Section 2.2.

5.3 – Results and Discussion

SDS-PAGE analysis of samples from the LtEc purification process is shown in Figure 5.2.

The 0.22 m TFF retentate and filtrate in lanes 2 & 3 show many of the same impurities, but a large MW impurity appears only in the 0.22 mm retentate sample. This large MW impurity may be the clotting aggregate observed in another worm, Eisenia foetida. Indeed, the 0.22 m

80 retentate was quite turbid. When it was diluted 10 and centrifuged at 18,000 g for 30 minutes, the supernatant was clear red and a solid dark pellet was observed. The pellet was resuspended and centrifuged several times with water, then dissolved in 4 M urea and loaded onto a PAGE gel (results not shown). The gel revealed a faint pair of bands around 40 kDa, which corresponds to the MW of the clotting proteins observed in E. foetida.4

It is important to note that the 0.22 m TFF retentate yielded a dark red color.

Therefore, fouling of the TFF membrane by the clotting aggregate limits transmission of LtEc through the filter and reduces the final yield of the process. Removing the 0.22 m filtration step did increase the yield, but the final product was significantly less pure and extremely viscous (data not shown). Therefore, an emphasis should be put on developing new ways to remove the clotting aggregate or inhibit clotting altogether. Attempts to remove the clotting aggregate with flocculating agents were unsuccessful. The clotting proteins are activated by a protease in the worm mucus, so we also tried adding a protease inhibitor to the Earthworm homogenate but the clotting aggregate was still observed.

The fourth lane in Figure 5.2 shows the purified LtEc product. Only the two expected bands for pure LtEc are seen around 15 kDa (LtEc globin monomers) and around 30 kDa (LtEc linker chains and globin dimers), indicating the sample is highly pure. In addition, Lanes 5-7 show the impurities that are removed during LtEc diafiltration. Many impurities are seen in the filtrate after the first round of diafiltration (lane 5), but no impurities are visible in the filtrate after the final round of diafiltration (lane 6). However, when the final filtrate sample is concentrated 25 (lane 7) two impurities are observed at approximately 75 and 125 kDa.

Therefore, there may be low concentrations of these two unidentified impurities in the purified

LtEc.

Final LtEc Yield (g) % metLtEc With 0.22 m TFF 4.8 + 2.8 3.3 + 0.7 Without 0.22 m TFF 7.8 + 1.4 3.5 + 1.6

Table 5.1 – Average yield and percent of oxidized heme (metLtEc) in purified LtEc samples

Figure 5.2 - SDS-PAGE analysis of samples from the LtEc purification process. Lane (1) Protein MW ladder, (2) 0.22 m TFF retentate, (3) 0.22 m TFF filtrate, (4) 500 kDa diafiltrated LtEc, (5) Initial 500 kDa TFF filtrate, (6) Final 500 kDa TFF filtrate, (7) Final 500 kDa TFF filtrate, concentrated 25.

The MALDI spectra of the TFF-purified LtEc product (Figure 5.3) agrees well with previously published results.5 A single distinct peak around 16 kDa is observed for the D1’ and

D2 monomers (MW -15,964 & 15,997 Da, respectively).6 The ABC trimer is linked together by disulfide bonds which are not disrupted during MALDI sample preparation. Therefore, a strong band is observed around 52 kDa as expected for the trimer (A + B + C = 17,525 + 16,254 + 17,289

= 51,068 Da).7, 8 Some residual ABCD tetramer (~67 kDa) is also observed around 67 kDa.

Finally, smaller peaks are seen from 25-32 kDa for each of the linker subunits (L1 = 27,422.64, L2

= 32,019.40, L3 = 26,834.85, and L4 = 24,256.05 Da).9, 10 No major protein impurities were observed in the purified LtEc sample. LC-MS of the TFF-purified sample also verified that each globin and linker subunit is present in the pure sample, but no other known impurities were identified.

OECs for the purified LtEc, HbA, and bHb are shown in Figure 5.4, while the O2 affinity

(P50) and Hill value (n) for each sample are shown in Table 5.3. The OEC for HbA is left shifted and its P50 (12.09 mm Hg) is significantly lower since 2,3 BPG is not present in solution. The O2 affinity of bHb (25.39 mm Hg), which is similar to human whole blood, is also similar to the O2 affinity of LtEc (26.09 mm Hg). The Hill values of HbA and bHb are similar (~2.44), but the Hill value of LtEc is much higher (~3.7), since there are many more subunit interactions within the dodecamer. Overall, these equilibrium data indicate that LtEc transports O2 similar to bovine or human whole blood.

Sample P50 (mm Hg) n LtEc 26.09 + 0.25 3.74 + 0.17 bHb 25.39 2.45 HbA 12.09 2.43

Table 5.2 – Oxygen affinity (P50) and cooperativity (n) of LtEc compared to HbA and bHb.

Figure 5.3 – MALDI-TOF analysis of TFF-purified LtEc. M, 2M – monomer and dimer D globin subunit, L1, L2, L3, L4 – linker chains, T, T2+ - singly and doubly charged ABC trimer, t – ABCD tetramer.

Figure 5.4 – OECs for LtEc, bHb, and HbA. The y-axis represents the fraction of hemes in the sample with bound O2, while the x-axis represents the partial pressure of O2 within the sample. Raw data are shown as dashed lines, while fits to the Adair model are shown as solid lines.

Ligand release rates for O2 (koff) and binding rates for CO (kon,CO) are shown in Table 5.3.

The oxygen release rates for LtEc, HbA, and bHb are all within a close range (30-39 s-1), but LtEc

-1 does have the slowest O2 release rate (30.29 s ). A significant difference in CO binding was observed between the mammalian Hbs (0.201-0.204 M-1s-1) and LtEc (0.097 M-1s-1). These differences in ligand interactions illustrate the effects of the smaller heme pocket of LtEc and suggest that LtEc has slightly increased ligand selectivity for O2. Experiments show that pure CO still displaces bound O2 in LtEc, but LtEc may be able to reduce O2 displacement at low concentrations of CO.

-1 -1 -1 Sample koff (s ) kon,CO (M s ) bHb 33.91 0.201 HbA 38.72 0.204 LtEc 30.29 0.097

Table 5.3 – Ligand binding kinetics for purified LtEc with HbA and bHb as controls.

The viscosity and COP of LtEc and human whole blood are shown in Table 5.4. The viscosity and COP of LtEc increase with increasing concentration. The viscosity and COP of LtEc at 10 g/dL (4.27 cP and 14.0 mm Hg) are only slightly lower than human blood (4.5 cP and 19-

24.5 mm Hg). Therefore, LtEc should have little to no effect on blood viscosity and COP when transfused into humans.

Sample Viscosity (cP) COP (mm Hg) LtEc (10 g/dL) 4.27 + 0.96 14.0 + 3.4 LtEc (5 g/dL) 2.10 + 0.36 10.0 + 2.0 Human whole blood 4.50 19-24.5

Table 5.4 – Viscosity and COP values of LtEc at 5 & 10 g/dL. Values for human whole blood are also included for reference.

UV-Vis spectra for LtEc and HbA with different ligands (O2, CO, NO) are shown in Figure

- 5.5, while the UV-Vis spectra of metLtEc and metHbA with and without bound nitrite (NO2 ) are shown in Figure 5.6. These spectra represent all of the ligands which bind to HbA and LtEc to form a stable intermediate. The maxima for each peak in the individual spectra are given in

Tables 5.5 and 5.6. For the most part, no significant differences are observed in the spectra for

LtEc and HbA. The maxima for each liganded species occur at essentially the same wavelengths, with differences of + 2 nm. There are large differences in the magnitude of each peak, but this difference may be due to error in the Bradford assay in assessing total protein concentration.

Therefore, binding of O2, CO, and NO appear to have similar effects on the UV-Vis spectra of LtEc and HbA.

The spectra of metLtEc and metHbA are significantly different. Whereas metHbA displays distinct peaks at 400, 500, 540, 577, and 630 nm, the metLtEc peaks are present yet much less well defined. There is also a large difference in the magnitude of the spectra, but this again may be due to errors in the Bradford assay. When metLtEc and metHbA bind nitrite, even more differences in the spectra are observed. While both samples display clear right shifts in their Soret bands and recovery of their Q-bands, only metHbA shows a defined peak at 627 nm.

MetLtEc completely lacks this peak. These large differences in the spectra of the oxidized samples suggest that the structure of the metLtEc heme pocket is significantly different than 86 metHbA. Crystallographic studies will need to be performed to determine the nature and possible effects of these differences.

Figure 5.5 – UV-Vis spectra of purified LtEc and HbA with various ligands. The y-axis shows the extinction coefficient (, AU/((mM heme)(cm))) at varying wavelengths (nm).

- Figure 5.6 – UV-Vis spectra of purified metLtEc and metHbA with and without bound NO2 . The y-axis shows the extinction coefficient (, AU/((mM heme)(cm))) at varying wavelengths (nm).

Q-bands Ligand Hb Soret   LtEc 428 (114.1) n/a n/a HbA 430 (151.3) 555 (11.3) LtEc 416 (114.) 540 (13.5) 576 (14.0) O 2 HbA 414 (163.5) 541 (18.5) 576 (19.6) LtEc 419 (165.2) 538 (12.7) 569 (12.4) CO HbA 418 (185.3) 538 (16.2) 568 (16.2) LtEc 417 (124.1) 544 (10.9) 575 (11.0) NO HbA 418 (148.5) 545 (12.0) 573 (12.0)

Table 5.5 – Maxima (nm)/Extinction Coefficients (, AU/((mM heme)(cm)) of LtEc and HbA spectra.

Ligand Hb Soret Peak #1 Peak #2 LtEc 401 (62.2) 491 (8.5) 631 (2.2) n/a HbA 405 (193) 500 (11.0) 630 (4.6) LtEc 416 (91.8) 537 (8.7) n/a NO - 2 HbA 412 (154) 538 (10.9) 627 (3.1)

Table 5.6 – Maxima (nm)/Extinction Coefficients (, AU/((mM heme)(cm)) of metLtEc and metHbA spectra.

The interactions of HbA and LtEc with NO are shown in Figure 5.7. When HbA and other mammalian Hbs are mixed with O2 and NO, they catalyze the NO dioxygenation reaction which

- forms NO3 and metHb. When metHb is formed, the Soret band shifts from 416 (oxy-Hb) to 405 nm (metHb). NO dioxygenation is traditionally monitored by tracking changes in the absorbance at 438 nm, which is on the right shoulder of the Soret band. Therefore, A438 should decrease when NO is mixed with Hb-O2 since the Soret band is shifting to the left. This effect is best illustrated in the left pane of Figure 5.7, which shows A438 decreasing immediately (14 ms) after mixing HbA-O2 and NO.

In contrast, A438 increases when LtEc-O2 is mixed with NO. This effect could be caused by a right shift in the Soret band, which may indicate NO binding. For example, when HbA or

LtEc are mixed with CO, the Soret band shifts to the right. At the very least, it appears that oxidation of the heme is not occurring. It is also important to note that the increase in A438 happens over a much longer time scale than NO dioxygenation, which would correlate with a ligand substitution reaction. Finally, the UV-Vis spectra of LtEc-O2 samples mixed with NO are highly similar to the UV-Vis spectrum of LtEc-NO (data not shown).

Figure 5.7 – Interactions of HbA and LtEc with NO. Changes in the absorbance at 438 nm suggest that HbA scavenges NO, while LtEc binds NO.

LtEc may also have a novel reaction with potassium cyanide (KCN, see Figure 5.8). Hb and metHb concentrations are usually measured using KCN and K3[Fe(CN)6] with the cyanomethemoglobin method. When KCN is mixed with HbA or metHbA, it tightly binds the CN- ion (red and green spectra). When HbA and K3[Fe(CN)6] are mixed, the HbA is instantly oxidized to metHbA (blue spectra).

In contrast, when LtEc and KCN are mixed, the LtEc-CN spectra slowly changes to resemble the metLtEc-CN spectra. Therefore, it appears that the cyanide ion is somehow oxidizing the LtEc. Excess cyanide then immediately binds to the metLtEc to give a spectrum that is identical to the spectrum obtained by adding KCN directly to metLtEc. More work will need to be done to determine the nature of this reaction and what happens to the CN- ligand.

However, this reaction might be utilized in bioremediation for eliminating cyanide.

Figure 5.8 – Effects of KCN and K3Fe(CN)6 on HbA (top) and LtEc (bottom). All spectra were monitored for 90 minutes and the final LtEc spectra are shown in bold black.

5.4 - Economic Considerations

The LtEc product seems to be highly pure and is an effective O2 transporter. However, in order to serve as a truly viable alternative to RBCs, it must also be available in large quantities at low cost. We have shown that each round of purification can yield several grams (~5 g) of

LtEc from 1,000 Earthworms. To put this number into context, a typical unit of donated RBCs contains approximately 40-80 g of HbA. Therefore, approximately 8,000-16,000 Earthworms would be required to produce enough LtEc to match a single unit of donated blood. The TFF process may be easily scaled up to meet such a demand, so the only limiting factor may be availability of Earthworms. However, thanks to the bait industry, a large infrastructure for

Earthworm production already exists and they are available at low cost (~$70.00/1000 worms).

Therefore, the base price of a unit of LtEc would be about $560-1020, which is close to the cost of a unit of donated blood after it has been processed and screened for infectious diseases (cost ranges from $200-$1000 per unit). Therefore, LtEc may indeed be an economically viable alternative to donated blood.

Capital costs for TFF processes are moderate (compared to other techniques like chromatography) and the maintenance costs are especially low. The 500 kDa TFF filters may be reused indefinitely with proper cleaning, however, it appears that the 0.22 m TFF filters are susceptible to fouling by the clotting proteins. Therefore, to make the process as cost-effective as possible, methods will have to be developed to either (A) remove the clotting protein and other aggregates which foul the membrane or (B) devise a better way to clean the membrane.

At any rate, the only other equipment required for the process was a low energy peristaltic pump and a benchtop centrifuge, making the process easily adaptable to most labs and easy to scale up for increased production, if necessary.

5.5 – Conclusion

Altogether, these results show that TFF can be used to quickly produce large amounts of highly pure LtEc. The economic analysis also shows that LtEc should be an economically viable

92 alternative to donated blood, since it is readily available and inexpensive to produce. The LtEc product effectively transports O2 and appears to lack NO dioxygenase activity. LtEc also binds

- CO, NO, and NO2 in a similar fashion to HbA. However, LtEc interacts with KCN in a novel way, forming metLtEc and presumably reducing cyanide.

All of these data support further investigation of LtEc as an HBOC. The interactions between LtEc, NO, and CN- must be investigated further in vitro to determine the exact nature of these phenomenon and anticipate any side-effects in vivo. Animal studies must also be done to determine the stability, oxidation rate, and any interactions with NO in vivo. Repeated transfusions must also be done to determine if LtEc creates an immune response.

5.6 – References

1. Fushitani K, Riggs AF. The extracellular hemoglobin of the earthworm, Lumbricus terrestris. Oxygenation properties of isolated chains, trimer, and a reassociated product. J Biol Chem 1991 Jun 5;266(16):10275-10281.

2. Hirsch RE, Jelicks LA, Wittenberg BA, Kaul DK, Shear HL, Harrington JP. A first evaluation of the natural high molecular weight polymeric Lumbricus terrestris hemoglobin as an oxygen carrier. Artif Cells Blood Substit Immobil Biotechnol 1997 Sep;25(5):429-444.

3. Canicatti C. Hemolysins: Pore-forming proteins in invertebrates. Cellular and Molecular Life Sciences 1990;46(3):239-244.

4. Valembois P, Roch P, Lassegues M. Evidence of plasma clotting system in earthworms. Journal of Invertebrate Pathology 1988;51(3):221-228.

5. Lamy J, Kuchumov AR, Taveau JC, Vinogradov SN, Lamy JN. Reassembly of Lumbricus terrestris hemoglobin: a study by matrix-assisted laser desorption/ionization mass spectrometry and 3D reconstruction from frozen-hydrated specimens. J Mol Biol 2000;298:633-647.

6. Xie Q, Donahue RA, Jr., Schneider K, Mirza UA, Haller I, Chait BT, et al. Structure of chain d of the gigantic hemoglobin of the earthworm. Biochim Biophys Acta 1997 Feb 8;1337(2):241- 247.

7. Fushitani K, Matsuura MS, Riggs AF. The amino acid sequences of chains a, b, and c that form the trimer subunit of the extracellular hemoglobin from Lumbricus terrestris. J Biol Chem 1988 May 15;263(14):6502-6517.

8. Jhiang SM, Riggs AF. The structure of the gene encoding chain c of the hemoglobin of the earthworm, Lumbricus terrestris. J Biol Chem 1989 Nov 15;264(32):19003-19008.

9. Suzuki T, Riggs AF. Linker chain L1 of earthworm hemoglobin. Structure of gene and protein: homology with low density lipoprotein receptor. Journal of Biological Chemistry 1993;268(18):13548-13555.

10. Kao WY, Qin J, Fushitani K, Smith SS, Gorr TA, Riggs CK, et al. Linker chains of the gigantic hemoglobin of the earthworm Lumbricus terrestris: primary structures of linkers L2, L3, and L4 and analysis of the connectivity of the disulfide bonds in linker L1. Proteins: Structure, Function, and Bioinformatics 2006;63(174-187):174.

CHAPTER 6: TRANSFUSION OF ERYTHROCRUORIN IN AN EXTREME HEMODILUTION MODEL

6.1 - Introduction

Chapters 4 and 5 have shown that earthworm hemoglobin (LtEc) is extremely stable, has a low rate of oxidation, and lacks NO dioxygenation activity, making it an attractive candidate for transfusion studies. Two previous animal studies have been performed using LtEc1 and

AmEc2 that have given some promising results.

Hirsch et al. showed that small amounts LtEc (0.2-1.0g%) effectively and safely transported O2 in a mouse model. The oxygen affinity (P50) of LtEc in mouse plasma was consistently 25-30 mm Hg, as desired (P50 of RBCs = 24-28 mm Hg). The mice showed no allergic responses or significant metLtEc formation and all of the mice (N = 8) were alive more than 10 months after infusion of LtEc. In addition, half of the blood of a single rat was also exchanged with a 0.71g% LtEc solution without any noticeable changes in health or behavior.1

Rousselot et al. also performed smaller scale experiments in which AmEc was infused into mice (600 mg AmEc/kg mouse weight) to detect any immunogenicity associated with AmEc.

AmEc was injected twice over a period of 7 days and blood samples were taken from each mouse to detect antibody formation (IgE or IgG2a). No significant antibody formation was observed in the mice after infusion of AmEc, while repeated infusions of the negative control

(ovalbumin) caused a significant antibody response and 33% death rate in the mice. In addition, the animals infused with AmEc were alive and healthy over 10 months after the experiment.

Since AmEc and LtEc have similar structures and sequences, the immunogenicity of LtEc may also be negligible.2

This study examines the use of LtEc in hemodilution experiments with hamsters to simulate conditions of severe blood loss in which a transfusion might be needed. Our main goals were to (1) assess O2 delivery by LtEc, (2) determine if LtEc causes hypertension and/or vasoconstriction, (3) measure the pharmacokinetics of LtEc, and (4) detect any allergic/immune responses to LtEc. Polymerized bovine hemoglobins (PolybHbs) were used as positive controls in each experiment. Overall, LtEc appears to be an effective O2 carrier that does not elicit hypertension or vasoconstriction and lacks any significant immune responses.

6.2 - Materials and Methods

Purification of LtEc

LtEc was prepared as described in Section 5.2.

Window Chamber animal Preparation

In vivo studies were performed in 55 - 65 g male Golden Syrian Hamsters (Charles River

Laboratories, Boston, MA) fitted with a dorsal skinfold window chamber. The hamster window chamber model is widely used for microvascular studies in the unanesthetized state. The complete surgical technique is described in detail elsewhere.3 Arterial and venous catheters filled with a heparinized saline solution (30 IU/ml) were implanted into the carotid and jugular vessels. Catheters were tunneled under the skin, exteriorized at the dorsal side of the neck, and securely attached to the window frame. Animal handling and care followed the NIH Guide for the Care and Use of Laboratory Animals. The experimental protocol was approved by the local animal care committee.

Inclusion Criteria

Animals were considered suitable for experiments if systemic parameters were within normal range, namely, heart rate (HR) > 340 beats/min, mean arterial blood pressure (MAP) >

80 mmHg, systemic Hct > 45%, and arterial O2 partial pressure (pAO2) > 50 mmHg. Additionally, animals were examined 3 to 4 days after implantation surgery, tissue was observed under 650 magnification, and only animals without signs of low perfusion, inflammation, edema or bleeding were included.

Systemic parameters

MAP and HR were monitored continuously (MP150, Biopac System Inc., Santa Barbara,

CA), except when: i) blood was sampled, and ii) during blood exchange/withdrawal. The Hct was determined from centrifuged arterial blood samples taken in heparinized capillary tubes.

Blood Chemistry and Biophysical Properties

Arterial blood was collected in heparinized glass capillaries (0.05 ml) and immediately analyzed for PaO2, PaCO2, base excess (BE) and pH (Blood Chemistry Analyzer 248, Bayer,

Norwood, MA).

COP, blood, and plasma viscosities

Blood samples for viscosity and colloid osmotic pressure measurements were quickly withdrawn into heparinized 5 ml syringes at the end of the experiment. Plasma COP was measured using a Wescor 4420 Colloid Osmometer (Wescor, Logan, UT).4 The viscosity was measured in a cone/plate viscometer DV-II plus with a cone spindle CPE-40 (Brookfield

Engineering Laboratories, Middleboro, MA) at a shear rate of 160 sec-1.

Experimental Setup

The unanesthetized animal was placed in a restraining tube with a longitudinal slit from which the window chamber protruded, then fixed to the microscopic stage for transillumination with the intravital microscope (BX51WI, Olympus, New Hyde Park, NY). Animals were given 20 min to adjust to the tube environment before any measurements were made. The tissue image was projected onto a charge-coupled device camera (4815, COHU, San Diego, CA) connected to a videocassette recorder and viewed on a monitor. Measurements were carried out using a 40

(LUMPFL-WIR, numerical aperture 0.8, Olympus) water immersion objective.

Figure 6.1 – Hemodilution protocol. The initial Hct of all animals was reduced to 18% with the plasma expander 6% w/v Dextran 70 kDa (Pharmacia, Uppsala, Sweden), then the test solutions were infused to reduce the final Hct down to 11%.

Acute Isovolemic Exchange Transfusion (Hemodilution) Protocol

Acute anemia was induced by two isovolemic hemodilution steps. This protocol was described in detail in our previous reports (Figure 6.1).5 Briefly, the volume of each exchange- transfusion step was calculated as a percentage of the blood volume (BV), estimated as 7% of

98 body weight (BW). The acute anemic state was induced by lowering systemic Hct to 18% by two steps of progressive isovolemic hemodilution using 6% w/v dextran 70 kDa (Pharmacia, Uppsala,

Sweden). Blood was simultaneously withdrawn at the same rate from the carotid artery catheter according to a previously established protocol.5, 6 The first exchange was 40% of the BV and second exchange was 35% of the BV, respectively. Moderate hemodiluted animals were randomly divided into three experimental groups.7 The exchange transfusion protocol was continued by exchanging 35% of the BV with the test solution, experimental groups are named accordingly. The duration of the experiments was 4 h. Each exchange and the respective observation time point post-exchange were completed in 1 h. Systemic and microcirculation data were taken after a stabilization period of 15 min.

Since mixed blood and dilution material was withdrawn during the exchanges, a 35% blood volume exchange was needed to reduce the functional Hct from 18% to 11%. Diluents were filtered (in-line, 0.22 µm filter) and infused into the jugular vein catheter at a rate of 100

l/min. Blood was simultaneously withdrawn at the same rate from the carotid artery catheter according to a previously established protocol. Blood samples were withdrawn at the end of the experiment for subsequent analysis of viscosity and colloid osmotic pressure. The duration of the experiments was 4h. Each exchange and the respective observation time point post- exchange were fully completed in 1h. Systemic and microcirculation data was taken after a 10 min stabilization period.

Test solutions and Experimental Groups

Anemic state (18% Hct) was induced with 6% Dextran 70 kDa (Pharmacia, Uppsala,

Sweden), and extreme hemodilution was induced LtEc solution at 10 gHb/dl or polybHb solutions polymerized at T-state PolybHb (PolyHbHighP50) and R-state PolybHb (PolyHbLowP50). The

99 biophysical properties of the LtEc and the PolybHb solutions are presented in Table 6.1. Animals were randomly divided into the following three experimental groups before the experiment: 1)

LtEc, 2) PolyHbHighP50, and 3) PolyHbLowP50.

Microvascular PO2 distribution

High resolution non-invasive microvascular PO2 measurements were made using

6, 8 phosphorescence quenching microscopy (PQM). PQM is based on the O2-dependent quenching of phosphorescence emitted by albumin-bound metalloporphyrin complex after pulsed light excitation. PQM is independent of the dye concentration within the tissue and is well suited for detecting hypoxia because its decay time is inversely proportional to the PO2 level, causing the method to be more precise at low PO2's. This technique is used to measure both intravascular and extravascular PO2 since the albumin-dye complex continuously

6, 8 extravasates the circulation into the interstitial fluid. Extravascular fluid PO2 (interstitial fluid) was measured in tissue regions in between functional capillaries. PQM allows for precise localization of the PO2 measurements without subjecting the tissue to injury. These measurements provide a detailed understanding of microvascular O2 distribution and indicate whether O2 is delivered to the interstitial areas.

O2 delivery and extraction

The microvascular methodology used in our studies allows a detailed analysis of O2 supply in the tissue. Calculations are made using equation 1 and 2:9

O2 delivery = [(RBCHb × γ × SA) + (PlasmaHb × γ × ŠA) ] × Q (1)

O2 extraction = [(RBCHb × γ × SA-V) + (PlasmaHb × γ × Š A-V] × Q (2)

Where, RBCHb is the Hb from RBCs, PlasmaHb is the acellular Hb [gHb/dlblood], γ is the O2 carrying capacity of saturated hemoglobin [1.34 mlO2/gHb], SA is the arteriolar blood O2 saturation and ŠA

100 is the arteriolar acellular Hb O2 saturation, A-V indicates the arteriolar/venular differences, and Q is the microvascular flow. O2 saturations were measured as described above.

Statistical analysis

Results are presented as mean ± standard deviation. The values are presented relative to the baseline. A ratio of 1.0 signifies no change from the baseline, whereas lower or higher ratios are indicative of changes proportionally lower or higher than the baseline. The Grubbs' method was used to assess closeness for all measured parameters values at baseline. This method quantifies how far each parameter value is from the other values obtained, computing a

P value supposing that all the values were really sampled from a Gaussian population. Data between interested time points in a same group was analyzed using analysis of variance for repeated measures (ANOVA) and followed by post hoc analyses with the Dunnett’s multiple comparison tests when appropriate. An unpaired t-test with two-tailed was performed to compare between groups at the time point of interest. All statistics were calculated using

GraphPad Prism 4.01 (GraphPad Software, San Diego, CA). Results were considered statistically significant if P < 0.05.

6.3 - Results

The properties of the test solutions are shown in Table 6.1. The viscosities of LtEc and

PolybHblow are roughly similar (~8 cP) and higher than blood (4.5 cP), but PolybHbhigh has the highest viscosity (11.4 cP). In contrast, PolybHbhigh has almost no COP (1 mm Hg) while LtEc and

PolybHblow have low COP values (6-7 mm Hg) which are much less than whole blood (19-25 mm

Hg). The most important difference between the HBOCs is their range of O2 affinities and cooperativities. PolybHblow has the highest O2 affinity (P50 = 1 mm Hg) and PolybHbhigh has the

101 lowest (P50 = 40 mm Hg), while the O2 affinity of LtEc (P50 = 28 mm Hg) is closer to that of whole human blood (P50 = 26 mm Hg). Finally, the PolybHb’s exhibit a have almost no cooperativity

(0.7-0.9) relative to whole blood (2.7) and LtEc (3.7).

MW [Hb]  COP P50 n (kDa) (g/dL) (cP) (mm Hg) (mm Hg) Plasma Expanders 6% Dextran 70 --- 2.8 50 ------

Hb-based O2 Carriers LtEc 3,600 10 8.2 6 28 3.7

PolybHbLowP50 26,330 10 7.8 7 1 0.7 PolybHbHighP50 16,590 10 11.4 1 40 0.9 Human Blood --- 14 4.5 19-25 26 2.7

Table 6.1 – Properties of infused HBOCs. MW = Molecular weight, [Hb] = hemoglobin concentration,  = viscosity, COP = colloid osmotic pressure, P50 = O2 affinity, n = cooperativity.

The systemic parameters of all animals involved in the study before and after hemodilution are shown in Table 6.2. The Hct of each animal was consistently reduced to 18.2% with dextran, then lowered further to 11.3-11.6% with the test solutions. The concentration of

HBOC in the plasma was constant at 2.7-2.9 g/dL and accounted for approximately 45% of the total hemoglobin in the blood stream. Arterial O2 pressure increased during hemodilution and continued to increase after injection of LtEc and PolybHbhigh (94-97 mm Hg), but increased significantly more after infusion of PolybHblow (126.2 mm Hg). In contrast, arterial CO2 pressure decreased during hemodilution but remained constant after injection of LtEc and PolybHbhigh while paCO2 continued to decrease after addition of PolybHblow.

102

Moderate Exchange Transfusion of: Baseline Hemodilution PolybHbLowP50 PolybHbHighP50 LtEc N (# animals) 18 18 6 6 6 Hct (%) 48.4 + 0.9 18.2 + 0.6† 11.4 + 0.4†‡ 11.3 + 0.5†‡ 11.6 + 0.7†‡ Total Hb (g/dL) 14.6 + 0.4 5.6 + 0.4 6.1 + 0.3†‡ 6.2 + 0.3†‡ 6.1 + 0.5†‡ Plasma Hb (g/dL) 2.8 + 0.4 2.9 + 0.3 2.7 + 0.4 † †‡ †‡ †‡ paO2 (mm Hg) 58.8 + 5.6 77.2 + 6.5 126.2 + 8.4 94.6 + 6.2 97.2 + 7.8 †‡• †‡ †‡ paCO2 (mm Hg) 53.3 + 5.1 46.5 + 5.4 37.6 + 6.3 44.3 + 4.5 45.1 + 4.8 pHa 7.336 + 0.020 7.372 + 0.022 7.387 + 0.026 7.369 + 0.024 7.372 + 0.029 †‡• †‡ †‡ BEa (mmol) 2.6 + 1.7 1.2 + 0.9 -2.2 + 1.7 0.6 + 0.8 0.8 + 0.9 Blood  (cP) 4.2 + 0.2 2.5 + 0.2† 2.9 + 0.3† 2.7 + 0.2† Plasma  (cP) 1.2 + 0.1 1.7 + 0.1†• 2.1 + 0.2† 2.0 + 0.2† COP (mm Hg) 18 + 2 17 + 2 17 + 1 16 + 2

Table 6.2 – Systemic parameters of all animals before and after exchange transfusion. † -

P<0.05 compared to baseline, ‡ - P<0.05 compared to MH, and § - P<0.05 compared to LtEc.

Blood viscosity was significantly lower than baseline (4.2 cP) after hemodilution with dextran and injection of all the test solutions (2.5-2.9 cP). The plasma viscosity, however, increased significantly after addition of the test solutions. It is important to note that PolybHblow had a significantly lower effect on plasma viscosity than the PolybHbhigh and LtEc. No significant change in COP was observed after infusion of any of the test solutions.

103

Figure 6.2 – Partial pressure of O2 (pO2) in arterioles, tissues, and venules after transfusion of the test solutions.

O2 delivery by the test solutions is illustrated in Figure 6.2. Each test solution significantly raised the pO2 in the arterioles, tissues, and venules relative to moderate hemodilution (18% Hct), indicating successful O2 transport by each HBOC. The effects of each

HBOC on functional capillary density (FCD) are shown in Figure 6.3. All animals showed significant decreases in FCD relative to baseline, as expected in hemodilution. PolybHblow exhibited the lowest FCD, but LtEc and PolybHbhigh had slightly higher values of FCD.

104

Figure 6.3 – Effects of test solutions on functional capillary density (FCD).

The effects of each HBOC on hypertension (measured by mean arterial pressure and heart rate) and vasoconstriction (measured by blood vessel diameter and blood flow) are shown in Figures 6.4 and 6.5, respectively. After moderate hemodilution, all animals experienced a significant reduction in MAP. Transfusion of LtEc caused a significant reduction in MAP, while the PolybHb’s elicited a significant increase in MAP relative to moderate hemodilution (MH). No significant changes in heart rate were observed during hemodilution or after transfusion of any of the HBOCs.

Figures 6.2-6.3 show that LtEc is able to restore O2 delivery after hemodilution just as well as, if not better than, existing HBOCs (PolybHbs). LtEc is also able to maintain some FCD, however, the FCD values in the LtEc subjects were significantly lower than baseline and moderate hemodilution.

105

Figure 6.4 – Effects of test solutions on mean arterial pressure (MAP) and heart rate (HR).

All animals experienced significant increases in arteriolar diameter and blood flow after moderate hemodilution (see Figure 6.5). Interestingly, animals transfused with LtEc exhibited the highest arteriole diameter and arterial blood flow, while the PolybHb subjects were similar to baseline. In contrast, no significant changes in venule diameter or blood flow were observed for any of the HBOCs.

106

Figure 6.5 – Effects of test solutions on arteriolar diameter and blood flow.

As expected, moderate hemodilution caused a significant decrease in MAP due to the reduction in circulation volume. A significant increase in MAP was observed after addition of the PolybHb’s, indicating some hypertension. In contrast, significant reductions in MAP were observed after addition of LtEc. Therefore, it seems that LtEc does not cause hypertension like

PolybHbs and other HBOCs. No significant changes in heart rate were observed for any of the test solutions.

Figure 6.5 also shows that LtEc does not elicit vasoconstriction. Instead, it appears to relax arterioles. This effect may indicate that LtEc is able to transport NO, instead of scavenging

NO like other HBOCs. However, it does not appear to have the same vasodilatory effect on

107 venules. Therefore, further studies will need to be done to conclusively determine the interactions between LtEc and endogenous NO in vivo. No significant changes were observed in arterial or venular blood flow, except a small increase in arterial blood flow after injection of

LtEc. The increase in blood flow is likely related to the increase in arterial diameter.

The pharmacokinetics of each HBOC after 40% exchange transfusion are shown in Figure

6.6. Each HBOC exhibits similar decays in concentration from 4 g/dL to <1 g/dL over a period of

48 hours. PolybHbhigh has the highest half life (15 hrs), with LtEc close behind (14.5 hrs) and

PolybHblow with the lowest half life (12.9 hrs). LtEc is also appears to have a lower rate of oxidation in vivo. The PolybHbs both have metHb levels around 30% after 48 hours, whereas the metLtEc level only increases to 10% after 48 hours.

108

Figure 6.6 – Pharmacokinetics and in vivo oxidation of each test solution.

The results in Figure 6.6 show that LtEc is eventually cleared from the blood stream, but has a half life (14.5 hrs) long enough to deliver a significant amount of O2. The measurements of

Fe2+ and Fe 3+ show that LtEc is able to resist oxidation much more effectively than the PolybHb samples. Therefore, while the clearance of LtEc and the PolybHb’s is mostly similar, a greater amount of the LtEc is still functional (Fe2+) than the PolybHb samples.

The effects of multiple injections of LtEc (10% exchange transfusions every day for 5 days) on MAP, HR, arteriolar diameter, and FCD are shown in Figure 6.7. All animals survived

109 the multiple injections without any complications. Likewise, no significant changes in MAP, HR, or arteriolar diameter were observed. The animals did exhibit significant decreases in FCD after the first injections of LtEc, but such an effect is expected during any exchange transfusion and

FCD stabilized after the 3rd day.

Figure 6.7 – Effects of LtEc on mean arterial pressure (MAP), heart rate (HR), arteriolar diameter, and functional capillary density (FCD) after consecutive injections over a period of 5 days.

LtEc does not appear to elicit any harmful side effects or immune/allergic responses.

Instead, animals appeared healthy (stable MAP, HR, and arteriolar diameter). These results are only preliminary and further immune studies over longer periods of time will need to be performed. We will also try to detect antibodies formed in response to LtEc injection.

6.4 - Conclusion

This study shows that LtEc is an effective HBOC which does not elicit any significant vasoconstriction or hypertension. Instead, its unique interactions with NO may endow it with a vasodilatory response. Future work will need to be done, both in vivo and in vitro, to determine

110 how LtEc interacts with NO and its effects on endogenous NO concentrations. We have also shown that LtEc has a circulation half life that is similar to other existing HBOCs and suitable for tranfusion. The in vivo oxidation rate of LtEc is also significantly lower than the PolybHbs, indicating that more LtEc remains functional for a longer period of time and does not increase tissue oxidative stress like other HBOCs. Most importantly, no immune or allergic responses were observed in any of the test animals, indicating that LtEc should be safe for further animal studies.

6.5 - References

1. Hirsch RE, Jelicks LA, Wittenberg BA, Kaul DK, Shear HL, Harrington JP. A first evaluation of the natural high molecular weight polymeric Lumbricus terrestris hemoglobin as an oxygen carrier. Artif Cells Blood Substit Immobil Biotechnol 1997 Sep;25(5):429-444.

2. Rousselot M, Delpy E, Drieu La Rochelle C, Lagente V, Pirow R, Rees JF, et al. Arenicola marina extracellular hemoglobin: a new promising blood substitute. Biotechnol J 2006 Mar;1(3):333-345.

3. Colantuoni A, Bertuglia S, Intaglietta M. Quantitation of rhythmic diameter changes in arterial microcirculation. Am J Physiol 1984 Apr;246(4 Pt 2):H508-517.

4. Webb AR, Barclay SA, Bennett ED. In vitro colloid osmotic pressure of commonly used plasma expanders and substitutes: a study of the diffusibility of colloid molecules. Intensive Care Med 1989;15(2):116-120.

5. Cabrales P, Sakai H, Tsai AG, Takeoka S, Tsuchida E, Intaglietta M. Oxygen transport by low and normal oxygen affinity hemoglobin vesicles in extreme hemodilution. Am J Physiol Heart Circ Physiol 2005 Apr;288(4):H1885-1892.

6. Tsai AG, Friesenecker B, McCarthy M, Sakai H, Intaglietta M. Plasma viscosity regulates capillary perfusion during extreme hemodilution in hamster skinfold model. Am J Physiol 1998 Dec;275(6 Pt 2):H2170-2180.

7. Altman DG, Bland JM. Statistics notes: How to randomise. BMJ 1999;319(7211):703- 704.

111

8. Kerger H, Groth G, Kalenka A, Vajkoczy P, Tsai AG, Intaglietta M. pO(2) measurements by phosphorescence quenching: characteristics and applications of an automated system. Microvasc Res 2003 Jan;65(1):32-38.

9. Cabrales P, Tsai AG, Intaglietta M. Microvascular pressure and functional capillary density in extreme hemodilution with low- and high-viscosity dextran and a low-viscosity Hb- based O2 carrier. Am J Physiol Heart Circ Physiol 2004 Jul;287(1):H363-373.

112

CHAPTER 7: EXPRESSION OF HEME PROTEINS IN E. COLI

7.1 – Introduction: A Recombinant Hb Review

The previous chapters have highlighted some intriguing differences between human and earthworm Hbs. Crystal structures and amino acids sequences can be used to identify specific residues which might be responsible for these differences. Site-directed mutagenesis may then be performed to verify which residues endow LtEc with a lower rate of autooxidation and NO dioxygenase activity than HbA.

HbA is one of the most extensively studied proteins in biochemistry. Recombinant HbA

(rHbA) has been expressed in a variety of cell lines and organisms, including Escherichia coli1, yeast2, insects3, mice4, and swine5. Most rHbA mutations have been made in E. coli, since they grow quickly and HbA does not require any post-translational modifications that would require a eukaryotic host. Nagai and Thogersen were the first to express the  subunit of HbA as a fusion protein in E. coli.1 Ho et al. expanded on this system by expressing native  and  subunits together under a T7 promoter. N-terminal methionines had to be added to each subunit express the subunits in E. coli, so a methionine amino peptidase (MAP) gene was also included in the vector.6 The MAP enzyme cleaved off the N-terminal methionines in rHbA to make its sequence identical to native HbA.7, 8 The codons in each rHbA gene were also optimized for use in E. coli, which caused a significant increase in yield over previous systems.9 Looker et al. developed a similar expression system in which two alpha subunits were genetically fused 113

(creating a single di-alpha gene) and the sequence of the  gene was mutated to lower its oxygen affinity. The recombinant HbA made in this system (rHb1.1) was immune to tetramer dissociation since its  subunits were covalently linked and rHb1.1 had an oxygen affinity similar to human blood.10 rHb1.1 was even used in clinical trials, but was withdrawn due to side effects associated with NO scavenging.11-16 A second generation rHbA (rHb2.0) was developed with several mutations to significantly reduce NO scavenging, but development of rHb2.0 was also terminated.15

Both the Ho and Looker systems have been used to make countless mutations in rHbA to determine which amino acids can be manipulated to influence its oxygen affinity17-20, resistance to autoxidation21, and NO dioxygenase activity.22, 23 However, even the rHbAs with the most beneficial mutations are still limited by low yields and much of the protein aggregates into inclusion bodies.19 Expression at lower temperatures24 and coexpression of a chaperonin for the a subunit (alpha hemoglobin stabilizing protein, AHSP)25, 26 have been shown to increase expression, but only up to 10-20 mg of rHbA per liter of culture. A new rHbA expression system was recently introduced by Domingues et al. in which the  subunit of rHbA is fused to a glutathione S-transferase demand through an enterokinase-cleavable linker. The  subunit is expressed from the same vector, but with an N-terminal methionine and no MAP gene. The rHbA expressed in this system can be highly purified in a single step by binding it to a glutathione resin, then harvested by cleaving rHbA with an enterokinase enzyme at yields of up to 20 mg rHbA per liter of bacterial culture.27

In this work, we present a novel expression system for rHbA in which each subunit is expressed individually as a fusion protein with a chitin binding domain and an intein that cleaves itself at low pH (6.0-6.5). To our knowledge, this is the first successful expression of individual 

114 subunits.9, 28 A sperm whale myoglobin (Mb) fusion protein was also expressed as a positive control, since Mb is known to express well as an individual subunit in E. coli. The fusion proteins are be easily purified in a single affinity chromatography step and the purified/cleaved subunits may be combined in an exact 1:1 ratio to produce an rHbA that is functionally similar to native

HbA. The yield of each fusion protein is high, but final yields of each purified globin are limited by slow intein cleavage. Nonetheless, this system is able to produce sufficient amounts of rHbA subunits to determine the effects of mutations of both individual subunits and a self-assembled rHbA tetramer.

7.2 - Materials and Methods

Vector Construction

Synthetic genes for the  and  subunits were designed using an E. coli K12 high expression codon bias and ordered from Integrated DNA Technologies (IDT, San Diego, CA). The

Mb gene was generously provided by Dr. John S. Olson from Rice University. All globin genes were cloned into the vector pCIX (provided by Dr. David Wood, from the Ohio State University) for expression in BL21 E. coli K12 cells. This vector has a pBR322 origin with a low copy number, a -lactamase gene for ampicillin resistance, and a T7 promoter/terminator that is regulated by a LacI repressor protein binding site. The vector expresses a fusion protein with a chitin binding domain followed by a pH-inducible self-cleaving intein and the target protein. A ribosome binding site (AAGGAGA) is located 7 bp upstream of the start codon to initiate translation and the stop codon TAA was added to the end of each globin gene to terminate translation. All vectors were sequenced twice to verify the sequence of the globin genes.

115

Figure 7.1 – Expression vector for the globin genes.

Expression in E. coli

BL21 E. coli were transformed with each vector and kept as glycerol stocks until needed for each fermentation. Small starter cultures (50 mL Luria Burtani media with 0.1 mg/mL ampicillin) were inoculated from these glycerol stocks and incubated overnight at 37oC (~10-16 hrs) and 225 rpm. The starter cultures were then used to inoculate larger cultures (four shaker

o flasks, each with 1L LB) which were incubated at 37 C and 225 rpm until the OD600 reached 2.0.

The temperature was then reduced to 15oC and IPTG and hemin were added to final concentrations of 0.5 mM and 40 M, respectively. The cultures were incubated for 24 hours, then harvested by centrifugation at 4,000 g for 25 minutes.

116

Figure 7.2 – Expression and purification scheme for recombinant globins.

Recombinant Protein Purification

Cells were resuspended in lysis buffer (20 mM Tris, 250 mM NaCl, 1 mg/mL benzamadine HCl, pH 8.3) and sparged with CO gas to stabilize the globin subunits. The cell suspension was lysed using a French Press and clarified by centrifugation at 18,000 g for 30 minutes. The lysate was also put through a 0.22 m filter to remove any particulates.

Chitin beads (New England Biolabs, Ipswitch, MA) were loaded onto an XK 20/50 column

(GE Life Sciences, Piscataway, NJ) and flushed with Buffer A (20 mM Tris, 250 mM NaCl, 10 mM

EDTA, pH 8.3) at 7.5 mL/min until the conductivity of the column effluent leveled off at 25 mS/cm. The bacterial lysate was then loaded onto the column and rinsed with Buffer A until the absorbance at 280 nm of the column effluent returned to zero. The chitin binding domain allows the fusion proteins to tightly bind to the resin, while the column buffer flushed away impurities in the mobile phase and the Cl- ions elute any electrostatically bound impurities. The red beads were then flushed with cleavage buffer (20 mM Tris, 10 mM EDTA, pH 6.2) until the conductivity of the column effluent decreased to zero. The beads were then transferred from

117 the column to an air-tight septum vial and flushed with CO for 30 minutes. The CO-saturated samples were then incubated in the dark at room temperature for 7 days to allow intein cleavage to occur. Cleaved globins were then collected by transferring the beads back to the column and flushing them with cleavage buffer. The globins were finally concentrated on a 10 kDa centrifugal filter and kept at -80oC until needed. All buffers were sparged with CO gas for 15 minutes immediately before use and the column temperature was maintained at 4oC.

The chitin beads were regenerated by flushing them with 1% sodium dodecyl sulfate

(SDS) to remove bound protein (the effluent was dark brown). The beads were then rinsed with distilled water, cleaned once more with 0.2 M NaOH and 10 mM EDTA, then flushed with 20% ethanol and stored at 4oC until needed.

Other Methods

PAGE, MALDI, UV-Vis, and stopped-flow kinetics were performed as described in section

2.2, with the following exceptions. PAGE samples were taken from the bacterial lysate, column flow through, beads immediately after purification (t = 0 hours) and after 7 days. Samples were not reduced with sodium dithionite prior to UV-Vis spectroscopy or kinetics experiments. The samples were bathed in bright white light and flushed with pure O2 gas in an ice bath for 30 minutes to remove bound CO prior to kinetics experiments. Elimination of CO was verified by a shift in the Soret band of each sample from 418 nm (CO form) to 414 nm (O2 form).

Recombinant  and  subunits were mixed in a 1:1 ratio to prepare r samples for kinetics.

7.3 – Results and Discussion

The purity of the, and Mb samples throughout the purification process is shown in

Figure 7.1. The fusion proteins are clearly visible in their respective lysates (lane 1), even though

118 they appear at a lower than expected MW. Each fusion protein should have a MW of approximately 41 kDa, but the fusion protein band appears at around 35 kDa (represented by a single asterisk). A significant amount of prematurely cleaved tag (chitin binding domain and intein, represented by **, MW 25,956 g/mol) is also seen in each lysate sample. Unfortunately, it appears that a significant portion of each fusion protein cleaves during expression, with Mb cleaving the most. Nonetheless, the flow through samples (lane 2) show that virtually all of the

E. coli impurities are washed away while the fusion proteins and the prematurely cleaved tag both bind the chitin beads.

The progression of intein cleavage and target protein release are illustrated by lanes 4 &

5. The initial bead samples (lane 4) show that a large amount of fusion protein is bound to the beads. After 7 days (lane 5), the fusion protein bands are only barely visible. It appears that intein cleavage is complete for the  subunit after 7 days, but cleavage of the  and Mb fusion proteins is not yet finished. Nonetheless, the purified globins all show dark bands at 15 kDa

(represented by ***) near their expected MWs. Native HbA is included in the last lane of each gel as a positive control. The MWs of the  and  subunits align nicely with the control, but additional impurity bands are also visible. It seems that some of the cleaved tag elutes along with globins. This is not surprising, since the chitin beads are known to slowly degrade at room temperature. Two more bands between 15-20 kDa are also seen in the purified samples which are not present in the native HbA control. These bands are also seen in the lysate and bead samples, but not in the column flow through samples. Therefore, these bands may be impurities which are able to bind the chitin beads and are released with the fusion tag as the chitin degrades. At any rate, these impurities should be easily removed by flowing the globin samples over fresh chitin beads.

119

Figure 7.3 – PAGE analysis of recombinant , and Mb samples. Lane 1 – Protein MW Ladder, 2 – E. coli lysate, 3 – Bead flow thru (waste impurities), 4 – Beads at t = 0 hrs, 5 – Beads after 7 days, 6 – Purified protein, 7 – Native HbA control. Asterisks mark the different forms of the fusion protein – (*) Full fusion protein, (**) Chitin binding domain & intein tag, (***) target protein.

The molecular weight (MW) of each globin was verified using MALDI-TOF (Figure 7.2).

The observed MWs of each globin ( = 15,123,  = 15,861, and Mb = 17,333 g/mol) are almost

120 identical to their expected MWs ( = 15,126,  = 15,867, and Mb = 17,330 g/mol). Long range scans of each sample reveal no major impurity peaks. Therefore, the unidentified bands in the

PAGE gels are either (1) degradation products of the globins formed during PAGE sample preparation or (2) impurities that are at concentrations too low to be detected by MALDI.

The PAGE and MALDI results indicate that each globin has the correct molecular weight and is highly pure. A few impurities might be present, but they might be easily and quickly removed with fresh chitin beads. Alternatively, if the intein cleaved at a faster rate, then the chitin beads would not have sufficient time to degrade and these impurities might not be in the final sample at all.

The rates of oxygen release (koff), carbon monoxide binding (kon,CO), and NO dioxygenation (kox,NO) for the recombinant globins are shown in Table 7.1. Each recombinant globin reacted with O2, CO, and NO as expected, but with some significant differences.

-1 -1 -1 -1 -1 koff (s ) kon,CO (M s ) kox,NO (M s ) r 28.5 + 2.1 1.94 44.2 r 16.4 + 1.4 2.67 49.1 r 25.0 + 2.5 0.26 --- HbA 37.7 + 1.5 0.22 40.9 rMb 12.4 + 0.01 0.83 11.2

Table 7.1 – Reaction rate constants of recombinant globins with O2, CO, and NO.

121

Figure 7.4 – MALDI analysis of purified , and Mb. Long range scans are shown from 10-115 kDa, while the insets show the observed molecular weights of each sample. Expected molecular weights for each protein are a = 15,126, b = 15,867, and Mb = 17,330 g/mol.

Figure 7.3 illustrates how each of the recombinant globins releases O2 in the presence of the O2 scavenger sodium dithionite. As oxygen is released, the absorbance at 438 nm (A438, y-

122 axis) of native HbA usually increases rapidly and plateaus. A 1:1 mixture of  and  subunits

(r) shows a similar behavior with a slight increase over time, but the behavior of the

individual subunit solutions is significantly different. The A438 of the  and  subunits sharply increases like the r mixture, but the A438 of the  subunit decreases over time and the A438 of the  subunit increases over time. These differences in behavior may reflect fundamental differences in the UV-Vis spectra of the individual recombinant globins relative to native HbA.

-1 The recombinant subunits also have much lower rates of O2 release than native HbA (37.7 s ).

Recombinant Mb, which is known to have a very high affinity for O2, has the lowest rate of

-1 oxygen release (12.4 s ). It seems that the individual globins have much higher O2 affinities than native HbA. It is important to note that a 1:1 mixture of r:r still has lower O2 affinity than native HbA. Therefore, there may also be some structural differences between the recombinant and native subunits that are affecting O2 affinity. 

Figure 7.5 – O2 release by recombinant globins. Raw data are represented by solid lines, whereas fitted curves used to calculate rate constants are represented by dashed lines.

123

The rate of CO binding for each recombinant globin is shown in Figure 7.4. The rate of

CO binding increases as CO concentration increases for each sample, as expected. However, the magnitudes of these reactions are significantly different. Both the native and recombinant HbA tetramers have rates around 0.22-0.26 M-1s-1. However, the rates of the individual r and r subunits are both an order of magnitude higher ( = 1.94 M-1s-1,  = 2.67 M-1s-1). The CO binding rate for myoglobin, the monomer control (0.83 M-1s-1), is also much higher than native

HbA. These results indicate that the individual subunits must have a higher affinity for CO than native HbA which accelerates their rates of CO binding.

Figure 7.6 – CO binding by recombinant globins. Raw data points are represented by crosses, whereas fitted curves used to calculate rate constants are represented by solid lines.

124

Figure 7.7 – CO binding time courses for each recombinant globin. Raw data are shown in blue, while fitted curves are shown in red.

Finally, Figure 7.5 shows the NO dioxygenation activity of each recombinant globin. In this case, the rates of NO dioxygenation for the individual subunits (a = 44.2 M-1s-1, b = 49.1

M-1s-1) and native HbA (50.6 M-1s-1) are quite similar. The rate for rMb, however, is much lower (11.2 M-1s-1). It is not surprising that rMb has a lower rate of NO dioxygenation, since rMb is exposed to much more NO than HbA.

Figure 7.8 – Rates of NO dioxygenation by recombinant globins.

125

Figure 7.9 – NO dioxygenation time courses for each recombinant globin. Raw data are shown in blue, while fitted curves are shown in red.

7.4 – Conclusion

The PAGE and MALDI analyses in Figures 7.1 and 7.2 show that all of the recombinant globins are highly pure. They may be contaminated with a few impurities (the chitin binding domain + intein tag and 1-2 other chitin-binding E. coli proteins), but these impurities might be easily removed by additional chitin beads in future experiments.

The ligand interactions of the recombinant globins are quite interesting. It seems that the oxygen affinities of the recombinant globins are significantly different than native HbA. This may be caused by a number of factors, including differences in structure or a “flipping” of the heme group within the heme pocket. The individual subunits also have a much higher affinity for CO than native HbA. However, the CO binding activity of the recombinant  and  subunits is quite similar to native HbA when they are combined in a 1:1 ratio. Therefore, the methods used to measure the O2 affinity of the individual subunits may be inadequate and giving false results. This would also explain why the O2 release curves of the  and  subunits in Figure 7.3 do not match native HbA or the r mixture.

The NO dioxygenase activity of the individual subunits and the native HbA are also quite similar. This is in contrast to their ligand affinities, which were significantly higher than native

126

HbA or r. This result is not surprising, since NO dioxygenation is an oxidation reaction instead of a ligand binding reaction.

Overall, it seems that this expression system is able to produce relatively large quantities (5-10 mg/L culture) of highly pure recombinant globins in a single affinity chromatography step. The yield and purity of the final products could be increased greatly if intein cleavage could be accelerated. Each recombinant globin also binds O2 and CO, as expected. The rates of O2 release and CO binding are somewhat different, but more in depth studies will be needed to determine the nature of these differences. The NO dioxygenation rates of the recombinant subunits seems to be similar to native HbA though. In conclusion, this expression system allows for convenient expression of fully functional recombinant Hb and Mb subunits, but it could be dramatically improved with a more efficient intein domain.

7.5 – References

1. Nagai K, Thogersen HC. Generation of beta-globin by sequence-specific proteolysis of a hybrid protein produced in Escherichia coli. Nature 1984 Jun 28-Jul 4;309(5971):810-812.

2. Wagenbach M, O'Rourke K, Vitez L, Wieczorek A, Hoffman S, Durfee S, et al. Synthesis of wild type and mutant human hemoglobins in Saccharomyces cerevisiae. Biotechnology (N Y) 1991 Jan;9(1):57-61.

3. Groebe DR, Busch MR, Tsao TY, Luh FY, Tam MF, Chung AE, et al. High-level production of human alpha- and beta-globins in insect cells. Protein Expr Purif 1992 Apr;3(2):134-141.

4. Grosveld F, van Assendelft GB, Greaves DR, Kollias G. Position-independent, high-level expression of the human beta-globin gene in transgenic mice. Cell 1987 Dec 24;51(6):975-985.

5. Swanson ME, Martin MJ, O'Donnell JK, Hoover K, Lago W, Huntress V, et al. Production of functional human hemoglobin in transgenic swine. Biotechnology (N Y) 1992 May;10(5):557- 559.

6. Shen TJ, Ho NT, Simplaceanu V, Zou M, Green BN, Tam MF, et al. Production of unmodified human adult hemoglobin in Escherichia coli. Proc Natl Acad Sci U S A 1993 Sep 1;90(17):8108-8112. 127

7. Ben-Bassat A, Bauer K, Chang SY, Myambo K, Boosman A, Chang S. Processing of the initiation methionine from proteins: properties of the Escherichia coli methionine aminopeptidase and its gene structure. J Bacteriol 1987 Feb;169(2):751-757.

8. Miller CG, Strauch KL, Kukral AM, Miller JL, Wingfield PT, Mazzei GJ, et al. N-terminal methionine-specific peptidase in Salmonella typhimurium. Proc Natl Acad Sci U S A 1987 May;84(9):2718-2722.

9. Hernan RA, Hui HL, Andracki ME, Noble RW, Sligar SG, Walder JA, et al. Human hemoglobin expression in Escherichia coli: importance of optimal codon usage. Biochemistry 1992 Sep 15;31(36):8619-8628.

10. Looker D, Abbott-Brown D, Cozart P, Durfee S, Hoffman S, Mathews AJ, et al. A human recombinant haemoglobin designed for use as a blood substitute. Nature 1992 Mar 19;356(6366):258-260.

11. Hayes JK, Stanley TH, Lind GH, East K, Smith B, Kessler K. A double-blind study to evaluate the safety of recombinant human hemoglobin in surgical patients during general anesthesia. J Cardiothorac Vasc Anesth 2001 Oct;15(5):593-602.

12. Siegel JH, Fabian M, Smith JA, Costantino D. Use of recombinant hemoglobin solution in reversing lethal hemorrhagic hypovolemic oxygen debt shock. J Trauma 1997 Feb;42(2):199-212.

13. Loeb AL, McIntosh LJ, Raj NR, Longnecker DE. Resuscitation after hemorrhage using recombinant human hemoglobin (rHb1.1) in rats: effects on nitric oxide and prostanoid systems. Crit Care Med 1998 Jun;26(6):1071-1080.

14. Sillerud LO, Caprihan A, Berton N, Rosenthal GJ. Efficacy of recombinant human Hb by 31P-NMR during isovolemic total exchange transfusion. J Appl Physiol 1999 Mar;86(3):887-894.

15. Raat NJ. Effects of recombinant-hemoglobin solutions rHb2.0 and rHb1.1 on blood pressure, intestinal blood flow, and gut oxygenation in a rat model of hemorrhagic shock. J Lab Clin Med 2005 Nov;146(5):304-305.

16. Lowe KC. Blood substitutes: from chemistry to clinic. Journal of Materials Chemistry 2006;16(43):4189-4196.

17. Reed CS, Hampson R, Gordon S, Jones RT, Novy MJ, Brimhall B, et al. Erythrocytosis secondary to increased oxygen affinity of a mutant hemoglobin, hemoglobin Kempsey. Blood 1968 May;31(5):623-632.

18. Jones RT, Osgood EE, Brimhall B, Koler RD. Hemoglobin Yakina. I. Clinical and biochemical studies. J Clin Invest 1967 Nov;46(11):1840-1847.

128

19. Weickert MJ, Curry SR. Turnover of recombinant human hemoglobin in Escherichia coli occurs rapidly for insoluble and slowly for soluble globin. Arch Biochem Biophys 1997 Dec 15;348(2):337-346.

20. Baudin-Creuza V, Vasseur-Godbillon C, Griffon N, Kister J, Kiger L, Poyart C, et al. Additive effects of beta chain mutations in low oxygen affinity hemoglobin betaF41Y,K66T. J Biol Chem 1999 Sep 3;274(36):25550-25554.

21. Eich R, Li T, Lemon DD, Doherty DH, Curry S, Aitken JF, et al. Mechanism of NO-induced oxidation of myoglobin and hemoglobin. Biochemistry 1996;35:6976-6983.

22. Minning DM, Gow AJ, Bonaventura J, Braun R, Dewhirst M, Goldberg DE, et al. Ascaris haemoglobin is a nitric oxide-activated 'deoxygenase'. Nature 1999 Sep 30;401(6752):497-502.

23. Olson JS, Eich RF, Smith LP, Warren JJ, Knowles BC. Protein engineering strategies for designing more stable hemoglobin-based blood substitutes. Artificial Cells, Blood Substitutes, and Immobilization Biotechnology 1997;25:227-241.

24. Weickert MJ, Pagratis M, Curry SR, Blackmore R. Stabilization of apoglobin by low temperature increases yield of soluble recombinant hemoglobin in Escherichia coli. Appl Environ Microbiol 1997 Nov;63(11):4313-4320.

25. Yu X, Kong Y, Dore LC, Abdulmalik O, Katein AM, Zhou S, et al. An erythroid chaperone that facilitates folding of alpha-globin subunits for hemoglobin synthesis. J Clin Invest 2007 Jul;117(7):1856-1865.

26. Vasseur-Godbillon C, Hamdane D, Marden MC, Baudin-Creuza V. High-yield expression in Escherichia coli of soluble human alpha-hemoglobin complexed with its molecular chaperone. Protein Eng Des Sel 2006 Mar;19(3):91-97.

27. Domingues E, Brillet T, Vasseur C, Agier V, Marden MC, Baudin-Creuza V. Construction of a new polycistronic vector for over-expression and rapid purification of human hemoglobin. Plasmid 2009 Jan;61(1):71-77.

28. Hoffman SJ, Looker DL, Roehrich JM, Cozart PE, Durfee SL, Tedesco JL, et al. Expression of fully functional tetrameric human hemoglobin in Escherichia coli. Proc Natl Acad Sci U S A 1990 Nov;87(21):8521-8525.

129

CHAPTER 8: FUTURE WORK

8.1 – Mammalian Hemoglobin Purification

The TFF and IMAC purification techniques presented in Chapters 2 and 3 each have their own advantages and disadvantages. TFF can be used to quickly purify large amounts of hemoglobins, but many impurities remain in the final Hb product. On the other hand, IMAC produces nearly 100% pure Hb. The final purity of the TFF samples might be improved by using

TFF cartridges with different pore sizes. For example, if the pore size of the stage II filters was smaller and the pore size of the stage III filters was slightly bigger, more large MW impurities would be retained during stage II and more small MW impurities would be removed during stage III. Alternatively, if the ionic strength of the Hb solution was increased and charged TFF membranes were used, transmission of some impurities might increase. Overall, however, TFF is not an exceptionally selective purification process and will probably never be able to produce ultrapure Hb.

The IMAC technique appears to be optimized for all Hbs, with the exception of cHb. The oxygen binding behavior of the cHbA and cHbD products was extremely inconsistent. Future work should focus on the nature of these functional differences between the samples to determine if the IMAC purification procedure is directly modifying or affecting cHb structure. It would also be interesting to characterize the impurities which elute during each IMAC wash.

130

Some rather concentrated impurities were observed during the different washes that might be useful and already partially purified. For example, if one of the bands were the antioxidant enzyme catalase, that crude catalase sample might be further purified with AEX or SEC and used in HBOC synthesis or formulation.

8.2 – LtEc Purification

The biggest problem in the LtEc purification process is membrane fouling caused by the alleged clotting proteins. Since clotting proteins have only been studied in Eisenia foetida, the clotting proteins from L. terrestris must first be isolated and identified. It should be easy enough to isolate the clotting proteins by centrifugation, then PAGE analysis and LC-MS should be able to provide a MW and amino acid sequence for the clotting protein(s). Once these proteins have been identified, strategies to prevent clotting can be devised. Since they are activated by a protease, protease inhibitors might be able to inhibit clotting. Alternatively, addition of the inexpensive chelator EDTA effectively inhibits the activity of metalloproteases and may inhibit clotting. It is also possible that some membrane fouling might be due to aggregation of LtEc itself. Since divalent cations are known to contribute to LtEc stability, minimizing (but not eliminating) the concentration of Ca2+ or Mg2+ might reduce LtEc aggregation.

Different ways to extend the life of the TFF cartridges should also be investigated.

Preliminary results show that aggressive cleaning of the 0.22 m filter (flow rate = 480 mL/min for 1 hour) with 0.5 M NaOH/20 mM EDTA greatly reduces membrane fouling after several consecutive runs. In my opinion, the best way to extend the life of the filters would be to prevent the clotting problem mentioned above.

131

From a clinical standpoint, it might also be beneficial to develop an additional polishing step to remove any trace impurities or metLtEc from the TFF-purified LtEc product. Both AEX and IMAC resins have extremely low binding capacities for LtEc, perhaps because it is so large.

Therefore, it may be possible to increase the binding capacity of these resins by increasing the diameter of the resin particles. LtEc is also easily precipitated in 20 mM Zn2+ solution and resuspended in 20 mM EDTA, so protein precipitation may be a simple way to quickly polish large amounts of LtEc.

8.3 – LtEc Characterization (in vitro)

The long term stability of LtEc during storage at 4oC and/or -80oC must be evaluated to determine the shelf life of LtEc. Our past experiences have suggested that LtEc may be susceptible to damage and precipitation during freeze-thaw cycles. The precipitate may be easily removed by centrifugation or filtration, but it would be best if this damage could be prevented altogether.

The interactions between LtEc and most ligands (O2, CO, NO, NO2-) appear to be similar to HbA, but there are a few notable differences which need to be studied further. First of all, the autoxidation rate of LtEc should be much lower than HbA, but no oxidation rates have been reported yet. We have obtained the pure spectra for LtEc-O2 and metLtEc which provide the means to perform spectral deconvolution on autoxidation data for LtEc and calculate values for kOX. We have also observed that addition of small amounts of H2O2 cause LtEc to oxidize much faster than HbA. This may be due to the presence of catalase in the HbA sample, but crude LtEc has a much slower rate of oxidation in the presence of H2O2. Earthworms are known to express their own form of catalase which would be removed during the purification process. Therefore,

132 oxidation experiments in the presence of oxidizing agents should be performed to determine the susceptibility of LtEc to oxidation by these agents. Alternatively, the beneficial effects of reducing agents (ascorbic acid, cucurmin, etc.) on the oxidation rate of LtEc should also be investigated.

The interactions between LtEc-O2 and NO are also quite interesting. It appears that LtEc is able to bind NO instead of undergoing NO dioxygenation like HbA. Our preliminary results must be verified using a more advanced type of stopped flow spectroscopy which will allow us to capture full spectra of LtEc-O2 + NO mixtures on millisecond time scales. Such data would definitively determine how LtEc is interactive with NO. Electron paramagnetic resonance (EPR)

2+ 3+ may also be used to detect formation of reaction intermediates (Fe -O2, Fe -NO, etc.) which can shed light on this process.

LtEc also appears to have a reduced rate of nitrite reduction. While HbA instantly oxidizes (turns brown) in the presence of trace amounts of nitrite, LtEc remains reduced (red) even at high concentrations of nitrite. UV-Vis spectroscopy also reveals that the spectra of metLtEc with bound nitrite does not have a peak at 630 nm like metHbA-NO2-. The absence of this peak may indicate that LtEc binds nitrite in a completely different way than HbA. Crystal structures of the metLtEc-NO2- complex should be obtained to definitively determine the nature of these spectral and functional differences.

Finally, LtEc appears to oxidize in the presence of excess cyanide (CN-) while HbA forms a stable intermediate with CN-. This result is quite interesting and unexpected. Studies should be performed to elucidate the exact mechanism of this reaction and to determine the benefits or applications of this unique ability, if any.

133

8.4 – Animal Studies with LtEc

Transfusion of LtEc into hamsters has produced extremely promising results that support further animal studies. The next step will be to transfuse LtEc into higher animals to determine its efficacy and safety. For example, hamsters and other rodents are able to produce and maintain higher concentrations of antioxidants in their blood than higher mammals, so the redox behavior of LtEc in a hamster model may be completely different than other animals. We have already begun preliminary transfusion studies with guinea pigs, whose blood has much less antioxidant activity than hamsters and is more similar to human blood. In the long run, LtEc will hopefully also be transfused in higher mammals (dogs, pigs, monkeys, etc.) and eventually enter clinical trials to determine its safety and efficacy in humans.

It may also be interesting to test erythrocruorins from other species. For example, E. foetida is commercially available and we have previously purified Ec from this worm as well. It would be interesting to determine if the different Ecs behave differently in vivo.

8.5 – Expression and Purification of Recombinant Globin Proteins

The recombinant globin expression system discussed in Chapter 7 is attractive because it provides a relatively inexpensive way to express sufficient amounts of recombinant HbA subunits and/or mutants for characterization. The bacteria may be grown in shaker flasks instead of large fermenters with undefined media, making the technique accessible to most labs. However, optimization of the expression process must be done. Fermenters may be used to increase cell density and possibly the target protein expression quite easily.

134

However, the two biggest problems associated with this technique are the chitin beads and the slow cleavage rate of the intein. The chitin beads are expensive ($200/100 mL) and degrade over time, especially at the room temperature required for intein cleavage. If a cheaper and more stable affinity resin could be made, then the affinity tag should be changed from a chitin binding domain to another type of binding domain to minimize process costs and prevent contamination of the fusion tag in the final purified sample.

The slow cleavage rate of the intein increases the metHb level of the subunits, degradation of the chitin resin, and reduces the final yield of the target protein. Several other inteins with faster cleavage rates have been discovered, but they usually cleave on both their C- and N-termini or tend to be rather large. Nonetheless, directed evolution of the intein to increase cleavage rates needs to be done to make this purification process viable.

The PAGE results in Chapter 7 indicate that most of the inteins have cleaved after 7 days, but the beads retain a deep red color when they are washed with column buffer. HbA is known to react with glucose to form a stable product called hemoglobin A1C. Since chitin is a sugar polymer, it is possible that inteins may be cleaving but the target proteins are covalently binding to the chitin beads. However, the red color may be easily removed by SDS and NaOH, indicating that LtEc must not be covalently bound to the beads. Therefore, it may be possible to increase the final yield of the target protein by finding a way to disrupt any interactions between the target proteins and the beads (i.e. using higher ionic strength to disrupt electrostatic interactions).

It is important to mention that I have been trying to use this system to express and purify LtEc globin subunits for over a year now. While I am able to produce high amounts of

135 fusion protein and get dark red chitin beads, the rate of intein cleavage is practically zero. Even after 3 months, the chitin beads retain their red color while the mobile phase is clear. I have been able to isolate trace amounts of each subunit, but nowhere near enough for characterization. Any of the changes suggested above (using a different affinity domain or intein) may be the solution to this problem. This issue should be given great priority since this system would be a great way to make mutations in LtEc subunits to elucidate the unique ways in which it resists oxidation and interacts with NO.

136

Comprehensive Bibliography

Chapter 1