Exploring Oxidative Reactions in Hemoglobin Variants Using Mass Spectrometry: Lessons for Engineering Oxidatively Stable Oxygen Therapeutics

Abstract

Significance:

Worldwide demand has driven the development of hemoglobin (Hb)-based oxygen carriers (HBOCs) as potential acellular oxygen therapeutics. HBOCs have the potential to provide an oxygen bridge to patients and minimize current problems associated with supply and storage of donated blood. However, to date, safety and efficacy issues have hampered the approval of viable HBOCs in the United States. These previous efforts have underscored the need for a better molecular understanding of toxicity to design safe and oxidatively stable HBOCs.

Recent Advances:

High-resolution accurate mass (HRAM) mass spectrometry (MS) has recently become a versatile tool in characterizing oxidative post-translational modifications that occur in Hb. When integrated with other analytical techniques, HRAM data have been invaluable in providing mechanistic insight into the extent of oxidative modification by quantifying oxidation in amino acids near the reactive heme or at specific “oxidative hotspots.”

Critical Issues:

In addition to providing a deeper understanding of Hb oxidative toxicity, HRAM MS studies are currently being used toward developing suitable HBOCs using a “two-prong” strategy that involves (i) understanding the mechanism of Hb toxicity by evaluating mutant Hbs identified in patients with hemoglobinopathies and (ii) utilizing this information toward designing against (or for) these reactions in acellular oxygen therapeutics that will result in oxidatively stable protein.

Future Directions:

Future HRAM studies are aimed at fully characterizing engineered candidate HBOCs to determine the most oxidatively stable protein while retaining oxygen carrying function in vivo. Antioxid. Redox Signal. 26, 777–793.

Introduction

Civilian (and military) blood transfusions are dependent, worldwide, on an extraordinary number of donors (approximately 92 million annually) who provide whole blood and packed red blood cells to roughly 15% of the world's population. While these transfusions are lifesaving, US hospitals consistently report blood shortages; this situation is much more severe in underdeveloped countries where blood donation rates are less than the United States (WHO summary report on Global Database on Blood Safety [GDBS] 2011) (102a). Another major issue related to this shortage is the imbalance between developing and developed countries in the level of access to safe blood. For example, 65% of the 92 million donations are collected in the 10 highest income countries. On the contrary, 43 countries in the poorer African Region, where blood-borne illness is more prevalent, collect less than 5% of global donation; this region makes up approximately 12% of the global population (GDBS) (102a). Global blood safety concerns impacting blood supply (most profound in developing countries) include the potential for costly diagnostics when processing blood obtained from patients with infectious disease (i.e., HIV) (6, 12, 29, 44).

The demand for blood donation alternatives has driven the commercial development of Hb-based oxygen carriers (HBOCs) or “blood substitutes” over the past 30 years (43). Initially, there was a high level of optimism for HBOCs to replace blood transfusion requirements. The benefits of such an accomplishment would potentially eliminate the problems associated with blood availability, reduce the risk of disease transmission, and improve oxygen delivery and need for long-term blood storage (34, 66, 88).

Despite these advantages, no viable HBOCs have been approved in the United States. The primary reason for this stems from multiple reports of adverse cardiovascular and cerebrovascular events associated with hypertension in individuals transfused with these products (79, 86). In hindsight, this should not be surprising since HBOCs (studied to date) are specifically acellular Hb-derived proteins; because they are not “actual blood” they do not perform other blood functions such as nutrient transport, immune response, and coagulation. Preliminary evaluation using animal models suggests that some HBOCs do oxygenate tissues shortly after infusion (3, 15, 86). Current-generation HBOCs are intended therefore to complement standard blood transfusions by acting as an “oxygen bridge” in extreme life-threatening situations, such as battlefield trauma and in surgical settings and/or where patients refuse blood transfusions due to religious reasons (3).

Historical Overview of HBOC Design Strategies

The first attempt to produce the so-called blood substitutes involved transfusion of stroma-free, unmodified acellular Hb into humans, which resulted in renal dysfunction, abdominal pain, and hypertension (64, 66, 81). Additional analysis of these early studies indicated that diluting Hb into blood plasma resulted in the accumulation of Hb dimers due to rapid tetrameric dissociation; dimeric Hb readily passed through the kidneys resulting in rapid clearance, which can be damaging to kidneys. In addition to this pathology, Hb dimers were characterized to be unstable and therefore susceptible to oxidation, heme loss, and precipitation; these events ultimately led to oxidative stress, blockage of kidney tubules, and ultimately renal failure. Finally, it was also observed that these acellular Hbs scavenge endothelial NO (a vessel dilator) thus negatively impacting vasoregulation and possess oxygen binding behavior vastly different than what occurs in the protective environment of red blood cells (5, 21, 45, 64).

To address these drawbacks, several different biochemical and recombinant strategies (Fig. 1) have been attempted with varying success. Among these included chemically crosslinking human- or bovine-derived Hb using a site-specific bifunctional crosslinker such as di-bromo bis-fumarate (DBBF/also known as diaspirin crosslinked Hb [DCLHb]), polymerization/conjugation with polyethylene glycol polymers (PEGylation), and genetically engineering α-crosslinked recombinant Hb (expressed in bacteria or yeast host systems) (3, 16, 36, 40, 47, 73, 97, 98). These modifications aim at either covalently linking dimers in a stable tetrameric form or in the case of polymerization/conjugation (such as PEGylation) to increase the surface area or hydrodynamic volume; all three modification strategies have been shown to increase intravascular retention and substantially reduce renal filtration of Hb dimers (4, 17, 25, 80).

FIG. 1.

Recombinant and biochemical strategies for the production of Hb-based oxygen carriers as blood substitutes are illustrated. (A) Tetrameric stabilization accomplished by joining two α/β heterodimers through intramolecular crosslinking between two α or two β subunits by either genetically engineering crosslinked recombinant Hb or a site-specific bifunctional crosslinker such as DBBF/DCLHb or NFPLP. (B) Effective molecular weight and surface area of Hb are increased by conjugating to polyethylene glycol (PEG-2000 or 5000). (C) Hb polymerization to create molecular weights significantly greater than native tetrameric (64 kDa) is produced when Hbs are reacted with polyfunctional crosslinking agents (e.g., glutaraldehyde, O-raffinose, O-adenosine). NFPLP, 2-nor-2-formyl pyridoxal-5′-phophate; DBBF, di-bromo bis-fumarate; DCLHb, diaspirin crosslinked Hb. To see this illustration in color, the reader is referred to the web version of this article at www.liebertpub.com/ars

To date, several different commercially developed HBOCs (using anyone of the methods in Fig. 1) have been tested in animal and human subjects. These Hbs, in general, demonstrated reasonable preclinical safety and efficacy in healthy volunteers in Phase 1 trials. However, as clinical development advanced into later stages, an undesirable safety and efficacy profile became clearer in the patient population studied leading to the termination of the clinical development of many first-generation HBOCs (8, 35, 52, 59, 73, 75, 83, 87). These HBOCs failed clinically, in large part, due to serious adverse events associated with their infusion in humans attributed to nitric oxide (NO) scavenging by extracellular Hb, heme-mediated oxidative reactions, and oxygen oversupply (100).

The Basis of HBOC Toxicity

Several recent in vitro and animal studies showed that Hb oxidation plays a key role in the genesis of many adverse events associated with Hb (chemically and nonchemically modified) (16, 64). Outside the RBC, acellular HBOCs are oxidatively toxic due to heme redox reactivity, which is ironically essential for the oxygen (O₂) carrying function. On O₂ binding, ferrous heme iron (HbFe²⁺-O₂) spontaneously auto-oxidizes to form a non-O₂ carrying ferric heme iron (HbFe³⁺ aka metHb). During this naturally occurring process, reactive oxygen species such as superoxide (O₂·⁻) and hydrogen peroxide (H₂O₂) are formed, which (in the case of H₂O₂) drives a catalytic pseudoperoxidase cycle that includes the transformation of HbFe²⁺-O₂ to a transient reactive oxyferryl Hb (HbFe⁴⁺ = O) intermediate. The HbFe⁴⁺ = O species autoreduces back to HbFe³ and in the presence of additional H₂O₂, the ferryl Fe⁴⁺ iron is regenerated. However, if the reaction with H₂O₂ begins with the HbFe³⁺, a ferryl protein cation radical (ċHbFe⁴⁺ = O) is formed (Equation 1 –3) (76). These oxidative processes are well controlled within the RBCs where Hb is continuously reduced to the functional ferrous form by several reductive enzymes (catalase, superoxide dismutase, methemoglobin reductase, etc.). However, when left unchecked (i.e., in noncompartmentalized environment), acellular Hb becomes oxidants that damage themselves and nearby biological molecules. These reactions are summarized by the following equations: \documentclass{aastex}\usepackage{amsbsy}\usepackage{amsfonts}\usepackage{amssymb}\usepackage{bm}\usepackage{mathrsfs}\usepackage{pifont}\usepackage{stmaryrd}\usepackage{textcomp}\usepackage{portland, xspace}\usepackage{amsmath, amsxtra}\pagestyle{empty}\DeclareMathSizes{10}{9}{7}{6}\begin{document} \begin{align*}{ \rm{HbF}}{{ \rm{e}}^{{ \rm{2 + }}}}{{ \rm{O}}_{ \rm{2}}}{ \rm{ + }}{{ \rm{H}}_{ \rm{2}}}{{ \rm{O}}_{ \rm{2}}} \to { \rm{HbF}}{{ \rm{e}}^{{ \rm{4 + }}}}{ \rm{ = O + }}{{ \rm{H}}_{ \rm{2}}}{ \rm{O + }}{{ \rm{O}}_{ \rm{2}}} \tag{1} \end{align*} \end{document} \documentclass{aastex}\usepackage{amsbsy}\usepackage{amsfonts}\usepackage{amssymb}\usepackage{bm}\usepackage{mathrsfs}\usepackage{pifont}\usepackage{stmaryrd}\usepackage{textcomp}\usepackage{portland, xspace}\usepackage{amsmath, amsxtra}\pagestyle{empty}\DeclareMathSizes{10}{9}{7}{6}\begin{document} \begin{align*}{ \rm{HbF}}{{ \rm{e}}^{{ \rm{4 + }}}}{ \rm{ = O}}^\cdot { \rm{ + }}{{ \rm{H}}_{ \rm{2}}}{{ \rm{O}}_{ \rm{2}}} \to { \rm{HbF}}{{ \rm{e}}^{{ \rm{3 + }}}}{ \rm{ + }}{{ \rm{O}}_{ \rm{2}}}{ ^{\cdot { \rm{-}}}}{ \rm{ + }}{{ \rm{H}}_{ \rm{2}}}{ \rm{O}} \tag{2} \end{align*} \end{document} \documentclass{aastex}\usepackage{amsbsy}\usepackage{amsfonts}\usepackage{amssymb}\usepackage{bm}\usepackage{mathrsfs}\usepackage{pifont}\usepackage{stmaryrd}\usepackage{textcomp}\usepackage{portland, xspace}\usepackage{amsmath, amsxtra}\pagestyle{empty}\DeclareMathSizes{10}{9}{7}{6}\begin{document} \begin{align*}{ \rm{HbF}}{{ \rm{e}}^{{ \rm{3 + }}}}{ \rm{ + }}{{ \rm{H}}_{ \rm{2}}}{{ \rm{O}}_{ \rm{2}}} \to \cdot { \rm{HbF}}{{ \rm{e}}^{{ \rm{4 + }}}}{ \rm{ = O + }}{{ \rm{H}}_{ \rm{2}}}{ \rm{O}} \tag{3} \end{align*} \end{document}

This oxidative toxicity is primarily due to the high midpoint redox potentials (E°_1/2∼1.0 V) of the ferryl heme and its associated protein cation radical (76). These internal reactions result in irreversible oxidation of amino acids, areas in the protein known as the “oxidative hotspots” (particularly the β Cys93 side chain) that cause the collapse of β subunits, Hb degradation, and the ultimate release of heme (39). Recent animal studies from sickle cell mice suggest that this resulting heme moiety gives rise to serious consequences, including endothelial cell activation and vaso-occlusion (9, 14). These studies have also shown that heme released from oxidizing Hb binds to Toll-like receptors (TLR4) that trigger a cascade of inflammatory reactions (9). Consequently, heme has been recognized as a “damage-associated molecular pattern” (DAMP) molecule on par with molecules such as lipopolysaccharides and other inflammatory stimulants (28). Clearly, these internal oxidative pathways (leading to heme release) would have to be curtailed to minimize these inflammatory reactions.

Two-Prong Strategy Toward Engineering Oxidatively Stable HBOCs

The challenge for HBOC developers is to design stable Hbs with optimal biochemical properties to reduce these oxidative toxicities. These include (i) enhancing globin stability for storage by reducing the rate of auto-oxidation and heme loss and (ii) adjusting oxygen affinity; this optimization will also reduce the oxidant potency of HBOCs. To date, one of the most versatile approaches for making candidate HBOCs has been a recombinant strategy which involves (in one case) genetically linked in-frame tandem copies of the α subunit gene to produce a di-α gene with a single glycine linker between the C-terminal end of one α subunit and the N-terminal end of the other in a low oxygen affinity variant (Presbyterian βN108K). This di-α gene is then coexpressed with V1 M β-subunit genes in Escherichia coli to produce the recombinant Hb rHb1.1 (47). Crystal structural data of rHb1.1 have indicated that the glycine linker does not perturb the overall tertiary, quaternary, or even local structure of Hb (Fig. 2) (13). The major advantage of using this recombinant strategy is that one can potentially engineer HBOCs with beneficial amino acid substitutions using mutagenesis strategies (98). To date, rational mutagenesis approaches developed from a “two-prong strategy” (as opposed to random mutagenesis where screening for multiple properties is not feasible) have provided some promising results (98). However, well-documented engineering problems associated with recombinant proteins, such as correct protein folding and heme orientation within the heme pocket, present additional hurdles that protein engineers have to overcome (25).

FIG. 2.

Graphic representation of a typical genetically engineered Hb (Optro (rHb 1.1 deoxy). This recombinantly expressed (in Escherichia coli) Hb-based oxygen carrier (Hb Presbyterian [βN108K], blue) is intramolecularly crosslinked via an αα glycine linker (red color). Highlighted (blue) ribbon represents the two linked amino (Val1 and Arg141) acids. Distances were measured and the theoretical model was generated using the PyMOL Molecular Graphics System distance measuring utility and mutagenesis wizard, respectively. To see this illustration in color, the reader is referred to the web version of this article at www.liebertpub.com/ars

In the search for favorable amino acid substitutions, our laboratory has relied on studying Hb variants identified in patients with hemoglobinopathies. Naturally occurring human Hb mutants potentially result in protein resistance to oxidative challenges, while others develop into a full circulatory disorder; in this regard, these mutants can be viewed as “Experiments in Nature” (27, 95). Exploring subunit-specific oxidative stability within these mutants may therefore lead to re-engineering of oxidatively stable HBOCs based on a combination of evolutionary stable mutations and rationally designed prototypes (Table 1 for summary of Hb mutants reviewed here). Our “two-prong” strategy involves (i) understanding the mechanism of Hb toxicity by evaluating mutant Hbs identified in patients with hemoglobinopathies and (ii) utilizing this information toward engineering acellular blood substitutes that are more oxidatively stable than native HbA. The implementation of this two-prong strategy for studying Hb variants will be highlighted in the section describing quantitative mass spectrometry for characterizing oxidative post-translational modifications (PTMs).

Table 1.

Select Human Hemoglobin Mutants and Relevance to Hemoglobin-Based Oxygen Carriers

Hb	Mutation position	Substitution	Clinical features	Biochemical properties	Relevance to HBOCs
Providence (1)	β82 (EF6)	Lys \documentclass{aastex}\usepackage{amsbsy}\usepackage{amsfonts}\usepackage{amssymb}\usepackage{bm}\usepackage{mathrsfs}\usepackage{pifont}\usepackage{stmaryrd}\usepackage{textcomp}\usepackage{portland, xspace}\usepackage{amsmath, amsxtra}\pagestyle{empty}\DeclareMathSizes{10}{9}{7}{6}\begin{document}$$\rightarrow$$ \end{document} Asn \documentclass{aastex}\usepackage{amsbsy}\usepackage{amsfonts}\usepackage{amssymb}\usepackage{bm}\usepackage{mathrsfs}\usepackage{pifont}\usepackage{stmaryrd}\usepackage{textcomp}\usepackage{portland, xspace}\usepackage{amsmath, amsxtra}\pagestyle{empty}\DeclareMathSizes{10}{9}{7}{6}\begin{document}$$\rightarrow$$ \end{document} Asp	Erythrocytosis	Oxidative stability	Hb Providence-based HBOC
Toms River (89)	γ67 (E11)	Val \documentclass{aastex}\usepackage{amsbsy}\usepackage{amsfonts}\usepackage{amssymb}\usepackage{bm}\usepackage{mathrsfs}\usepackage{pifont}\usepackage{stmaryrd}\usepackage{textcomp}\usepackage{portland, xspace}\usepackage{amsmath, amsxtra}\pagestyle{empty}\DeclareMathSizes{10}{9}{7}{6}\begin{document}$$\rightarrow$$ \end{document} Met \documentclass{aastex}\usepackage{amsbsy}\usepackage{amsfonts}\usepackage{amssymb}\usepackage{bm}\usepackage{mathrsfs}\usepackage{pifont}\usepackage{stmaryrd}\usepackage{textcomp}\usepackage{portland, xspace}\usepackage{amsmath, amsxtra}\pagestyle{empty}\DeclareMathSizes{10}{9}{7}{6}\begin{document}$$\rightarrow$$ \end{document} Asp	Anemia; cyanosis	Oxidative instability; Met \documentclass{aastex}\usepackage{amsbsy}\usepackage{amsfonts}\usepackage{amssymb}\usepackage{bm}\usepackage{mathrsfs}\usepackage{pifont}\usepackage{stmaryrd}\usepackage{textcomp}\usepackage{portland, xspace}\usepackage{amsmath, amsxtra}\pagestyle{empty}\DeclareMathSizes{10}{9}{7}{6}\begin{document}$$\rightarrow$$ \end{document} Asp (PTM)	Comparison of α, β, and γ to probe oxidative stability
Bristol-Alesha (89)	β67 (E11)	Val \documentclass{aastex}\usepackage{amsbsy}\usepackage{amsfonts}\usepackage{amssymb}\usepackage{bm}\usepackage{mathrsfs}\usepackage{pifont}\usepackage{stmaryrd}\usepackage{textcomp}\usepackage{portland, xspace}\usepackage{amsmath, amsxtra}\pagestyle{empty}\DeclareMathSizes{10}{9}{7}{6}\begin{document}$$\rightarrow$$ \end{document} Met \documentclass{aastex}\usepackage{amsbsy}\usepackage{amsfonts}\usepackage{amssymb}\usepackage{bm}\usepackage{mathrsfs}\usepackage{pifont}\usepackage{stmaryrd}\usepackage{textcomp}\usepackage{portland, xspace}\usepackage{amsmath, amsxtra}\pagestyle{empty}\DeclareMathSizes{10}{9}{7}{6}\begin{document}$$\rightarrow$$ \end{document} Asp	Hemolytic anemia; reticulocytosis	Oxidative instability; Met \documentclass{aastex}\usepackage{amsbsy}\usepackage{amsfonts}\usepackage{amssymb}\usepackage{bm}\usepackage{mathrsfs}\usepackage{pifont}\usepackage{stmaryrd}\usepackage{textcomp}\usepackage{portland, xspace}\usepackage{amsmath, amsxtra}\pagestyle{empty}\DeclareMathSizes{10}{9}{7}{6}\begin{document}$$\rightarrow$$ \end{document} Asp (PTM)	Comparison of α, β, and γ to probe oxidative and structural stability
Evans (89)	α62 (E11)	Val \documentclass{aastex}\usepackage{amsbsy}\usepackage{amsfonts}\usepackage{amssymb}\usepackage{bm}\usepackage{mathrsfs}\usepackage{pifont}\usepackage{stmaryrd}\usepackage{textcomp}\usepackage{portland, xspace}\usepackage{amsmath, amsxtra}\pagestyle{empty}\DeclareMathSizes{10}{9}{7}{6}\begin{document}$$\rightarrow$$ \end{document} Met	Hemolytic anemia	No Met \documentclass{aastex}\usepackage{amsbsy}\usepackage{amsfonts}\usepackage{amssymb}\usepackage{bm}\usepackage{mathrsfs}\usepackage{pifont}\usepackage{stmaryrd}\usepackage{textcomp}\usepackage{portland, xspace}\usepackage{amsmath, amsxtra}\pagestyle{empty}\DeclareMathSizes{10}{9}{7}{6}\begin{document}$$\rightarrow$$ \end{document} Asp (PTM)	Investigate unusual α stability
HbS (42)	β6 (E1)	Glu \documentclass{aastex}\usepackage{amsbsy}\usepackage{amsfonts}\usepackage{amssymb}\usepackage{bm}\usepackage{mathrsfs}\usepackage{pifont}\usepackage{stmaryrd}\usepackage{textcomp}\usepackage{portland, xspace}\usepackage{amsmath, amsxtra}\pagestyle{empty}\DeclareMathSizes{10}{9}{7}{6}\begin{document}$$\rightarrow$$ \end{document} Val	Sickle cell disease; vaso-occlusion hemolytic anemia; and so on	Oxidative instability; deoxy HbS polymerizes	Investigate oxidative instability
HbE (90)	B26 (E8)	Glu \documentclass{aastex}\usepackage{amsbsy}\usepackage{amsfonts}\usepackage{amssymb}\usepackage{bm}\usepackage{mathrsfs}\usepackage{pifont}\usepackage{stmaryrd}\usepackage{textcomp}\usepackage{portland, xspace}\usepackage{amsmath, amsxtra}\pagestyle{empty}\DeclareMathSizes{10}{9}{7}{6}\begin{document}$$\rightarrow$$ \end{document} Lys	Minimal microcytosis; anemia: severe when coupled to β-thalassemia	Oxidative instability accelerated heme loss?	Investigate oxidative instability and to understand impact on β-thalassemia

Approaches to Characterizing HBOCs

Several analytical approaches are used to characterize Hb variants and candidate HBOCs when implementing this strategy. One example of the classic methods for assaying auto-oxidation and denaturation involves incubating rHb variants in an air equilibrated buffer and spectrophotometrically following auto-oxidation and precipitation over time and/or by chemically inducing oxidation (at pH 7 at 37°C) (11, 39, 63, 84). This review will focus on the use of high-resolution accurate mass (HRAM) mass spectrometry (MS) for characterizing oxidative PTMs that occur in Hb. HRAM analysis is new to the field of characterizing Hb oxidative stability and when integrated with other more familiar methods (i.e., spectrophotometric, stopped-flow spectrophotometric assays) is invaluable in providing mechanistic insight into the extent of oxidative toxicity by quantifying oxidation in amino acids near the reactive heme or at specific oxidative hotspots. A basic introduction to mass spectrometry will be presented followed by examples of HRAM characterization of Hb variants. This review will also discuss promising results for a candidate HBOC designed from principles learned by studying Hb variants.

Introduction to HRAM Mass Spectrometry for the Study of Oxidative Changes and PTMs in Hemoglobins

The postgenomic era has had a huge impact on how biologists and biochemists study proteins. Scientists are now using high-throughput strategies that are versatile for characterizing enriched proteins, protein complexes, and whole proteomes, which not only identify and quantify proteins but can also identify a wide variety of PTMs, under a variety of experimental conditions (54). The most commonly used and extremely powerful technique for the high-throughput characterization of proteins, peptides, and PTMs is liquid chromatography coupled online with mass spectrometry (LC-MS) (38, 54, 85). By combining the power of bioinformatics, chromatography, and HRAM mass spectrometry, LC-MS is a versatile tool capable of comprehensive protein identification and in-depth characterization of PTM changes from complex biological samples (55).

Major Components of Mass Spectrometer

Mass spectrometers (MS) are analytical instruments that measure ionized analytes in the gas phase. The basic setup for a mass spectrometer consists of an ion source, a mass analyzer that measures the mass-to-charge ratio (m/z) of gas-phase ions, and a detector that registers the number of ions at each m/z value (Fig. 3A). For electrospray ionization (ESI-MS), this MS configuration is generally coupled online to an ultra-performance or high-performance liquid chromatography system (UPLC and HPLC, respectively), which separates the peptide analyte; as peptides are eluted from the column (see section “HPLC Online Peptide Separation”), they are ionized by the electrospray source and then directly emitted into the mass spectrometer (at the spray needle or emitter tip).

FIG. 3.

Basic setup for a mass spectrometer and mass analyzers currently used with ESI-based mass spectrometers. (A) The basic setup for a mass spectrometer consists of an ion source, a mass analyzer that measures the mass-to-charge ratio (m/z) of gas-phase ions, and a detector that registers the number of ions at each m/z value. Most MS configurations are generally coupled online to a UPLC or HPLC system. In this illustration, ESI is shown where the ESI source located at the spray needle. (B) Commonly used mass spectrometer configurations (for protein/PTM identifications) with one or more of these analyzers are illustrated. (1) The ion-trap first captures ions (for a short time interval) and subjects them to either MS or fragmentation analysis (MS/MS). (2) The Q-TOF mass spectrometer typically involves ion selection via a quadrupole mass filter (Q) followed by fragmentation in a collisional cell (q2); the fragment ions then travel through the TOF analyzer at a rate proportional to their m/z ratio. (3) The FT-ICR analyzer captures ions under high vacuum in a high magnetic field; the ions orbit within the cyclotron at a frequency proportional to their m/z ratio. (4) The Q-Exactive mass spectrometer involves ion selection with a quadruple mass filter, ion capture in the Orbitrap analyzer followed by fragmentation in a HCD cell, and analysis in the Orbitrap. ESI, electrospray ionization; FT-ICR, Fourier transform ion cyclotron; HPLC, high-performance liquid chromatography; PTM, post-translational modification; TOF, time-of-flight; UPLC, ultra-performance liquid chromatography. To see this illustration in color, the reader is referred to the web version of this article at www.liebertpub.com/ars

Electrospray and Matrix-Assisted Laser Desorption/Ionization

The most common ion sources for peptide and protein identification utilize either ESI, which is displayed in Figure 3A, or matrix-assisted laser desorption/ionization (MALDI). These soft ionization methods (developed in the 90s and early 2000s) revolutionized the field of mass spectrometry as they allowed for the nondestructive ionization of macromolecular structures (i.e., peptides or proteins) (23, 41). Before their development, hard ionization (i.e., electron and chemical ionization) limited the utility of mass spectrometry to the analysis of small volatile molecules stable at temperatures required for vaporization (99). However, ESI and MALDI ionization convert macromolecular structures (including peptides, proteins, lipids, carbohydrates, and nucleic acids) into ions with minimal chemical degradation (23, 41, 99).

Specifically, ESI involves applying high voltage, via ionization source at spray needle (Fig. 3A), to a liquid aqueous/organic/acid mixture to create an aerosol of droplets that are then emitted into the mass spectrometer through a heated capillary carrying a potential difference in atmospheric pressure (due to vacuum). Immediately after emission, the droplet surface area decreases due to solvent evaporation; as the droplet surface area decreases, the positively charged peptide ions eventually experience a columbic explosion (due to charge repulsion) resulting in desorption of solvated ions (67). MALDI, in contrast, desorbs (via sublimation) and ionizes the analyte from a crystalline matrix using a directed laser pulse. Traditionally, MALDI is used to analyze relatively simple peptide mixtures. The application of ESI for studying proteins far exceeds the use of MALDI because ESI ionizes analytes out of a solution and as a result is more amenable to liquid-based chromatographic separation tools. The primary focus of this review will, therefore, be on MS methods that utilize ESI coupled online with liquid chromatography (LC-MS).

Mass Analyzers

The mass analyzer is a central component for HRAM MS methods. In the context of peptide/protein and PTM identifications, its key parameters are sensitivity, dynamic range, duty cycle, resolution, mass accuracy, and the ability to generate sequence information by virtue of fragmenting peptides (tandem or MS/MS spectra) (2, 70). There are four major types of mass analyzers currently used with ESI-based mass spectrometers. These include the linear ion trap, time-of-flight (TOF), Fourier transform ion cyclotron (FT-ICR), and the Orbitrap analyzers. All four (uniquely different in design and performance) can be the principle analyzer but are generally featured in tandem as part of hybrid mass spectrometers that combine the strengths of each analyzer.

Commonly used mass spectrometer configurations (for protein/PTM identifications) with one or more of these analyzers are illustrated in Figure 3B. Briefly, the linear ion-trap (Fig. 3B-1) first captures ions (for a short time interval) and subjects them to either MS or fragmentation analysis (MS/MS) via collisional-induced dissociation (CID) (32). Linear ion traps are a common component of HRAM hybrid mass spectrometers such as the LTQ-Orbitrap (50, 51, 82). TOF analyzers (traditionally coupled with MALDI ion sources) are most often used with the HRAM hybrid Q-TOF (quadrupole-time-of-flight). A typical configuration for a Q-TOF mass spectrometer (Fig. 3B-2) involves ion selection via a quadrupole mass filter (Q) followed by fragmentation in a collisional cell (q2); the fragment ions then travel through the TOF analyzer at a rate proportional to their m/z ratio (19, 46). The FT-ICR analyzer captures ions under high vacuum in a high magnetic field (similar to an NMR instrument); the ions orbit within the cyclotron at a frequency proportional to their m/z ratio (56, 57). The FT-ICR mass spectrometer (Fig. 3B-3) has the best resolving power for analyzing intact proteins. Finally, the Orbitrap analyzer (type of ion trap analyzer) consists of an outer barrel-like electrode and a coaxial inner spindle-like electrode that trap ions in an orbital motion around the spindle (49). Two major Orbitrap configurations include coupling the Orbitrap to a linear ion trap analyzer (LTQ-Orbitrap) or quadrupole mass filter (Q-Exactive) (50). A typical LTQ-Orbitrap configuration involves ion capture in the linear trap, analysis in the Orbitrap analyzer, and then ion fragmentation (MS/MS) via CID in the linear trap (50). The Q-Exactive mass spectrometer (Fig. 3B-4) involves ion selection (and isolation) with a quadruple mass filter, ion capture in the Orbitrap analyzer followed by fragmentation in an HCD cell, and analysis in the Orbitrap (26).

Step 1: LC/MS/MS Analysis: The Bottom-Up Approach (LC/MS/MS) for Identifying Proteins and PTMs

Two HRAM methods are the frequently applied “bottom-up” (also known as shotgun proteomics) and the less widely used “top-down” strategies. An essential component of the quantitative methods described in this review, the bottom-up strategy involves enzymatic digestion of intact proteins into peptides that are then analyzed by LC-MS. The resulting peptides are fragmented by CID or HCD (higher energy C-trap dissociation) to generate MS/MS spectra that are then analyzed by database search algorithms (such as Mascot/SEQUEST) for sequence information (22, 71, 103). The ion current from reproducibly matched peptides (or fragment ions) can then be used for label-free quantification of Hb oxidation (discussed in Step 2). Key features of the bottom-up strategy (Fig. 4) are discussed in the following sections. The alternative top-down strategy involves direct analysis and/or fragmentation of intact proteins by the mass spectrometer to potentially identify protein isoforms (from complex mixtures) resulting from translational start and stop sites, mRNA splice variants, and PTMs of expressed gene products.

FIG. 4.

Schematic diagram of the bottom-up strategy (LC/MS/MS analysis). This strategy involves enzymatic digestion (top left corner) of intact proteins into peptides that are then analyzed by the mass spectrometer (top right corner). The resulting peptides are fragmented by CID (or HCD) (lower right corner) to generate MS/MS spectra that are then analyzed by database search algorithms (such as Mascot/SEQUEST) (lower left corner) for sequence information, in this case the Hb β gene sequence. To see this illustration in color, the reader is referred to the web version of this article at www.liebertpub.com/ars

A major advantage of the bottom-up method is that peptide MS/MS spectra are much more efficiently analyzed than top-down fragmentation spectra of intact proteins; to date, the bioinformatic software required to interpret top-down spectra lags far behind algorithms designed for interpreting peptide MS/MS spectra. The bottom-up strategy is therefore much more effective at systematically identifying a large number of proteins (via identified peptides) and their corresponding PTMs from complex biological samples. There are two different methods for conducting bottom-up MS proteomic measurement. The first implementation uses one-dimensional (1D) or two-dimensional (2D) gel electrophoresis followed by in-gel digestion and either peptide mass fingerprinting or tandem mass spectrometry (1D LC MS/MS) to identify peptides that are matched to their corresponding proteins. The second more high-throughput approach involves in solution direct proteolytic digestion of complex mixtures (enriched protein, protein complexes, or whole proteomes) followed by 1D or 2D liquid chromatography coupled online with ESI MS/MS. The latter 1D LC MS/MS implementation is more commonly used in the quantitative examples described later in this report.

Enzymatic Digestion of Intact Proteins into Peptides

As stated above, peptide MS/MS spectra are more efficiently analyzed (relative to fragmentation spectra of intact proteins) for PTM identifications. Before LC-MS analysis, it is therefore critical to optimally digest protein samples toward maximizing the number of identifiable peptide isoforms (a peptide can exist unmodified or contain PTMs) in a given sample (Fig. 4). Important to MS/MS fragmentation is the use of site-specific proteases during the enzymatic digestion of a protein sample. Although a variety of different digestive enzymes are available for generating peptides, trypsin is the most commonly used protease because it cleaves C-terminally at lysine and arginine with a remarkable degree of specificity. The frequency of lysine and arginine in proteins generally results in a decent number of peptides that are the ideal size for gas-phase fragmentation (between 5 and 30 amino acids) (92). Furthermore, trypsin cleavage results in peptides with positive C-terminally protonated ends, beneficial for CID fragmentation; trypsin digestion results in a large frequency of +2 charged peptides (and +3 charge) that fragment well to produce product ion spectra that are very informative. These characteristics of trypsin generally increase the number of identifiable peptides (reflected by the primary sequence covered) per LC-MS analysis.

While LC-MS analysis of trypsin digestions most often reveal oxidative hotspots containing Hb peptides (see section “Database Searches of MS/MS Fragmentation Spectra”), alternative digestion strategies are commonly utilized for the purpose of mapping PTM identifications. One such strategy involves digestion with multiple enzymes; this approach increases the likelihood of generating additional peptides that are missed by one of the other proteases and increases the confidence of PTM identification when the same PTM is identified by overlapping peptides generating from more than one protease (48). In scenarios where the protein of interest has a low frequency of lysine or arginine residues (i.e., the transmembrane spanning domains), other cleavage strategies can be used, including the use of cyanogen bromide cleavage or the use of other proteases such chymotrypsin (cleaves C-terminally at aromatic amino acid residues) or pepsin (cleaves nonspecifically) (92).

HPLC Online Peptide Separation

Due to the complex nature of biological samples, online chromatographic separation is important for improving sensitivity and dynamic range for bottom-up peptide identifications. Dynamic range here refers to the ability to detect low-abundance peptides/proteins in the presence of high-abundance peptides/proteins. The goal of HPLC separation (before MS and MS/MS analyses) is to therefore simplify the mixture of peptides presented to the mass spectrometer at any given point in time. Indeed, direct infusion (without online chromatography) of a peptide mixture into the mass spectrometer would result in limited detection of only the most abundant peptides and those with the highest ionization efficiency (53). The separation of this complex mixture of peptides via HPLC allows for the mass spectrometer to interrogate a much wider range of peptides and thus provide sequencing and PTM detection of low-abundance peptides (from low-abundance proteins or low-abundance PTM containing peptides).

In a typical LC/MS/MS experiment, the peptide mixture is loaded onto an HPLC column (typically a C18 reversed-phase column that separates peptides based on hydrophobicity) via an autosampler. Peptides are eluted from the column via gradient elution and subsequently ionized into the mass spectrometer via an electrospray emitter (directly connected to the HPLC column or part of the HPLC column to minimize dead volume). During the entire chromatographic separation, the mass spectrometer is acquiring full-scan intact peptide masses and subsequently isolating and fragmenting the peptides to produce MS/MS spectra. Because ESI efficiency is greatly enhanced at lower flow rates (low flow rates generate much smaller initial droplets), modern online HPLC system components are scaled down to provide flow rates at the optimal nl/min range (61, 102). For single-dimension chromatography (1D), online separation columns most often consist of reversed-phase (RP) stationary phases. RP columns (typically C18) are hydrophobic; 1D peptide separations therefore involve the use of an organic mobile phase (acetonitrile) to elute peptides as a function of hydrophobicity. For 2D chromatographic separation, the 2nd dimension column generally consists of a strong cation exchange (SCX) column and RP in the 1st. A 2D configuration (generally only used for extremely complex samples) requires incremental ionic mobile-phase elution (off the SCX column) followed by organic mobile-phase elution for the 1D.

Database Searches of MS/MS Fragmentation Spectra

MS/MS fragmentation spectra provide sequence information that is searched against genomic databases (described in the next section). As stated, HRAM mass spectrometers utilize collision-induced dissociation CID (or HCD) to fragment peptides. CID fragmentation energizes peptide ions (precursor ion) via collisions with gas molecules (such as helium or nitrogen) to produce predominantly y and b ions (although fragmentation also results in some a/x and c/z ions) (69, 94). A typical MS/MS spectrum therefore represents y and b ion series (among other less populated ion types) of a fragmented peptide at various positions along the peptide backbone (103). Experimentally derived MS/MS fragmentation spectra (Fig. 4) are searched by database algorithms such as SEQUEST, Mascot, MyriMatch, and X-tandem (just to name a few) against predicted protein databases (generally created from genome annotation) to identify which peptide they represent (18, 22, 71, 91). The basic function of a database algorithm is to interrogate an MS/MS spectrum and identify the most likely peptide sequence matching (based on the y and b fragment ions) the spectrum from a provided translated genomic database.

There are four main processes involved in peptide identification from MS/MS spectra via database search algorithms: in silico peptide list generation from the database, candidate sequence generation, theoretical spectrum prediction, and scoring. Depending on the database and enzyme specificity, the above database algorithms first create a list of peptide sequences. For example, if the researchers used trypsin on a human blood sample (to generate a complex mixture of tryptic peptides), they would configure the algorithm to generate a complete list of all tryptic peptides from the predicted human proteome database (such databases can be easily downloaded from UniProt and other online genome databases). The algorithm next takes MS/MS spectra, extracts m/z values, and then compares these m/z values to calculated m/z values derived from peptides produced by the in silico list. A theoretical spectrum is then generated and compared to the experimentally derived spectrum and a score is assigned (18, 93). It is up to the researcher to decide what scores are considered probable matches. The researcher then utilizes additional filtering software (i.e., Scaffold software) that groups all high-scoring peptide matches to their corresponding protein.

An important step before PTM quantification (discussed in the next sections) is to first reproducibly identify which peptides contain PTMs; this of course will reflect where on the protein sequence the modification occurs. Database search algorithms are also used to identify peptides with PTMs; to do this the researcher specifies the change in mass associated with the modified form on a particular peptide residue (such as 48 Da shift in mass associated with tri-oxidation of cysteine residues). When the algorithm generates candidate sequences, it considers either of two masses possible for the specified residues. This allows the algorithm to match and score peptides that would otherwise be missed by looking for MS/MS that matched to peptides predicted by genomic sequence. Peptides identified with PTMs can then be further studied using quantitative methods.

Step 2: Quantitative Approach for Characterizing Hb Oxidative PTMs

This section illustrates how quantitative HRAM mass spectrometry has recently been used to provide mechanistic insight into the extent of oxidative changes within hemoglobin variants. First, we describe recent studies evaluating several mutants identified in patients with hemoglobinopathies and show how information gleaned from these studies was utilized toward the design of oxidatively stable HBOCs (blood substitutes). Table 1 provides a list of Hb variants recently investigated in our laboratory (1, 42, 62, 89). Information from several of these studies has been utilized toward the design of candidate acellular blood substitutes (discussed at the end of this section). To date, Hb mass spectrometry laboratories (with an interest in Hb oxidation studies) utilize a quantitative strategy similar to the one shown in Figure 5 (or some minor variation of it) (10, 74, 89, 96, 104). For example, our laboratory generates extracted ion chromatograms of oxidative hotspot containing peptides directly from full-scan data; however, a targeted approach using select ion monitoring can also be used to quantify peptides identified from the bottom-up approach (96). As previously stated, the first step in all Hb quantitative experiments involves implementation of the bottom-up strategy (LC/MS/MS analysis) to identify modified/oxidized Hb residues. After identifying oxidatively modified peptides, the second step involves a label-free method of quantifying the abundance of the oxidized peptide relative to the unmodified peptide; comparative differences in this ratio (relative to native HbA) provide insight into oxidative stability.

FIG. 5.

Schematic diagram of label-free quantitative mass spectrometry strategy. Schematic diagram illustrating the basic workflow for quantitative mass spectrometry utilized toward characterizing Hb-based oxygen. Peptides are chromatographically separated, ionized via electrospray, and analyzed (top left corner) during a gradient run. The total ion chromatogram represented in the top right corner consists of thousands of MS (full scan) and MS/MS spectra that are first analyzed by Step 1 (bottom-up strategy) to identify oxidized Hb peptides followed by Step 2, a label-free method of quantifying the abundance of oxidized peptides relative to the unmodified peptide. To see this illustration in color and view labels, the reader is referred to the web version of this article at www.liebertpub.com/ars

HRAM Analysis Provides Mechanistic Insight into Subunit-Specific Post-Translational Transformation of Methionine to Aspartate

In a recent report from our laboratory, Strader et al. (89) provide an example of how this quantitative mass spectrometric approach was used to characterize the impact of positionally equivalent E11 methionine (E-helix is adjacent to the reactive iron in the distal heme pocket) mutations in three recombinant Hb variants; Toms River (γVal67Met) (γV67 M), Alisha Bristol (βVal67Met) (βV67 M), and Hb Evans (αVal67Met) (αV62 M) (89) (Table 1). The origin of this study began with a focus on characterizing the fetal Hb variant Hb Toms River (γV67 M) from an infant girl in Toms River, New Jersey; the patient was heterozygous for the V67 M mutant allele. In addition to identifying the γV67 M isoform, 1D/LC/MS/MS bottom-up analysis (step one in Fig. 5) also confirmed the presence of an unexpected γV67D variant in the patient hemolysate (Fig. 6A, B); based on DNA sequencing, the presence of this variant could only occur as a consequence of post-translational oxidation of γM67.

FIG. 6.

Mass spectrometric characterization of Toms River's patient blood and the recombinant Hb γ Val67Met counterpart. Displayed above the panels is the primary sequence of V67 M showing the tryptic peptide (highlighted in red) with position M67 (highlighted in blue). All charge states of this peptide sequence were the target of our quantitative proteomic approach. Fragmentation spectra of the singly charged +1 γV67 M (A) and γV67D tryptic peptides representing residues 67–76 from patient hemolysates are shown in (B). Both spectra show singly charged y and b fragment ions. (C) Isotopic profile of the singly charged γV67D tryptic peptide listed above. (D) Extracted ion chromatogram (XIC) of γV67D tryptic peptide (residues 67–76). XICs were generated from the most abundant monoisotopic peak represented in the isotopic profile of each targeted peptide. (E) Histogram representing the fold-change differences of V67D for experimental conditions 1–6 (1: CO ligand control, 2: control (air equilibrated buffer), 3: 1:1 H₂O₂, 4: 2.5:1 H₂O₂, 5: 5:1 H₂O₂, 6: Ferric/5:1 H₂O₂). These data indicate M to D conversion increases with increasing H₂O₂. (F) The isotopic profile represented by labeled isotopologous peptides after oxidation with H₂ ¹⁸O_2. The most abundant isotopic peak (M+H) of 1036 m/z represents the γV67D tryptic peptide with 2 ¹⁸O labeled carboxyl oxygen molecules. The 1034.56 m/z peak represents the γV67D tryptic peptide with a mixture of ¹⁸O/¹⁶O labeled carboxyl oxygen, and the lowest abundance 1032.56 m/z represents the unlabeled peptide (M + H). ¹⁸O monosubstituted and unlabeled γV67D tryptic peptide, respectively. This research was originally published in a study by Strader et al. (89) © the American Society for Biochemistry and Molecular Biology.

We reasoned that Met transformation to Asp was driven by an oxidative event that emanated from the nearby heme group. As outlined in the second step of Figure 5, an integrated quantitative mass spectrometric approach was then utilized to explore the relationship between increasing H₂O₂ levels and γMet-67 post-translational conversion to V67D; the results established a direct linkage between accelerated auto-oxidation (the rate was seven times faster than native HbA) and the Met(E11) modification. Figure 6C and D shows how HRAM data from a hybrid Q-Exactive Orbitrap mass spectrometer (illustrated in Fig. 3B-4) were utilized to generated extracted ion chromatograms (XICs) for quantifying the impact of increasing H₂O₂ on γMet-67 (the example in Fig. 3C, D corresponds to the singly charge D67 containing peptide DLTSLGDAIK). Specifically, XICs were generated from the most abundant monoisotopic peak of all Met-67 (thioether, sulfoxide, and sulfone) and Asp-67 peptide isotopic profiles using a designated mass tolerance of 0.01 Da, and mass precision set to three decimals. For relative quantification, the ratio of each Asp-67 or Met-67 peptide isoform was calculated based on the sum of the XIC peak area from all forms, which was normalized to be 100%. The experimental conditions and results from these XIC studies show that increasing H₂O₂ concentrations resulted in higher V67D levels (Fig. 6E). It is important to note that precursor isotopic resolution (that HRAM instruments provide) is required for the type of XIC construction described above; generating XICs from the most abundant isotopic peak results in XICs with superb signal to noise ratio and eliminates the potential for ion current from coeluting ions to interfere with data analysis. High-resolution data were also required in H₂ ¹⁸O₂ follow-up experiments aimed at further examining the direct role of H₂O₂ in V67D formation. Impressively, experimental conditions with labeled H₂ ¹⁸O₂ and HbF γV67 M resulted in Met-67 and Asp-67 (peptide isotopic profile shown in Fig. 6F) containing peptides with incorporated ¹⁸O, which correlates with the results shown in Figure 6E and confirms that the source of oxygen for the Asp modification comes from H₂O₂ (a proposed mechanism is discussed in the article (89).

This study was expanded to characterize the impact of positionally equivalent E11 methionine mutations in recombinant βV67 M (Hb Bristol-Alesha) and αV62 M (Hb Evans) variants. Met→Asp conversion was identified in the β(E11) but not in the α(E11) variants under similar experimental conditions. The initial crystal structures of the three recombinant Met(E11) HbCO forms show similar active site thioether side-chain orientations (Fig. 7A, B), with the γ side-chain carbon atoms all being roughly 5.0–5.3 Å from the heme iron. However, in agreement with solution experiments, auto-oxidation of the crystals over a period of weeks revealed electron density maps indicative of Asp(E11) formation in the active sites of the HbA β V67 M mutant but not in the active site of the HbA αV62 M variant (Fig. 7C, D). These findings demonstrate the unique redox reactivity of the γ and β heme iron atoms, which leads to post-translational oxidative conversion of amino acid side chains within these subunits, but not within the α subunits of human Hb. We went on to show that Asp was formed by an oxidative mechanism involving the ferryl heme complexes. Interestingly, ferryl Hb complexes were seen in Hb Bristol-Alesha (Fig. 7C, D), however, in the case of Hb Evans, the internal electron pathway may have minimized the accumulation of sufficient quantities of ferryl heme iron to drive this transformation.

FIG. 7.

Structures of the active sites of Hb Toms River γ V67 M, Hb Bristol-Alesha β V67 M, and Hb Evans α V62 M. (A) Electron density of γ MetE11 in HbCO Toms River 2Fo-Fc, refined electron density (4MQK). (B) Comparisons of MetE11 side-chain orientations in refined models for the CO forms of Hb γV67 M (4MQK), Hb βV67 M (4MQG), and Hb αV62 M (4MQC). (C) Oxidation of Hb βV67 M crystals (HbCO form, 4MQG; metHb form after 4 weeks, 4MQI). The electron density suggests the appearance of D at E11 position due to oxidation. The ligand electron density also suggests a ferryl complex with a single coordinated distal atom rather than aquomet complex with a spherical water O atom about 2 Å away from the central iron atom. (D) Oxidation of αV62 M crystals (HbCO form, 4MQC; metHb form after 7 weeks, 4MQH). There was no change in the electron density for the αE11 side chain in the oxidized Hb indicating no modification of the M62 residue. However, the coordination in the wild-type met-β active site again suggests a significant fraction of ferryl complex. This research was originally published in a study by Strader et al. (89) © the American Society for Biochemistry and Molecular Biology.

Comparative HRAM Analysis of Oxidation in HbS and HbA

As previously stated, internal oxidative pathway reactions result in irreversible oxidation of amino acid in oxidation hotspots, regions (particularly the β Cys93 side chain) that cause the collapse of β subunits, Hb degradation, and the ultimate release of heme (39, 72). The quantitative strategy outlined in Figure 5 has also been recently used to quantify the levels of irreversible oxidation of hotspot residues, including βCys93 in the presence of H₂O₂ (10, 72, 74, 96, 104). This surface amino acid has previously been shown to be an important endpoint for free radical-induced protein oxidation (72).

HbS (β6Glu→Val substitution, linked to sickle cell disease [SCD]) is known to be unstable in vitro, is less stable than HbA (on exposure to heat, oxidants), and autoxidizes at faster rates than HbA (7, 33, 58). In a recent study by our group, Kassa et al. (42) indicated that chemically induced oxidation of HbS resulted in higher accumulation of the highly reactive ferryl HbFe⁴⁺ (due to slower autoreduction of ferryl to the ferric form) in HbS solutions (42). HRAM quantitative MS analysis was also used in this HbS study to compare differences in the post-translational oxidation of β subunit hotspot residues for both HbS and HbA under increasing H₂O₂ conditions (42). Amino acid residues identified by LC-MS/MS analysis in this study correlated well with previously published data; the most prevalent oxidative changes were found to be restricted to peptides containing βCys93 (10, 72, 74, 96, 104). XICs for the βCys93 containing peptide GTFATLSELHCDKLHVDPENFR (Fig. 8A) were used to construct the XIC in Figure 8B. Because βCys93 residue can either be oxidized or unoxidized, the percentages of both isoforms were calculated based on the sum of the XIC peak area from all charged forms of βCys93 peptides. The HRAM quantitative results from this study indicated that HbS βCys93 is substantially more oxidized (>2-fold difference when the ratio of H₂O₂ to heme equivalent is 5:1) in the presence of excess H₂O₂ compared to HbA. This dose-dependent increase correlated (with previous reports) with the oxidation kinetic data providing further evidence that HbS oxidation results in a higher accumulation of the highly reactive ferryl HbFe⁴⁺ (and protein radical).

FIG. 8.

Isotopic profile and XIC of β Cys93 tryptic peptide are illustrated. XICs were generated from the most abundant monoisotopic peaks of each isotopic profile. (A) Representative isotopic profile of the triply charged βC93 tryptic peptide GTFATLSELHCDKLHVDPENFR. (B) XIC of βC93 tryptic peptide (residues 83–104). (C) LC-MS/MS deconvoluted spectrum of βC93 tryptic peptide from HbS with DMPO adducts. The reaction solution of HbS with H₂O₂ in the presence of DMPO was digested with trypsin and analyzed by LC-MS/MS. The fragmentation spectra of the triply charged DMPO-labeled βC93 tryptic peptide GTFATLSELHCDKLHVDPENFR are shown. These spectra show singly charged y and b fragment ions. *Corresponds to y and b ions shifted in mass by 111 Da (from expected mass) confirming the location of the DMPO adducts to be Cys93. This research was originally published in a study by Kassa et al. (42) © the American Society for Biochemistry and Molecular Biology. XIC, extracted ion chromatogram.

To further confirm that ferryl ions and globin-centered radicals are more abundant in HbS, HRAM quantitative experiments were performed using the spin trap label 5,5-dimethyl-1-pyrroline N-oxide (DMPO). The exposure of Hbs to H₂O₂ has previously been shown to initially produce a porphyrin cation radical (and ferryl ion) that oxidizes cysteinyl and tyrosyl amino acids to form globin-centered radicals (20, 31, 60, 96). The spin trap DMPO reacts with these modified amino acids to form a nitroxide radical that is further oxidized to the corresponding globin radical-derived nitrone adduct by the ferryl moiety (20, 31, 60). These DMPO-derived adducts are stable under fragmentation conditions used in LC-MS/MS analysis. Analyses of LC/MS/MS and full MS data in the presence of H₂O₂ (5- and 10-fold excess relative to heme) identified DMPO-labeled Cys112 and Tyr42α peptides with the addition of H₂O₂. Among the three DMPO-labeled amino acids, only Cys93 was substantially different between HbA and HbS. DMPO Cys93 labeling (Fig. 8C) in the absence of H₂O₂ was at negligible levels (∼1% for both HbA and HbS). The degree of DMPO labeling was considerably more abundant for HbS than HbA (>2-fold when the ratio of H₂O₂ to heme equivalent is 5:1). These labeling results further substantiate the observation that ferryl levels are more prevalent in HbS and likely account for the observed accelerated oxidation and oxidative changes within the protein. We have postulated that dominant Cys93 oxidation in sickle cell Hb is due, in part, to the persistence of its ferryl and radical species. Thus, highly oxidizing ferryl Hb and heme may be central to the evolution of vasculopathy in SCD and suggest therapeutic modalities that interrupt heme-mediated inflammation (42).

The Role of Unpaired Alpha Subunits in Destabilizing HbE

β-thalassemia (caused by chromosome 11 β-globin gene alterations) is characterized by reduction or absence of β-globin chain synthesis (101). HbE (β26Glu→Lys) is an oxidatively unstable human Hb often coinherited together with a β-thalassemia allele; the resulting HbE/β-thalassemia condition is often severe requiring multiple transfusions (24, 37). Because of HbE instability, a higher level of unpaired α subunits within HbE containing red cells has been proposed to cause the pathogenesis associated with HbE/β thalassemia (77, 78). However, it is experimentally challenging to determine the effects of free uncomplexed α subunits in the presence of another Hb tetramer (i.e., pairs of α2β2 of either HbA or HbE). Because we have previously identified βCys93, as the main target of heme radical chemistry, we recently performed a comparative study that monitors the effects of excess α subunits on βCys93 oxidation within both HbE and HbA (90). Using the same HRAM quantitative mass spectrometric strategies described for HbS characterization (including the spin trap, DMPO), we quantified irreversible oxidization of βCys93 to reflect oxidative instability of β subunits. In the presence of free α subunits and H₂O₂, βCys93 oxidation increased with higher H₂O₂ concentrations for both HbA and HbE. In the presence of alpha-hemoglobin stabilizing protein (AHSP), which stabilizes the α-subunit in a redox-inactive hexacoordinate conformation (thus the α subunit is unable to undergo the redox ferric/ferryl transition), Cys93 oxidation was substantially reduced in both proteins. These experiments establish two important features that may have relevance to the mechanistic understanding of these two inherited hemoglobinopathies, that is, HbE/β thalassemia: First, a ferryl and ferryl protein radical target βCys93 of both proteins but results in the destabilization of HbE due to its persistent ferryl species. Second, in the presence of excess free α-subunit and under the same oxidative conditions, these events are substantially increased for HbE compared to HbA and may therefore create an oxidative milieu affecting the already unstable HbE.

Step 3: Engineering Oxidative Stability in Human Hemoglobin Based on Rationally Designed Proteins

As stated, the first part of the two-prong strategy is to utilize information from Hb variant studies toward the rational design of acellular blood substitutes that are more oxidatively stable than native HbA (and chemically modified Hbs). We have begun to engineer HBOC prototypes based on naturally occurring human mutant Hbs and on rational design (65, 98).

The first of these variant studies involves the variant Hb Providence (Table 1); a mutant identified in a patient at Providence hospital in Washington DC. The patient was heterozygous for the allele expressing the βK82 N variant; due to spontaneous deamidation of asparagine to aspartic acid, two β-subunit variants with the single amino acid mutations βLys82→Asn (βK82 N) and βLys82→Asp (βK82D) were identified in the patient hemolysate. We have previously separated Hb Providence into its three component fractions and contrasted oxidative reactions of these β-mutant fractions with HbA. The Hb Prov. βK82 N variant oxidation reaction kinetics and ligand binding properties were similar to HbA; however, the βK82D fraction was more effective in consuming H₂O₂ and appeared more oxidatively stable. The spectrophotometric data in Figure 9 show that ferryl βK82D, unlike native HbA, reverts back to the ferric state more readily (HbA shows oxidative damage), suggesting that βK82D safely internalizes radicals (through the ferric/ferryl pseudoperoxidase cycle) better than HbA (1). The βK82D amino acid substitution has recently been reported to improve expression levels of soluble recombinant Hb in E. coli (30). Taken together, the structural and oxidative stability of Hb Prov. βK82D may assist in reducing auto-oxidation, oxidative modifications, and heme loss; such features are important goals in the design of oxidatively stable HBOCs. In collaboration with Dr. John Olson at Rice University in Houston, TX (using the rational engineering approach discussed earlier in this review), a genetically crosslinked di-α gene (see section describing two-prong strategy) βK82D Hb (rHb0.1) has recently been engineered. Preliminary kinetic, crystal structure and HRAM quantitative mass spectrometry data of βK82D rHb0.1 indicate that this candidate HBOC is structurally similar to and autooxidizes at similar rates to native HbA yet maintains exceptionally higher oxidative stability in the face of higher levels of peroxide (Strader M, unpublished data). Additional studies are underway to determine how the βK82D substitution impacts oxygen carrying function and if the observed oxidative stability reduces the rate of heme loss relative to control HbA.

FIG. 9.

Pseudoperoxidase catalytic of Hb. Spectral analysis H₂O₂ oxidation reactions: Absorbance spectra obtained at intervals of 2 min for 1 h on treatment of (A) HbA and (B) βK82D with 30:1 H₂O₂/heme equivalents. Note that while HbA undergoes the transition from ferrous to ferryl heme, spectral changes indicative of oxidative modification and some denaturation were evident, whereas in the case of Hb Providence (βK82D), H₂O₂ was consumed more efficiently as ferryl reverted back to ferric with little or no oxidative changes. (A, B) Reprinted with permission from Abraham et al. (1).

Conclusion

Blood substitutes have long been a dream for physicians, researchers, and industrial scientists alike, because they can be stored for a long time, potentially save lives without waiting for typing, and substantially reduce the risk of infection. Despite these advantages, no viable HBOCs have been approved in the United States due, in large part, to clinical trial reports of serious adverse events (79, 86). To address these toxicological responses, several different biochemical and recombinant strategies have been attempted with limited success. A major factor for this limited success is that outside the RBC, acellular HBOCs are oxidatively toxic due to the reactive heme, which is ironically essential for oxygen binding. The two-prong strategy that we have described here is an approach aimed at exploring ways of mitigating Hb oxidative toxicity. Studying oxidative pathways in select Hb variants (with diverse oxidative profiles) has enabled us to develop criteria toward the rational design of a new series of oxidatively stable HBOC candidates. Future HRAM studies will therefore be aimed at characterizing these new candidate HBOCs to determine which are the most oxidatively resistant toward losing heme, while retaining oxygen carrying function.

Footnotes

Acknowledgments

We thank Dr. Tigist Kassa (Laboratory of Biochemistry and Vascular Biology, Center for Biologics Evaluation and Research, Food and Drug Administration) for assistance in designing the molecular graphics illustrated in . We thank Dr. Nathan C. Verberkmoes (University of Texas, El Paso) for reviewing and suggesting edits to the article. Work in this review was supported by the National Institutes of Health (NIH/NHLBI) grant P01-HL110900 and a grant from the U.S. Food and Drug Administration (MODSCI).

Abbreviations Used

References

Abraham

, Hicks

, Jia

, Baek

, Miller

, and Alayash

. Isolated Hb Providence beta82Asn and beta82Asp fractions are more stable than native HbA(0) under oxidative stress conditions. Biochemistry, 50: 9752–9766, 2011.

Aebersold

and Goodlett

. Mass spectrometry in proteomics. Chem Rev, 101: 269–295, 2001.

Alayash

. Blood substitutes: why haven't we been more successful?. Trends Biotechnol, 32: 177–185, 2014.

Alayash

. Hemoglobin-based blood substitutes: oxygen carriers, pressor agents, or oxidants?. Nat Biotechnol, 17: 545–549, 1999.

Alayash

. Oxygen therapeutics: can we tame haemoglobin?. Nat Rev Drug Discov, 3: 152–159, 2004.

Alayash

. Setbacks in blood substitutes research and development: a biochemical perspective. Clin Lab Med, 30: 381–389, 2010.

Asakura

, Adachi

, Sono

, Friedman

, and Schwartz

. Mechanical stability of hemoglobin subunits: an abnormality in betaS-subunits of sickle hemoglobin. Biochem Biophys Res Commun, 57: 780–786, 1974.

Azari

, Rohn

, and Picken

. Diaspirin crosslinked hemoglobin (DCLHb): characterization of the process and the product manufactured under GMP requirements for clinical studies. Artif Cells Blood Substit Immobil Biotechnol, 22: 701–708, 1994.

Belcher

, Chen

, Nguyen

, Milbauer

, Abdulla

, Alayash

, Smith

, Nath

, Hebbel

, and Vercellotti

. Heme triggers TLR4 signaling leading to endothelial cell activation and vaso-occlusion in murine sickle cell disease. Blood, 123: 377–390, 2014.

10.

Bosman

, Lasonder

, Groenen-Dopp

, Willekens

, and Werre

. The proteome of erythrocyte-derived microparticles from plasma: new clues for erythrocyte aging and vesiculation. J Proteomics, 76 Spec No.: 203–210, 2012.

11.

Brantley

Jr. , Smerdon

, Wilkinson

, Singleton

, and Olson

. The mechanism of autooxidation of myoglobin. J Biol Chem, 268: 6995–7010, 1993.

12.

Brooks

. Reengineering transfusion and cellular therapy processes hospitalwide: ensuring the safe utilization of blood products. Transfusion, 45: 159S–171S, 2005.

13.

Brucker

. Genetically crosslinked hemoglobin: a structural study. Acta Crystallogr D Biol Crystallogr, 56: 812–816, 2000.

14.

Buehler

, Abraham

, Vallelian

, Linnemayr

, Pereira

, Cipollo

, Jia

, Mikolajczyk

, Boretti

, Schoedon

, Alayash

, and Schaer

. Haptoglobin preserves the CD163 hemoglobin scavenger pathway by shielding hemoglobin from peroxidative modification. Blood, 113: 2578–2586, 2009.

15.

Buehler

and Alayash

. Toxicities of hemoglobin solutions: in search of in-vitro and in-vivo model systems. Transfusion, 44: 1516–1530, 2004.

16.

Buehler

, Vallelian

, Mikolajczyk

, Schoedon

, Schweizer

, Alayash

, and Schaer

. Structural stabilization in tetrameric or polymeric hemoglobin determines its interaction with endogenous antioxidant scavenger pathways. Antioxid Redox Signal, 10: 1449–1462, 2008.

17.

Bunn

, Esham

, and Bull

. The renal handling of hemoglobin. I. Glomerular filtration. J Exp Med, 129: 909–923, 1969.

18.

Craig

and Beavis

. TANDEM: matching proteins with tandem mass spectra. Bioinformatics, 20: 1466–1467, 2004.

19.

de Hoffmann

ESV

. Mass Spectrometry: Principles and Applications. Toronto, Canada: John Wiley & Sons, Ltd., 2003.

20.

Deterding

, Ramirez

, Dubin

, Mason

, and Tomer

. Identification of free radicals on hemoglobin from its self-peroxidation using mass spectrometry and immuno-spin trapping: observation of a histidinyl radical. J Biol Chem, 279: 11600–11607, 2004.

21.

Eich

, Li

, Lemon

, Doherty

, Curry

, Aitken

, Mathews

, Johnson

, Smith

, Phillips

Jr. , and Olson

. Mechanism of NO-induced oxidation of myoglobin and hemoglobin. Biochemistry, 35: 6976–6983, 1996.

22.

Eng

, McCormack

, and Yates

. An approach to correlate tandem mass spectral data of peptides with amino acid sequences in a protein database. J Am Soc Mass Spectrom, 5: 976–989, 1994.

23.

Fenn

, Mann

, Meng

, Wong

, and Whitehouse

. Electrospray ionization for mass spectrometry of large biomolecules. Science, 246: 64–71, 1989.

24.

Frischer

and Bowman

. Hemoglobin E, an oxidatively unstable mutation. J Lab Clin Med, 85: 531–539, 1975.

25.

Fronticelli

and Koehler

. Design of recombinant hemoglobins for use in transfusion fluids. Crit Care Clin, 25: 357–371, Table of Contents, 2009.

26.

Gallien

, Duriez

, Crone

, Kellmann

, Moehring

, and Domon

. Targeted proteomic quantification on quadrupole-orbitrap mass spectrometer. Mol Cell Proteomics, 11: 1709–1723, 2012.

27.

Garrod

. An Address on THE PLACE OF BIOCHEMISTRY IN MEDICINE. Br Med J, 1: 1099–1101, 1928.

28.

Gladwin

and Ofori-Acquah

. Erythroid DAMPs drive inflammation in SCD. Blood, 123: 3689–3690, 2014.

29.

Goodnough

, Brecher

, Kanter

, and AuBuchon

. Transfusion medicine. Second of two parts—blood conservation. N Engl J Med, 340: 525–533, 1999.

30.

Graves

, Henderson

, Horstman

, Solomon

, and Olson

. Enhancing stability and expression of recombinant human hemoglobin in E. coli: Progress in the development of a recombinant HBOC source. Biochim Biophys Acta, 1784: 1471–1479, 2008.

31.

Gunther

, Tschirret-Guth

, Witkowska

, Fann

, Barr

, Ortiz De Montellano

, and Mason

. Site-specific spin trapping of tyrosine radicals in the oxidation of metmyoglobin by hydrogen peroxide. Biochem J, 330 (Pt 3): 1293–1299, 1998.

32.

Hager

. A new linear ion trap mass spectrometer. Rapid Commun Mass Spectrom, 16, 2002.

33.

Hebbel

, Morgan

, Eaton

, and Hedlund

. Accelerated autoxidation and heme loss due to instability of sickle hemoglobin. Proc Natl Acad Sci U S A, 85: 237–241, 1988.

34.

Henkel-Honke

and Oleck

. Artificial oxygen carriers: a current review. AANA J, 75: 205–211, 2007.

35.

Hsia

, Song

, Er

, Wong

, Keipert

, Gomez

, Gonzales

, Macdonald

, Hess

, and Winslow

. Pharmacokinetic studies in the rat on a o-raffinose polymerized human hemoglobin. Biomater Artif Cells Immobil Biotechnol, 20: 587–595, 1992.

36.

, Prabhakaran

, Acharya

, and Manjula

. Influence of the chemistry of conjugation of poly(ethylene glycol) to Hb on the oxygen-binding and solution properties of the PEG-Hb conjugate. Biochem J, 392: 555–564, 2005.

37.

Itano

, Bergren

, and Sturgeon

. The abnormal human hemoglobins. Medicine, 35: 121–159, 1956.

38.

Jensen

. Modification-specific proteomics: characterization of post-translational modifications by mass spectrometry. Curr Opin Chem Biol, 8: 33–41, 2004.

39.

Jia

, Buehler

, Boykins

, Venable

, and Alayash

. Structural basis of peroxide-mediated changes in human hemoglobin: a novel oxidative pathway. J Biol Chem, 282: 4894–4907, 2007.

40.

Johnson

, Adamson

, and Mauk

. Functional comparison of specifically cross-linked hemoglobins biased toward the R and T states. Biophys J, 75: 3078–3084, 1998.

41.

Karas

and Hillenkamp

. Laser desorption ionization of proteins with molecular masses exceeding 10,000 daltons. Anal Chem, 60: 2299–2301, 1988.

42.

Kassa

, Jana

, Strader

, Meng

, Jia

, Wilson

, and Alayash

. Sickle cell hemoglobin in the Ferryl state promotes betaCys-93 oxidation and mitochondrial dysfunction in epithelial lung cells (E10). J Biol Chem, 290: 27939–27958, 2015.

43.

Kim

and Greenburg

. Artificial oxygen carriers as red blood cell substitutes: a selected review and current status. Artif Organs, 28: 813–828, 2004.

44.

Klein

. Will blood transfusion ever be safe enough?. JAMA, 284: 238–240, 2000.

45.

Liao

, Hein

, Vaughn

, Huang

, and Kuo

. Intravascular flow decreases erythrocyte consumption of nitric oxide. Proc Natl Acad Sci U S A, 96: 8757–8761, 1999.

46.

Loboda

, Krutchinsky

, Bromirski

, Ens

, and Standing

. A tandem quadrupole/time-of-flight mass spectrometer with a matrix-assisted laser desorption/ionization source: design and performance. Rapid Commun Mass Spectrom, 14: 1047–1057, 2000.

47.

Looker

, Abbott-Brown

, Cozart

, Durfee

, Hoffman

, Mathews

, Miller-Roehrich

, Shoemaker

, Trimble

, Fermi

, and et al. A human recombinant haemoglobin designed for use as a blood substitute. Nature, 356: 258–260, 1992.

48.

MacCoss

, McDonald

, Saraf

, Sadygov

, Clark

, Tasto

, Gould

, Wolters

, Washburn

, Weiss

, Clark

, and Yates

3rd . Shotgun identification of protein modifications from protein complexes and lens tissue. Proc Natl Acad Sci U S A, 99: 7900–7905, 2002.

49.

Makarov

. Electrostatic axially harmonic orbital trapping: a high-performance technique of mass analysis. Anal Chem, 72: 1156–1162, 2000.

50.

Makarov

, Denisov

, Kholomeev

, Balschun

, Lange

, Strupat

, and Horning

. Performance evaluation of a hybrid linear ion trap/orbitrap mass spectrometer. Anal Chem, 78: 2113–2120, 2006.

51.

Makarov

, Denisov

, Lange

, and Horning

. Dynamic range of mass accuracy in LTQ Orbitrap hybrid mass spectrometer. J Am Soc Mass Spectrom, 17: 977–982, 2006.

52.

Malhotra

, Kelly

, Miller

, Hartman

, Fabian

, and Proctor

. Resuscitation with a novel hemoglobin-based oxygen carrier in a Swine model of uncontrolled perioperative hemorrhage. J Trauma, 54: 915–924, 2003.

53.

Mann

, Hendrickson

, and Pandey

. Analysis of proteins and proteomes by mass spectrometry. Annu Rev Biochem, 70: 437–473, 2001.

54.

Mann

and Jensen

. Proteomic analysis of post-translational modifications. Nat Biotechnol, 21: 255–261, 2003.

55.

Mann

and Kelleher

. Precision proteomics: the case for high resolution and high mass accuracy. Proc Natl Acad Sci U S A, 105: 18132–18138, 2008.

56.

Marshall

, Hendrickson

, and Jackson

. Fourier transform ion cyclotron resonance mass spectrometry: a primer. Mass Spectrom Rev, 17: 1–35, 1998.

57.

Martin

, Shabanowitz

, Hunt

, and Marto

. Subfemtomole MS and MS/MS peptide sequence analysis using nano-HPLC micro-ESI fourier transform ion cyclotron resonance mass spectrometry. Anal Chem, 72: 4266–4274, 2000.

58.

Marva

and Hebbel

. Denaturing interaction between sickle hemoglobin and phosphatidylserine liposomes. Blood, 83: 242–249, 1994.

59.

Matheson

, Kwansa

, Bucci

, Rebel

, and Koehler

. Vascular response to infusions of a nonextravasating hemoglobin polymer. J Appl Physiol, 93: 1479–1486, 2002.

60.

McArthur

and Davies

. Detection and reactions of the globin radical in haemoglobin. Biochim Biophys Acta, 1202: 173–181, 1993.

61.

Mitulovic

and Mechtler

. HPLC techniques for proteomics analysis—a short overview of latest developments. Brief Funct Genomic Proteomic, 5: 249–260, 2006.

62.

Mollan

, Abraham

, Strader

, Jia

, Lozier

, Olson

, and Alayash

. Familial secondary erythrocytosis due to increased oxygen affinity is caused by destabilization of the T state of hemoglobin Brigham (alpha(2)beta(2)(Pro100Leu)). Protein Sci, 21: 1444–1455, 2012.

63.

Mollan

, Banerjee

, Wu

, Parker Siburt

, Tsai

, Olson

, Weiss

, Crumbliss

, and Alayash

. alpha-Hemoglobin stabilizing protein (AHSP) markedly decreases the redox potential and reactivity of alpha-subunits of human HbA with hydrogen peroxide. J Biol Chem, 288: 4288–4298, 2013.

64.

Moore

. Blood substitutes: the future is now. J Am Coll Surg, 196: 1–17, 2003.

65.

Mozzarelli

, Ronda

, Faggiano

, Bettati

, and Bruno

. Haemoglobin-based oxygen carriers: research and reality toward an alternative to blood transfusions. Blood Transfus, 8 Suppl 3: s59–s68, 2010.

66.

Ness

and Cushing

. Oxygen therapeutics: pursuit of an alternative to the donor red blood cell. Arch Pathol Lab Med, 131: 734–741, 2007.

67.

Nguyen

and Fenn

. Gas-phase ions of solute species from charged droplets of solutions. Proc Natl Acad Sci U S A, 104: 1111–1117, 2007.

68.

This reference has been deleted.

69.

Paizs

and Suhai

. Fragmentation pathways of protonated peptides. Mass Spectrom Rev, 24: 508–548, 2005.

70.

Pandey

and Mann

. Proteomics to study genes and genomes. Nature, 405: 837–846, 2000.

71.

Perkins

, Pappin

, Creasy

, and Cottrell

. Probability-based protein identification by searching sequence databases using mass spectrometry data. Electrophoresis, 20: 3551–3567, 1999.

72.

Pimenova

, Pereira

, Gehrig

, Buehler

, Schaer

, and Zenobi

. Quantitative mass spectrometry defines an oxidative hotspot in hemoglobin that is specifically protected by haptoglobin. J Proteome Res, 9: 4061–4070, 2010.

73.

Privalle

, Talarico

, Keng

, and DeAngelo

. Pyridoxalated hemoglobin polyoxyethylene: a nitric oxide scavenger with antioxidant activity for the treatment of nitric oxide-induced shock. Free Radic Biol Med, 28: 1507–1517, 2000.

74.

Ratanasopa

, Strader

, Alayash

, and Bulow

. Dissection of the radical reactions linked to fetal hemoglobin reveals enhanced pseudoperoxidase activity. Front Physiol, 6: 39, 2015.

75.

Rebel

, Ulatowski

, Kwansa

, Bucci

, and Koehler

. Cerebrovascular response to decreased hematocrit: effect of cell-free hemoglobin, plasma viscosity, and CO2. Am J Physiol Heart Circ Physiol, 285: H1600–H1608, 2003.

76.

Reeder

. The redox activity of hemoglobins: from physiologic functions to pathologic mechanisms. Antioxid Redox Signal, 13: 1087–1123, 2010.

77.

Rees

, Clegg

, and Weatherall

. Is hemoglobin instability important in the interaction between hemoglobin E and beta thalassemia?. Blood, 92: 2141–2146, 1998.

78.

Rees

, Williams

, Maitland

, Clegg

, and Weatherall

. Alpha thalassaemia is associated with increased soluble transferrin receptor levels. Br J Haematol, 103: 365–369, 1998.

79.

Reid

. Hb-based oxygen carriers: are we there yet?. Transfusion, 43: 280–287, 2003.

80.

Sanders

, Ackers

, and Sligar

. Engineering and design of blood substitutes. Curr Opin Struct Biol, 6: 534–540, 1996.

81.

Savitsky

, Doczi

, Black

, and Arnold

. A clinical safety trial of stroma-free hemoglobin. Clin Pharmacol Ther, 23: 73–80, 1978.

82.

Scigelova

and Makarov

. Orbitrap mass analyzer—overview and applications in proteomics. Proteomics, 6 Suppl 2: 16–21, 2006.

83.

Sehgal

, Sehgal

, Rosen

, Gould

, DeWoskin

, and Moss

. Characteristics of polymerized pyridoxylated hemoglobin. Biomater Artif Cells Artif Organs, 16: 173–183, 1988.

84.

Shikama

. The Molecular Mechanism of Autoxidation for Myoglobin and Hemoglobin: A Venerable Puzzle. Chem Rev, 98: 1357–1374, 1998.

85.

Silva

, Vitorino

, Domingues

, Spickett

, and Domingues

. Post-translational modifications and mass spectrometry detection. Free Radic Biol Med, 65: 925–941, 2013.

86.

Silverman

, Weiskopf

, and Planning Committee and the Speakers. Hemoglobin-based oxygen carriers: current status and future directions. Anesthesiology, 111: 946–963, 2009.

87.

Standl

, Freitag

, Burmeister

, Horn

, Wilhelm

, and Am Esch

. Hemoglobin-based oxygen carrier HBOC-201 provides higher and faster increase in oxygen tension in skeletal muscle of anemic dogs than do stored red blood cells. J Vasc Surg, 37: 859–865, 2003.

88.

Stowell

. What happened to blood substitutes?. Transfus Clin Biol, 12: 374–379, 2005.

89.

Strader

, Hicks

, Kassa

, Singleton

, Soman

, Olson

, Weiss

, Mollan

, Wilson

, and Alayash

. Post-translational transformation of methionine to aspartate is catalyzed by heme iron and driven by peroxide: a novel subunit-specific mechanism in hemoglobin. J Biol Chem, 289: 22342–22357, 2014.

90.

Strader

, Kassa

, Meng

, Wood

, Hirsch

, Friedman

, and Alayash

. Oxidative instability of hemoglobin E (beta26 Glu—>Lys) is increased in the presence of free alpha subunits and reversed by alpha-hemoglobin stabilizing protein (AHSP): relevance to HbE/beta-thalassemia. Redox Biol, 8: 363–374, 2016.

91.

Tabb

, Fernando

, and Chambers

. MyriMatch: highly accurate tandem mass spectral peptide identification by multivariate hypergeometric analysis. J Proteome Res, 6: 654–661, 2007.

92.

Tabb

, Friedman

, and Ham

. Verification of automated peptide identifications from proteomic tandem mass spectra. Nat Protoc, 1: 2213–2222, 2006.

93.

Tabb

, Shah

, Strader

, Connelly

, Hettich

, and Hurst

. Determination of peptide and protein ion charge states by Fourier transformation of isotope-resolved mass spectra. J Am Soc Mass Spectrom, 17: 903–915, 2006.

94.

Tabb

, Smith

, Breci

, Wysocki

, Lin

, and Yates

3rd . Statistical characterization of ion trap tandem mass spectra from doubly charged tryptic peptides. Anal Chem, 75: 1155–1163, 2003.

95.

Thom

, Dickson

, Gell

, and Weiss

. Hemoglobin variants: biochemical properties and clinical correlates. Cold Spring Harbor Perspect Med, 3: a011858, 2013.

96.

Vallelian

, Garcia-Rubio

, Puglia

, Kahraman

, Deuel

, Engelsberger

, Mason

, Buehler

, and Schaer

. Spin trapping combined with quantitative mass spectrometry defines free radical redistribution within the oxidized hemoglobin:haptoglobin complex. Free Radic Biol Med, 85: 259–268, 2015.

97.

Vandegriff

, Malavalli

, Minn

, Jiang

, Lohman

, Young

, Samaja

, and Winslow

. Oxidation and haem loss kinetics of poly(ethylene glycol)-conjugated haemoglobin (MP4): dissociation between in vitro and in vivo oxidation rates. Biochem J, 399: 463–471, 2006.

98.

Varnado

, Mollan

, Birukou

, Smith

, Henderson

, and Olson

. Development of recombinant hemoglobin-based oxygen carriers. Antioxid Redox Signal, 18: 2314–2328, 2013.

99.

Vestal

. Methods of ion generation. Chem Rev, 101: 361–375, 2001.

100.

Weiskopf

. Hemoglobin-Based Oxygen Carriers: Disclosed History and the Way Ahead: The Relativity of Safety. Anesth Analg, 119: 758–760, 2014.

101.

Williams

, Maitland

, Bennett

, Ganczakowski

, Peto

, Newbold

, Bowden

, Weatherall

, and Clegg

. High incidence of malaria in alpha-thalassaemic children. Nature, 383: 522–525, 1996.

102.

Wilm

and Mann

. Analytical properties of the nanoelectrospray ion source. Anal Chem, 68: 1–8, 1996.

103.

World Health Organization. WHO Global database on Blood Safety: Summary Report 2011. Available at www.who.int/bloodsafety/global_database/en/ Accessed September 9, 2015 .

104.

Wysocki

, Tsaprailis

, Smith

, and Breci

. Mobile and localized protons: a framework for understanding peptide dissociation. J Mass Spectrom, 35: 1399–1406, 2000.

105.

Xiang

, Weisbach

, Heinricht

, Seebahn

, Bussmann

, Zimmermann

, and Becker

. Oxidative stress-induced posttranslational modifications of human hemoglobin in erythrocytes. Arch Biochem Biophys, 529: 34–44, 2013.