CA2219242A1 - Modified hemoglobin-like compounds and methods of purifying same - Google Patents

Abstract

The present invention relates to modified hemoglobin-like compounds. The novel compounds include a globin-like polypeptide containig at least two dialpha domains and multimeric hemoglobin-like proteins having a core hemoglobin-like moeity directly attached to at least two other hemoglobin-like moieties. The invention also relates to nucleic acid molecules encoding the novel polypeptides. Methods of making and purifying the multimeric hemoglobin-like proteins are also provided, as well as compositions containing the proteins.

Description

W O 96/40920 PCT~US96/10420 l~ HEMOGLOBIN-LIKE COMPOUNDS
AND METHODS OF PURIFYING SAME

Back~round of the Invention The present invention is directed to modified hemoglobin-like compounds, and more particularly to modified hemoglobin-like polypeptides and ~roleiuls. The present invention is directed also to methods of purifying such modified hemoglobin-like compounds.
Hemoglobin ( referred to herein as "Hb") is the oxygen-carrying component of blood. Hemoglobin circulates through the bloodstream inside small enudeate cellscalled ~lyL~,Locytes (red blood cells). Hemoglobin is a ~rol~ill constructed from four associated polypeptide chains, and bearing prosthetic groups known as hemes. Thee~yl~uocyte helps maintAin hemoglobin in its reduced, fi1nctiQnAl form. The hemeiron atom is susceptible to oxidation, but may be reduced ao ain by one Q~ tw~e~y~.e systems within the ~lyL~u~ocyte, the cytochrome b5 and gllltAthione reduction systems.
Hemoglobin binds oxygen at a l~s~i,dlc"~/ surface (skin, gills, trachea, lung, etc.) and transports the oxygen to inner tissues, where it is released and used for metabolism. In nature, low molecular weight hemoglobins (16-120 kilodaltons) tend to be enclosed in circulating red blood cells, while the larger polymeric hemoglobins circulate freely in the blood or hemolymph.
The structure of hemoglobin is well known as described in Bunn & Forget, eds., Hemoglobin: Molecular, Genetic and Clinical Aspects (W.B. Saunders Co., Phila-i~lphia, PA: 1986) and Fermi & Perutz "Hemoglobin and Myoglobin," in Phillips and Richards, Atlas of Molecular Structures in Biology (Clarendon Press: 1981).
About 92% of normal adult hl1mAn hemolysate is Hb Ao (~ ignAt~l alpha2 beta2 because it comprises two alpha and two beta chains). In a hemoglobin tetramer, each alpha subunit is associated with a beta subunit to form a stable alpha/betadimer, two of which in turn associate to form the tetramer. The subunits are noncovalently associated through Van der Waals forces, hydrogen bonds and salt bridges. The amino acid sequences of the alpha and beta globin polypeptide chains of Hb Ao are given in Table 1 of PCT Publication No. WO 93 / 09143. The wild-type alpha chain consists of 141 amino acids. The iron atom of the heme (L~ ,plol~,~ol~hyrin IX) group is bound covalently to the imiriA7(~le of ~ 87 (the W O ~6/~0320 PCT~US96/10420 "proximal histidine"). The wild-type beta chain is 146 resi~ s long and heme is bound to it at His 92.
The human alpha and beta globin genes reside on chromosomes 16 and 11, respectively. Bunn and Forget, infra at 172. Both genes have been cloned and seqll~n~ e~l, Liebhaber, et al., PNAS 77: 7054-58 (1980) (alpha-globin genomic DNA);
Marotta, et al., J. Biol. Chem., 252: 5040-53 (1977) (beta globin cDNA); Lawn, et al., Cell, 21:647 (1980) (beta globin genomic DNA).
Hemoglobin exhibits cooperative binding of oxygen by the four subunits of the hemoglobin molecule (the two alpha globins and two beta globins in the case of Hb 10 Ao)~ and this cooperativity greatly facilitates efficient oxygen transport.
Cooperativity, achieved by the so-called heme-heme interaction, allows hemoglobin to vary its affinity for oxygen. Cooperativity can also be ~ " l; ~ ~ed using the oxygen dissociation curve (described below) and is generally reported as the Hill coefficient, "n" or "nma~2-" Hemoglobin reversibly binds up to four moles of oxygen per mole of 15 hemoglobin.
Oxygen-carrying compounds are frequently cull-ydL~d by means of a device known as an oxygen dissociation curve. This curve is obtained when, for a given oxygen carrier, oxygen saturation or content is graphed against the partial pressure of oxygen. For Hb, the perc~lLLdge of saturation increases with partial pressure 20 acc~,ldillg to a .sigmnit1Al r~l~ti-)nship. The Pso is the partial pressure at which the oxygen-carrying species is half saturated with oxygen. It is thus a measure of oxygen-binding affinity; the higher the Pso, the more readily oxygen is released.
The ability of hemoglobin to alter its oxygen affinity under physiological conditions, increasing the efficiency of oxygen transport around the body, is largely 25 dependent on the presence of the metabolite 2,3-diphosphoglycerate (2,3-DPG). The oxygen affinity of hemoglobin is lowered by the presence of 2,3-DPG. Inside the ~lyLhl~cyte 2,3-DPG is present at a concentration nearly as great as that of hemoglobin itself. In the absence of 2,3-DPG ''COllV~ innA1" hemoglobin (hemoglobin Ao) binds oxygen very strongly at physiological oxygen partial pressures 30 and would release little oxygen to res~il;l-g tissue. Accordingly, any substitute for hemoglobin must somehow correct the oxygen arLniLy and/or the Hill coefficient to physiologically meaningful levels (see e.g., Rausch, C. and Feola, M., US Patents 5,084,558 and 5,296,465; Sehgal, L.R., US Patents 4,826,811 and 5,194,590; E~offmAn et al., WO 90/ 13645; Hoffman and Nagai, US Patent 5,028,588; Anderson et al., WO
35 93/ 09143; Fronticelli, C. et al., US Patent 5,239,061; and De Angelo et al., WO
93/08831 and WO 91/16349).
It is not always practical or safe to transfuse a patient with donated blood. Inthese situations, use of a red blood cell ("RBC") substitute is desirable. When human blood is not available or the risk of transfusion is too great, plasma expanders can be W O ~G/10~0 PCT~US96/10420 administered. However, plasma expanders, such as colloid and crystalloid solutions, replace only blood volume, and not oxygen carrying capacity. In ~ Ation~ where blood is not available for transfusion, a red blood cell substitute that can transport oxygen in A~liti~n to providing volume replacement is desirable.
To address this need, a number of red blood cell substitutes have been developed (Winslow, R.M.(1992) Hemoglobin-based Red Cell Substitutes, The Johns Hopkins Univ~ ity Press, B~lhmore 242 pp). These sub~liLul~s inwude synthetic perfluorocarbon solutions, (Long, D.M. European Patent 0307087), stroma-free hemoglobin solutions, both whemically ~os~ k-o~ and unwos~lil ked, derived from a 10 variety of mAmm~ n red blood cells (Rauswh, C. and Feola, M., US Patents 5,084,558 and 5,296,465; Sehgal, L.R., US Patents 4,826,811 and 5,194,590; Vlahakes, G.J. et al., (l990)J. Thorac. Cardiovas. Surg. 100: 379 - 388) and hemoglobins ~ r~ssed in and purified from genetically engineered organisms (for example, non-t ly Lhlocyte cells such as bacteria and yeast, ~offmAn et al., WO 90/ 13645; bacteria, Anderson et al., 15 WO 93/09143, bacteria and yeast Fronticelli, C. et al., US Patent 5,239,061; yeast, De Angelo et al., WO 93/08831 and WO 91/16349; and transgenic m~mmAl~, Logan et al., WO 92/22646; Townes, T.M and McCune, S.L., WO 92/11283). These red blood cell substitutes have been designed to replace or A~ nt the volume and the oxygen carrying capability of red blood cells.
However, red blood cell replacement solutions that have been A~lmini~tered to AnimAl~ and hllmAn~ have exhibited certain adverse events upon administration.
These adverse reactions have included l.~ ion, renal failure, neurotoxicity, andliver toxicity (Winslow, R.M., (1992) Hemoglobin-based Red Cell Subshtutes, The Johns Hopkins Unlv~ ity Press, RAltimore 242 pp.; Biro, G.P. et al., (1992) Biomat., Art. Cells 25 ~ Immob. Biotech. 20: 1013-1020). In the case of perfluorocarbons, hypertension, activation of the reticulo-endothelial system, and complement activation have been observed (Reichelt, H. et al., (1992) in Blood Substitutes and Oxygen Carriers, T.M.
Chang (ed.), pg. 769-772; Bentley, P.K. supra, pp. 778-781). For hemoglobin-based oxygen carriers, renal failure and renal toxicity are the result of the formation of 30 hemoglobin c~/ ~ dimers. The form~ti~ n of dimers can be prevented by chemically crosslinking (Sehgal, et al., US Patents Nos. 4,826,811 and 5,194,590; Walder, J.A. US
Reissue Patent RE34271) or genetically linking (Hoffman, et al., WO 90/13645) the hemoglobin dimers so that the tetramer is ~l~v~ d from dissociating.
Prevention of dimer formation has not alleviated all of the adverse events 35 associated with hemoglobin administration. Blood pressure changes and ga:jLui~ slil.al effects upon administration of hemoglobin solutions have been attributed to vasoconstriction resulting from the binding of endothelium derivedrelaxing factor (EDRF) by hemoglobin (Spahn, D. R. et al., (1994) Anesth. Analg. 78:
1000-1021; Biro, G.P., (1992) Biomat., Art. Cells ~ Immob. Biotech., 20: 1013-1020;

W O ~G/~0320 PCTAJS96/10420 Van~egnff, K.D. (1992) Biotechnology and Genetic Engineering Reviews, Volume 10: 404-453 M. P. Tombs, Editor, Lll~1C~L Ltd., Andover, England). Endothelium derived relaxing factor has been identified as nitric oxide (NO) (Mon~ Ac~A, S. et al., (1991) Pharmacol. Rev. 43: 109-142 for review); both inducible and constitutive NO are primarily produced in the endothelium of the vac~llAtllre and act as local mo-llllAtors of vascular tone.
When hemoglobin is contained in red blood cells, it cannot move beyond the boundaries of blood vessels. Therefore, nitric oxide must diffuse to the hemoglobin in an RBC before it is bound. When hemoglobin is not contained within an RBC, such as 10 is the case with hemoglobin-based blood substitutes, it may pass beyond the endothelium lining the blood vessels and penetrate to the extravascular space (extravasation). Thus, a possible mechanism causing adverse events associated with administration of extracellular hemoglobin may be excessive inactivation of nitric oxide due to hemoglobin extravasation. Furthermore, NO is constitutively 15 synthesized by the vascular endothelium. Inactivation of NO in the endothelium and extravascular space may lead to vasoconstriction and the pressor response observed after infusions of cell-free hemoglobin. Larger hemoglobins may serve to reduce hy~el L~l sion associated with the use of some extracellular hemoglobin solutions.
In addition to the effects noted above, the dosage of non-polymeric 20 extracellular hemoglobin that can be A~lmini~tered may be limite~l by the colloidal osmotic ~les~lre (COP) of the solution. Administration of an extracellular hemoglobin composed of hemoglobin tetramers that would have the same grams of hemoglobin as a unit of packed red blood cells might result in a significant influx of water from the cells into the blood stream due to the high colloid osmotic pressure of 25 the hemoglobin solution. Polymeric hemoglobin solutions can be administered at higher effective hemoglobin dosages, because as the molecular weight increases, the number of the individual molecules is decreased, resulting in rerll1ce-1 COP (Winslow, R.M., (1992) Hemoglobin-based Red Cell Substitutes, The Johns Hopkins UniversityPress, Baltimore, pp 34-35).
Some higher m(~le~llAr weight hemoglobins occur in nature. For example, there are three mlltAnt~ of human hemoglobin that are known to polymerize as a result of formation of intermolecular (first tetramer to second tetrarner) disulfide bridges. Tondo, Biochem. Biophys. Acta, 342:15-20 (1974) and Tondo, An. Acad. Bras.
Cr., 59:243-251 (1987) describe one such mutant known as Hb Porto Alegre. Hb ~
35 Mi~ ippi is charActeri7e~ by a cysteine substitution in place of Ser CD3(44)~ and is believed to be composed of ten or more hemoglobin tetramers accordillg to Adams et al., Hemoglobin, 11(5):435-542 (1987). Hemoglobin Ta Li is charActeri7e~ by a ,1383(EF7)Gly Cys mutation, which showed slow mobility in starch gel electrophoresis, indicating that it too was a polymer.

W O 96/40920 PCT~US96/10420 There are a few known naturally occurring mutants of non-polyntf~ri7ing human hemoglobins that have a cysteine mutation that do not polymerize (Harris et al., Blood, 55(1):131-137 (1980)(Hemoglobin Nigeria); Greer et al., Nature [New BiologyJ, 230:261-264 (1971) (Hemoglobin Rainier). Hemoglobin Nunobiki (a 141 Arg ~ Cys) also features a non-polymerizing cysteine substitution. In both Hb Rainier and Hb Nunobiki, the mutant cysteine r~si~ s are surface cysteines.
Polymeric hemoglobins have also been reported in various v~ dtes and il~V~l l~Lates. Murine polymeric hemoglobins are described in Bonaventura & Riggs (Science, (1967)149:800-802) and Riggs (Science, (1965)147:621-623). A polym~ri7ing hemoglobin variant in macaque monkeys is reported in TAk~n~k~ et al., Biochem.
Biophys. Acta, 492:433-444 (1977); Ishimoto et al., J. Anthrop. Soc. Nippon, 83(3):233-243 (1975). Both amphibians and reptiles also possess polym~ri7ing hemoglobins (Tam et al., J. Biol. Chem, (1986) 261:8290-94).
Some inv~l l~l dte hemoglobins are also large multi-subunit ylOL~ s. The extracellular hemoglobin of the earthworm (Lumbricus l~r 1 e~ll is ) has twelve subunits, each of which is a dimer of structure (abcd)2 where "a", "b", "c", and "d" denote the major heme containing chains. The "a", "b", and "c" chains form a disulfide-linked trimer. The whole molecule is composed of 192 heme-containing chains and 12 non-heme chains, and has a molecular weight of 3800 kDa. The brine shrimp Artemza produces three polyrneric hemoglobins with nine genetically fused globin subunits (Manning, et al., Natzlre, (1990) 348:653). These are formed by variable association of two different subunit types, a and b. Of the eight intersubunit linkers, six are 12 residues long, one is 11 residues and one is 14 r~si~ s.
Non-polym~ri7ing crosslinked hemoglobins have been artificially produced.
For example, hemoglobin has been altered by chemically crosslinking the alpha chains between the Lys99 of alphal and the Lys99 of alpha2 (Walder, U.S. Patent Nos.
4,600,531 and 4,598,064; Snyder, et al., PNAS (USA) (1987) 84: 7280-84; ChaterJee, et al., J. Biol. Chem., (1986) 261: 9927-37). The beta chains have also been chemically crosslinked (Kavanaugh, et al., Biochemistry, ( 1988) 27: 1804-8). U.S. Patent No.
5,028,588 suggests that the T state of hemoglobin (collesponding to deoxygenatedhemoglobin) may be stabilized by intersubunit (but intratetrameric) disulfide ssli~lks resulting from substitution of cysteine residues for other residues.
Hemoglobin has also been artificially crosslink~l to form polymers. For example, U.S. Patent No. 4,001,401, U.S. Patent No. 4,001,200, U.S. Patent No.
4,777,244 and U.S. Patent No. 4,053,590 all relate to polym~ri7~tion of red blood cell-~ derived hemoglobin by chemical crnsslinking. The crosslinking is achieved with the aid of bifunctional or polyfunctional crosslinking agents, especially those reactive with exposed amino groups of the globin chains. Aldehydes such as glutaraldehydeand glycolaldehyde have been used to crosslink hemoglobin both intramolecularly WO ~'4~320 PCT~US96/10420 (within a tetramer) and intermolectll~rly (between tetramers). Intramolecular crosslinks serve to pl~Vellt dim~ri7~ti~n into alpha/beta dimers and may also alter oxygen affinity, while int~rmolecular cros~link~ create polymers of tetrameric hemoglobin. Polymeric hemoglobins may result in reduced ~Ldvdsation because of theirincreasedsize. ~P~tlce~lextravasationmay,inturn,leadto re~ ce~ ess~
effects resulting from infused hemoglobin solutions.
The result of these polym~ri~tion chemistries that have been used to ~losslillk hemoglobins is a polydisperse composition of covalently cros~link~-7 aggregates.Bucci, U.S. Patent No. 4,584,130, at col. 2, commPnt~ that "the polyhemoglobin 10 reaction products are a heterogeneous mixture of various molecular species which differ in size and shape. The molecular weights of these polyhemoglobins range from 64,500 to 600,000 Daltons. The separation of individual molecular species from the heterogeneous mixture is virtually impossible. In addition, although longer ret~n*~ n times in vivo are obtained using polyhemoglobins, the oxygen affinity thereof is15 higher than that of stroma-free hemoglobin."
It is well recognized that random polym~ri7~ti~ n is difficult to control and that a number of different polymers can be obtained, commonly between two and ten tetramers per polymer. For example, accol~ing to Tye, U.S. Patent No. 4,529,179,polym~ri7e(1 pyridoxylated hemoglobin has "a profound chemical heterogeneity 20 making it difficult to study as a ph~rm~celltical agent."
Furthermore, once hemoglobin is polym~ri~e~l, purification of specific m~ lec~ r weight fractions can be ~ccomplished using only molecular weight separation techniques. For example, tangential flow separation techniques can beused to s~aldl~ certain size ranges of polymerized hemoglobins. However the 25 membranes that are available for such separations are available only in a limited number of size ranges which allow the production of hemoglobins less than 100 kDa or greater than 300 kDa. In addition, such membranes are cumbersome, expensive, difficult to dean and the separation can be very slow.
Size exdusion chromatography (also known as, for example, gel filtration 30 chromatography or gel p~rm~tion chromatography) has also been used in the past to S~d dL~ hemoglobin molecular weight fractions. However, this technique is not suitable for large scale operation, and furthermore, does not provide good resolution for separation of molecular weight fractions (Simoni et al., (1993) Anal. Chim. Acta, 279: 7~88).
Simoni (1993, infra) also report the use of ion exchange chromatography to separate different molecular weight fractions of hemoglobin polymers. However, these workers noted that this kind of separation required differences in net charges.
In addition, they used a salt gradient elution to separate the different molecular W O 9~'4C3Z0 PCTAUS96/10420 weight fractions, and they did not demonstrate any significant resolution of tetramer, octamer and ~lerAm~r.
Correlations of m~le~llAr weight with serum half life for various ~roL~ills, such as IL-2, demonstrate that a significantly longer half life may be expected as the moleclllAr weight of a ~roLeil- increases, particularly above the renal filtration limit of 50-70 kDa. The use of cros~link~rs that can inhibit the degradation of hemoglobin tetramers into dimers that are readily cleared can also lead to increased serum half life.
Accordingly, a need exists for additional hemoglobin-like compounds having 10 these desired characteristics. In addition, a need exists for simple methods of creating specific molecular weight distributions in high molecular weight hemoglobin mixtures. The present invention sAti~fies these needs and provides related advantages.

Snm~ y of the Invention The present invention relates to modified hemoglobin-like coll,~o~ ds. In one aspect, the invention is directed to globin-like polypeptides having multiple dialpha 20 ~1OmAin~. Such polypeptides can cul~ . two dialpha domains, also referred to herein as "di-dialpha" domains, or more. These globin-like polypeptides can be linked by a peptide linker having at least five amino acids between the dialpha domAin~, ~Ler~ldbly at least seven amino acids. Preferably, the linkers are encoded by a peptide linker having Ser-Gly-Gly as a repeat unit, such as the amino acid sequences:
25 Ser-Gly-Gly-Ser-Gly-Gly-Ser (SEQ.ID.NO.1); Gly-Gly-Ser-Gly-Gly-Ser-Gly-Gly-Ser-Gly-Gly-Ser-Gly-Gly (SEQ.ID.NO. 2) and Ser-Gly-Gly-Ser-Gly-Gly-Ser-Gly-Gly-Ser-Gly-Gly-Ser-Gly-Gly-Ser (SEQ.ID.NO. 3). The globin-like polypeptides can be recombinantly ~ressed in a host cell, such as E. coli. Di-dibeta globin-like polypeptides are analogously defined, and are a further aspect of the instant 30 invention.
The invention also relates to nucleic acid molecules having a nucleic acid sequence encoding such globin-like polypeptides. In one embodiment, the nucleic acid molecules encode a globin-like polypeptide having two dialpha domains and aseparate polypeptide having a single beta domain or a di-beta domain.
In another aspect, the present invention relates to a multimeric hemoglobin compound that comprises two dialpha globins that are connected through a peptidelinker, wherein only one of the four alpha globin domains COll~illS a non-naturally occurring cysteine residue (mono-cys di-dialpha). In a further aspect of this invention, such mono-cys di-dialpha-containing hemoglobin composition can be WO ~G/~0~20 PCTAUS96/10420 crosslinked directly or indirectly to another identical mono-cys di-dialpha or any other suitable hemoglobin-like mQle~ll~ Mono-cys di-dibeta m~le~ll~s are analogously ~l~finf~l and also can be cros~linkPr~ as described herein.
rn another aspect, the present invention also provides mlll*m~ric hemoglobin-5 like ploleills in which a first hemoglobin-like moiety is dile~Lly attached to two or more other hemoglobin-like moieties. Compositions conLai~ g such multimeric hemoglobin-like ~LoL~ills are also provided.
In a further aspect, the present invention relates to methods for m~king the multimeric hemoglobin-like ~roL~ills. The methods are ~ccomrlished by:
(a) obtaining a first hemoglobin-like moiety having amino acids capable of hing to one end of a heterobifunctional linker to form a core hemoglobin-like moiety;
(b) obtaining at least two other hemoglobin-like moieties having an amino acid capable of attaching to the other end of the heterobifunctional linker;
(c) contacting the heterobifunctional linker to the first hemoglobin-like moiety;
and (d) adding the other hemoglobin-like moieties to form the mlll*m~ric hemoglobin-like pluL~
In a still further aspect, the present invention relates to methods for separation of molecular weight fractions of polym~ri7e~ hemoglobin or hemoglobin-like molecules to obtain substantially monodisperse hemoglobin solutions. Such methods are ~cc-mplished by:
(a) con~c*ng a polydisperse mixture of polym~ri7e~1 hemoglobin-like mole~ll~c: with an ion exchange matrix;
(b) washing the ion exchange matrix with a first buffer;
(c) eluting the ion exchange matrix with a second buffer which may be the same or different than said first buffer to obtain a substantially monodisperse hemoglobin-like solution.

Detailed Description of &e Invention The present invention generally relates to hemoglobin-like compounds comprised of novel globin-like polypeptides or hemoglobin-like ~roL~ills. These compounds contain various modifications to the naturally-occllrring hemoglobins,particularly human Hb Ao. In a further aspect, the present invention relates to methods of purifying such hemoglobin-like molecules and other polymeric hemoglobin-like mcle~lles.
As described above, most naturally-occurring human hemoglobins are constructed of four non-covalently linked polypeptide chains: two chains cont~ining W O ~6/~03~0 PCTAJS96/10420 identical alpha domains and two chains containing identical beta ~7~mAin~. The novel globin-like polypeptides of the present invention, however, cc,. ,1 ~ at least two dia7pha (or two dibeta) ~l~)m~in.~ in a single polypeptide chain. A "dialpha domain"
(or "dibeta ~7omAin") consists of two alpha (or beta) domains (or polypeptide 5 sequences) connected between the C-terminus of a first alpha domain (or beta m~in) and the N-~ s of a second alpha domain (or beta domain) as described in PCT Publication No. WO 93/ 09143, incol~oldted herein by reference. Thus, thenovel globin-like polypeptides have as a,..i..i...l7m four alpha (or beta) domains per polypeptide.
As used herein, the term "globin-like polypeptide" means a polypeptide having a domain that is substantially homologous with a globin subunit of a natura ly occLlrring hemoglobin. For example, a globin-like polypeptide containing two dialpha domains means that each of the four a pha domains is substantially homologous to a native a7pha globin or a llllll~lt thereof differing from the native 15 sequence by one or more substitutions, deletions or insertions, while rPmAining substantia7.1y homologous with the native alpha globin and l~ldilullg its ability to associate with a beta globin. As used herein, the term "alpha domain" is intended to include but not be limited to naturally occurring alpha globins, including normal human alpha globin, and mutants thereof. A "beta domain" is analogously defined.20 Subunits of V~:l L~l ate and invt l L~l ate hemoglobins or mutants thereof which are sufficiently homologous with human a7pha or beta globin are embraced by the terms "alpha or beta ~om~in~." For example, the subunits of bovine hemoglobin are within the scope of these terms.
In ~7~l~., . . i . Ullg whether an alpha or beta globin contemplated by the present 25 invention is substantially homologous to a particular wild-type a7pha or beta globin, sequence ~imil~rity is an important but not exclusive criterion. Sequence ~imilArity may be d~l~l...il.ed by col-velllional a g~ llls~ which typically allow introduction of a small number of gaps in order to achieve the best fit. An alpha domain contemplated for use in the present invention will typically have at least about 75~O
30 sequence identity with wild-type human alpha globin, and greater homology with human alpha globin than with human beta globin. However, a polypeptide having analpha domain of lesser sequence identity may still be considered "substantially homologous" with a wild-type alpha globin if it has a greater sequence identity than would be expected from chance and also has the charA~ tPri~tic higher structure (e.g., 35 the "myoglobin fold") of alpha globin.
- ~lltAti~n~ can be introduced to alter the oxygen affinity (or cooperativity, or activity with respect to pH, salt, Leln~ldlule, or other ellvilul~lllental parameters) or stability (to heat, acid, allcali, or other denaturing agents) of the hemoglobin, to facilitate genetic fusion or cros~linking~ or to increase the ease of expression and assembly of the individual chains. Glli~Ance as to certain types of mlltAtinnc is provided, for example, in U.S. Patent No. 5,028,588 and PCT Publi~ Atinn No. WO
93/09143, both incorporated herein by reference. The present invention further includes molecules which depart from those taught herein by gratuitous mutations5 that do not substantially affect biological activity.
The dialpha (or dibeta) domains of the novel globin-like polypeptides can be ~onneclP~l by various means known in the art. For example, the domains can be coupled by a peptide linker between any two dialpha domains. A discussion of suitable distances is also provided in WO 93/09143, incolyclldLed herein by reference.
10 With knowledge of these distances, one skilled in the art can readily ~le~;,~e, for example through mole~llAr modeling, the useful lengths of suitable peptide linkers.
Particularly useful peptide linkers have at least five amino acids, ~lereldbly at least seven amino acids. The peptide linker can have an amino acid sequence that contains Ser-Gly-Gly as a repeating unit, as in the following illu~LLdLive amino acid sequence:
Ser-Gly-Gly-Ser-Gly-Gly-Ser (SEQ.ID.NO. 1). Examples of other amino acid sequences useful as peptide linkers containing this repeating unit include: Gly-Gly-Ser-Gly-Gly-Ser-Gly-Gly-Ser-Gly-Gly-Ser-Gly-Gly (SEQ.ID.No. 2) and Ser-Gly-Gly-Ser-Gly-Gly-Ser-Gly-Gly-Ser-Gly-Gly-Ser-Gly-Gly-Ser (SEQ.ID.No. 3).
The multiple dialpha domains and the peptide linkers of the globin-like polypeptides can be g~n~tit Ally fused through recombinant methods known in the art or as described, for example, in WO 93 / 09143 or in the Examples below. The ~yaldLion of a single dialpha globin as an int~rme~liAt~ product is also described in this publication.
The globin-like polypeptides can be used to ~/l'e~dl~ hemoglobin-like pseudomers. Such pseuflQm~ric Hb-like ~LoLeins are described in WO 93/09143.
Pseudomeric hemoglobin-like ~ruL~ins have at least one more domain than the number of polypeptide chains, i.e., at least one polypeptide chain contains two or more globin-like ~lomAin~
It is also possible to introduce non-naturally occurring cysteine residues into one alpha subunit of a dialpha ~lom~in or one alpha subunit of a di-dialpha domain or larger dialpha domains to ~~ other pcell~lnm~ri~ hemoglobin-like yroL~ins.
Preferably these non-naturally occtlrring cysteine residues are asymmehic, that is they occur in only one alpha domain of the longer di-dialpha polypeptide. Such mutations can also be incorporated in an analogous fashion in di-dibeta globins. The asymmett ic cysteine residues can then be used to form direct disulfide bridges connecting the dialpha (or dibeta) ~lnmAin~ or .losslil ked by coupling reagentsspecific for cysteine residues to produce the larger psel-~ m~ric Hb ylOL~il s .The hemoglobin-like psell~lnm~rs can be purified by any suitable purification method known to those skilled in the art. Useful purification methods for the hemoglobin-like ~rol~ s of the present invention are taught in PCT PublicAtion WO
95/ 14038, incorporated herein by reference. Briefly, the methods described therein involve an immobilized metal affinity chromatography resin charged with a divalent metal ion such as zinc, followed by anion exchange chromatography. Accc,l .ling to this publication, the solution cullldil~illg the desired Hb-c~ illillg mAt~riAl to be purified can first be heat treated to remove protoporphyrin IX-conldillil,g Hb. This basic purification method can be further followed by a sizing column (S-200), then another anion exchange column. Alternatively, this solution can be s~dldl~d intomolecular weight fractions using ion exchange chromatography accoLdil,g to the methods of the instant invention. The resulting solution can then be buffer exchanged to the desired formulation buffer.
The invention further provides nucleic acids encoding the novel polypeptides of the present invention. Those skilled in the art can readily derive a desired nucleotide sequence based on the knowledge of published nucleotide or amino acidsequences of known hemoglobin subunits with selection of codons and control elements specific for the org~ni~m used for e,~les~ion, using methods known in the art. For example, the amino acid sequence of the dialpha domain and the beta domain of a synthetic hemoglobin can be used to derive the nudeic acids of the present invention, both of which are identified in Figure 12 of PCT Publication WO 90/ 13645, incorporated herein by reference, with the following corrections to the nucleotide sequence: bases 55, 56 and 57 (codon 19) should read GCG and bases 208 and 209 (the first two bases of codon 70) should read GC. The following changes to the amino acid sequence of this figure would yield the pseudoleLdll,er, rHbl.l: the gly-gly bridge at residues 142 and 143 of the dialpha .11 mAin can be changed to a single gly residue bridging a I and a2 domains; residues 54 and 97 of the dialpha domain should read Gln; residue 70 of the beta subunit should read Asn; and residue 107 of the betasubunit should read Lys. The pseudoLeLldlller, rHbl.l is also described in Looker et al., Nature, 356:258-260 (1992), incorporated herein by reference.
The nucleic acids of the present invention can be used to construct plasmids to be inserted into a~ Liate recombinant host cells accol ding to conventional methods or as described in the Examples below. Any suitable host cell can be used to express the novel polypeptides. Suitable host cells include, for example, bacterial, yeast, nlAmmAliAn and insect cells. E. coli cells are particularly useful for expressing the novel polypeptides. Preferably, when multiple subunits are expressed in bacteria, it is desirable, but not required, that the subunits be co-expressed in the same cell - polycistronically as described in WO 93/09143. The use of a single promoter is eLe.dble in E. coli to drive the expression of the genes encoding the desired ~rol~ills.
The present invention is also directed to novel multimeric hemoglobin-like L~illscontaining at least three hemoglobin-like moieties, of which at least one is directly attached to the other moieties. The term "hemoglobin-like moiety" includes tetramers having four globin-like domains composed of two alpha ~ m~inc and two beta domains and pse~ om ~ri c hemoglobin-like ~,oLe, ~ ~s as previously ~l~fin~l The hemoglobin-Iike moiety that is directly attached to the other hemoglobin-like m~ieti~s 5 is referred to herein as the "core hemoglobin-like moiety" or "core moiety" while the other hemoglobin-like moieties are referred to as the "surrounding hemoglobin-like moieties" or "SL LloLLlLding moieties."
In one embo-lim~nt, the core moiety is different from the surrounding hemoglobin-like moieti~s, which in turn can be the same or different from each other.
10 Such multimeric hemoglobin-like ~Lol~ins are referred to as heL~lol..llltimeric hemoglobin-like ~rol~ins (or heteromers). For example, the core moiety can be rHbl.1, while the surrounding moieties can be mutants reL~l.ed to as K158C. The pseudotetramer, rHbl.1, is described in WO 90/ 13645, incu.~u- dted herein by reference. K158C is a mutant moiety of rHbl.1 and is composed of three 15 polypeptides, one l~o~ g two alpha domains (a dialpha) and the other two eachcQ,.~ g a single beta domain. A single lysine to cysteine substitution in the second alpha domain of the dialpha component appears at amino acid residue 158 of the K158C dialpha sequence. Note that because rHbl.1 consicts of a dialpha molecule (two alpha subunits, each 141 amino acids in length, connected by a single glycine) 20 mutations in the second subunit are denoted by the position with respect to the N
1~.... i l lll s of the dialpha, and not the alpha subunit. Thus the mutation at position 158 is a mutation in the second alpha globin ~ m~in, co..es~onding to position 16 innormal alpha globin. A general method for obLdLLLillg a moiety having one or more asymmetrical cysteine mutations and the desirability of such asymmetrical crosslinked mutants are provided in WO 93/09143, which is specifically incorporated herein by reference. The publication also provides guidance for s~lerting other candidate sites for substitution on the alpha or beta domains.
The core and surrolm~ling moieties can be directly attached by any means known in the art, induding without limitAtil~n the use of chemical crosslinkers. Such linkers are ~isrllsse~l in Wang, S.S. (1993) Chemistry of Protein Conjugation and Cross-lin7ang, CRC Press. Other suitable crosslinking methods are described, for example in Var -l~griff, K.D.(1992) Biotechnology and Genetic Engineering Reviews, Volume 10: 404-453 M. P. Tombs, Editor, Intercept Ltd., Andover, England; and Winslow, R.M.(1992) Hemoglobin-based Red Cell Substitutes, The Johns Hopkins Ul-iv~ iLy Press, Baltimore 242 pp. Such crosslinking chemistries are generally linkers conLd.l~ g two or more fimction~l groups. These functional groups can be the same or different (i.e., homobifunctional linkers, heterobifi-nction~l link~rs, homopolyfunctional linkers, or heL~ ~olyflm~tion~l linkers and can furth~rmc-re be d~n~rim~ric, branched or contain armed cores) and include, for example, bis-imidoesters, bis-sllcrinimi~yl W O 9f'10920 PCTrUS96/10420 esters, oxidized ring structures of sugars or nucleotides, crosslinkers conLdil~ing haloacetyl or vinyl sulfone functional groups, and dialdehyde and polyaldehyde crosslinkers, such as glycolaldehyde and glutaraldehyde.
For heL~ivil~ultimeric hemoglobin-like ~roL~ins, a heterobifurlc~ionAl chemical crosslinker is ~r~L~lled, such as a sucrinimidyl 4-(N-maleimidomethyl) cyclohexane-l-carboxylate (SMCC) or N-y-maleimidobutyrloxysuccinimide ester (GMBS).
Preferably, the heterobifunctional chemical cros~link~r is one that does not elicit a sigrlificant immunogenic response. Other useful heterobifunctional crosslinkers are described in WO 93/09143, incorporated herein by reference.
In the case of GMBS or SMCC, for example, the sllc-inimide of these compounds can be used to attach to the lysine residue of the non-cysteine mutated hemoglobin-like ~roL~ins, such as rHbl.l (the core moiety). The maleimide can beused to attach to the cysteine of the hemoglobin-like ~roL~ill collLdi.lillg a cysteine mutation, such as K158C. By first reacting linkers with the core moiety, then A~lcling the desired amount of cysteine-~ Ldi~ling mutant, various forms of these multimeric hemoglobin-like ~roL~ins can be made, for example a trimeric, tetrameric, pentameric and higher order multimeric ~ruLeins. Factors that constrain the number of hemoglobin-like moieties that can be AttA~h~l to the core moiety include steric hindrance as additional s~ ounding moieties are added and the number of residues20 that are available for AttArhing to the crosslinkers. Methods for identifying and using such crosslinkers are known to those skilled in the art or as described in the Examples below.
In a further embodiment, the core moiety and the surrounding moieties can be the same moiety, which are referred to herein as "homomultimeric hemoglobin-like25 ~roL~ills." An example of a homomultimeric Hb-like ~LoLei" is one which is composed of only K158C mllt~nt~
For m~king the mtlltim~ric hemoglobin-like ~roL~ins of this embodiment, the formAtion of substantial amounts of polymPri7~ oL~ills is ~,~L~,dbly avoided.
Polym~ri7e~l ~r~,L~:il-s collLdin Hb-like moieties that are indirectly attached to the core moiety through AttA~hm~nt to an inL~l vel~ng hemoglobin-like moiety and are generally formed by uncontrolled crosslinking reactions. Accordillg to the methods of the instant invention, such rAnciom polymerization is reduced by coupling of specific reactive sites on the core hemoglobin ~roleill to certain sites on the surrounding ~ hemoglobin-like molecules. Any method known in the art can be used in which site specific attachment can be achieved.
The present invention also provides methods for making homomultimeric hemoglobin-like ~l~,teills, that is a mllltim~ric hemoglobin-like ~roL~in composed of a core molecllle that is the same as the surrollntling molecllles. These methc ~1~ are Arcomrlished by the use of a h~L~lobir mctional crosslinker and a protective moiety, for example, borate. Alternatively, reaction conditions with any of the crosslinkers can be modified by altering for example, CQnc~l ~1. dLions, temperatures or reaction time such that the degree of polymr-ri~tion is constrained.
Through the use of an a~plo~Liate amine/sulfhydryl heterobiflmction~l 5 crosslinker, a desired hemoglobin-like moiety, for example, rHbl.1, can be modified so that it will subsequently react with several rHbl.1 molecules bearing surfacecysteine mutations as described in WO 93/09143. This reaction is achieved, for example, by first reacting the amine function~liti~ on unmodified rHbl.1 with the surrinimide moiety of a heterobif~lnctio~7~l crosslinker in a sodium borate buffer at 10 pH 8.5. Reaction with lysine residues on rHbl.1 leads to loss of the sllcrinimide group of the heterobifunctional crosslinker by the formation of a stable amide linkagebetween the crosslinker and the hemoglobin. The unreacted maleimide residues of the heterobifunctional crosslinker are highly reactive towards sulfhydryl groups. The intrinsic sulfhydryl groups of rHbl.1 are prevented from reacting with the maleimide 15 moiety of the heterobifunctional cro~link~r by either their in~cr~sihility or by forming a complex with borate. After reaction with the sllrrinimide group of theheterobifunctional crosslinker, the hemoglobin molecule can be considered to be "activated" at multiple surface lysine residues towards rf~ctinn with the surface sulfhydryl residue of, for example, a K158C hemoglobin mutant because the core 20 moiety now has reactive m~ imide residues attached to it.
By using an ~r~ ;ate conc~l.LldLion of crosslinker and reaction time, which can be ~ i empirically by those skilled in the art, the reaction with surface cysteine-containing hemoglobin (e.g., K158C) with the activated core hemoglobin molecule yields higher molecular weight hemoglobins. The polymers that are formed 25 by reaction of the activated hemoglobin and the cysteine-cont~ining hemoglobin mutants have a distribution of apparent molecular weights. However, the distribution of molecular weights can be constrained to a certain extent by the extent of initial activation with the heterobifunctional ~losslinker coupled with the use of certain moieties, such as, for example, K158C. The site-directed nature of the reaction 30 with, for example, K158C limits the molecular weight distribution to pre~nmin~ntly p~nt;~m~ric hemoglobin. It is believed that the manipulation of reactivity, such as sulfhydryl reactivity, through formation of a reversible complex with a suitableL~:~Liv~ buffer, such as borate/boric acid for certain mutants, is a novel method for controlling reactivity, such as sulfhydryl reactivityj in f~rming the multimeric35 hemoglobin-like ~ eills of the present invention.
AccoL.lingly, the present invention further provides methods for making a mllltimeric hemoglobin-like t~ ill. The methods are ~--complished by:

-W O 96/40920 PCTrUS96/10420 (a) oL,Ldil il.g a first hemoglobin-like moiety having an amino acid capable of attaching to one end of a heterobifunctional linker to form a core hemoglobin-like moiety;
(b) obldillillg at least two other hemoglobin-like moieties having an amino acidcapable of attaching to the other end of the heterobifunctional linker;
(c) contacting the heterobifunctional linker to the first hemoglobin-like moietyto form a linked moiety; and (d) coI~t~cting the other hemoglobin-like moieties to the linked moiety to form the multimeric hemoglobin-like ~loleill.
The invention further provides compositions containing the novel mllltimenc hemoglobin-like ~l oteil,s of the present invention and the globin-like polypeptides, including ~lol~-s containing such polypeptides. In compositions containing the multimeric hemoglobin-like ~loL~ills, a polydisperse composition containing various multimeric ~loL~il,s can be obtained, i.e., differing species of trimerics, tetramerics, 15 pentamerics and so forth. In addition, these compositions collldinil,g the multimeric hemoglobin-like ~loLeills are ~leL~ldbly substantially free of polymerized ~loL~ins, although they need not be completely free depending on the intended use of the desired ~rol~il.s. As used in this context, "substantially free" means the presence of polymerized ~ leills will not adversely affect the desired function of the multimeric hemoglobiniike ~loLeil.s. Furth~nn( re, these multimeric hemoglobin-like ~roL~ sare substantially monodisperse. As used herein, "substantially monodisperse" means that there is less than 30% hemoglobin that is not the desired molecular weight.Accc.ldil.gly, in a substantially monodisperse high molecular weight hemoglobin solution, less than 30% of the hemoglobin is not the target high molecular weight hemoglobin that is desired. Note that the target monodisperse high molecular weight hemoglobin can comprise mixtures of high molecular weight, such as trimers, tetramers and pPnt~mPr~. Likewise, in a substantially monodisperse pentahemoglobin solution, less than 30% of the hemoglobin in the solution is notpentahemoglobin. Preferably, a monodisperse high molecular weight hemoglobin solution collLdills less than 25~o non-target hemoglobin, more ~ler~idbly less than 20 non-target hemoglobin.
After ~os~linking, regardless of ~osslinking technology that is utilized, ion exchange chromatography is used to se~d~dL~ hemoglobin polymers by molecular weight according to the methods of the instant invention. Typically ion exchangechromatography is used to se~a~dL~ pn~L~il-s according to differences in isoelectric - point. Surprisingly, the inventors have found that ion exchange technology can be used to separate hemoglobins that have no measurable difference in isoelectric point.
For example, the isoelectric point of rHbl.1 that has been ~lvsslinked with glutaraldehyde is approximately 7.05 for mono-hemoglobin (1 tetramer) di-WO ~/10320 PCT~US96/10420 hemoglobin (2 tetramers), tri-hemoglobin, and higher order m~lltimPrs. Nevertheless, such hemoglobins were resolved using the methods of the instant invention (see, for example, Example 12 herein).
According to the instant invention, the purification of the polydisperse hemoglobin solution is accomrlished as follows. The polydisperse hemoglobin solution is transferred into a buffer compatible with the ion exchange matrix ifrequired. A suitable buffer is, for example, 20mM Tris, pH 8.0-8.9 at 8~C. The polydisperse hemoglobin solution is then loaded onto an ion exchange matrix. Such ion exchange matrices can be any suitable support, for example membranes or resins 10 that are anion or cation exchange matrices. A particularly suitable exchange resin can be, for example, a Q-SEPHAROSE fast flow anion-exchange column (PhArmA~-a Biotech, Uppsala, Sweden). Alternate anion-exchange resins include, for example,Super Q 650 C or Toyopearl QAE-550C (TosoHaas Inc., Mul~L~,olllery, PA) or Macro-Prep Q Support (Bio-Rad Inc., Hercules, CA). The amount of ~r~ l~in that can be 15 loaded on the column can be varied depending on the binding capacity of the column and the mix of molecular weights desired. The flow rate of the column will depend on the type of column and resin used for the chromatography. Typically, for a 450ml resin bed packed in an XK-50 column (Pharmacia Biotech, Uppsala, Sweden) a flow rate of 200cm / hr is used.
After the column is loaded with the polydisperse hemoglobin solution, it is washed with sufficient column volumes of buffer to remove unbound ~>r~)leil- from the column matrix. Such washes can be, for example, 2-3 column volumes (CV's) of 20 mM Tris buffer, pH 8.9 at 8~C. Alternatively, the column can be washed until thedesired ~rot~ concentration in the eIuent is rPAt he~l This can be ~lei f l l l l i .~ed by the 25 absorbance at 215 or 280nm, or by other suitable monitoring techniques. Next, the column is washed with the desired buffer system. This buffer system can include combinations of buffer, buffer concentrations and/or salt to elute the desired ~rol~ill.
Elution can occur utilizing any suitable elution scheme, for example by isocratic elution, stepwise isocratic elution, stepwise gradient elution or gradient elution. A
30 particularly suitable elution scheme is by stepwise isocratic elution. Det~rminAtion of suitable washes, elution buffers and elution s~-h~m ~s can be readily ~L.-. ., . ; . led by one of skill in the art using the g~ Ance provided herein.
For the purpose of removing glutaraldehyde crosslinke~ ~rolt:ills with m~lec~lAr weights ~190kDa the column can be washed with, for example, 11 CV's of35 20mM Tris buffer, pH 7.6 at 8~C. Using this system, the desired hemoglobin polymer fraction is then eluted with 20mM Tris, pH 7.4 at 8~C. Likewise, monomeric hemoglobin can be removed from a p~ntAh~moglobin solution formed using a polyfunctional crosslinkf~r (see Example 18) using, for example, 7-8 CV wash 25 mM
Bis-Tris/Tris pH=7.5 at 8~C followed by elution with 25 mM Bis-Tris/Tris, 100mM

W O 96/40920 PCT~US96/10420 NaCl pH=7.5 at 8 ~C. The hemoglobin molecular weight fraction of intt:le~l can then formulated as desired, or futher purified by, for example, ultrafiltration.
The hemoglobin-like ylOL~ s and compositions cc,l~ g the globin-like polypeptides or the multimeric hemoglobin-like ~ L~ s (collectively "hemoglobins") 5 can be used for in vitro or in vivo applications. Such in vitro applications include, for example, the delivery of oxygen by compositions of the instant invention for theenhancement of cell growth in cell culture by maintaining oxygen levels in vitro(DiSorbo and Reeves, PCT publication WO 94/22482, herein incorporated by reference). M~l~uvt:l, the hemoglobins of the instant invention can be used to remove 10 oxygen from solutions requiring the removal of oxygen (Bonaventura and Bonaventura, US Patent 4,343,715, incorporated herein by rererence) and as rerelence standards for analytical assays and instrumentation (Chiang, US Patent 5,320,965, incol~oldL~d herein by reference) and other such in vitro applications known to those of skill in the art.
In a further embodiment, the hemoglobins of the present invention can be formulated for use in th~ldy~:uLic applications. Example formulations suitable for the hemoglobin of the instant invention are described in Milne, et al., WO 95 / 14038 and Gerber et al., PCT/ US95/ 10232, both herein incorporated by reference.
Ph~rmAcel1tical c~,ll,yositions of the invention can be ~lmini~tered by, for example, subcutaneous, intravenous, or il~Ldll~Llscular injection, topical or oral ~-lmini~tration, large volume parenteral solutions useful as blood substitutes, etc. Ph~rm~relltical compositions of the invention can be ~lmini~tered by any col-vel-Lional means such as by oral or aerosol ~lmini~tration, by transdermal or mucus membrane adsorption, or by injection.
For example, the hemoglobins of the present invention can be used in compositions useful as substitutes for red blood cells in any application that red blood cells are used or for any application in which oxygen delivery is desired. Such hemoglobins of the instant invention formulated as red blood cell substitutes can be used for the treatment of h~m(~rrhages, traumas and surgeries where blood volume is lost and either fluid volume or oxygen carrying capacity or both must be replaced.
Moreover, because the hemoglobins of the instant invention can be made ph~rm~reutically acceptable, the hemoglobins of the instant invention can be used not only as blood sub~LiLuL~ that deliver oxygen but also as simple volume expandersthat provide oncotic pressure due to the presence of the large hemoglobin ylol~im- leclllP In a further embodiment, the crosslinked hemoglobin of the instant invention can be used in situations where it is desirable to limit the extravasation or reduce the colloid osmotic pressure of the hemoglobin-based blood substitute. The hemoglobins of the present invention can be synthe~i 7e~ with a high molecular weight. Thus the hemoglobins of the instant invention can act to transport oxygen as WO 96/40920 PCT~US96/10420 a red blood cell substitute, while reducing the adverse effects that can be associated with excessive extravAcAti- n A typic~ dose of the hemoglobins of the instant invention as an oxygen delivery agent can be from 2 mg to 5 grams or more of extracellular hemoglobin per 5 kilogram of patient body weight. Thus, a typical dose for a hllmAn patient might be from a few grams to over 350 grams. It will be appreciated that the unit l~o~ l of active ingredients rontAine-7 in an individual dose of each dosage form need not in itself constitute an effective amount since the n~c~.cs~ y effective amount could be reached by Aclmini~ctration of a plurality of A~lminictrations as injections, etc. The 10 selection of dosage depends upon the dosage form utilized, the condition being treated, and the particular purpose to be achieved according to the d~l~. ",;"Ation of the skilled artisan in the field.
Administration of the hemoglobins of the instant invention can occur for a period of seconds to hours depending on the purpose of the hemoglobin usage. For15 example, as an oxygen delivery vehicle, the usual time course of administration is as rapid as possible. Typical infusion rates for hemoglobin solutions as blood replAc~m~nts can be from about 100 ml to 3000 ml/hour.
In a futher embo~iim~nt~ the hemoglobins of the instant invention can be used to treat anemia, both by providing A~l~litionAl oxygen carrying capacity in a patient 20 that is suffering from anemia, and/or by stimlllAting h~mAt( poiesis as described in PCT publication WO 95/24213. When used to stimlllAt~ hematopoiesis, Atlminictration rates can be slow because the dosage of hemoglobin is much smaller than dosages that can be required-to treat hemorrhage. Therefore the hemoglobins of the instant invention can be used for applications requiring A~minictration to a patient 25 of high volumes of hemoglobin as well as in situations where only a small volume of the hemoglobin of the instant invention is A~minictered.
Because the distribution in the vAcclllAtllre of the hemoglobins of the instant invention is not limited by the size of the red blood cells, the hemoglobin of the present invention can be used to deliver oxygen to areas that red blood cells cannot 30 penetrate. These areas can include any tissue areas that are located downstream of obstructions to red blood cell flow, such as areas downstream of thrombi, sickle cell occlusions, arterial occlusions, angioplasty balloons, surgical instrumentation, any tissues that are suffering from oxygen starvation or are hypoxic, and the like.
Additionally, any types of tissue ischemia can be treated using the hemoglobins of the 35 instant invention. Such tissue ischemias include, for example, stroke, emerging stroke, transient ischemic attacks, myocardial stunning and hibernation, acute or unstable angina, emerging angina, infarct, and the like. Because of the broad distribution in the body, the hemoglobins of the instant invention can also be used to deliver drugs and for i~z vivo imaging.

The hemoglobins of the instant invention can also be used as replacement for blood that is removed during surgical procedures where the patient's blood is removed and saved for reinfusion at the end of ~ y or during recc~v~:.y (acute normovolemic hemodilution or h~mo~llgmPnt~tion). In addition, the hemoglobins of5 the instant invention can be used to increase the amount of blood that can be predonated prior to SULg~ly, by acting to replace some of the oxygen carrying capacity that is donated.
Under n~lrm~l physiological conditions, nitric oxide is not produced in excess amounts. However, certain disease states are associated with excess nitric oxide10 production. Such conditions include septic shock and hypotension. In these cases, the crosslinked hemoglobin of the present invention can be used to remove excessnitric oxide from the vasculature or to remove any other ligand that is found in toxic excess and that can be bound to the hemoglobins of the instant invention.
The following Examples are int~n~le-l to illustrate, but not limit, the present 15 invention.

Example 1 Production of Protein Solution Containing Modified Hemoelobin A. Construction of a Bacte~al Svstem for the Production of Modified rHbl.1 On January 20, 1994 E. coli strain SGE1661 was deposited with the American Type Culture ~~ollecfion (ATCC Arc~ion Number 55545). Note that Strain SGE1661 carrying the plasmid pSGE705 was denoted SGE1662. pSGE705 was a medium copy 25 number plasmid because it resulted in a~ro~ ately 100 copies of the plasmid per cell. The pl~mi~l~ used in ~r~d~ g pSGE705 are identified in Table 1, which alsoprovides a brief description of each.
Materials. pBR322, pUC19 and pNEB193 were purchased from New England 30 Biolabs, Beverly, Massa~ hll~ett~. Oligonucleotides were synthesized on an Applied Bio:iy~ ..ls DNA Synthesizer Model 392. The oligonucleotides used in preparing pSGE705 are listed in Table 2. Restriction endonucleases were purchased from NewEngland Biolabs (Beverly, Massachusetts) and used accor-li--g to manllf~chlrer's~ specifications. T4 DNA Ligase was purchased from either New England Biolabs or 35 Gibco-BRL (Gaith~l~urg, Maryland) and used accolding to manl1f~ hlrer's specifications. Pfu polymerase was purchased from Shratagene (La Jolla, California) and used acc~..di..g to m~nllf~chlrer's specifil :~hon~
Media used to culture the shrains are described in J. H. Miller, Experiments in Molecular Genehcs. (Cold Spring Harbor Press 1972). and J. H. Miller, A Short Course in Bacterial Genetics. (Cold Spring Harbor Press 1992). Addine orange, ampicillin and kanamycin sulfate were purchased from Sigma Chemical Co. (St. Louis, Missouri).
Tetracycline was purchased from Aldrich Chemicals (Milwaukee, Wisconsin).

Table 1. Plasmids PLASMID DESCRIPTION
pSGEl.lE4 rHbl.l ~ ression plasmid coL.ldi~ g dialpha and beta genes pSGEl.lE5 like pSGEl.lE4 but ampicillin resistant instead of tetracycline resistant pSGE490 pUCl9 lacI on a Bam HI-Hind III fragment pSGE491 pUC19 a on an Eco Rl-Xba I fr~gment pSGE492 pNEB193 Ptac- a pSGE493 pUCl9 ~ on an Xba I-Hind m fragment pSGE500 pUCl9 a ,~ on a Bam HI-Hind III fr~gm~nt pSGE504 pSELECT-1 replace Sty I with a Pme I site pSGE505 pSGE504 rrnB T1 transdptional terminator in the Eco RI-Cla I
sites pSGE507 ColE1 ori and tet, 2213 bp pSGE509 ColEl ori tet lacI, 3425 bp pSGE513 ColE1 ori tet lacI a ,B, 4386 bp pSGE515 ColEl ori tet lacI dia ,B, 4812 bp pSGE700 pTZ18U+dia,B frompSGE515 pSGE705 modified rHbl.l expression plasmid, ColE1 ori, tet, lacI, dialpha and beta genes pTZ18U a phagemid derivative of pUCl9, for oligonucleotide directed mutagenesis pDLII-9lF pGEM1 + a mi~sing valine in 2nd position (Des-val) pNEB193 Like pUCl9 but has more restriction sites in the multi cloning sites pBR322 ColEl ori tet amp pRGl pACYC177 lacIq W O 96/40920 PCTrUS96/10420 Table 2. Oli~onl-~leoiides OT TGO SFOUFN~-F (5~-3~) DF~;CRIPTION
EV18 CGGGAATACGGTCTAGATCATTAA C-term of ~1 gene, SEQ. ID #4 CGGTAmCGAAGTCAGAACG Xba I site EV27 GATCCGAGCTGTTGACAATTAATCATCGGCT tac ~lullloL~l SEQ. ID #5 CGTATAATGTGTGGAATTGTGACGGATAACAA sequence, Bam HI-mCACACAGGAAATTAATTAATGCTGTCTCC Eag I sites EV28 GGCCGGAGACAGCATTAATTAAmCCTGT tac yl~Jllloh, SEQ. ID #6 GTGAAATTGTTATCCGCTCACAATTCCACA sequence, Bam HI-CATTATACGAGCCGATGATTAATTGTCAAC Eag I sites, AGCTCG complement of EV27 EV29 TCGGATTCGAATTCCAAGCTGTTGGATCCTTA 5' end of a with Eco RI, SEQ. ID #7 GATTGAACTGTCTCCGGCCGATAAAACCACCG Bam HI andEag I sites 20 EV30 CGGAAGCCCAATCTAGAGGAAATAATATAT 5' end of ~ with SEQ. ID #8 GCACCTGACTCCGGAAGAAAAATCC Xba I site EV31 CCCGAAACCAAGCTTCATTAGTGA 3' end of the ~ gene SEQ. ID #9 GCTAGCGCGTTAGCAACACC with Hind III site MW007 TTTAAG~ l lCATTAGTGGTATT llluld~j~lleaia reverse primer SEQ. ID #10 TGTGAGCTAGCGCGT replaces last 3 codons of missing in pSGE515 30 MW008 CAGCATTAATTAACCTCCTTA mutagenesis reverse SEQ. ID #11 GTGAAATTGTTATCCG primer to .~Lllli~: a ribozyme binding site (RBS) MWOO9 GGTGCATATAmACCTCCTT ll,uLdgel,esis reverse p}imer 35 SEQ. ID #12 ATCTAGATCATTAACGGTAmCG to ~J~>LillL~ RBS; remove 2nd Bgl II
TG14 GGmAAACC Pme I linker SEQ. ID #13 TG59 GGCGAATAAAAGCTTGCGGCCGCG UpsLl~edll, of lacI gene, has SEQ. ID #14 TTGACACCATCGAATGGCGCAAAA Hind III and Not I site CCTTTCGCGG- u~ aLIedlll of promoter 45 TG60 GGGCAAATAGGATCCAAAAAAAAG DowllaLl~:dlll side of lacI
SEQ. ID #15 CCCGCTCATTAGGCGGGCTTTAT gene with trp Lldns~ Lional CACTGCCCG~ l-lCCAGTCGGG L~--llilldL~J- and Bam HI site TG62 CCCCGAAAAGGATCCAAGTA u~aLl~dlll primer for pBR322 50 SEQ. ID #16 GCCGGCGGCCGCGTTCCACTG ori positions 3170-3148 AGCGTCAGACCCC w/Bam HI and Not I site TG63 GGCGGTC(~ -lAAACGCT duwllaLledlll primer for SEQ. ID #17 GCGCTCGGTCGTTCGGCTGCGG pBR322 ori posili~,l,s 2380-2404 w/ Pme I site W O ~6/109~0 PCTAJS96/10420 Genetic and Molecular Biological Procedures. Standard bAct~riAl genetic procedures are described in J. H. Miller, E~rime~Ll~ in Molecular Genetics, (Cold Spring Harbor Press 1972) and J. H. Miller, A Short Course in Bacterial Genetics (Cold Spring Harbor Press, 1992 ). Standard m-)leclllAr biology procedures were performed as described in Sambrook et al., Molecular Cloning, (Cold Spring Harbor Press, 1989).
Plasmid DNA Transformation. DNA transformations were performed by the procedure described in Wensick et al., CeZI 3: 315-325 (1974). Briefly, cells were grown to mid log phase and then pelleted, resuspended in an equl volume of 10 rnM
10 MgS04 and incubated on ice for 30 mimlt~. The cells were centrifuged and the pellet resuspended in 1/2 original volume of 50 mM CaCl2 and placed on ice for 20 minutes. The cells were centrifuged again and then resuspended in 1 / 10 ~riginAl volume of 50 mM CaC12. Plasmid DNA was added to the competent cells in a solution of 10 mM Tris-HCl pH 8.0, 10 mM MgCl2 and 10 mM CaCl2. The mixture 15 was incubated on ice for 15 minutes and then incubated at 37~C for 5 minutes. One millilit~r of LB medium was added and the mixture incubated with sh~ing for 30-60 1-.i"~ The culture was then centrifuged, resuspended in 0.1 ml of LB medium and plated on the a~rululiate selective medium.
Purification of DNA. DNA fragments were purified from an agarose gel using the Genedean system (Bio 101, Inc. La Jolla, CA) accul ling to the method provided with product. PCR products were ~re~aled and cleaved with restriction ~n~onlldeases using the Double G~n~rleAn system. (Bio 101, Inc. La Jolla; methodprovided with product.) Briefly, the PCR product was purified away from the PCR
25 primers, then the PCR product was cleaved with restriction endonudease(s) andpurified from the restriction endonudease and buffer. The PCR product was then ready for a ligation r~Action Annealing of oligonucleotides. Compl~mf~ntAry oligonucleotides were 30 annealed accoldil,g to the following procedure. Equimolar amounts of each oligonucleotide were mixed in 15-25 ,ul of 10 mM Tris-HCl pH 8.0/ 1 mM EDTA and incubated at 65~C for 30 minutes. The sample was transferred to a 37~C water bath for 30 mintTte~ Finally, the sample was incubated on ice for 60 minutes or in the refri~;~ldLc,l overnight.
Oligonucleofide directed mutagenesis. Oligonucleotide directed mutagenesis was ~ l.ed with the Muta-gene phagemid in vitro mutagenesis kit (Bio-Rad, Hercules, ~~AliforniA) accordil,g to mAnllfActnrer's instructions which are based on the method of Kunkel (Kunkel, T. A. (1985) Proc. Na~l. Acad. Sci. USA 82: 488; Kunkel et 40 al., (1987) Me~hods Enzymol. 154: 367). The rHbl.1 region of pSGE515 was doned into W O 96/40920 PCT~US96/10420 pTZ18U (Bio-Rad, Hercules, CA or U.S. Biochemical, Cleveland, OH) on a BamE~-Hind~ fragment to create pSGE700. Three oligon1lc leotides, MW007, MW008 and MW009 were used to ~im1llt~neously introduce multiple changes in a single r~ction Pre~al ation of pBR322 ori. PCR primers were designed to amplify the pBR322 origin of replication. These primers, TG62 and TG63, annealed to the positions 2380-~ 2404 and 3170-3148 on the pBR322 DNA sequence (Sutcliffe, J. G. (1979) Cold Spring Harbor Symp. Quant. Biol. 43: 77-90). The PCR product was digested with NotI andPmeI. The DNA fragment was purified acco~ g to the G~n~rle~n procedure.
Preparation of tet gene fragment. The source for the tet gene was pSELECT-1 (Promega Corp., Madison, WI). This plasmid has a number of restriction en-lonllrlease sites, such as BamHI, Hindm, Sall and SphI removed from the tet gene (Lewis and Thompson (1993) Nucleic Acids Res. 18:3439-3443). A PmeI linker was inserted into the StyI site of pSELECT-1. This plasmid was designated pSGE504.
Oligonucleotides TG71 and TG72 were annealed and ligated to the EcoRI - ClaI
fragment of pSGE504. This pl~mi~l, pSGE505, was shown to have the expected restriction endonudease sites and to have lost the sites present in the multicloning site of pSELECT-1. pSGE505 was digested with NotI and PmeI. The 1417 bp fragment was purified accol.ling to the G~n~rle~n protocol.
Ple~dlalion of lad gene. The lad gene was isolated by amplifying the gene sequence from pRG1 (Dana-Farber Cancer Inst., Boston) that carried the lacI gene.
The PCR primers, TG59 and TG60 were designed to generate a wild type lacI
promoter (Farabaugh, P. J. (1978) Nature 274:765), u ~Lealll of the gene and to place the trp terminator sequence (Christie et al., (1981) Proc. Natl. Acad. Sci. USA 78:4180-4184) downstream of the gene. The same step could be carried out using Y1089 (Promega) or d~romosomal DNA from any E. coli strain carrying the lac region, such as MM294 (ATCC 33625.) The PCR product was gel purified and isolated accoL.ling to the Geneclean procedure and cloned into BamHI-Hincm~ digested pUC19 DNA to make pSGE490.
Construction of pSGE515. PCR primers EV29 and EV18 were chosen to amplify the alpha gene from pDLII-9lF (~offm~n et al., WO 90/ 13645). The purified PCR product was deaved with the restriction endonucleases EagI and XbaI.
To create a plasmid that rontAin~l Ptac-a, the alpha gene (from above) and the - tac promoter, whidh was prepared by ~nn-~ling EV27 and EV28, were mixed with Eco RI-Xba I-cleaved pUC19 DNA. The mixture of the three DNA fr~gm~nt~, in approximately equimolar ratio, was treated with T4 DNA Ligase. After incubation the ligation mixture was used to transform SGE476 and ampicillin r~ t~nt WO ~f'40320 PCT~US96/10420 transforrnants were selected. (TrAn~forrnatiQn into Strain MM294 (ATCC 33625) yields equivalent results.) An isolate with the correct restriction endonucleasefragments was ~ ignat~l pSGE492. The a gene and the tac promoter DNA
sequences were verified by DNA sequencing.
Primers EV30 and EV31 were used to amplify the ,~ gene from pSGE1.lE4 by PCR. The purified ,13 gene fragrnent was digested with XbaI and Hindm and then mixed with XbaI-Hindm digested pUC19 DNA and treated with T4 DNA ligase. The ligation mixture was used to transform competent SGE476 (equivalent to MM294, ATCC 33625) and transfor~nant~ were s~le~ t~l on LB + ampicillin (100 ,ug/ml) plates.
10 An isolate that contained the a~ l;ate restriction endonuclease fragments waschosen and designated pSGE493. The ,13 gene was cQllr;l~l~e~ by DNA sequencing.
The ,B gene was isolated from pSGE493 by restriction with XbaI and Hindm followed by pllrifi~ Ation according to the G~ne- l~An method. This DNA fragTn~nt was then ligated to XbaI-Hindm ~estricted pSGE492 DNA and transformed into 15 SGE713. (Any dam- strain such as JM110 (ATCC 47013) or GM119 (ATCC 53339) could also be used.) An ampicillin resistant transformant that carried a plasmid that had the a~ro~liate restriction fragments was chosen and designated pUC19a,13 (pSGE500).
The BamHI-Hind m fragment that contained the a and ,B genes of pSGE500 was purified accoldil,g to the Geneclean method. An XhoI fragrnent that carried a portion of the di-a gene cul~t~ining the glycine linker region was gel purified from pSGE1.lE5. pSGE1.lE5 (described in Hoffman et al., United States Serial Number 789,179, filed November 8, 1991) IS a tetracycline sensitive analogue of pSGE1.lE4 (Hoffman et al., WO 90/ 13645), which could also have been used.
The pBR322 origin of replication region (pBR322 ori, above) was ligated to the tet gene fragrnent (above) and the ligation mixture was transformed into SGE476.(Transformation into MM294, above would yield equivalent results.) Tetracycline resistant transformants were selected and plasmid DNA was isolated and analyzed.An isolate that c- ntAinP~I the a~plu~l;ate restriction endonudease frAgmPnt~ was chosen and ~le~ign~te~ pSGE507.
Next, pSGE507 and pSGE490 were digested with BamHI and NotI and the ~ru~liate frA~n~nt~ were pllrifie~l The two purified fragments were ligated together and the ligation mixture was used to transforrn competent SGE713. (Any dam~ strain could also be used; see above.) Tetracycline resistant transforrnants were s~lect~-l, and plasmid DNA was isolated and analyzed. A plasmid that had the a~r.,~liate restriction fragments was chosen and designated pSGE509.
The purified BamHI-Hind~I fragment of pSGE500 that contained the a and ,~
genes was ligated to BamHI-HindIII digested pSGE509. The ligation mixture was used to trAn.~forrn pSGE713 (see above for equivalent strains) and tetracycline resistant W O g-/~0320 PCTAUS96/10420 Ld,~s~"~lants were selected and chara~ri7~1 ~n is~late yjeldin" th~ cc,r.ect size plasrnid with the expected r~stri~i~n Pn~lon~ P~ce frA~m~nt~ was choser. and designated pSGE513.
The XhoI fragment of pS~ (described in ~c-ffm~n et al., United States SerialNumber 07/789,179, filed November 8, 1991, in~u,~o,dl~d hereinby reference) that5 .~ .l the di-a glycine linker sequence was ligated to XhoI ~ig~te-~ pSGE513 to create a plasrnid that r. nt~inP~l the di-a gene. SGE753 was h ,- ~.ru. ~ ~1 with the li~3*on ~ ~e and tetracycline r~ t~nt h,...~ro. .~.~nt~ were s~ t~ (Tr~ r~. ~ticn into SGE800 would have yielded equivalent results.) Isolates were screened to identify those that ront~in~-l the XhoI fragment i, s~, Lc:d into pSGE513 in the correct ori~nt~finn An isolate that cu. .l ,.; . .f.~l 10 the correct configuration of the di-a gene, as ~t~rmin~-l by r~ctrirtion Pn~- nllrl~e analysis with EagI, was ~ign~t~-l pSGE515.

Modi~cation of pSGE515 to ~eate pSGE70S. The DNA sequence record used to design PCR
primers for the amplifir~tinrl of the ,B gene did not contain the C-t~rn~in~l three amino acids.
15 Oligonucleotide directed mutagenesis was used to add these nine ml~ oti~lPs to the DNA
sequence of the ,B gene. In the same r~l tion~, modifiications were introduced to o~Li lli;Ge the ribosome binding sites for the di-a and ,B genes, and to remove a Bgm site near the end of the di-a gene. The Hindm-BarnHI fragment from pSGE515 was subcloned into pTZ18U, creating pSGE700. pSGE700 was then used as a source of ssDNA for site-directed 20 mnt~n~cic The following are the changes that were made with the oligonudeotides MW008 and MW009 to optimize ribosomal binding sites ar.d to remove a BglI restriction ~n~ionllt le~ce site.
25 di a1Dha before - CAATTTCAC--AGGA~ATTAATTAATGCTG ( SEQ.ID.NO.25) 111111111**1111**1111111111111 after - CAATTTCACTAAGGAGGTTAATTAATGCTG ( SEQ.ID.NO.26) Four nucleotide changes, shown above, in~ rling the insertion of two nucleotides, were introduced with MW008 to c,~ e the ribosome binding site for dialpha. ( I - in~ t~s identity, *- in-Jir~tf~ a change) beta before - ~AA~-A~CTAGA---G~AAA~AA-TATATGCAC (SEQ.ID.NO.27) 111*11111111***111**111*111111111 after - TAATGATCTA~-ATAAGGAGGTAAA~A~A~GCAC (SEQ.ID.NO.28) RECTIFIED SHEET (RULE 91) IS~P

W O ~6/~03Z0 PCTrUS96/10420 The six nllrleotir~ ~ changes shown above. in-ln~ing tt._ nsertion of '.:our n~.cleotides, were introduced with MW009 to optimize the ribosome binding site for beta. rr~e iower case "a"
on the before strand was a T to A mlltation in the construction of the alpha gene that introduced a Bgl II site into the sequence. This was removed so that there would only be a 5 single Bgl II site in pSGE705. ( I - in~ir~t,~c identity, * - in~1ir~tPC a change) End of Beta bef ore - CTCGCTCAC TAATGA~ (SEQ.ID.NO.29) 111111111*********1111111 after -- ~ 'A~Ir~CCACTA~TGAA (SEQ.ID.NO.30) MW007 introduced the coding sequence for the last three amino acids of the beta gene as shown above. ( I - indicates identity, * - indicates a change) PIlLdLive mlltantc were screened for loss of a Bgm restrirtion Pn~lonllrlP~ce cleavage site (introduced by MW008). Seventeen of 24 had lost the site and were further charact~ri7e~ by DNA seqnPnring at the other two mutagenized sites. One of the 17 had incu~ dl~d all the modifications from the three oligonucleotides. These changes were 20 verified by DNA seqllpnring and the rHbl.1 genes were cloned into BamHI-Hindm digested pSGE509. An isolate that had the correct restriction ~n-lonllrleaqe fragmPnts was ~Cign~t'~ pSGE705.
A new sequence u~hea~-, of the a gene minimi7r~ the distance between the tac promoter (De Boer et al., Proc. Natl. Acad. Sci. USA 80, 21-25, 1983) and the first codon of 25 the alpha gene. The il~L~ lliC region between the di-a gene and the ,B gene was also cignf.rl to contain the ...;..;..---... sequence that rr~ntainP~l a rPstrirtirn endonuclease site and the ribosome binding site for the ~ gene. A plasmid map of pSGE705 is shown in Figure 1. The plasmid map inr~ir~tr~s many of the restriction endonuclease cleavage sites.
pSGE705 is smaller than its ~.:OLIllL~ dlL pSGE1.lE4, and the plarPm~nt of its restriction 30 sites f~rilitat~s modular alterations of the sequence. An unused antibiotic r~?cictanrP marker was removed, and a promoter was added to the lacI gene that would allow tighter control of rHbl.1 ~ ,.es:,ion. pSGE705 was the base plasrnid used in all manipulations described in the Examples set forth below.

35 General F~ uLrO~ Protocol Hemoglobin was ~ sed in the strains described herein using any one of the f ~rml~ntati :~n protocols described below. First, a f~rm~?nh~r inoculum was grown from seed stock. An optional 2 liter flask fr?rm~nta*r~n was then ~. r..., ..L~rl prior to lLLlLl~
to a 15 liter f?rmr~ntor and induction. ~lt~rn~t-ively~ 100 liter fr-rm~nt~tion~ were 40 used. If the latter approach was used, then a f?rm~ntnr inoculum was grown from RECTIFIED SHEET (R~LE~ 91) ~S~J~P

W O g~/~0320 PCT~US96/10420 seed stock 2 liter shake flasks. Four of these shake flasks were then used to inoculate the 100 liter fermentors. The details of the fermentation process are described below.
Any suitable fermentation and pre-purification scheme (purification prior to the ion exchange molecular weight separation) can be used for the production of the mAtPri,~l of the instant invention.

Seed Stock - all fermentations Seed stock was grown up in LB broth containing 10 g/L BactoTryptoneTM, 5 g/L yeast extract, 5 g/L NaCl, 0.2 g/L NaOH, and 10 ug/ml tetracycline to an optical 10 density of 1.5 -1.7 at 600 nm. The solution was then made up to 10% glycerol and stored at -80~C until required.

15 liter Fermentation ProtocQI

Fermentor Inoculum (500 ml broth in 2 L shake flasks - seed flasks) To ~r~ the f~nrt~ntt r inoculum, seed stock was thawed and 0.1-0.4 ml of seed stock were inoculated into 500 ml of a solution (DM-1) collL~ilullg approximately 20 4.1 g/L KH2PO4, 7 g/L K2HPO4, 2 g/L (NH4)2SO4, 1 g/L Na3 citrate-2H2 O, 153 mg/L MgSO4 7H2O, 2.3 g/L of L-proline, 2 g/L yeast extract, 4.8-5.5 g/L glucose, 320 mg/L thiamine HCl, 10 mg/L tetracycline, and 3 ml/L of a trace metal solution c~llLdilullg 32.5 mg/L FeC13 6H20, 1.6 mg/L ZnC12, 2.4 mg/L CoCl2-6H20, 2.4 mg/LNa2MoO4-2H20, 1.2 mg/L Caa2 2H2O, 1.5 mg/L CU(II)so4-5H2 O, 0.6 mg/L H3BO3, 25 and 120.2 ml / L HCl. This culture is allowed to grow for 8-10 hours at 37~C on a shaker. Two flasks were combined and used to inoculate the 15L fermentors if no intermediate "2 Liter" fermentation was ~lL~lmed. Alternatively, an intPrme~ te seed fermentation in two liter fermentors was performed prior to the 15 liter fermentation. .
Ferrnentor (2 L volume - seed fermentation) As an optional int~rrrte~ tt~ step, the cells were grown in a 2 liter fermentation.
400 mL of the seed ferrrt~nt~tiQn was then aseptically transferred to a 2-liter New Brunswick fermentor c~ g ~roxilllately 1700 mL of a solution col l l, i . . i t~g 35 ~roxill~ately: 2.2 g/L KH2P04, 4 g/L K2HPO4~ and 2-2 g/L (NH4)2SO~
The medium in the fermenter also rc~nt,tine-l 1.2 g/L trisodium citrate, 1.2 g/LMgSO4-7H20, 2.5 g/L proline, 3.1 g/L of the trace metal solution described above, 0.1 mg/ L tetracycline in 50% ethanol solution, 345 mg/ L t~ti,-tmin~ Ha in purified water, sterile filtered solution, 200 g/L of 70% glucose, 50 + 10 g/L of 30% NH40H, and 2 ml 40 PPG 2000 (polyethylene glycol 2000).

WO 96/40920 PCT~US96/10420 Cells were grown in the fermentor for approximately 10 hours. The pH was maintained at 6.8 - 6.95 by A~ iti(~n of 15% to 30% NH40H, dissolved oxygen was maintained at or above 20%, and 50-70% glucose was added throughout the growth period, sufficient to mAint~in low but adequate levels of glucose in the culture (2 g/L-5 10 g/L). The culture was grown at approximately 30~C to an OD60o ~ 2-5.

15L Fermentor (14 L volume in 20 L Fermentor - "lSL") Either 800mls of the seed flask or 400mls of the "2 liter" seed fermentor were then aseptically transferred to a 20-liter fermentor COl lailling 8 liters of the following media (DM4-RP): 1.3 g/L KH2P04, 2.4 g/L K2HPO4, 1.3 g/L (NH4)2SO4, 195 mg/L
thiamine HCl, 6.1 mg/L tetracycline, 1.8 g/L proline, and 2.2 ml/L of the trace metal solution described above. Note that masses of added reagents are cal~llAt~l using the final volume of fermentation (11.5 liters) and are approximate within measurement error. The pH was maintained at 6.8 to 6.95 by addition of 15% to 30~3fO
NH40H, dissolved oxygen was maintained at or above 20%, and 50 to 70% glucose was added throughout the growth period, sufficient to mAintAin low but adequate levels of glucose in the culture (2 g/L-10 g/L). Dissolved oxygen was maintained as close to 20% as possible. The culture was grown between 28 and 32~C until an OD60o of 30 was reached. Induction was Accrlmplished by the ~ ition of 10-1000,uM
is~ o~yl thiog~lActnsi~le (IPTG). Upon induction of hemoglobin synthesis, the E. coli heme biosynthesis was supplemented by addition of hemin dissolved in 1 N NaOH, either by addition at induction of the total mass of hemin required, by CC11LLi11UOUS
addition of hemin throughout the induction period, or by periodic addition of hemin dissolved in 50 mM to 1 M NaOH (e.g. one third of the total mass of hemin to be added to the fermentor was added at induction, another third was added after 1 /4 of the total time after fermentation had elapsed, and the last third was added half-way through the induction period). Total hemin added ranged from 50 to 300 mg/L. Thefermentor was allowed to continue for 8-12 hours post-inrlll~-hc-n.

100 liter F~",.~rL~LzLion Protocol Fermentor Inoculum (500 mL broth in 2 L shake flaslcs) To prepare the fermentor inoculum, seed stock was thawed. Seed stock (lOOml) was grown up in 500 ml of DM59 in an Erlenmeyer flask at 37~C in a 1 inch rotary shaker (275 to 300 rpm) for 8 to 10 hours. DM59 media is: 3.34 g/L KH2PO4, 5.99 g/L K2HPO4, 1.36 g/L NaH2PO4 H20, 1.95 g/L Na2HPO4, and 1.85 g/L
~H4)SO4 which are ct~rili7e~l After st~rili7~tion, 12.20 ml/L of a trace metal solution was added. The trace metal soll~*c-n col-ldil.ed: 134.2 g/L tripotassium citrate, 32.2 g/L trisodium citrate, 27 g/L FeC13 6H2o~ 2.2 g/L ZnCl2~ 0.3 g/L CoCl2 ~6H2O, 0.3 g/L Na2MoO4 ~2H2O, 2.73g/L MnCl2~ 6.6 g/L Cacl2-2H2o~ 1.5 g/L
Cu(II)SO4-5H2 O, and 15 ml/L 85% H3PO4. In A~lclition, the following coll.pul.ents 5 were added to the media after s~rili 7A*Qn to achieve the final conc~ . dlionsin~lic Ate~l 10 mg/L tetracydine and 320 mg/L thi~mine Poly~ ylene glycol 2000 was added if a foaming problem was observed.

Fermentor (100 L volume) 2000 mL of the Fermentor Inoculum was then aseptically transferred to a 100-liter BioLafitte fermentor contairling 54 L of DM59 medium described above.
The fermentor was run at 30 + 1~C, controlling dissolved oxygen at 20% and glucose between 0-6 g/L. At OD 30 + 2, inclllcti~n occurred by lowering the lell-~eldture of the fermentor to 26 "C, adding 43.5 mL of 100 mM IPTG and 73 mL of 50 mg/mL hemin. At 3 hours post induction, 96 mL of 50 mg/mL hemin was added, at 6 hours post induction, 125 mL of 50 mg/mL hemin was added, at 9 hours post induction, 125 mL of 50 mg/mL hemin was added and at 12 hours post induction, 125 mL of 50 mg/mL hemin was added. Harvest and further pllrifi~ ~tion occurred at 16 hours post induction. Cells were either immerliAt.oly purified or frozen for later purification.

Purification If required, frozen cells were partially thawed in warm water for approximately 20-30 minutes. Cells were chopped into small bits in a steel beaker using break buffer (40 mM Tris base, lmM benzamidine) as nee~le~l- The chopped cells and break buffer at a ratio of 2 mL break buffer per 1 gram of frozen cells were placed in a Waring Industrial Blender and homogenized for 1 - 5 minutes on the low setting. The solution was allowed to settle for 5 minutes after homogenization and any foamed mA~riAl was removed.
A Niro PandaTM cell disruption device (Niro Hudson, Inc. Hudson, WI) was ~aled for homog~ni7Ati~ n by passing 200-300 mL of break buffer through the system. Cells were lysed by one or two passages of the homogenized cell solutionthrough the Niro set at 850 bar. The pH of the lysate was adjusted to approximately 8 with sodium hydroxide, and sufficient Zn(OAc)2 was added to make the solution 2 - 4 mM in Zn(OAc)2. The solution was then spun at 10,000 rpm in a JA-10 rotor at 4~C for 60 mimlt~ in a Be~ km~n centrifuge. The supernatant was collected and was optionally diluted 1:1 with distilled water. When using this protocol to purify K158, care should be taken to keep levels of oxygen as low as possible.

WO ~G/~_320 PCT~US96/10420 Chromato~raphy:
All solutions were 4~C and were adjusted to the correct pH at 4~C. 500 mL of ChPlA~;ng SEPHAROSE fast flow resin (Ph~rm~ , Piscataway, New Jersey) was ple~ared by washing with 4 column volumes of distilled water. Flow through the 5 column for all steps was 200 mL/min. The resin was charged with 2 to 3 column volumes of 2mM Zn(OAc)2 followed by 2 - 3 column volumes of 200 mM NaCl. The lysate was loaded onto the column and washed with 4 to 6 column volumes of 20 mMTris, 500 mM NaCl, pH 8.5, 7 - 8 cc Inmn volumes of 240 mM Tris, pH 8.5, and 7 - 8 column volumes of 20 mM Tris, pH 8.5. Hemoglobin was eluted with 15 mM EDTA, 20 mM Tris, pH 8.5 and collected into 200 mL of well oxygenated 20 mM Tris, pH 8.5.
The column was then rinsed with an additional 3 - 4 column volumes of 15 mM
EDTA, 20 mM Tris, pH 8.5, regenerated with 4 column volumes of 200 mM NaCl and stored in 0.2 N NaOH.
The solution was then buffer exchanged 5 times into 20 mM Tris, pH 8.5 prior to loading onto 200 mL of a SEPHAROSE Q column. The colllmn had been prepared by rinsing with 4 column volumes of distilled water, 4 column volumes of 1 M NaCl, 4 additional column volumes of distilled water and equilibrating with 3 to 4 column volumes of 20 mM Tris, pH 8.5. After loading the sample, the column was washed with 2 to 3 column volumes of 20 mM Tris, pH 8.5 and eluted with 20 mM Tris, pH
7.6. Fractions were collected and pooled if the As7s/As40 ratio was greater than or equal to 1.03. The column was then ~ lp~n~rT with 3 - 4 column volumes of 1 M NaCl, 4 column volumes of distilled water, 2 -3 column volumes of 50% acetic acid, 4 column volumes of distilled water and finally 2 -3 column volumes of 0.2 N NaOH for storage. The column was run at 30 mL/min flow rate. The resultant hemoglobin wasstored at -80~C or in liquid nitrogen.

Example 2 Construction of di-dialpha Gene Construct linked by a 7 amino acid linker (SGE 939) A. Construction of pTZ19U/705 Mutants rHbl.1 genes were cloned as a BamH~/ HindJ:~ DNA fragment into pTZ19U
(BioRad, Hercules, California). This constru* was then transformed using a modified process of the Hanahan protocol (H~nz~h;~n, J. Mol. Biol., 166:557 (1983)) into CJ236 E.
coli strain (BioRad). The ~n:~h~n transformation buffer contained 45 mM MnC12, 60 mM CaCl2, 40 mM KOAc, 620 mM sucrose, 15~ glycerol and 100 mM rubidium l~hlnric~e A 5 ml culture of an E.coli strain was started in 2x TY broth from an isolated colony and cultured ovPmight Then, 200 ml of 2x TY broth was inocllT~tP~T with 2 ml W O 96/40920 PCT~US96/10420 of the overnight culture and incubated at 37~C with vigorous shaking for 2.5 hours.
The culture was then transferred to two 300 ml centrifuge tubes and placed on ice for 15 minllt~. Cells were pelleted in a centrifuge at 8000 rpm, 4~C, for 10 ~ Les and the supernatant was poured off. The cells were gently but thoroughly resuspended in 80 ml transformation buffer. The cells were again pelleted at 8000 rpm, 10 ~ -ules at 4~C. The cells were gently resuspended in 20 ml of ice-cold transf ~rma*l n buffer and left on ice for 30-60 minlltos Cells were aliquoted in buffer into twenty 1 ml tubes.
The cells were quickly frozen on dry ice and stored at -80~C.
Single-stranded DNA containing uracil substitutions was isolated and 10 oligonudeotide-directed mutagenesis was pelLolllled using the Muta-gene Kit (BioRad, Hercules, CA) and standard protocols accul.lil~g to the mar~l~f~tl1rer's instructions. Two pTZ19U/705 clones were ~le~aLed as follows.
The first pTZ19U/705 clone was ~le~ared using oligonudeotide JD29 (ACC
GTT CTG ACT AGT AAA TAC CGT TAA TGA [SEQ. ID. NO. 18]). This 15 oligonucleotide created a unique SpeI site in the end of the dialpha ~m~in~ Asecond pTZ19U/705 done was ~le~aled using oligonucleotides JD28 (5'-GGA GGT
TAA TTA ATG CTG TCT CCT GCA GAT-3' [SEQ. ID. NO. 19]) and JD30 (5'-CTG
GTG GGT AAA GTT CTG GTT TGC GTT CTG-3' [SEQ. ID. NO. 20]). The resulting done incol~oldled a unique PstI site in the dialpha genes and removed an SpeI site in 20 the beta domain.

B. Assemblv of the di-dialpha gene construct The assembly of di-dialpha gene construct was accomplished by removal of a dialpha gene cassette from the first pTZ19U/ 705 clone using BamE~I/ SpeI enzymes 25 and gel purification of the DNA fr~gm~rlt A second pTZ19U/705 done was cut with PstI/BglII enzymes to give a second dialpha gene cassette with the 5' end of the beta gene, which was also purified. These were then further ligated together with armealed oligonucleotides JA113 and JA114 to create a di-dialpha cassette with a 7 amino acid fusion peptide linker linking the two dialpha globins.
JA113: 5'-CT AGT AAA TAC CGA TCG GGT GGC TCT GGC GGT TCT GTT CTG TCT CCT GCA-3' (SEQ. ID. NO. 21).
JA114: 5'-GG AGA CAG AAC AGA ACC GCC AGA GCC ACC CGA TCG GTA TTT A-3' 35 (SEQ. ID. NO. 22).
- This di-dialpha cassette was then ligated as a BamHI/ Bgm fragment into pSGE705 (described in PCT publication number WO 95 / 14038, herein incorporated by reference) that had the rHbl.l genes removed as a BamHI/BglII fr~grn~nt The resulting di-dialpha plasmid (pSGE1000) was transformed into SGE1661 (also W O 96/40920 PCT~US96/10420 described in PCT publication number WO 95/ 14038) using the modified Hanahan's protocol described above to create SGE939. Two other plasmids were also constructed using the same methods described above, pSGE1006 and pSGE1008.
pSGE1006 c~,lle~onds to pSGE1000, except that the linker linking the two dialpha5 regions was excised as an SpeI/PstI frA~m~nt and replaced with a synth~ci7e~1 region encoding a 14 amino acid linker of the following sequence:

GlyGlySerGlyGlySerGlyGlySerGlyGlySerGlyGly (SEQ.ID.NO.2) pSGE1008 was created in the same fashion as pSGE1006, except that the replA~ ~m~nt linker was a 16 amino acid linker of the following sequence:

SerGlyGlySerGlyGlySerGlyGlySerGlyGlySerGlyGlySer (SEQ.ID.NO.3) Example 3 Construction of a High Copy Plasmid The construction of pSGE720 was p~rfnrmPrl in two stages. First, the pUC
origin of replication was introduced to create plasmid pSGE715, which is similar to pSGE705 in that it inc lt~ s the lad gene. Then, the lacI gene was ~lPlet~ from the plasmid and replaced with a short oligonucleotide containing several c~l~vel~ient restriction sites to create plasmid pSGE720.

A. Construction of pSGE715 The pUC origin of replication was introduced to create plasmid pSGE715 through pSGE508, which is identical to pSGE509 with the exception of a single basepair substihl~ion at base 602 (G A). The substitution changes the pBR322 origin of replication to a pUC19 origin of replication.
pl~mi~ls pSGE508 and pSGE705 were digested to completion with restriction enzymes BamE~I and Hindm, accoldillg to the manllf~c~lrer's instructions (New ~nglAn~l Biolabs.). The plasmid, pSGE508, was digested first with BamHI to completion, then Hindm was added, and the digestion col.lil.~ ed. The pSGE705 digest was purified with Promega Magic DNA aean-up protocols and reagents (Promega, Madison, WI) and further digested to completion with BgZI, accol~lillg to the manllfAr~lrer's instructions (New England Biolabs). The enzymes in both pSGE508 and pSGE705 digests were inactivated by heating at 67~C for 10 mintlt~s,then the DNA was pooled and purified together using Promega Magic DNA Clean-up protocols and reagents. The DNA was suspended in ligation buffer, T4 DNA ligase W O 96/40920 PCT~US96/10420 was added to one aliquot, and the DNA was incubated overnight at 16~ C. SGE1661 cells were made competent by the method of Hanahan, using rubidium chloride (Hanahan, D., In DNA Cloning; A Practical Approach (Glover, D. M., ed.) vol. 1, pp.109-135, IRL Press, Oxford, 1985), and transformed with the ligation mix acc~.ldil.g to the 5 Hanahan protocol. Transformants were selected by plating the cells on LB plates containing 15,ug/ml tetracycline. Candidates were screened by restriction digestion to ~let~rmine the presence of the rHbl.1 genes, and sequencing to detect the pUC origin of replication. Several candidates were identified, and the resulting plasmid was named pSGE715, and pSGE715 in SGE1661 was called SGE1453.
The copy number of pSGE715 is about four-fold higher than pSGE705, measured to be about 460 plasmids per cell. As noted above, the diLLerence between pSGE705 and pSGE715 is a single basepair change in the origin of replication region, which has been confirme~l by sequencing.

15 B. Construction of pSGE720 The lacI gene was deleted from pSGE715, replacing it with a short 'oligonucleotide containing several co nveluent restriction sites, by the following steps.
First, plasmid pSGE715 was digested to completion with restriction enzymes BamHIand NotI, accoldil.g to the manllhctllrer's instructions (New England Biolabs). The 20 pSGE715 digest was purified with Promega Magic DNA Clean-up protocols and reagents. The DNA was mixed with armealed, kinased oligon~ leotides, CBG17 +
CBG18, and suspended in ligation buffer.

CBG17 = 5'-GGCCGCCTTAAGTACCCGGGmCTGCAGAAAGCCCGCCTA
25 ATGAGCGGG~: l l l''l''l"l' l l CCTTAGGG-3' (SEQ. ID. NO.: 23) CBG18 = 5'-GATCCCCTAAGGAAAAAAAAGCCCGCTCATTAGGCGGGCTTT
CTGCAGAAACCCGGGTACTTAAGGC-3' (SEQ. ID. NO.: 24) T4 DNA ligase was added to one aliquot, and the DNA was incubated overnight at 16~C. SGE1821 cells were made competent by the method of Hanahan, using Rubidium C~hloritle, and transformed with the ligation mix accol lillg to the Hanahan protocol. SGE1821 con~in~ pRG1 pl~mi~l~ in addition to pSGE720. pRG1 is a low copy number plasmid Cullldilli,.g LacIq. Transformants were selected byplating the cells on LB plates collldillillg 15g/ml tetracycline. C~~n~ tes werescreened by restriction digestion using PstI and SmaI to detect the presence of the new linker and the absence of the lacI gene, and sequenced to detect the pUC origin of replication and the absence of the Zad gene. Several candidates were identified, and WO 9~'103~0 PCTAUS96/10420 the resulting plasmid was named pSGE720. The plAcmi~, pSGE720 in SGE1675 was denoted SGE1464.

Example 4 Hi~;h copy Di-dialpha Construct A second plasmid c~ i . lg the di-dialpha hemoglobin genes was created using pSGE720 as the vector. The di-dialpha gene ~ A~ett~ was removed as a BamHI/Hindm fragment and gel pllrifie~l The vector pSGE720 was also cut with BamHI/HindIII and the rHbl.l genes removed. The vector was gel pllrifie~l The di-dialpha ~A~sette was ligated into the pSGE720 vector, resulting in a new vector pSGE1004. This new vector was then transformed into E.coli strain SGE1675 using the modified Hanahan method as described below to produce strain SGE946.
Example 5 CharAct~ri7Ation of SGE939 and SGE946 Globins Several 15 liter f~rrn~ntA*~ nq were p~.r.,l~l,ed on both strains SGE939 and SGE946 and soluble vs. insoluble western blots were performed using conv~ntic nAI
me~hods. This data coupled with purification yields in(li~Ate~ that more solubleL~ could be obtained from SGE946 (250-300 mg/ L by the BioCAD assay (BioRad). The data obtained shows that both strains make di-dialpha globin and beta globin ~.oL~ins, but that the SGE946 strain makes a larger amount of total ~ L~in and soluble ~r~Lt~
The SGE939 hemoglobin-like ~. ~JL~:in was first eluted from a Q-SEPHAROSE
column and then from a S-SEPHAROSE column on an FPLC. Fractions were collert~rl by eluting with a pH gradient. By SDS-PAGE analysis, there appeared to be a population of degradation products since these cross-reacted with anti-rHb antibodies. The deanest fractions were pooled and analyzed by C4 HPLC. A
chromatogram of SGE939 showed the beta globin eluting at 43.7 minutes as expected, and the di-dialpha peak eluting at 61.8 minutes. Dialpha globin normally eluted at about 55 minlltf~q under these conditions. There was also a peak at 56.2 miml~q and a large shoulder on the di-dialpha peak. The peaks were collert~ and analyzed by mass spectroscopy. The beta globin peak gave the expected molecular weight of 15,910 daltons, while the di-dialpha peak gave a molecular weight of 61,088 ~Alton~.
The ~AlclllAtPrl m( leclllAr weight for beta globin is 15,913, while the f~Al~ll~te~l moleclllAr weight of di-dialpha globin is 61,107.8 daltons. These results indicate that the ~loL~il. ex~.~ssed from SGE939 ~ cntAined the expected di-dialpha polypeptide.

CA 022l9242 l997-l0-27 W O ~/10320 PCT~US96/10420 The ~lol~)~orphyrin IX co~ nt was shown to be below 3~. The P50 averaged to be 24.7 and the nma~ was 1.75.

Example 6 Tetra-Dialpha A. Construction of di-dialpha vector colll~il~illg the K15~C mutation.
Replacing the lysine residue at position 158 of dialpha globin allows ch~mi~Al cross-linking of rHbl.1 molecllles to form a dimeric hemoglobin molecule rt:l~,led to as K158C. This mutation can be inserted into the di-dialpha ~ res~ion plasmid (pSGE1000), to produce a mutant genetically linked di-hemoglobin that can be chemically cross-linked to form a tetra-hemoglobin. The modification will place the K158C mutation in the fourth (3'-tPrmin~l) alpha globin coding sequence of the di-dialpha pl~mi~7 The K158C mutation is a 3 base change in the coding sequence, and can be transferred among dialpha - c(Jnldil-il-g vectors on an Eag I-Bgl II restriction fragment. Because there are multiple Eag I sites in pSGE1000, an intPrme(li~tp cloning step in the plasmid "pFusion II" is required. The doning steps are as follows:

1. Isolate an EagI - Bgl IIfr~gmPntcolll~ilL~ng the K158C mutation from pSGE 1.1E4 2. Isolate large Eag I - Bgl II fragment from plasmid pFusion II, which removes the comparable "wild type" fragment from the second alpha gene 3. Ligate above fragments to form the intPrme~i~tP pFusion II - based vector collLdillillg the K158C mutation 4. Replace the Pst I - Bgl Il fragment in pSGE1000 with the Pst I -Bgl II fr~gm~nt conldil illg the K158C mutation.

E~. Development of a cloning strategy for genetically linked tetra dialpha.
Expression of a genetically linked tetra-hemoglobin mnle~lle requires construction of a plasmid ~ nnt~ining coding sequences for four dialpha hemoglobin 30 genes, connect~~ by coding sequences for peptide linkers, and one beta globin gene.
A plasmid with these characteristics can be based on pSGE1000, which is currently being used to express a genetically linked di-hemoglobin. The following steps will be required to generate this plasmid:
~5 1. Generate a modified vector with a new restriction site at the 5' end of the di-dialpha coding sequence;
2. Generate a second modified vector with a new restriction site at the 3' end of the di-dialpha coding sequence;

WO 96/40920 PCT~US96/10420 3. Design an amino acid sequence suitable for linking the di-dialpha m(~lec~ s in such a way that a tetra-hemoglobin can ~s~mhle and design the DNA sequence required to encode the peptide linker; and 4. Assemble a new plasmid cc,., I ,.i . .; . .g the two modified di-dialpha sequences, the linker sequence, and either a 705 or 720 plasmid background.

Silent mutations in were irl~ntifie-l in the di-dialpha sequence that will generate restriction sites unique to di-dialpha in either the 705 or 720 (low and high-copy) plasmid backgrounds, near the 5' and 3' ends of the di-dialpha coding sequence.
10 A restriction site for one of the enzymes, AatII, is also ~res~nt in the beta globin gene;
the site in the beta gene will be removed to facilitate cloning. A preliminary ~ ning strategy has been generated for construction of a tetra-hemoglobin ~res~ion vector as follows:

15 1. Create an Aat II site at the 3' end of a dialpha gene in pFusion II by site directed mutagenesis to create a fr~gm~nt denoted "A1."
2. Subclone A1 into di-dialpha on a PstI/ BglJI restriction fr~gm~nt to create "A2."
3. Remove the AatII site from the beta globin gene in pFusion II by site directed mutagenesis to create "B1."
20 4. Subclone "B1" into a second di-dialpha construction on a PstI/ BglII fr~rn~nt to create "B2."

5. Geate a BlpI site at the 5'end of the dialpha gene in pFusion I by site directed mutagenesis to create "C1.~' 6. Subclone C1 on a BamHI/ Spe I fragment into the modified di-dialpha plasmid (B2) to create "D1."

7. Isolate the Bam~/ AatII fr~grn~nt from A2, and the BlpI/Hind~I fr~grn~nt fromD1; ligate with a new synthetic sequence encoding a peptide linker containing AatII and BlPI ends, in a cc,l.v~luent plasmid background to form a tetrahemoglobin coding sequence.

Example 7 General Tldllsfull--ation Procedure A modified Hanahan protocol was used to produce competent E. coZi cells. The ~n~h~n Transformation buffer contains 45 mM MnCI2, 60 mM CaCI2, 40 mM KOAc, 620 mM sucrose, 15% glycerol and 100 mM rubidium chloride. A 5 ml culture of an E.coli strain was started in 2x TY broth from an isolated colony and cultured overnight. Then, 200 ml of 2x TY broth was ino~ll~t~l with 2 ml of the overnight CA 022l9242 l997-l0-27 culture and incubated at 37~C with vigorous shaking for 2.5 hours. The culture was then transferred to two 300 ml centrifuge tubes and placed on ice for 15 millul~.
Cells were pelleted in a centrifuge at 8000 rpm's, 4~C, for 10 minlltPç and the supernatant was poured off. The cells were resuspended gently, but thoroughly in 80 ml transformation buffer. The cells were again pelleted at 8000 rpm for 10 min1ltPS at 4~C. The cells were gently resuspended in 20 ml of ice-cold transfcrm~tion buffer and left on ice for 30-60 minutes. Cells were aliquoted in buffer into twenty 1 ml tubes.
The cells were quickly frozen on dry ice and stored at -80~C.

Example 8 Pl~dldtion of BMH-crosslinked di-alphaK158C (Di-hemoglobin) Di-hemoglobin was produced by ~1..s~ king dialpha hemoglobin containing a 15 K158C mutation in the second alpha globin ~inm~in using biçm~l~imidohexane (BMH, Pierce Chemical Co., Rockford, Illinois). BMH is a homobifunctional m~lPimide crnççlinkPr, and its primary reactivity is towards sulfhydryl residues. The linkage is v~ ible once formed. The alkane spacer between the m~lPimide residues is hexane (six calbolls) and the mnle~lle has poor solubility in b.lLLeled aqueous 20 solutions. The nominal length of the crrs~linkPr is 16.1A.
K158C was conc~:llLldLed to 60 mg/mL in 20 mM Tris buffer pH 8, and deoxyg~llated by gas exchange with humid oxygen free nitrogen in a rotating glass flask (ROTOVAP RE111, Bri~kn-~nn, Inc., Cuntiague Road, W~LLully, NY). K158C
was maintained in the deoxy form in order to limit the reaction of BMH with the 25 intrinsic sulfhydryls of hemoglobin, especially residue Cys93 in the beta subunit. The reactivity of this residue with sulfhydryl reactive reagents is generally at least 50 fold slower in the deoxy form than in liganded forms of hemoglobin. The reactivity of the surface K158C residue is not ~ffertP~l significantly by the heme ligation state.A solution of BMH was prepared in pure dimethyl sulfoxide (DMSO) at 10 30 mg/mL. An aliquot of this solution was added to the deoxyHb solution (0.6 moles of BMH per mole of Hb, mdinLi~ilullg deoxy conditions) with swirling to mix, and the sample was allowed to react for 1 hour on ice. Following reaction, the hemoglobin solution was centrifuged or filtered (0.2 micron) to remove any precipitated material, diluted to 25 mg/mL and then chromaL~,ldphed on SEPHACRYL S-200 HR
35 (Ph~rmAri;~, Uppsala, Sweden) to resolve the rlih~?mnglobin fraction from theunreacted morch~mnglobin and the small amount of trihemoglobin formed during the reaction. Two S-200 HR columns (Ph~rm~riA BPP 113, ca. 6L of resin each, 11.3 cm mPtpr x 60 cm) were used in series to give acceptable resolution and volume ~n~lling capabilities. The yield of coupling was typically 60~o, and about 50~O of the WO 96/40920 PCTrUS96/10420 starting hemoglobin was recovered following size exclusion chromatography.
Following chromatography on Q-SEPHAROSE to remove endotoxin, the ~-iih~moglobin was submitted for several routine analyses and the results are reported below. Methods for these analyses are described in PCT publication WO 95 / 14038.
5 Average molecular weight was ~-iel~ eri by size exclusion chromatography using a SUPEROSE 12 column using Bio-Rad molecular weight standards (Bio-Rad, Hercules, California).

Assay Result Endotoxin (LAL assay) 0.6 EU/mL
E. coli ~roL~i l . below rletection Protop~ hy~ IX 1.14~,~o p50, Torr at 37 ~C 32.7 Nmax. 2.09 Average molecular weight 128 kDa Example 9 LAL ~ssay for Endotoxin Fifty microliters of endotoxin standard, blank diluent, or hemoglobin solution (rHbl.1) was rnixed with 50 ul of LAL lysate (BioWhittaker, Inc., Walkersville, MD) in a well of a 96-well, pyrogen-free microtiter plate, accoLdil-g to the m~nllf~cturer's instructions. The mixture was allowed to incubate for 30 minl1t~s in a 37~ C water bath. One hundred microliters of acetate-Tle-Glu-Gly-Arg- conjugated to para-l.iLLoaniline (chromogenic substrate) was added to each well and the plate allowed to incubate for an additional 16 to 60 minllt~s at 37~ C. The reaction was stopped by the addition of 50 ul 25~ glacial acetic acid, and the samples transferred to HPLC sample vials for analysis.
Twenty microliters of each sample was injected onto a Vydac C4-reversed phase chromatography column (2mm x 250mm), pre-equilibrated at 40~ C, 5~O
Solvent B. (Solvent A is 20~ ...ii. ;le in water with 0.1~o TFA and Solvent B is100% ~c~ic-l~ilLile with 0.1~o TFA). The chromatographic system was run at a flow of 1 ml/min. Separation was achieved as follows: a 1 minute hold in 95% Solvent A/5%
Solvent B after injection, a 4 minute ramp to 50% Solvent A/50~o Solvent B, a 2 minute increase to 100~O Solvent B, a 3 ", i ,~ wash in 1005'O Solvent B, a return to 955'O
Solvent A/5% Solvent B over 1 minllte and an equilibration at 95% Solvent A/5~o Solvent B for 4 minlltes. The separation was mo~ li lu~ ed at 405 nm.
The pealc areas of the standard solutions were used to construct a standard curve against which test samples were me~sllred. A series of curves were generated W O ~G/403 0 PCT~US96/10420 from the analysis of standard solutions ranging in conc~l~L dLions from 0.5 EU/ml to 0.0005 EU/ml. T in~r~ty was achieved when the standards were analyzed in groups accordi~g to the time of incubation. One curve was generated from analysis of samples incubated with chromogenic substrate for 16 minutes, others were generated 5 from analysis of samples incubated with chromogenic substrate for 30 ~ and 60 minllt~. Therefore, a standard curve for use in a particular circumstance depended on the sensitivity of the endotoxin measurement that was required.

Example 10 Production of penta-hemoglobin rHbl.1 containing a K158C mutation in the dialpha globin (hereinafter referred to as K158C) was ex~res~ed and purified as described above. rHbl.1 was ~ ,essed and purified as described in PCT publication number WO 95/ 14038, filed November14, 1994, entitled "Purification of Hemoglobin." The penta-hemoglobin was then formed by reacting K158C with a core rHbl.1 molecule (that did not collLdill theK158C mutation) activated as described below.
The core rHbl.1 molecllle was activated by re~c*ng with sulfosllcrinimidyl 4-(N-mal~imi-l~ methyl)cyclohexane-1-carboxylate (sulfo-SMCC) (Pierce Chemicals, Rockford, Illinois). Sulfo-SMCC is a water soluble heterobifunctional crosslinker that reacts with both amine and sulfhydryl functional groups. Reaction with lysine residues on rHbl.1 leads to loss of the sulfosllc.-inimide group with the fnrm~ti~n of a stable amide linkage between the ~oL~il- and the succinimide moiety. The m~l~imide residues are highly reactive towards sulfhydryl groups. Therefore, following reaction with sulfo-SMCC, the rHbl.1 has been "activated" at multiple surface lysine residues towards re~ction with the surface sulfhydryl residue of K158C. The N-(4-carboxy-cyclohexylmethylm~l~imi~le) residues are particularly stable to hydrolysis and the "activated" rHbl.1 can thus be manipulated extensively prior to ~ iti~n of K158C.
The desired extent of modification of rHbl.1 was ~leL~ ed empirically by reaction with K158C following activation. The initial reaction with sulfo-SMCC was modulated by altering the conc~l-L dLion of sulfo-SMCC and reaction time, until a covalent Hb polymer of the desired size range was achieved upon subsequent reaction with K158C. Once determined, these conditions were used throughout. In 35 ~ P.;~;g these conditions, the stability of the polymer was monitored.
To activate the rHbl.1 that formed the core of the penta-hemoglobin m~le~ll~, a solution of sulfo-SMCC, 10 mg/ mL in 100 mM sodium borate buffer pH 8.5, was added to a solution of rHbl.1 (30 mg/mL) at a molar ratio of 35: 1 under oxy conditions at 22~C. This was allowed to react for 35 minutes with gentle mixing. The .

succinimide reactive portion of the crosslinker was then quenched by addition ofglycine at a molar ratio of 25: 1 (25 M glycine to 1 M crnsslink~r).
The reaction mixture was then chromatographed on Sephadex G-25 equilibrated in 50 mM Tris-HCl buffer, pH 8.0 to remove quenched crosslinker and5 borate ions and to buffer exchange the activated rHbl.1. Following buffer exchange, the activated rHbl.1 was conc~llL~dL~d to 15 mg/mL and col.v~ d into the deoxy form by deoxygenating on a rotary evaporator (ROTOVAP RE111, BrinkmAnn, Inc., Cllnfi~sgue Road, Westbury, NY) flushed with humidified niLLogen.
Following activation of the rHbl.1 molecule, penta-hemoglobin was 10 synthesized as follows. Activated hemoglobin was concentrated to 15 mg/mL and deoxygenated as above. Deoxy activated rHbl.1 was added to previously deoxygenated K158C (60 mg/ml) at a molar ratio of 1 to 5.5 in drop wise fashion.Crosslinking was allowed to proceed spnntAneously at room l~llL~eldLule (22~C) for 3 hours with gentle mixing. The mixture was then cooled to 4~C and cysteine was 15 added to a final concentration of 8.5 mM and allowed to react for 15 llullul~s to quench the maleimide portion of the ~ros.~ k~r. The resultant product contained a mixture of monohemoglobin, dihemoglobin, trihemoglobin, tetrahemoglobin, pentahemoglobin, hexahemoglobin and higher order mllltim~rs. Approximately 20~o of the mixture was mono- and di-hemoglobin. A~loxilllately 65% of the mixture 20 was most likely tetra-, penta- and hexa-hemoglobin.
The desired molecular weight fraction was resolved from the mixture by size exclusion chromatography on Sephacryl S-200 HR and S-300 HR. Two columns (PhArmAt~iA BPP113, 60 cm length,-one of each resin type) were used in series toachieve the desired frActi-)nA1ion Both columns were equilibrated with phosphate25 buffered saline, pH 7.5. ~lt~rnAtively, the molecular weight fractions were separated using ion exchange chromatography as described below.
The pentaHb fraction exhibited the following equilibrium oxygen binding properties: P50= 32 Torr and NmaX = 2.1 on average for multiple ~ ....;..Ations Example 11 Formation of penta-hemoglobin using sulfo-GMBS crosslinker Sulfo-GMBS (N-y-maleimidobutyrloxy)sucrinimi~le ester) was dissolved at 10 mg/ml in 100 mM sodium borate, pH 8.5. All other steps were perfor~ned identically to the steps disclosed in exarnple 10. The final penta hemoglobin with GMBS
crosslinking was produced in d~ ,u,nately the same yield as in example 10.

W O 9~'40920 PCTrUS96/10420 The pentaHb fraction produced using the GMBS linker exhibited the following equilibrium oxygen binding ~luluel Lies: P50= 30 Torr and Nm~ = 2.1 for multiple~l~l~....i.~ations.

Example 12 Formation of penta-hemo~;lobin - K158C core An entire multimeric Hb-like luroLein can be ~Lelu~ed using only K158C
10 tetramers. The procedure described in Example 10 can be followed identically.Excess crosslinker is removed by, for example, gel filtration or tangential flowultrafiltration in the continued presence of borate buffer. Borate buffer should be maintAined while sulfhydryl reactive crosslinker is being removed. Following adequate removal of the (amine) quenched ~ l;. .k~r, the borate buffer is exchanged 15 completely for another suitable buffer, such as Tris-HCl buffer (using, for example, gel filtration or tangential flow ultrafiltration). This readies the mAt~riAl for the final crosslinking step as described in Example 10 in which pentaHb is produced.

Example 13 Purification of ~lutaraldehyde ~los~lil,ked hemoglobin by anion exchange chromato~graphy Recombinant hemoglobin ~rHbl.1) was expressed, lul eluared and purified as described in PCT publication number WO 95/ 14038, filed November 14, 1994, entitled "Purification of Hemoglobin" herein incorporated by lefelel.ce in its entirety.
This hemoglobin (24g) was concentrated to ~150mg/ml and deoxygenated in a lL
round bottom flask by purging for 5 hours with humidified nitrogen on a BrinkmAnn ROTOVAP RE111 (BrinkmAnn, Inc., Cuntiague Road, Westbury, NY). and crosslinked at 25~C with a 6:1 molar ratio of glutaraldehyde:rHbl.1 (glutaraldehyde was a 10% aqueous solution, diluted from 25% aqueous solution, Sigma Chemical Company, St. Louis, MO). The reaction was t~rminAted after 4 minutes with a 3:1 molar ratio of sodium borohydride: glutaraldehyde then buffer exchanged with ultrafiltration into a 20mM Tris, pH 8.9 (8~C) buffer. The ~:losslinked hemoglobin (21g) was then lûaded onto a 450ml SEPHAROSE-Q ion-exchange resin. After the column was loaded it was washed 8CV's of 20mM Tris, pH 7.6 (8~C) followed by elution with 20mM Tris, pH 7.4 (8~C).

W O 96/40920 PCT~US96/10420 Table Three.
Protein distribution displayed as % of total in eluted fraction 65 kDa 128 kDa 190 kDa>230 kDa Load 27.7 17.7 13.3 40.5 (230-5000 kDa) Breakthrough 99.3 (200-5000 kDa) pH 7.6 Wash 60.9 25.6 7.1 2.5 (250 535 kDa) pH 7.4 wash 1.0 6.6 13.1 79.2 (250-2000 kDa) peak = 410 kDa Example 14 Selective purification of glutaraldehyde ~l~sslil-ked hemoglobin molecular weight fractions using ~H elution Glutaraldehyde ~os~ . .k~l hemoglobin (~ lg) prepared as described in 10 Example 13 was loaded onto a 50ml bed volume SEPHAROSE-Q ion-exchange resin.
The column was washed with loading buffer as described in the previous example followed by elution of the bound ~rol~ with a stepwise pH gradient beginning with a 20mM Tris buffer, pH 7.8 (8~C). The pH steps were decreased in 0.1 pH unit increments with only selected fractions shown here for illustration. The use of very 15 small pH increments il~ Jved resolution of the different molecular weight fractions.

Table Four Protein distribution displayed as ~fo of total in eluted fraction 65 kDa 128 kDa 190 kDa> 230 kDa Load 30.7 19.0 13.5 36.7 (260 4200kDa) Load Break 98.0 (pH 8.9) g260 4200k a) pH 7.8 wash 92.6 5.7 pH 7.7 wash 38.8 54.6 5.1 pH7.5wash 7.4 20.3 29.441.7 (250-2000kDa) peak=285kDa pH 7.3 wash 3.5 8.2 11.2 76.6 (250-4200kDa) peak=4531cDa 4?

CA 022l9242 l997-l0-27 W O 96/40920 PCTrUS96/10420 Example 15 Effect of p~ concentration on separation efficiency Hemoglobin was crosslinked as described in Example 13 and loaded onto a 50ml bedvolume SEPHAROSE-Q ion-exchange resin. The column loads were sequentially increased (Table Five). Loading procedures and elution of the yroLt:ill was the same as in Example 13. As noted in Table Five below, increasing the hemoglobin load yrov~d the efficiency of separation, particularly in the region of 230 - 800 knamolecular weights.
0 Table Five Protein distribution displayed as % of total 1,l ol~ilL in eluted fraction 65kDa 128kDa 190kDa 230- >800kDa 800kDa Load 31.6 19.0 13.8 24.8 10.7 10g/L resin 2.7 6.6 9.3 50.1 31.2 25g/ L resin 2.6 7.3 9.9 54.2 25.9 50g/ L resin 1.9 7.8 11.8 57.7 20.6 80g/ L resin 1.2 6.1 11.4 57.2 23.9 Example 16 Effect of column size on s~al dLion~5 Hemoglobin was cro~link~(l as described in Example 13 and loaded onto either a 50ml or 2100ml bed volume SEPHAROSE-Q ion-exchange resin. The column loads were 30g/L resin and 14.7g/L resin for the 50ml and 2100ml columns respectively. Protein distributions in each column load were similar to those 20 described previously. Loading and elution of the yroLeill was the same as in Example 14. As noted in Table Six below, there was only a minimal effect of column size on efficiency of separation. Therefore, this methodology can be applicable to any scale of separation.

Table Six Pr~,L~il. distribution displayed as ~o of total ylul~ in eluted fraction 651cDa 128 kDa 190 kDa ~230 kDa 2.6 9.7 15.1 72.3 50ml Q-SEPHAROSE (2504200kDa) column peak = 328kDa 3.9 9.1 13.0 73.6 2.1L Q-SEPHAROSE (2504200kDa) column peak = 4261<Da W O g~/103~0 PCT~US96/10420 Ex~mple 17 Selective purification of ~ hemoglobin molecular weight fr~cfion~ by ion ~yrh~nge Super Q 650 C (TosoHaas) was equilibrated with 5 CV's of 20 mM Tris pH=8.9. The column was sized at binding of 15 grams pl~oleil- per liter of resin. The pH and the conductivity of the l,rolein sample were adjusted to match the equilibration buffer and loaded onto the column at a~L~illlately 4.5 grams for an approximately 300 ml column. The column was then washed with 2 column volumes of 20 mM Tris pH=8.9, followed by 7-8 CV wash 25 mM Bis-Tris/Tris pH=7.5, which allowed for removal ofm~nf-m(~ric hemoglobin. The column was eluted using 25 mM Bis-Tris/Tris, 100mM
NaCl pH=7.5. After this purification, only approximately 3~O monomeric hemoglobin remained in the purified pentahemoglobin solution, indicating a 5 - 6 fold purification across the anion exchange step.

The foregoing description of the invention is exemplary for purposes of illustration and explanation. It will be a~ t to those skilled in the art that changes and modifications are possible without departing from the spirit and scope of the invention. It is intended that the following daims be il ll~ Led to embrace all such dhanges and modifications.

W O 96/40920 PCTrUS96/10420 SEQu~ LISTING
(1) GENERAL lN~uKMATION
(i) APPLICANTS: Trimble, Stephen P.
Mathews, Antony J.
Anderson, David C.
Anthony-CAh;ll, Spencer Marquardt, David A.
Madril, Dominic G.
Kerwin, Bruce A.
Epp, Janet K.

(ii) TITLE OF lNv~llON: Modified Hemoglobin-like Proteins and Methods of Purifying Same (iii) NUMBER OF SEQ~N~S: 24 (iv) COR~ SPON~ ADD~ SS:
(A) ADD~ SSEE: Somatogen, Inc.
(B) ~lK~l: 2545 Central Avenue, Suite FDl (C) CITY: Boulder (D) STATE: Colorado (E) ZIP: 80301 (V) CO~ 'U'l'~iK ~T~'AnART.F~ FORM:
(A) MEDIUM TYPE: Diskette, 3.50 inch, 1.4 Mb storage (B) COMPUTER: Apple Macintosh (C) OPERATING SYSTEM: System 7.5 (D) SOFTWARE: Microsoft Word 5.0a (Vi) ~UK~N'l' APPLICATION DATA:
(A) APPLICATION NUMBER:
(B) FILING DATE: 6 June 1996 (C) CLASSIFICATION: not known (vii) PRIOR APPLICATION DATA:
(A) APPLICATION NUMBER: 08/240,712 (B) FILING DATE: November 6, 1992 (vii) PRIOR APPLICATION DATA:
(A) APPLICATION NUMBER: 08/487,431 (B) FILING DATE: June 7, 1995 (viii) AllOKN~Y/AGENT INFORMATION:
(A) NAME: Novelli, Marianne F.
(B) ~ GISTRATION NUMBER: 38S71 (C) ~K~N~/DOC~ T NUMBER: 62/PCT
(viii) AllVK~Y/AGENT INFORMATION:
(A) NAME: Brown, Theresa A.
(B) ~ GISTRATION NUMBER: 32,547 (C) ~K~N~/DOCKET NUMBER: 62/PCT

WO 96/40920 PCT~US96/10420 (ix) TELECOMMUNICATION INFORMATION:
(A) TELEPHONE: 303-541-3324 (B) TELEFAX: 303-444-3013 (2) INFORMATION FOR SEQ ID NO:1:
(i) SEQu~N~ CHAR~CTERISTICS:
(A) LENGTH: 7 (B) TYPE: amino acid (D) TOPOLOGY: unknown to applicant (ii) MOLECULE TYPE: protein (iii) HYPO~ CAL: no (xi) SEQUENCE DESCRIPTION: SEQ ID NO:l:
Ser Gly Gly Ser Gly Gly Ser (2) INFORMATION FOR SEQ ID NO:2:
(i) SEQu~:N~ ~R~TERISTICS:
(A) LENGTH: 14 (B) TYPE: amino acid (D) TOPOLOGY: unknown to applicant (ii) MOLECULE TYPE:
(iii) HYPO~l~llCAL: no (xi) ~QU~:N~ DESCRIPTION: SEQ ID NO:2:

Gly Gly Ser Gly Gly Ser Gly Gly Ser Gly Gly Ser Gly Gly (2) lN~OKMATION FOR SEQ ID NO:3:
(i) SEQu~N~ CHARACTERISTICS:
(A) LENGTH: 16 (B) TYPE: amino acid (D) TOPOLOGY: unknown to applicant (ii) MOLECULE TYPE:
(iii) HYPOl~llCAL: no (xi) SEQu~N~ DESCRIPTION: SEQ ID NO:3:

Ser Gly Gly Ser Gly Gly Ser Gly Gly Ser Gly Gly Ser Gly Gly W O 96/40920 PCT~US96/10420 Ser ( 2 ) lN ~O~MATION FOR SEQ ID NO:4:
(i) ~U~N~ CHARACTERISTICS:
(A) LENGTH: 45 (B) TYPE: nucleic acid (C) STRAN~N~SS: single (D) TOPOLOGY: l;neAr (ii) MOLECULE TYPE: C-term of a gene, Xba I site (iii) HYPO~l~n~llCAL: no (xi) ~QU~:N-~ DESCRIPTION: SEQ ID NO:4:
cGGr~ rG GTCTAGATCA TTAACGGTAT TTCGAAGTCA GAACG 45 ( 2 ) lN~O~ATION FOR SEQ ID NO:5:
( i ) ~yU~N~ ~: CHARACTERISTICS:
(A) LENGTH: 95 (B) TYPE: nucleic acid (C) STRAN~N~SS: single (D) TOPOLOGY: 1;neAr (ii) MOLECULE TYPE: tac promoter sequence, Bam HI-Eag I
sites (iii) HYPOl~n~llCAL: no (xi) SEQu~ DESCRIPTION: SEQ ID NO:5:
GATCCGAGCT GTTr~Ac~A~T AATCATCGGC TCGTA~TG TGTGGAATTG 50 ( 2 ) INFORMATION FOR SEQ ID NO:6 (i) SEQ~N~ CHARACTERISTICS:
(A) LENGTH: 96 (B) TYPE: nucleic acid (C) STR~ :.SS: single (D) TOPOLOGY: l;neAr (ii) MOLECULE TYPE: tac promoter, Bam HI - Eag I sites (iii) HYPOln~llCAL: no (xi) SEQu~N~: DESCRIPTION: SEQ ID NO:6:

INFORMATION FOR SEQ ID NO:7:
(i) SEyu~:N~ CHARACTERISTICS:
(A) LENGTH: 64 (B) TYPE: nucleic acid (C) STRAN~N~SS: single (D) TOPOLOGY: l;neAr (ii) MOLECULE TYPE: 5' end of alpha gene,with EcoR1, BamHl and Eagl sites (iii) ~Y~O~ CAL: no (xi) SEQu~N~: DESCRIPTION: SEQ ID NO:7:
TCGGATTCGA ATTCCAAGCT GTTGGATCCT TAGATTGAAC l~l~lCCGGC 50 C~.~A~C ACCG 64 (2) INFORMATION FOR SEQ ID NO:8:
(i) SEQu~N~: CHARACTERISTICS:
~A) LENGTH: 55 B) TYPE: nucl.eic acid C) STRA-N~ .SS: single (D) TOPOLOGY: l; n~Ar (ii) MOLECULE TYPE: 5' end of beta with Xba I site (iii) HYPOl~llCAL: no (xi) SEQu~N-~ DESCRIPTION: SEQ ID NO:8:

CGGAAGCCCA ATC~A~'-~t~ ~TZ~ GCACCTGACT CCGt~ 50 (2) lN~OKMATION FOR SEQ ID NO:9:
(i) SEQu~N~: CHARACTERISTICS:
(A) LENGTH: 44 (B) TYPE: nucleic acid (C) STRA-N~N~:SS: single (D) TOPOLOGY: l; neAr (ii) MOLECULE TYPE: 3' end of the beta gene with Hind III
site (iii) HYPO~lHr;~ CAL: no (xi) SEQUENCE DESCRIPTION: SEQ ID NO:9:

cccr~cr~ AGCTTCATTA GTGAGCTAGC GCGTTAGCAA CACC 44 (2) lNrOKMATION FOR SEQ ID NO:10:
(i) SEQur;~ CHARACTERISTICS:
(A) LENGTH: 37 (B) TYPE: nucleic acid (C) STRANvr;vNr;SS: single (D) TOPOLOGY: l; n~r (ii) MOLECULE TYPE: mutagenesis reverse primer (iii) HYPOl~r;~l~lCAL: no (xi) ~r;Q~r;NCE DESCRIPTION: SEQ ID NO:10:

TTTAAGCTTC ATTAGTGGTA lll~l~AGCT AGCGCGT 37 (2) INFORMATION FOR SEQ ID NO:11:
(i) SEQur;N~r; r~R~rTERISTICS:
~A) LENGTH: 37 B) TYPE: nucleic acid C) STRANvr;vNr;SS: single D) TOPOLOGY: l;neA~
(ii) MOLECULE TYPE: mutagenesis reverse primer (iii) HYPOl~r;llCAL: no (xi) SEQur;N~r; DESCRIPTION: SEQ ID NO:11:

(2) INFORMATION FOR SEQ ID NO:12:
(i) SE~ur;N~ r~R~rTERIsTIcs:
(A LENGTH: 45 (B TYPE: nucleic acid (C STRA-Nvr;vNr;SS: single (DJ TOPOLOGY 1; n~
(ii) MOLECULE TYPE: mutagenesis reverse primer (iii) HYPOl~r;~l~lCAL: no CA 022l9242 l997-l0-27 W 0 96/~320 PCT~US96/10420 (xi) SEQUENCE DESCRIPTION: SEQ ID NO:12:

GGTG~A~A~A TTTACCTCCT TATCTAGATC ATTAACGGTA TTTCG 45 (2) INFORMATION FOR SEQ ID NO:13:
(i) SEQ~N~: ~ARArTERISTICS:
(A'l LENGTH: 10 (B TYPE: nucleic acid (C ST~AN~ N~ S single (D/ TOPOLOGY: l;neAr (ii) MOLECULE TYPE: Pme I linker (iii) HYPO~ CAL: no (xi) SEQUENCE DESCRIPTION: SEQ ID NO:13:

(2) lN~O~MATION FOR SEQ ID NO:14:
(i) SEQu~N~ r~ARA~TERISTICS:
'A) LENGTH: 58 B) TYPE: nucleic acid C) ST~ANI~ )N~:.~s single ~D) TOPOLOGY: l; n~r (ii) MOLECULE TYPE: oligonucleotide upstream o~ lacI gene (iii) HYPOl~ lCAL: no (xi) SEQu~N~: DESCRIPTION: SEQ ID NO:14:

GGC~.AA~AA~ AGCTTGCGGC CGC~ll~ACA CCATCGAATG GCG~AAAACC 50 (2) INFORMATION FOR SEQ ID NO:15:
(i) SEQ~N~ CHARACTERISTICS:
(A) LENGTH: 69 (B) TYPE: nucleic acid (C) STRAN~N~SS: single (D) TOPOLOGY: l; ne~r (ii) MOLECULE TYPE: downstream side of lacI gene (iii) HYPOl~rillCAL: no (xi) SEQ~riN~r; DESCRIPTION: SEQ ID NO:15:

GGG~-AAA~AG GATC~AAAAA AAAGCCCGCT CATTAGGCGG G~~ ATCAC 50 (2) lNrOKMATION FOR SEQ ID NO:16:
(i) SEQur;N~ CHARACTERISTICS:
(A) LENGTH: 54 B) TYPE: nucleic acid C) STRANDEDNESS: single ~D) TOPOLOGY: l;n~r (ii) MOLECULE TYPE: primer for pBR322 ori positions 3170-(iii) HYPOl~rillCAL: no (xi) SEQuriN~r; DESCRIPTION: SEQ ID NO:16:
CcccrAAAAG GATCCAAGTA GCCGGCGGCC GC~llCCACT GAGCGTCAGA 50 (2) lNrO~MATION FOR SEQ ID NO:17:
(i) ~r;Qur;N~r; CHARACTERISTICS:
~A'l LENGTH: 42 ~B TYPE: nucleic acid C, STRANvrwNr;SS: single D) TOPOLOGY: l; n~r (ii) MOLECULE TYPE: primer for pBR322 ori positions 2380-(iii) HYPOl~r;llCAL: no (xi) SEQu~N~r; DESCRIPTION: SEQ ID NO:17:
GGCGGTCCTG TT~AAA~GCT GCGCTCGGTC GTTCGGCTGC GG 42 (2) INFORMATION FOR SEQ ID NO:18:
(i) ~r;Qur;N~r; CHARACTERISTICS:
(A' LENGTH: 30 (B TYPE: nucleic acid (C STRAN~ N~:~S single (D) TOPOLOGY: unknown to applicant (ii) MOLECULE TYPE: Other nucleic acid -W O 9''~0320 PCTAUS96/10420 (iii) HYPO-l~H~llCAL: no (xi) ~:Q~NCE DESCRIPTION: SEQ ID NO:18:

ACC~Ll~l~A CTAGT~A~T~ CCGTTAATGA 30 (2) INFORMATION FOR SEQ ID NO:l9:
(i) ~QU~: CHARACTERISTICS:
(A) LENGTH: 30 (B) TYPE: nucleic acid (C) STRANDEDNESS: single (D) TOPOLOGY: unknown to applicant (ii) MOLECULE TYPE: Other nucleic acid (iii) HYPOl~:~llCAL: no (xi) SEQ~N~ DESCRIPTION: SEQ ID NO:19:

(2) INFORMATION FOR SEQ ID NO:20:
(i) SE~u~N~ C~AR~cTERIsTIcs:
(A) LENGTH: 30 (B) TYPE: nucleic acid (C) STRAN~N~SS: single (D) TOPOLOGY: unknown to applicant (ii) MOLECULE TYPE: Other nucleic acid (iii) HYPOl~n~llCAL: no (xi) SEQU~N~: DESCRIPTION: SEQ ID NO:20:
~lG~l~GGTA AA~ll~lG~l~ TTGC~ll~lG 30 (2) INFORMATION FOR SEQ ID NO:21:
(i) ~:yu~N~ C~RA~TERISTICS:
(A) LENGTH: 50 (B) TYPE: nucleic acid (C) STR~ : single (D) TOPOLOGY: unknown to applicant (ii) MOLECULE TYPE: Other nucleic acid (iii) nY~Ol~ CAL: no (xi) ~EQ~N~ DESCRIPTION: SEQ ID NO:21:

W O 96110920 PCT~US96/10420 CTAG~A~A~A CCGATCGGGT GG~l~lGGCG ~ GTcTc~l~cA 50 (2) INFORMATION FOR SEQ ID NO:22:

(i) ~u~N~ C~R~CTERISTICS:
~A) LENGTH: 42 B) TYPE: nucleic acid C) STR~N~ N~:SS: single ~D) TOPOLOGY: unknown to applicant (ii) MOLECULE TYPE: Other.nucleic acid (iii) HYPOl~ CAL: no (xi) SEQUENCE DESCRIPTION: SEQ ID NO:22:

Gr.~ A~CGCC AGAGCCACCC GATCGGTATT TA 42 (2) INFORMATION FOR SEQ ID NO:23:
(i) SEQu~:N~ C~R~TERISTICS:
~A) LENGTH: 67 IB) TYPE: nucleic acid ,C) STRANDEDNESS: single ~D) TOPOLOGY: unknown to applicant (ii) MOLECULE TYPE: Other nucleic acid (iii) HYPOl~H~:llCAL: no (xi) SEQUENCE DESCRIPTION: SEQ ID NO:23:

GGCCGCCTTA AGTACCCGGG ~ ~lGCAGA AAGCCCGCCT AATGAGCGGG 50 (2) lN ~MATION FOR SEQ ID NO:24:
( i ) ~U~N~ CHARACTERISTICS:
'A) LENGTH: 67 B) TYPE: nucleic acid C) STRAN~ )N~:SS: single ~D) TOPOLOGY: unknown to applicant (ii) MOLECULE TYPE: Other nucleic acid (iii) HYPOl~ lCAL: no WO 96/40920 PCTrUS96/10420 (xi) SEQU~N~: DESCRIPTION: SEQ ID NO:24:

GATCCCCTAA G~-AAAAAAAA GCCCG~lCAT TA~GCGGGCT TTCTGCAGAA 50

Claims (29)

What is claimed is:

1. A globin-like polypeptide comprising two dialpha domains.

2. The globin-like polypeptide of claim 1, wherein the two dialpha domains are coupled by a peptide linker.

3. The globin-like polypeptide of claim 2, wherein the peptide linker comprises an amino acid sequence of at least seven amino acids.

4. The globin-like polypeptide of claim 2, wherein the peptide linker comprises Ser-Gly-Gly.

5. The globin-like polypeptide of claim 3, wherein the amino acid sequence is SEQ.ID.No.1.

6. The globin-like polypeptide of claim 3, wherein the amino acid sequence is SEQ.ID.No.2.

7. The globin-like polypeptide of claim 3, wherein the amino acid sequence is SEQ.ID.No.3.

8. The globin-like polypeptide of claim 3, wherein the peptide linker comprises an amino acid sequence of at least fourteen amino acids.

9. The globin-like polypeptide of claim 1, wherein the polypeptide is expressed by a recombinant host cell.

10. The globin-like polypeptide of claim 9, wherein the host cell is E. coli.

11. The globin-like polypeptide of any of claims 1 to 8, wherein said polypeptide contains a non-naturally occurring cysteine residue.

12. The globin-like polypeptide of claim 11, wherein said non-naturally occurring cysteine residue is asymmetric.

13. A hemoglobin-like molecule comprising at least two connected globin-like polypeptides of any of claims 1-8 or 11-12, wherein said connection consists of direct connection through a disulfide bond or indirect connection through a chemical crosslinker selected from the group consisting of homobifunctional linkers, heterobifunctional linkers, homopolyfunctional linkers and heteropolyfunctional linkers.

14. A nucleic acid molecule comprising a nucleic acid sequence encoding the globin-like polypeptide of any of claims 1-8 and 11-12.

15. A multimeric hemoglobin-like protein comprising a core hemoglobin-like moiety to which each of at least two other hemoglobin-like moieties are directlyattached and for which substantially all non-core hemoglobins are attached only to the core hemoglobin-like moiety.

16. The multimeric hemoglobin-like protein of claim 15, wherein the core hemoglobin-like moiety is directly attached to at least four other hemoglobin-like moieties.

17. The multimeric hemoglobin-like protein of claim 15, wherein the core hemoglobin-like moiety is directly attached to four other hemoglobin-like moieties.

18. The multimeric hemoglobin-like protein of claim 15, wherein the core hemoglobin-like moiety is different than the other hemoglobin-like moieties.

19. The multimeric hemoglobin-like protein of claim 15, wherein said other hemoglobin-like moieties contain an asymmetric crosslinkable cysteine residue for attachment to the core hemoglobin-like moiety.

20. The multimeric hemoglobin-like protein of claim 19, wherein the other hemoglobin-like moieties are attached to the core hemoglobin-like moiety by a chemical crosslinker.

21. The multimeric hemoglobin-like protein of claim 20, wherein the chemical crosslinker is a succinimidyl 4-(N-maleimidomethyl)cyclohexane-1-carboxylate.

22. The multimeric hemoglobin-like protein of claim 20, wherein the chemical crosslinker is a N-.gamma.-maleimidobutyrloxysuccinimide ester.

23. The multimeric hemoglobin-like protein of either claim 21 or 22, wherein the core hemoglobin-like moiety is rHb1.1 and the other hemoglobin-like moieties are K158C.

24. The multimeric hemoglobin-like protein of claim 15, wherein the core hemoglobin-like moiety is the same as the other hemoglobin-like moieties.

25. The multimeric hemoglobin-like protein of claim 24, wherein the core hemoglobin-like moiety is K158C.

26. A method for making a multimeric hemoglobin-like protein, comprising:
(a) obtaining a first hemoglobin-like moiety having an amino acid capable of attaching to one end of a heterobifunctional linker to form a core hemoglobin-like moiety;
(b) obtaining at least two other hemoglobin-like moieties having an amino acid capable of attaching to the other end of the heterobifunctional linker;
(c) contacting the heterobifunctional linker to the core hemoglobin-like moiety to form a linked moiety; and (d) contacting the other hemoglobin-like moieties to the linked moiety to form the multimeric hemoglobin-like protein.

27. A composition comprising the multimeric hemoglobin-like protein having a core hemoglobin-like moiety to which each of at least two other hemoglobin-like moieties are directly attached.

28. A method for separation of molecular weight fractions of polymerized hemoglobin-like molecules to obtain substantially monodisperse hemoglobin solutions comprising:
(a) contacting a polydisperse mixture of polymerized hemoglobin-like molecules with an ion exchange matrix;
(b) washing the ion exchange matrix with a first buffer;
(c) eluting the ion exchange matrix with a second buffer which can be the same or different from said first buffer to obtain a substantially monodisperse hemoglobin-like solution.

29. The method of claim 24 wherein the ion exchange matrix is an anion exchange matrix.