Stability of Maleimide-PEG and Mono-Sulfone-PEG Conjugation to a Novel Engineered Cysteine in the Human Hemoglobin Alpha Subunit.

In order to use a Hemoglobin Based Oxygen Carrier as an oxygen therapeutic or blood substitute, it is necessary to increase the size of the hemoglobin molecule to prevent rapid renal clearance. A common method uses maleimide PEGylation of sulfhydryls created by the reaction of 2-iminothiolane at surface lysines. However, this creates highly heterogenous mixtures of molecules. We recently engineered a hemoglobin with a single novel, reactive cysteine residue on the surface of the alpha subunit creating a single PEGylation site (ßCys93Ala/αAla19Cys). This enabled homogenous PEGylation by maleimide-PEG with >80% efficiency and no discernible effect on protein function. However, maleimide-PEG adducts are subject to deconjugation via retro-Michael reactions and cross-conjugation to endogenous thiol species in vivo. We therefore compared our maleimide-PEG adduct with one created using a mono-sulfone-PEG less susceptible to deconjugation. Mono-sulfone-PEG underwent reaction at αAla19Cys hemoglobin with > 80% efficiency, although some side reactions were observed at higher PEG:hemoglobin ratios; the adduct bound oxygen with similar affinity and cooperativity as wild type hemoglobin. When directly compared to maleimide-PEG, the mono-sulfone-PEG adduct was significantly more stable when incubated at 37°C for seven days in the presence of 1 mM reduced glutathione. Hemoglobin treated with mono-sulfone-PEG retained > 90% of its conjugation, whereas for maleimide-PEG < 70% of the maleimide-PEG conjugate remained intact. Although maleimide-PEGylation is certainly stable enough for acute therapeutic use as an oxygen therapeutic, for pharmaceuticals intended for longer vascular retention (weeks-months), reagents such as mono-sulfone-PEG may be more appropriate.

Engineering hemoglobin to enable homogenous PEGylation without modifying protein functionality.

In order to infuse hemoglobin into the vasculature as an oxygen therapeutic or blood substitute, it is necessary to increase the size of the molecule to enhance vascular retention. This aim can be achieved by PEGylation. However, using non-specific conjugation methods creates heterogenous mixtures and alters protein function. Site-specific PEGylation at the naturally reactive thiol on human hemoglobin (ßCys93) alters hemoglobin oxygen binding affinity and increases its autooxidation rate. In order to avoid this issue, new reactive thiol residues were therefore engineered at sites distant to the heme group and the α/ß dimer/dimer interface. The two mutants were ßCys93Ala/αAla19Cys and ßCys93Ala/ßAla13Cys. Gel electrophoresis, size exclusion chromatography and mass spectrometry revealed efficient PEGylation at both αAla19Cys and ßAla13Cys, with over 80% of the thiols PEGylated in the case of αAla19Cys. For both mutants there was no significant effect on the oxygen affinity or the cooperativity of oxygen binding. PEGylation at αAla19Cys had the additional benefit of decreasing the rates of autoxidation and heme release, properties that have been considered contributory factors to the adverse clinical side effects exhibited by previous hemoglobin based oxygen carriers. PEGylation at αAla19Cys may therefore be a useful component of future clinical products.

Engineering tyrosine residues into hemoglobin enhances heme reduction, decreases oxidative stress and increases vascular retention of a hemoglobin based blood substitute.

Hemoglobin (Hb)-based oxygen carriers (HBOC) are modified extracellular proteins, designed to replace or augment the oxygen-carrying capacity of erythrocytes. However, clinical results have generally been disappointing due to adverse side effects, in part linked to the intrinsic oxidative toxicity of Hb. Previously a redox-active tyrosine residue was engineered into the Hb ß subunit (ßF41Y) to facilitate electron transfer between endogenous antioxidants such as ascorbate and the oxidative ferryl heme species, converting the highly oxidizing ferryl species into the less reactive ferric (met) form. We inserted different single tyrosine mutations into the α and ß subunits of Hb to determine if this effect of ßF41Y was unique. Every mutation that was inserted within electron transfer range of the protein surface and the heme increased the rate of ferryl reduction. However, surprisingly, three of the mutations (ßT84Y, αL91Y and ßF85Y) also increased the rate of ascorbate reduction of ferric(met) Hb to ferrous(oxy) Hb. The rate enhancement was most evident at ascorbate concentrations equivalent to that found in plasma (< 100 µM), suggesting that it might be of benefit in decreasing oxidative stress in vivo. The most promising mutant (ßT84Y) was stable with no increase in autoxidation or heme loss. A decrease in membrane damage following Hb addition to HEK cells correlated with the ability of ßT84Y to maintain the protein in its oxygenated form. When PEGylated and injected into mice, ßT84Y was shown to have an increased vascular half time compared to wild type PEGylated Hb. ßT84Y represents a new class of mutations with the ability to enhance reduction of both ferryl and ferric Hb, and thus has potential to decrease adverse side effects as one component of a final HBOC product.

Comparison of the oxidative reactivity of recombinant fetal and adult human hemoglobin: implications for the design of hemoglobin-based oxygen carriers.

Hemoglobin (Hb)-based oxygen carriers (HBOCs) have been engineered to replace or augment the oxygen carrying capacity of erythrocytes. However, clinical results have generally been disappointing, in part due to the intrinsic oxidative toxicity of Hb. The most common HBOC starting material is adult human or bovine Hb. However, it has been suggested that fetal Hb may offer advantages due to decreased oxidative reactivity. Large-scale manufacturing of HBOC will likely and ultimately require recombinant sources of human proteins. We, therefore, directly compared the functional properties and oxidative reactivity of recombinant fetal (rHbF) and recombinant adult (rHbA) Hb. rHbA and rHbF produced similar yields of purified functional protein. No differences were seen in the two proteins in: autoxidation rate; the rate of hydrogen peroxide reaction; NO scavenging dioxygenase activity; and the NO producing nitrite reductase activity. The rHbF protein was: less damaged by low levels of hydrogen peroxide; less damaging when added to human umbilical vein endothelial cells (HUVEC) in the ferric form; and had a slower rate of intrinsic heme loss. The rHbA protein was: more readily reducible by plasma antioxidants such as ascorbate in both the reactive ferryl and ferric states; less readily damaged by lipid peroxides; and less damaging to phosphatidylcholine liposomes. In conclusion in terms of oxidative reactivity, there are advantages and disadvantages to the use of rHbA or rHbF as the basis for an effective HBOC.

Engineering tyrosine electron transfer pathways decreases oxidative toxicity in hemoglobin: implications for blood substitute design.

Hemoglobin (Hb)-based oxygen carriers (HBOC) have been engineered to replace or augment the oxygen-carrying capacity of erythrocytes. However, clinical results have generally been disappointing due to adverse side effects linked to intrinsic heme-mediated oxidative toxicity and nitric oxide (NO) scavenging. Redox-active tyrosine residues can facilitate electron transfer between endogenous antioxidants and oxidative ferryl heme species. A suitable residue is present in the α-subunit (Y42) of Hb, but absent from the homologous position in the ß-subunit (F41). We therefore replaced this residue with a tyrosine (ßF41Y, Hb Mequon). The ßF41Y mutation had no effect on the intrinsic rate of lipid peroxidation as measured by conjugated diene and singlet oxygen formation following the addition of ferric(met) Hb to liposomes. However, ßF41Y significantly decreased these rates in the presence of physiological levels of ascorbate. Additionally, heme damage in the ß-subunit following the addition of the lipid peroxide hydroperoxyoctadecadieoic acid was five-fold slower in ßF41Y. NO bioavailability was enhanced in ßF41Y by a combination of a 20% decrease in NO dioxygenase activity and a doubling of the rate of nitrite reductase activity. The intrinsic rate of heme loss from methemoglobin was doubled in the ß-subunit, but unchanged in the α-subunit. We conclude that the addition of a redox-active tyrosine mutation in Hb able to transfer electrons from plasma antioxidants decreases heme-mediated oxidative reactivity and enhances NO bioavailability. This class of mutations has the potential to decrease adverse side effects as one component of a HBOC product.

The hydrogen-peroxide-induced radical behaviour in human cytochrome c-phospholipid complexes: implications for the enhanced pro-apoptotic activity of the G41S mutant.

We have investigated whether the pro-apoptotic properties of the G41S mutant of human cytochrome c can be explained by a higher than wild-type peroxidase activity triggered by phospholipid binding. A key complex in mitochondrial apoptosis involves cytochrome c and the phospholipid cardiolipin. In this complex cytochrome c has its native axial Met(80) ligand dissociated from the haem-iron, considerably augmenting the peroxidase capability of the haem group upon H2O2 binding. By EPR spectroscopy we reveal that the magnitude of changes in the paramagnetic haem states, as well as the yield of protein-bound free radical, is dependent on the phospholipid used and is considerably greater in the G41S mutant. A high-resolution X-ray crystal structure of human cytochrome c was determined and, in combination with the radical EPR signal analysis, two tyrosine residues, Tyr(46) and Tyr(48), have been rationalized to be putative radical sites. Subsequent single and double tyrosine-to-phenylalanine mutations revealed that the EPR signal of the radical, found to be similar in all variants, including G41S and wild-type, originates not from a single tyrosine residue, but is instead a superimposition of multiple EPR signals from different radical sites. We propose a mechanism of multiple radical formations in the cytochrome c-phospholipid complexes under H2O2 treatment, consistent with the stabilization of the radical in the G41S mutant, which elicits a greater peroxidase activity from cytochrome c and thus has implications in mitochondrial apoptosis.

An investigation into a cardiolipin acyl chain insertion site in cytochrome c.

Mitochondrial cytochrome c associates with the phosphoplipid cardiolipin (CL) through a combination of electrostatic and hydrophobic interactions. The latter occurs by insertion into cytochrome c of an acyl chain, resulting in the dissociation of the axial Met-80 heme-iron ligand. The resulting five coordinate cytochrome c/CL complex has peroxidatic properties leading to peroxidation of CL and dissociation of the complex. These events are considered to be pre-apoptotic and culminate with release of cytochrome c from the mitochondria into the cytoplasm. Two distinct surface regions on cytochrome c have been suggested to mediate CL acyl chain insertion and this study has probed one of these regions. We have constructed a series of alanine mutants aimed at disrupting a surface cleft formed between residues 67-71 and 82-85. The physicochemical properties, peroxidase activity, CL binding, and kinetics of carbon monoxide (CO) binding to the ferrous cytochrome c/CL complex have been assessed for the individual mutants. Our findings reveal that the majority of mutants are capable of binding CL in the same apparent stoichiometry as the wild-type protein, with the extent to which the Met-80 ligand is bound in the ferrous cytochrome c/CL complex being mutant specific at neutral pH. Mutation of the species conserved Arg-91 residue, that anchors the cleft, results in the greatest changes to physicochemical properties of the protein leading to a change in the CL binding ratio required to effect structural changes and to the ligand-exchange properties of the ferrous cytochrome c/CL complex.

Disruption of the M80-Fe ligation stimulates the translocation of cytochrome c to the cytoplasm and nucleus in nonapoptotic cells.

Native cytochrome c (cyt c) has a compact tertiary structure with a hexacoordinated heme iron and functions in electron transport in mitochondria and apoptosis in the cytoplasm. However, the possibility that protein modifications confer additional functions to cyt c has not been explored. Disruption of methionine 80 (M80)-Fe ligation of cyt c under nitrative stress has been reported. To model this alteration and determine if it confers new properties to cyt c, a cyt c mutant (M80A) was constitutively expressed in cells. M80A-cyt c has increased peroxidase activity and is spontaneously released from mitochondria, translocating to the cytoplasm and nucleus in the absence of apoptosis. Moreover, M80A models endogenously nitrated cyt c because nitration of WT-cyt c is associated with its translocation to the cytoplasm and nucleus. Further, M80A cyt c may up-regulate protective responses to nitrative stress. Our findings raise the possibility that endogenous protein modifications that disrupt the M80-Fe ligation (such as tyrosine nitration) stimulate nuclear translocation and confer new functions to cyt c in nonapoptotic cells.

EPR and optical spectroscopic studies of Met80X mutants of yeast ferricytochrome c. Models for intermediates in the alkaline transition.

The ferric forms of Met80X mutants of yeast iso-1-cytochrome c (X = Ala, Ser, Asp, and Glu) display EPR and optical spectra that are strongly pH dependent. At low pH values (pH approximately 5) the sixth coordination sites are filled by H(2)O that, on elevating the pH, is replaced by OH(-) in the cases of Met80Ala and -Ser (pK approximately 5.6 and 5.9, respectively) and by a lysine amino group in the cases of Met80Asp and -Glu (pK approximately 9.3 and 11.6, respectively). The ligand sets and the pK values of the transitions are rationalized in terms of the structure of the heme pocket, and a possible mechanism of the "trigger" in the alkaline transition of the native protein is suggested.