REACTIVE CYSTEINE RESIDUES

BACKGROUND

Field of the Invention

[0001] The present invention generally relates to fibroblast growth factor-1 (FGF-1). Related Art

[0002] Buried free Cys residues can contribute to an irreversible unfolding pathway that promotes protein aggregation, increases immunogenic potential, and significantly reduces protein functional half-life. As such, mutation of buried free Cys residues may result in significant improvement in properties of storage, reconstitution, and pharmacokinetics, and is therefore of significant practical interest for protein-based therapeutics. SUMMARY

[0003] According to a first broad aspect, embodiments of the present inventions provide a fibroblast growth factor-1 (FGF-1) mutant protein comprising an amino acid sequence of SEQ ID NO: 5; wherein the amino acid sequence of SEQ ID NO: 5 differs from the sequence of wild-type human FGF-1 of SEQ ID NO: 1 by a combination of mutations at positions 16, 66, and 117; and wherein the combination of mutations at position 16, 66, and 117 comprises substitutions of Cys at position 16 by Ser, Ala at position 66 by Cys, and Cys at position 117 by Val.

[0004] According to a second broad aspect, embodiments of the present invention provide an FGF-1 mutant protein comprising an amino acid sequence of SEQ ID NO: 6; wherein the amino acid sequence of SEQ ID NO: 6 differs from the sequence of wild-type human FGF-1 of SEQ ID NO: 1 by a combination of mutations at positions 12, 16, 66, 117, and 134; and wherein the combination of mutations at position 12, 16, 66, 117, and 134 comprises substitutions of Lys at position 12 by Val, Cys at position 16 by Ser, Ala at position 66 by Cys, Cys at position 117 by Val, and Pro at position 134 by Val. [0005] According to a third broad aspect, embodiments of the present invention provide an FGF-1 mutant protein comprising an amino acid sequence of SEQ ID NO: 2, wherein the amino acid sequence of SEQ ID NO: 2 differs from the sequence of wild-type human FGF-1 of SEQ ID NO: 1 by a mutation at position 16, and wherein the mutation at position 16 comprises a substitution of Cys at position 16 by Ser.

[0006] According to a fourth broad aspect, embodiments of the present invention provide an FGF-1 mutant protein comprising an amino acid sequence of SEQ ID NO: 3, wherein the amino acid sequence of SEQ ID NO: 3 differs from the sequence of wild-type human FGF-1 of SEQ ID NO: 1 by a mutation at position 117, and wherein the mutation at position 117 comprises a substitution of Cys at position 117 by Val.

[0007] According to a fifth broad aspect, embodiments of the present invention provide an isolated nucleic acid comprising a sequence encoding an FGF-1 mutant protein; wherein the encoded FGF-1 mutant protein comprises an amino acid sequence of SEQ ID NO: 5, wherein the amino acid sequence of SEQ ID NO: 5 differs from the sequence of wild-type human FGF-1 of SEQ ID NO: 1 by a combination of mutations at positions 16, 66, and 117; and wherein the combination of mutations at position 16, 66, and 117 comprises substitutions of Cys at position 16 by Ser, Ala at position 66 by Cys, and Cys at position 117 by Val.

[0008] According to a sixth broad aspect, embodiments of the present invention provide an isolated nucleic acid comprising a sequence encoding an FGF-1 mutant protein; wherein the encoded FGF-1 mutant protein comprises an amino acid sequence of SEQ ID NO: 6; wherein the amino acid sequence of SEQ ID NO: 6 differs from the sequence of wild-type human FGF-1 of SEQ ID NO: 1 by a combination of mutations at positions at positions 12, 16, 66, 117, and 134; and wherein the combination of mutations at position 12, 16, 66, 117, and 134 comprises substitutions of Lys at position 12 by Val, Cys at position 16 by Ser, Ala at position 66 by Cys, Cys at position 117 by Val, and Pro at position 134 by Val.

[0009] According to a seventh broad aspect, embodiments of the present invention provide an isolated nucleic acid comprising a sequence encoding an FGF-1 mutant protein, wherein the encoded FGF-1 mutant protein comprises an amino acid sequence of SEQ ID NO: 2, wherein the amino acid sequence of SEQ ID NO: 2 differs from the sequence of wild-type human FGF-1 of SEQ ID NO: 1 by a mutation at position 16, and wherein the mutation at position 16 comprises a substitution of Cys at position 16 by Ser.

[0010] According to an eighth broad aspect, embodiments of the present invention provide an isolated nucleic acid comprising a sequence encoding an FGF-1 mutant protein, wherein the encoded FGF-1 mutant protein comprises an amino acid sequence of SEQ ID NO: 3, wherein the amino acid sequence of SEQ ID NO: 3 differs from the sequence of wild-type human FGF-1 of SEQ ID NO: 1 by a mutation at position 117, and wherein the mutation at position 117 comprises a substitution of Cys at position 117 by Val. BRIEF DESCRIPTION OF THE DRAWINGS

[0011] The accompanying drawings, which are incorporated herein and constitute part of this specification, illustrate exemplary embodiments of the invention, and, together with the general description given above and the detailed description given below, serve to explain the features of the invention.

[0012] FIG. 1 is a graph of isothermal equilibrium denaturation profiles for mutant and reference proteins that have Ala, Ser, Thr, and Val mutations at buried free Cys position 16 (Cys16).

[0013] FIG. 2 is a graph of isothermal equilibrium denaturation profiles for mutant and reference proteins that have Ala, Ser, Thr, and Val mutations at buried free Cys position 83 (Cys83).

[0014] FIG. 3 is a graph of isothermal equilibrium denaturation profiles for mutant and reference proteins that have Ala, Ser, Thr, and Val mutations at buried free Cys position 117 (Cys117).

[0015] FIG. 4 is a set of relaxed stereo diagrams of the X-ray crystal structures of wild- type FGF-1 (PDB accession 1JQZ) centered at position 16 with structural features within 4.0 Å radius (CPK model in light grey).

[0016] FIG. 5 is a set of relaxed stereo diagrams of the X-ray crystal structures of a human FGF-1 mutant protein that has the substitution of Cys16ĺ Ala (PDB accession 4Q91, CPK model in light grey) overlaid onto the X-ray crystal structures of wild-type FGF-1(dark grey), respectively, centered at position 16 with structural features within 4.0 Å radius, [0017] FIG. 6 is a set of relaxed stereo diagrams of the X-ray crystal structures of a human FGF-1 mutant protein that has the substitution of Cys16ĺ Ser (PDB accession 4Q9G, CPK model in light grey) overlaid onto the X-ray crystal structures of wild-type FGF-1 (dark grey), respectively, centered at position 16 with structural features within 4.0 Å radius.

[0018] FIG. 7 is a set of relaxed stereo diagrams of the X-ray crystal structures of a human FGF-1 mutant protein that has the substitution of Cys16ĺ Thr (PDB accession 4Q9P, CPK model in light grey) overlaid onto the X-ray crystal structures of wild-type FGF-1 (dark grey), respectively, centered at position 16 with structural features within 4.0 Å radius.

[0019] FIG. 8 is a set of relaxed stereo diagrams of the X-ray crystal structures of wild- type FGF-1 (PDB accession 1JQZ, CPK model in light grey) centered at position 83 with structural features within 4.0 Å radius.

[0020] FIG. 9 is a set of relaxed stereo diagrams of the X-ray crystal structures of a human FGF-1 mutant protein human having the substitution of Cys83ĺ Ala (PDB accession 3FJH, CPK model in light grey) overlaid onto the X-ray crystal structures of wild-type FGF-1 (dark grey), respectively, centered at position 83 with structural features within 4.0 Å radius.

[0021] FIG. 10 is a set of relaxed stereo diagrams of the X-ray crystal structures of human FGF-1 mutant protein having the substitution of Cys83ĺ Ser (PDB accession 3FJE, CPK model in light grey) overlaid onto the X-ray crystal structures of wild-type FGF-1 (dark grey), respectively, centered at position 83 with structural features within 4.0 Å radius.

[0022] FIG. 11 is a set of relaxed stereo diagrams of the X-ray crystal structures of human FGF-1 mutant protein having the substitution of Cys83ĺ Thr (PDB accession 3FJF, CPK model in light grey) overlaid onto the X-ray crystal structures of wild-type FGF-1 (dark grey), respectively,

[0023] FIG. 12 is a set of relaxed stereo diagrams of the X-ray crystal structures of human FGF-1 mutant protein having the substitution of Cys83ĺ Val (PDB accession 3FJJ, CPK model in light grey) overlaid onto the X-ray crystal structures of wild-type FGF-1 (dark grey), respectively, centered at position 83 with structural features within 4.0 Å radius. [0024] FIG. 13 is a set of relaxed stereo diagrams of the X-ray crystal structures of wild- type FGF-1 (PDB accession 1JQZ, CPK model in light grey) centered at position 117 with structural features within 4.0 Å radius.

[0025] FIG. 14 is a set of relaxed stereo diagrams of the X-ray crystal structures of a human FGF-1 mutant protein (PDB accession 4QAL, CPK model in light grey) having the substitution of Cys117ĺ Ala overlaid onto the X-ray crystal structures of wild-type FGF-1 (dark grey), respectively, centered at position 117 with structural features within 4.0 Å radius.

[0026] FIG. 15 is a set of relaxed stereo diagrams of the X-ray crystal structures of a human FGF-1 mutant protein having the substitution of Cys117ĺ Ser (PDB accession 4QC4, CPK model in light grey) overlaid onto the X-ray crystal structures of wild-type FGF-1 (dark grey), respectively, centered at position 117 with structural features within 4.0 Å radius.

[0027] FIG. 16 is a set of relaxed stereo diagrams of the X-ray crystal structures of a human FGF-1 mutant protein having the Cys117ĺ Thr substitution (PDB accession 4QBC, CPK model in light grey) overlaid onto the X-ray crystal structures of wild-type FGF-1 (dark grey), respectively, centered at position 83 with structural features within 4.0 Å radius.

[0028] FIG. 17 is a set of relaxed stereo diagrams of the X-ray crystal structures of a human FGF-1 mutant protein having the substitution of Cys117ĺ Val (PDB code 1JY0, CPK model in light grey) overlaid onto the X-ray crystal structures of wild-type FGF-1(dark grey), respectively, centered at position 117 with structural features within 4.0 Å radius.

[0029] FIG. 18 is an image of the appearance of human corneal endothelial cells (HCECs) at passage 4.

[0030] FIG. 19 is a graph showing the effect of mutant proteins M1, C3, and C2V3 on cell proliferation.

[0031] FIG. 20 is a set of images of the appearance of HCECs showing the effect of mutant protein C3 on cell proliferation.

[0032] FIG. 21 is a graph showing the effect of mutant protein C2V3 on cell proliferation. DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS Definitions

[0033] Where the definition of terms departs from the commonly used meaning of the term, applicant intends to utilize the definitions provided below, unless specifically indicated.

[0034] For purposes of the present invention, it should be noted that the singular forms, “a,”“an” and“the” include reference to the plural unless the context as herein presented clearly indicates otherwise.

[0035] For purposes of the present invention, the term“buried free cysteine” refers to a cysteine amino acid in a protein wherein the cysteine is not normally involved in a disulfide (i.e. half a cysteine), and is located within the solvent excluded interior of the protein (i.e. is not normally surface accessible).

[0036] For purposes of the present invention, the term“comprising”, the term“having”, the term“including,” and variations of these words are intended to be open-ended and mean that there may be additional elements other than the listed elements.

[0037] For purposes of the present invention, the term“correspond” and the term “corresponding” in reference to a protein sequence refer interchangeably to an amino acid position(s) of a protein, such as a mutant FGF protein, that is equivalent or corresponds to an amino acid position(s) of one or more other protein(s), such as a wild-type FGF protein, according to any standard criteria known in the art. An amino acid at a position of a protein may be found to be equivalent or corresponding to an amino acid at a position of one or more other protein(s) based on any relevant evidence, such as the primary sequence context of the each amino acid, its position in relation to the N-terminal and C-terminal ends of its respective protein, the structural and functional roles of each amino acid in its respective protein, etc. For proteins having similar or nearly identical polypeptide sequences, a “corresponding” amino acid(s) and corresponding amino acid position(s) between proteins may be determined or deduced by sequence alignment and comparison. However, “corresponding” amino acid(s) for two or more proteins may have different amino acid position numbers or numbering (e.g., when counted from the N-terminus of each protein) since the two or more proteins may have different lengths and/or one or more substitutions, insertions, deletions, etc. For example, related proteins may have deletions or insertions in relation to each other that offset the numbering of their respective or corresponding amino acid sequences (i.e., based on their primary structure or sequence). An amino acid position(s) of a protein, such as a mutant FGF protein, may“correspond” to an amino acid position(s) of one or more other protein(s) if the amino acid positions are structurally equivalent or similar when comparing the three-dimensional structures (i.e., tertiary structures) of the respective proteins. A person skilled in the art would be able to determine“corresponding” amino acids and/or“corresponding” amino acid positions of two or more proteins based on their protein sequences and/or protein folding or tertiary structure.

[0038] For purposes of the present invention, the term“deletion” refers to the absence of an amino acid residue from the polypeptide sequence of a mutant protein.

[0039] For purposes of the present invention, the term“disulfide bond” refers to a covalent bond that is usually derived by the coupling of two thiol groups. The linkage is also called an S-S bond or disulfide bridge. Disulfide bonds are usually formed from the oxidation of sulfhydryl (-SH) groups. In proteins, disulfide bonds are formed between the thiol groups of cysteine residues by the process of oxidative folding. Disulfide bonds play an important role in the folding and stability of some proteins.

[0040] For purposes of the present invention, the term“free cysteine” refers to a cysteine amino acid in a protein wherein the cysteine is not normally involved in a disulfide (i.e. half a cysteine).

[0041] For purposes of the present invention, the term“functional half-life” of a FGF protein refers to the amount of time it takes for the activity or effect of a FGF protein (e.g., a mutant FGF protein) to become reduced by half. For example, the functional half-life may be based on the activity of a FGF protein over time in inducing growth, proliferation, and/or survival of cells, such as according to a cultured fibroblast proliferation assay. The functional half-life of a protein may be different than the thermostability of the same protein since these are separable properties of proteins. For example, a protein may be mutated such that the thermostability of the protein is decreased while its functional half-life is increased.

[0042] For purposes of the present invention, the term“gauche” refers to conformational isomers (conformers) where two vicinal groups are separated by a 60° torsion angle. According to IUPAC, gauche are groups that have a "synclinal alignment of groups attached to adjacent atoms." In stereochemistry, gauche interactions hinder bond rotation. For example, sighting along the C2-C3 bond in staggered butane, there are two possible relative potential energies. The two methyl groups can be in an anti-bonding relationship, or offset at sixty degree dihedral angles. In the latter configuration, the two methyls are said to be in a gauche relationship, and the relative potential energy of each methyl-methyl gauche interaction is 0.9 kilocalories per mole (4 kJ/mol). In general a gauche rotamer is less stable than an anti-rotamer.

[0043] For purposes of the present invention, the term“growth factor” refers to a naturally occurring substance capable of stimulating cellular growth, proliferation and cellular differentiation. It is usually a protein or a steroid hormone and is important for regulating a variety of cellular processes.

[0044] For purposes of the present invention, the term“half-life” refers to the time required for half the quantity of a substance such as a protein or a drug deposited in a living organism to be metabolized or eliminated by normal biological processes.

[0045] For purposes of the present invention, the term“hydrogen bond” refers to an electrostatic attractive force between the hydrogen attached to an electronegative atom of one molecule and an electronegative atom of a different molecule. Usually the electronegative atom is nitrogen, oxygen, or fluorine, which has a partial negative charge. The hydrogen then has the partial positive charge. The name hydrogen bond is something of a misnomer, as it is not a true bond but a particular strong diploe-dipole attraction, and should not be confused with a covalent bond. Intramolecular hydrogen bonding is partly responsible for the secondary and tertiary structure of proteins and nucleic acids. It plays an import role in the structure of polymers, both synthetic and natural.

[0046] For purpose of the present invention, the term“mitogen” refers to a compound having mitogenic potency.

[0047] For purposes of the present invention, the term“mitogenic potency” refers to an ability to stimulate cell division, typically quantified using fibroblast cells such as NIH 3T3 fibroblasts, etc. The ability to stimulate cell division is often quantified by counting the number of cells after an appropriate period of time (e.g. typically 48-72 hrs) after application of a mitogen.

[0048] For purposes of the present invention, the term“mutation” refers to a change in the polypeptide sequence of a protein.

[0049] For purposes of the present invention, the term“polyhistidine-tag” and the term “His-tag” refer to an amino acid motif in proteins that consists of at least six histidine(His) residues, often at the N- or C-terminus of the protein. It is also known as hexa histidine- tag, 6xHis-tag, His6 tag and by the trademarked name His-tag (registered by EMD Biosciences).

[0050] For purposes of the present invention, the term“polypeptide” refers to a polymer chain of amino acids joined together by peptide that is unbranched.

[0051] For purposes of the present invention, the term“protein aggregation" refers to newly synthesized proteins may not fold correctly, or properly folded proteins can spontaneously misfold. In these cases, if the unfolded protein is not assisted in re-folding or degraded, the unfolded protein may aggregate. In this process, exposed hydrophobic portions of the unfolded protein may interact with the exposed hydrophobic patches of other unfolded proteins, spontaneously leading to protein aggregation. Protein aggregation is intimately tied to protein folding and stability.

[0052] For purposes of the present invention, the term“residue” and the term“amino acid residue” refer to an amino acid within a polypeptide or a protein.

[0053] For purposes of the present invention, the term“substitution” refers to the replacement of an amino acid residue at a specific position along the polypeptide sequence of a mutant protein.

[0054] For purposes of the present invention, the term“sulfhydryl group” refers to a functional group consisting of a sulfur bonded to a hydrogen atom. The sulfhydryl group, also called a thiol, is indicated in chemistry nomenclature by "-thiol" as a suffix and "mercapto-" or "sulfanyl" as a prefix. Thiols have great affinity for soft metals. Sulfhydryls play an important role biochemistry, as disulfide bonds connect necessary amino acids together for functional purpose in secondary, tertiary, or quaternary proteins structures. [0055] For purposes of the present invention, the term“thiol” refers to an organosulfur compound that contains a carbon-bonded sulfhydryl (-C-SH or R-SH) group where R represents an alkane, alkene, or other carbon-containing group of atoms.

[0056] For purposes of the present invention, the term“vector” refers to a vehicle used to transfer genetic material such as DNA sequences from the donor organism to a target cell of the recipient organism. A vector may further refer to a plasmid vector, a binary vector, a cloning vector, a viral vector, a shuttle vector, an expression vector, etc.

[0057] For purposes of the present invention, the terms“identical” or“identity” refer to the percentage of amino acid residues of two or more polypeptide sequences having the same amino acid at corresponding positions. For example, a protein that is at least 90% identical to a polypeptide sequence will have at least 90% of its residues that are the same as those in the polypeptide sequence at corresponding positions.

[0058] For purposes of the present invention, the terms“thermodynamic stability” or “thermostability” of a protein refer interchangeably to the ability of a protein to maintain its tertiary structure (i.e., to resist denaturation or unfolding) at a given temperature or in the presence of a denaturant. The thermostability of a protein (e.g., a mutant FGF protein) may be expressed in terms of the Gibb’s free energy equation relative to a standard (e.g., wild-type FGF protein) according to known methods.

[0059] For the purposes of the present invention, the term“epithelial-mesenchymal transition (EMT)” refers to a process of transdifferentiation of epithelia cells into motile mesenchymal cells. In other words, it refers to a process by which epithelia cells lose their cell polarity and cell to cell adhesion, and gain migratory and invasive properties to become mesenchymal stem cells. EMT is essential for numerous developmental processes including mesoderm formation and neural tube formation. EMT is integral in development, wound healing, stem cell behavior, and contributes pathologically to fibrosis and cancer progression.

[0060] For the purposes of the present invention, the term“half maximal effective concentration (EC ₅₀ )” refers to the concentration of a composition such as a drug which induces a response halfway between the baseline and maximum after a specified exposure time. It is commonly used as a measure of the composition’s potency. The EC ₅₀ of a graded dose response curve therefore represents the concentration of a composition where 50% of its maximal effect is observed. The EC ₅₀ of a quantal dose response curve represents the concentration of a composition where 50% of the population exhibit a response,[3] after a specified exposure duration.

Description

[0061] Embodiments disclosed herein provide compositions encompassing mutants proteins of FGF-1, wherein buried free cysteine residues at positions 16, 83, and 117 in wild- type human FGF-1 are substituted respectively by serine, valine, alanine, and/or threonine. Embodiments disclosed herein also provide FGF-1 mutant proteins that encompass combinations of amino acid substitutions at multiple positions in wild-type human FGF-1.

[0062] Among the least abundant amino acids found in proteins, ¹ Cys satisfies a number of unique and critically important structural and functional roles—including disulfide bonds, structurally important H-bonds, metal coordination, activity regulation via posttranslational modification, and redox-active as well as non-redox catalysis. ^1,2 A statistical analysis of Cys residues from a set of 378 non-redundant protein X-ray structures reported that Cys occupies 0.73% of all positions. ³ Free Cys residues (i.e., those not involved in disulfide bonds) are approximately evenly distributed between solvent-exposed and buried positions within proteins. ¹ Free Cys residues bear a reactive sulfhydryl group, which in buried free Cys residues is structurally protected from redox reactions that can result in thiol adducts or other oxidized derivatives. However, due to the dynamic nature of protein structures (e.g., transient global unfolding or localized structural“breathing”) buried free Cys residues can be exposed to solvent and participate in redox chemistry. Oxidized derivatives of free Cys residues, including formation of thiol adducts or cysteic acid, can effectively block refolding--resulting in an irreversible pathway that, by Le Chatelier’s principle, will continuously drive protein unfolding. Consequently, buried free Cys residues can play a key role in effectively regulating protein functional half-life. ^4,5,6,7,8 [0063] The presence of buried free Cys residues in protein biopharmaceuticals, notably monoclonal antibodies, is problematic because they can effectively limit shelf-life as well as induce immunogenic aggregates. ^9,10,11,12 A recent study of IgG1 and IgG2 antibodies showed that roughly one third of recombinant antibody molecules from a standardized industrial preparation contain a single open disulfide bond (that is, two free Cys residues that should form a disulfide bond in the native state structure). ^9,12 In addition, all natively free Cys residues within monoclonal IgG’s are buried in the hydrophobic core and mediate aggregation in response to agitation stress (as can occur during shipment). ¹² [0064] Cyso Ser mutation is a common approach to substitute Cys because Ser is isosteric to Cys (i.e., substituting the Ser OJ for the Cys SJ). Several engineered protein therapeutics have successfully introduced Cyso Ser mutations to improve shelf life, including interleukin- 2 (Proleukin ^® ), interferon E1b (Betaseron ^® ), and granulocyte colony-stimulating factor (Neulasta™). ¹³ Studies have shown that while Cyso Ser mutations can be substantially destabilizing, the mutant proteins nonetheless exhibit substantially increased functional half- life compared to the wild-type protein. ^4,8 [0065] However, Cys residues have a comparatively high hydrophobicity index, similar to Met, Phe and Tyr, which suggests chemical compatibility if buried within the hydrophobic core environment. ³ Thus, it is not clear whether Cyso Ser mutation is the best de facto substitution to effectively eliminate buried free Cys residues, or whether small hydrophobic amino acids such as Ala or Val might be more appropriate substitutions.

[0066] Although free Cys residues are commonly eliminated by mutation to Ser (an isosteric substitution) or Ala (to preserve hydrophobicity), it is unknown whether there is a universally preferred substitution for buried free Cys residues, or whether each situation must be considered as uniquely different. Embodiments of the present invention provide solutions for selecting substitution position for buried Cys residues.

[0067] Some embodiments of the present invention are based on the study of three uniquely different buried free Cys residues in fibroblast growth factor-1 (FGF-1)--a protein with therapeutic potential to accelerate wound healing but having poor thermostability and a proclivity to unfold, form mixed thiols, and aggregate. ^{14,15,16,17,18} [0068] Wild-type FGF-1 has three free cysteine (Cys) residues at positions 16, 83, and 117 (residue numbers described in this application are all based on the 140 amino acid numbering scheme of wild-type human FGF-1 of SEQ ID NO: 1). The side chains of Cys16 and Cys83 are fully buried in protein interior, while Cys117 exhibits two conformers in which one is buried and the other is partially exposed to solvent (~12 Å ² accessible surface area). ^19,20 Recognizing the comparatively small size of Cys, and that it is known to occupy both hydrophilic and hydrophobic environments, a conservative set of amino acids to investigate Cys substitution includes alanine (Ala), serine (Ser), threonine (Thr), and valine (Val).

[0069] Embodiments of the present invention provide mutations of each of the three buried free Cys residues at positions Cys16, Cys83, and Cys117 in the amino acid sequence of wild-type human FGF-1, wherein each of the three buried free Cys residues is substituted by Ala, Ser, Thr, or Val, respectively. The resulting mutant proteins are subjected to X-ray structural and thermodynamic characterization. The characterization results of thermodynamic and X-ray structural properties of Ala, Ser, Thr, and Val substitutions at positions Cys16 and Cys117, combined with the characterization results of thermodynamic and X-ray structural properties of Ala, Ser, Thr, and Val substitutions at positions Cys83, ²¹ provide insight into the diverse structural and stability properties of buried free Cys residues.

[0070] According to embodiments of the present invention, mutation for each of the three buried free Cys residues is almost universally destabilizing, indicating a general structural optimization for each Cys within the protein core. However, each of the three positions, Cys16, Cys83, and Cys117, is uniquely different with regard to both tolerance to Cys mutation and the most stable mutant amino acid. In this regard, prediction of the response to buried Cys mutation is surprisingly difficult, involving structural changes to maintain, or effectively substitute, a critical set of local H-bond interactions.

[0071] Optimization of Cys mutation may be described by the set of local H-bond interactions involving the buried Cys, combined with a general structural rigidity– such that the uniquely long H-bond interactions involving Cys cannot be maintained by substitution of other small polar amino acids such as Ser or Thr. The protein response to Cys mutation follows one of two general solutions to maintain local H-bond interactions: the first involves structural collapse to shorten H-bond distances, and the other is expansion to permit introduction of a buried solvent to serve as novel H-bond partner. In this regard, expansion appears to be more energetically costly as regards mutant protein stability.

[0072] According to discovery of the present invention, substitutions for the buried Cys result in different structural effects, and the most stable mutation is different for each of the three buried Cys positions in FGF-1. The least destabilizing average substitution at each position is Ala, and not isosteric Ser. Thus, the best average mutant as regards thermostability is Ala.

[0073] Embodiments of the present invention provide FGF-1 mutant proteins that have mutations at Cys16 and/or Cys117. In one embodiment of the present invention, an FGF-1 mutant protein has an amino acid sequence of SEQ ID NO: 2. The amino acid sequence of SEQ ID NO: 2 differs from the amino acid sequence of wild-type human FGF-1 of SEQ ID NO: 1 by a single substitution of Cys16 by Ser (C16S). One embodiment of the present invention provide an FGF-1 mutant protein that has an amino acid sequence of SEQ ID NO: 3. The amino acid sequence of SEQ ID NO: 3 differs from the amino acid sequence of wild- type human FGF-1 of SEQ ID NO: 1 by a single substitution of Cys117 by Val (C117V).

[0074] Since the mutation at each buried free Cys is almost universally destabilizing, to construct stable mutants, other mutations that increase the half-life of FGF-1 may be further introduced to the FGF-1 mutant proteins which already have substitutions at Cys16 and/or Cys117.

[0075] Mutations at both positions 12 and 134 contribute to increased stability, with positions 12 mutations primarily increasing the rate of folding and position 134 mutations primarily decreasing the rate of unfolding. The combination of the substitution of Lys at position 12 by Val (K12o V) and the substitution of Pro at position 134 by Val (P134o V) also exhibits about 30-fold increase in mitogenic potency.

[0076] An FGF-1 mutant protein is constructed to encompasses a combination of triple mutations at positions 12, 117, and 134, wherein the mutations encompass substitutions of Lys12 by Val (K12o V), Cys117 by Val (C117o V), and Pro134 by Val (P134o V) and result in a mutant protein FGF K12V/C117/P134V (M1) with an amino acid sequence of SEQ ID NO: 4. ²⁷ The mutant protein M1 is about 18.6 kJ/mol more stable than wild-type FGF-1 and thus permits folding of highly destabilizing mutations.

[0077] Positions 12, 117, and 134 are distal to site 16 and form no direct packing interactions. Therefore, the position 12, 117, and 134 stabilizing mutations are considered essentially independent of the position 16 and are further introduced in combination with the mutation at position 16. Additionally, it is discovered that a substitution of Ala66 by Cys in human FGF-1 (amino acid number is based on the amino acid sequence of SEQ ID NO: 1 for wild-type human FGF-1) encourages the formation of a disulfide bond between Cys66 and Cys83.

[0078] Accordingly, embodiments of the present invention provide FGF-1 mutant proteins that, in addition to the substitution of C16S, further encompass mutations at positions 12, 66, 117, and/or 134.

[0079] One embodiment of the present invention provides a mutant protein FGF-1 C16S/A66C/C117V with an amino acid sequence of SEQ ID NO: 5. The amino acid sequence of SEQ ID NO: 5 differs from the amino acid sequence of wild-type human FGF-1 of SEQ ID NO: 1 by a combination of mutations of mutations at positions 16, 66 and 117, wherein the mutations comprise the substitutions of Cys16 by Ser (C16S), Ala66 by Cys (A66C), and Cys117 by Val (C117V). It is discovered that a substitution of Ala66 by Cys in human FGF-1 (amino acid number is based on the amino acid sequence of SEQ ID NO: 1 for wild-type human FGF-1) encourages the formation of a disulfide bond between Cys66 and Cys83. Biophysical characterization shows that a new disulfide bond is formed between the Cys66 and Cys 83 (the 66-83 disulfide bond) in this mutant protein FGF C16S/A66C/C117V. The new covalent bond reduces the entropy of the denatured state without significantly affecting the entropy of the native state. Thus, the disulfide bond stabilizes the mutant protein FGF C16S/A66C/C117V and a mutant protein FGF that has the only 66-83 disulfide bond is more stable than wild-type human FGF-1. In addition, the removal of the other two Cys residues at positions 16 and 117 creates a protein that has no free Cys residues and thus is minimally reactive to oxidation.

[0080] One embodiment disclosed herein provides a mutant protein C3, wherein the C3 encompasses a His-tag linked to the N-terminus of the mutant protein FGF C16S/A66C/C117V. The His-tag at the N-terminus of the mutant protein C3 will not alter the stability and biological activity of the mutant protein C3 and can be used for the purification of the mutant protein C3 after expression of the mutant protein C3. It is discovered that the mutant protein C3 has both mitogenic effect and morphological effect on cultured cells. Thus, the mutant protein C3 may be used to stimulate the proliferation of cells and/or to reverse the process of epithelial-mesenchymal transition (EMT). [0081] According to one embodiment disclosed herein, an FGF-1 mutant protein may combine the mutations at positions 12, 16, 66, 117, 134, wherein the mutations comprise the substitutions of Lys12 by Val (K12V), Cys16 by Ser (C16S), Ala66 by Cys (A66C), Cys117 by Val(C117V) and Pro134 by Val(P134V) and result in a mutant protein FGF K12V/C16S/A66C/C117V/P134V with an amino acid sequence of SEQ ID NO: 6. The amino acid sequence of SEQ ID NO: 5 differs from the amino acid sequence of wild-type human FGF-1 of SEQ ID NO: 1 by a combination of mutations of K12V, C16S, A66C, C117V and P134V. The mutant protein FGF K12V/C16S/A66C/C117V/P134V combines the amino acid substitutions made in mutant protein M1 and in mutant protein FGF C16S/A66C/C117V. Same as the mutant protein FGF C16S/A66C/C117V, the mutant protein FGF K12V/C16S/A66C/C117V/P134V has no free Cys residue and thus is minimally reactive to oxidation. The mutant protein FGF K12V/C16S/A66C/C117V/P134V has only one disulfide bond formed between Cys66 and Cys83 and is more stable than wild-type human FGF-1. Since the mutant protein C2V3 include the substitutions of K12V/C117/V/P134V as that in the mutant protein M1, the mutant protein C2V3 not only has all the properties of the mutant protein C3 but also has the stabilizing property of the mutant protein M1.

[0082] One embodiment of the present invention also provides a mutant protein C2V3, wherein the mutant protein C2V3 encompasses a His-tag linked to the N-terminus of the mutant protein FGF K12V/C16S/A66C/C117V/P134V. The His-tag at the N-terminus of the mutant protein C2V3 will not alter the stability and biological activity of the mutant protein C2V3 and can be used for purification of the mutant protein C2V3 after expression of the mutant C2V3. According to the discovery of the present invention, the mutant protein C2V3 has mitogenic effect on cultured cells and thus may be used for stimulating cell proliferation.

[0083] One embodiment of present invention provides an isolated nucleic acid that comprises a sequence encoding the FGF-1 mutant protein, and the encoded FGF-1 mutant protein has the amino acid sequence of SEQ ID NO: 2. An isolated nucleic acid may further comprise a sequence encoding a mutant protein having a His-tag linked to the N-terminus of the FGF-1 mutant protein having the amino acid sequence of SEQ ID. NO: 2. [0084] In another embodiment of the present invention, an isolated nucleic acid provided has a sequence encoding the human FGF-1 mutant protein that has the amino acid sequence of SEQ ID NO: 3. An isolated nucleic acid may further comprise a sequence encoding a mutant protein having a His-tag linked to the N-terminus of the huamn FGF-1 mutant protein having the amino acid sequence of SEQ ID. NO: 3.

[0085] According to one embodiment disclosed herein, an isolated nucleic acid may also have a sequence encoding an FGF-1 mutant protein that has the amino acid sequence of SEQ ID. NO: 5. Another embodiment disclosed herein provides an isolated nucleic acid having a sequence encoding an FGF-1 mutant protein that has the amino acid sequence of SEQ ID. NO: 6. An isolated nucleic acid may further comprise a sequence encoding a mutant protein having a His-tag linked to the N-terminus of the mutant protein FGF C16S/A66C/C117V. An isolated nucleic acid may also comprise a sequence encoding a mutant protein having a His-tag linked to the N-terminus of the mutant protein FGF K12V/C16S/A66C/C117V/P134V.

[0086] Each of the isolated nucleic acid identified above can also be cloned into a vector and stored, amplified and expressed in compatible host cells. One skilled in the art knows how to purified the isolated nucleic acid from compatible host cells and how to purified mutant proteins of FGF expressed in compatible host cells.

[0087] Embodiments disclosed herein are not limited to mutations applied to wild-type human FGF-1. The mutations disclosed herein apply to any member of the human/mouse FGF family, including FGF-1, FGF-2, FGF-3, FGF-4, FGF-5, FGF-6, FGF-7, FGF-8, FGF-9, FGF-10, FGF-11, FGF-12, FGF-13, FGF-14, FGF-15, FGF-16, FGF-17, FGF-18, FGF-19, FGF-20, FGF-21, FGF-22, and FGF-23, and FGF-24. The numbers and locations of mutations may vary among the members of the FGF family but are corresponding to the numbers and locations based on the 140 amino acid number scheme of wild amino acid residues wild-type human FGF-1 of SEQ ID NO: 1.

[0088] Mutations at the buried free Cysteine residues in FGF-1 applies to all proteins having buried free Cysteines. The 22 members of the mouse/human FGF family contain a conserved cysteine residue at a position corresponding to the positon 83 in the amino acid sequence of SEQ ID NO:1. Accordingly, to increase the stability of an FGF family member, Ala at a position corresponding to Ala 66 of the amino acid sequence of SEQ. ID NO: 1 may be substituted by Cys, resulting a formation of a disulfide bond corresponding to the 66-83 bond based on the numbering scheme of sequence of SEQ ID NO:1,

[0089] The present invention is further defined in the following examples, in which all parts and percentages are by weight and degrees are Celsius, unless otherwise stated. It should be understood that these examples are given by way of illustration only. From the above discussion and these examples, one skilled in the art may ascertain the essential characteristics of embodiments of the present invention. Without departing from the spirit and scope thereof, one skilled in the art may make various changes and modifications of the invention to adapt it to various usages and conditions. All publications, including patents and non-patent literature, referred to in this specification are expressly incorporated by reference herein. EXAMPLES

Example 1

Systematic mutational analysis to substitutions of buried free cysteines in FGF-1

[0090] In this example, each of three buried free Cys residues in FGF-1, Cys16, Cys83, and Cys117 is substituted by Ala, Ser, Thr, or Val, respectively, and structural and thermodynamic properties of the resulting mutant proteins are analyzed.

Materials and Methods

Protein mutagenesis and expression

[0091] A codon-optimized synthetic gene encoding 140 amino acids form of human FGF- 1 (SEQ ID NO: 1) with an N-terminal six His tag is cloned into the pET21a(+) expression vector (the His tag has shown no influence upon stability and mitogenic activity). ²² Residue Cys16 and Cys117 are mutated respectively to Ala, Ser, Thr, or Val using the QuikChange ^TM site-directed mutagenesis protocol (Agilent Technologies, Santa Clara, CA), and each mutation is confirmed by DNA sequencing (Biomolecular Analysis Synthesis and Sequencing Laboratory, Florida State University). Recombinant wild-type (WT) FGF-1 and mutant proteins are expressed from pET21a(+)/BL21(DE3) E.coli. After induction with 1mM Isopropyl-E-D-thio-galactoside (IPTG), the incubation temperature is decreased from 37°C to 16°C for overnight expression. The expressed protein is purified utilizing sequential column chromatography on Ni-NTA affinity resin (Qiagen, Valencia CA) and heparin Sepharose resin (GE Life Sciences, Pittsburgh PA). Purified protein is buffer exchanged into "crystallization buffer" (50 mM sodium phosphate, 100 mM sodium chloride, 10 mM ammonium sulfate, and 2 mM dithiothreitol (DTT) pH 7.5) using an 8 kDa molecular weight cutoff membrane tubing (Spectrum Industries Inc., Chippewa Falls, WI).

Isothermal equilibrium denaturation

[0092] FGF-1 contains a single buried tryptophan residue (Trp107) whose fluorescence is internally quenched in the native state, and quenching is subsequently released upon denaturation and providing a spectroscopic probe of unfolding. Isothermal equilibrium denaturation by guanidine hydrochloride (GuHCl) is quantified by probing fluorescence intensity. Wild-type (WT) FGF-1 and mutant protein samples are equilibrated overnight in crystallization buffer at 298 K in 0.1 M increments of GuHCl with final protein concentration of 5 M. A Varian Eclipse fluorescence spectrophotometer (Varian Medical Technologies, Palo Alto CA) is used for data collection. Triplicate scans is collected and averaged; buffer background is collected and subtracted from the protein scans. Fluorescence scans are integrated to quantify the total fluorescence as a function of denaturant concentration. Data are analyzed using a six parameter two-state model: ²³ [0093] where [D] is the denaturant concentration, F _0N and F _0D are the 0M denaturant intercepts for the native and denatured state baselines, respectively, and S _N and S _D are the slopes of the native and denatured state baselines, respectively. ' G ₀ and m describe the linear function of the unfolding free energy versus denaturant concentration. The effect of a given mutation upon the stability of the protein (ΔΔG) is calculated by taking the difference between the C _m values for wild-type and mutant proteins and multiplying by the average of the m values, as described by Pace and Scholtz: ²⁴

[0094] where a negative value indicates the mutation is stabilizing in comparison to the wild type protein.

Crystallization, X-ray data collection and structural refinement of FGF-1 mutant proteins

[0095] FGF-1 mutant protein in crystallization buffer is concentrated to 6-10 mg/ml for crystallization trials. Crystals is grown in a 25 °C incubator using the hanging drop method (combining 5.5 l of protein and 2.5 μl of mother liquor). Diffraction quality crystals of mutant Cys16ĺ Thr, Cys117ĺ Ala, and Cys117ĺ Ser are obtained from a reservoir solution containing 0.7-1.0 M sodium citrate and 0.1 M imidazole pH 8.0-pH 8.5. Mutant Cys16ĺ Ala and Cys16ĺ Ser are crystallized from a reservoir solution containing 3.4 M and 1.8 M sodium formate. Mutant Cys117ĺ Thr is crystallized from two different reservoir solutions: one containing 0.9 M sodium citrate with 0.1 M imidazole pH 8.0, and the other containing 4.0 M sodium formate with 0.45 M ammonium sulfate.

[0096] Crystals are briefly dipped in their reservoir solution formulated with 25% glycerol as a cryoprotectant, subsequently mounted using Hampton Research (Aliso Viejo, CA) cryoloops and frozen either directly in liquid nitrogen or in a stream of nitrogen gas at 100 K. Diffraction data for all mutant proteins except for mutant Cys117ĺ Ala are collected at the Southeast Regional Collaborative Access Team (Ser-CAT) 22-BM beam line at the Advanced Photon Source at Argonne National Laboratory using a MarCCD225 detector (MarUSA, Evanston IL). Diffraction of mutant Cys117ĺ Ala is performed using an in-house Rigaku RU-H3R rotating-anode X-ray source (Rigaku, The Woodlands, TX) equipped with Osmic Blue confocal mirrors and a Rigaku R-axis IV++ image plate detector (Rigaku, The Woodlands, TX). Diffraction data are indexed, integrated and scaled using the HKL2000 software package.

[0097] Wild-type FGF-1 (PDB code 1JQZ) is used as search model in molecular replacement for all mutant X-ray data using the PHENIX ²⁵ software package. Structure refinements utilize the PHENIX software package with 5% of the data in the reflection files set aside for R _free calculations. Model building and visualization utilize the COOT ²⁶ molecular graphics software package. Results

Isothermal equilibrium denaturation

Cys16→ Xxx mutations

[0098] As shown in FIG. 1 and Table 1, the Cys16ĺ Xxx (Xxx=Ala/Ser/Thr/Val) mutants constructed in the wild-type FGF-1 background are significantly destabilizing in each case, as evidenced by substantial precipitation during purification, and are therefore constructed within a stabilizing background mutant form of FGF-1. The stabilizing background form of FGF-1 utilized is the mutant protein M1 that has triple mutations of Lys12ĺ Val/Cys117ĺ Val/Pro134ĺ Val 27. This mutant is about 18.6 kJ/mol more stable than wild-type FGF-1 and thus permits folding of highly destabilizing mutations. Furthermore, positions 12, 117, and 134 are distal to site 16 and form no direct packing interactions; thus, the position 12, 117, and 134 stabilizing mutations are considered essentially independent of the position 16 Ala, Ser, Thr, and Val mutations. ΔΔG values for the Cys16ĺ Xxx mutations ranged between about +9.1 to about +13.8 kJ/mol, significantly destabilizing in each case, and with a stability differential of about 4.7 kJ/mol between the set of mutants (Table 1). Of the four Ala, Ser,Thr, and Val substitutions, the least destabilizing substitution of Cys16 is Ser, while the most destabilizing is Thr. The folding cooperativity m- value of the position 16 mutations is essentially conserved in comparison to the Lys12ĺ Val/Cys117ĺ Val/Pro134ĺ Val reference protein, the exception being the Val mutation whose m-value decreased significantly from about 18.3 to about 15.8 kJ mol ^-1 M ^-1 (Table 1).

Cys83→ Xxx mutations

[0099] As shown in FIG. 2 and Table 1, the Cys83ĺ Xxx (Xxx=Ala/Ser/Thr/Val) mutations have previously been described ²⁸ and are included here for purposes of comparison to mutations at positions Cys16 and Cys117. The Cys83 mutations are destabilizing in each case (ΔΔG ranging from about +4.1 to about +11.1 kJ/mol; Table 1), and with a stability differential of about 7.0 kJ/mol between this set of mutants. Soluble folded protein could be purified, for each mutant, when constructed in the wild-type FGF-1 background. Of the four Ala, Ser, Thr, and Val substitutions, the least destabilizing substitution at position 83 is Thr, which is essentially isoenergetic to Ala; the most destabilizing mutation is Val. The folding cooperativity m-value of the position 83 mutations is essentially conserved, in each case, in comparison to the wild-type FGF-1 reference protein.

Cys117→ Xxx mutations

[0100] As shown in FIG. 3 and Table 1, the Cys117ĺ Xxx mutants exhibited generally minor effects upon stability (ΔΔG ranging from a slightly destabilizing about +1.3 kJ/mol to slightly stabilizing about 1.8 kJ/mol, Table 1) and with a stability differential of about 3.1 kJ/mol between this set of mutants. Of the four Ala, Ser, Thr, and Val mutations, the Cys117ĺ Ala mutation results in the most stable mutant, being slightly more stable than the wild-type FGF-1 reference protein (ΔΔG =-1.8 kJ/mol). The most destabilizing mutation was Cys117ĺ Thr (ΔΔG =+1.3 kJ/mol). The folding cooperativity m-value of the position 117 mutations is essentially conserved, in each case, in comparison to the wild-type FGF-1 reference protein.

[0101] Relaxed stereo diagrams of the X-ray crystal structures of wild-type FGF-1 (PDB accession 1JQZ) and FGF-1 mutant protein having substitutions at position 16, centered at position 16 with structural features within 4.0 Å radius are shown in FIG. 4, FIG. 5, FIG. 6, and FIG. 7 (CPK models). [0102] Crystals are grown for the Cys16ĺ Ala, Cys16ĺ Ser and Cys16ĺ Thr mutants. The Cys16ĺ Val mutant protein exhibits reduced solubility– with a practical concentration limit of approximately 1.3 mg/mL in crystallization buffer. No crystallization condition is found for the Cys16ĺ Val mutant despite screening several hundred conditions. The aberrantly low m-value and low solubility of the Cys16ĺ Val mutant suggests the possibility of a region of non-native structure, or partial unfolding, in response to this mutation. The Cys16ĺ Ala and Cys16ĺ Ser mutants crystallize in space group C2221, isomorphous with the His-tagged form of wild-type FGF-1 (PDB accession 1JQZ), while the Cys16ĺ Thr mutant crystallize in a novel P41212 space group. High resolution diffraction data sets ( 1.80 Å), with excellent completion, are collected for each mutant. A molecular replacement search using the wild-type FGF-1 X-ray coordinates (PDB accession 1JQZ) is successful in each case, and all structures are refined to acceptable stereochemistry and crystallographic residual (Table 2).

Cys83→ Xxx mutations

[0103] X-ray structures for FGF-1 mutant proteins having amino acid residue substitutions of Cys83ĺ Ala, Cys83ĺ Ser, Cys83ĺ Thr and Cys83ĺ Val, respectively, with resolution ranges of about 1.90– 2.10 Å and each crystallizing isomorphous to the wild-type C2221 space group, have previously been described. ²⁸ [0104] FIG. 8, FIG. 9, FIG. 10, FIG. 11, and FIG. 12 show the X-ray crystal structures of wild-type FGF-1 (PDB accession 1JQZ) and FGF-1 mutant protein having amino acid residue substitutions at position 83, ²¹ centered at position 83 with structural features within 4.0 Å radius.

Cys117→ Xxx mutations

[0105] FIG.13, FIG. 14, FIG. 15, FIG. 16, and FIG. 17 show the X-ray crystal structures of wild-type FGF-1 (PDB accession 1JQZ) and FGF-1 mutant protein having amino acid residue substitutions at position 117, centered at position 117 with structural features within 4.0 Å radius.

[0106] Crystals aree grown for the Cys117ĺ Ala, Cys117ĺ Ser, and Cys117ĺ Thr mutants; the crystal structure of the Cys117ĺ Val mutant has previously been reported. ²⁹ All Cys117 mutants crystallize in space group C2221, isomorphous with the His-tagged form of wild-type FGF-1. The Cys117ĺ Thr mutant crystallizes both from formate and citrate precipitants. High resolution diffraction data sets ( 1.52 Å), with excellent completion, are collected for each mutant (the previously reported Cys117ĺ Val mutant structure is solved to 1.70Å). A molecular replacement search using the wild-type FGF-1 X-ray coordinates is successful in each case, and all structures are refined to acceptable stereochemistry and crystallographic residual (Table 2).

[0107] Cys residues are capable of participating in H-bond interactions, either as a donor or acceptor. Due to the larger atomic radius of the SȖ, Cys H-bond interaction distances are characteristically longer than those among N, O, and solvent groups. The average Cys SȖ H- bond donor distance to a main chain carbonyl, and Cys H-bond acceptor distance to a main chain amide, are both about 3.50 ± 0.25 Å. ³⁰ This distance is approximately 0.7 Å longer than ideal N, O H-bond distances (i.e., about 2.80 ± 0.35 Å). In the absence of neutron diffraction– which can distinguish H-atom positions– the identification of donor or acceptor for the Cys SȖ is assigned based upon the specific neighbor atom, or is ambiguous. In the case of an interaction with a main chain carbonyl, the Cys SȖ is the donor; in the case of an interaction with a main chain amide, the Cys SȖ is the acceptor. Interactions with solvent and hydroxyl groups are ambiguous with regard to donor/acceptor identification. In describing the general local structure perturbation in response to mutation, r.m.s.d. values involving CĮ groups within a 4 Å radius are reported. An r.m.s.d. value <0.25 Å is considered“minimal”, 0.25 to 0.50 is considered“minor”, and >0.50 is considered“modest” structural perturbation. Position 16

Wild-type FGF-1

[0108] The wild-type Cys16 residue is fully buried and participates in three H-bond interactions: 1) as a donor to the main chain carbonyl of Ala129, 2) as a donor to the main chain carbonyl of Leu111, and 3) as an acceptor to the main chain amide of Asn18 (FIG. 4, Table 3). Cys16 is the sole H-bond partner for the main chain amide of Asn18. The main chain carbonyl of Ala129 has a second H-bond partner with the main chain amide of Leu111. The main chain carbonyl of Leu111 has a second H-bond partner with the side chain NG2 of Asn18.

Cys16→Ala mutation (ǻǻG = +11.1 kJ/mol)

[0109] Substitution of the wild-type Cys16 residue by Ala results in minor positional shifts of the local structure; the r.m.s.d. value for all main chain CĮ groups within a 4.0 Å sphere is about 0.35 Å. Of the four Ala, Ser, Thr, and Val substitutions, the most significant structural change of the local protein atoms is the mutant Ala16 CE which shifts about 0.8 Å. This shift is not consistent with a collapse towards the former Cys16 SJ– rather, movement away from this position. This shift of the mutant Ala16 CE facilitates the introduction of a novel buried water (Sol 65) overlaying approximately the position of the former Cys16 SJ (although 1.6 Å distal). Sol65 makes H-bond interactions with the main chain carbonyl of Ala129 and the main chain amide of Asn18; thus, effectively regenerating two of the three former H-bond interactions involving the wild-type Cys16 SJ (FIG. 4, Table 3). However, Sol65 is too distal (about 4.15 Å) to make an H-bond interaction with the main chain carbonyl of Leu111; thus, the H-bond requirements of this group are unsatisfied with the Cys16o Ala mutation.

Cys16→Ser mutation (ǻǻG = +9.1 kJ/mol)

Substitution of the wild-type Cys16 residue by Ser results in minor positional shifts of the local structure; the r.m.s.d. value for all main chain CĮ groups within a 4.0 Å sphere is about 0.39 Å. The Ser mutation at position 16 adopts a similar rotamer as the wild-type Cys; however, there is positional adjustment of the Ser16 CĮ-Cȕ vector essentially identical to that observed in the above-described Ala16 mutation. This shift, combined with the shorter Cȕ-OȖ bond length (about 1.43 Å) in Ser compared to the Cȕ-SȖ bond length (about 1.82 Å) in Cys, results in the introduced Ser OȖ atom residing about 1.4 Å distal to the position of the original Cys16 SȖ atom. Consequently, the mutant Ser16 is only able to participate in an H-bond interaction with the main chain amide of Asn18; and former interactions with the main chain carbonyls of Leu111 and Ala129 are lost (FIG. 4, Table 3). However, as with the Ala mutation, a novel buried water is observed (Sol 17)--making the same set of H-bond interactions as Sol 65 in the Ala mutation. Sol 17 and Ser16 OȖ together provide a bifurcated H-bond for the main chain amide at position Asn18. The introduced Ser16 also makes an H-bond with adjacent Sol38 (equivalent to Sol204 in the wild-type structure).

[0110] Substitution of the wild-type Cys16 residue by Thr results in more substantial positional shifts of the local structure; in comparison to the Ala or Ser mutations the r.m.s.d. value for all main chain CĮ groups within a 4.0 Å sphere is about 0.59 Å. The rotamer of the mutant Thr16 residue overlays its OȖ2 atom with the wild-type Cys SȖ (being about 1.6 Å distal). There is also positional adjustment of the Thr16 CĮ-Cȕ vector; however, in this case it is a shift opposite in direction as observed with the Ala and Ser mutations. Consequently, the Thr16 OȖ1 atom is closer (about 0.88 Å) to the position of the novel Sol 65/Sol 17 in the Ser/Ala mutant structures than it is to the Cys SȖ (about 1.16 Å) in the wild-type structure. The mutant Thr16 H-bonds with the main chain amide of position Asn18 (FIG. 4, Table 3). However, both H-bonds with the carbonyl of Ala129 and the carbonyl of Leu111 are lost. This position of the mutant Thr16 side chain precludes the introduction of a novel solvent to recover H-bonds lost by Cys mutation; thus, the former H-bonds with the main chain carbonyls of Leu111 and Ala129 are both compromised. In response to introduction of the bulkier Thr side chain the neighboring residues Leu131 and Phe132 shift away from position 16 by about 0.7 Å; consequently a novel buried water (Sol30) intercalates between the main chain carbonyl of Leu131 and the main chain amide of Ser17– thus, maintaining this H-bond interaction as a water-mediated H-bond.

Position 83

Wild-type FGF-1

[0111] Cys83 is completely buried and is present in the gauche+ rotamer. Cys83 makes H-bond interactions with the main chain amide of Asn80, the side chain Oį 1 of Asn80, and the main chain carbonyl of Thr78. Cys83 is the exclusive H-bond partner to the main chain amide of Asn80; whereas, both the side chain Oį 1 of Asn80, and the main chain carbonyl of Thr78 have additional H-bond partners (FIG. 5, Table 4).

Cys83→Ala mutation (ǻǻG = +4.5 kJ/mol)

[0112] The Cys83ĺ Ala mutation is accommodated with limited structural perturbation; the main chain CĮ atoms within a 4 Å radius of position 83 overlay the wild-type FGF-1 structure with an r.m.s.d. value of about 0.27 Å. However, the majority of this average structural deviation is contributed by the main chain region of residue positions 78-80. The most substantial movement is that of adjacent Pro79--whose main chain CĮ moves about 0.53 Å towards the location of the former Cys83 SȖ thereby effectively preventing cavity formation due to the deleted SȖ. No new solvent structure is formed in response to the deletion of the Cys83 SȖ; consequently, the H-bond interactions with the main chain amide of Asn80, the side chain Oį 1 of Asn80, and the main chain carbonyl of Thr78 are lost (FIG. 5, Table 4). As mentioned above, the side chain Oį 1 of Asn80, and the main chain carbonyl of Thr78 each have secondary H-bond partners; however, the main chain amide of position Asn80 appears unsatisfied in its H-bond potential in the Ala mutation.

Cys83→Ser mutation (ǻǻG = +5.6 kJ/mol)

[0113] The Cys83ĺ Ser mutation adopts the same gauche+ rotamer as the wild-type Cys83. The r.m.s.d. value of main chain CĮ groups within 4 Å is about 0.37 Å--indicating minor structural changes. The mutant OȖ group overlays within 0.19 Å of the wild-type Cys83 SȖ--indicating that a minor positional shift of position 83 has occurred that tends to conserve the position of the H-bonding Ȗ atom. However, the largest positional shifts are observed for the Asn80 main chain amide (0.25 Å) and Asn80 side chain Oį 1 (0.95 Å) which both move towards the mutant Ser OȖ– indicating a detectable structural collapse in response to the Ser mutation. These positional shifts result in the H-bonds between the Asn80 main chain amide and Asn80 side chain Oį 1 with the position 83 Ȗ atom being effectively maintained with the Ser83 mutation (FIG. 5, Table 4). In this regard, the largest change in H- bond distance is with Asn80 Oį 1: in the wild-type Cys83 the H-bond distance is 3.43 Å, and in the mutant Ser83 this is reduced to 2.68 Å. Although the above-described structural shift effectively retains two H-bond interactions, the former H-bond between Cys83 and the main chain carbonyl of Thr78 is too distal (about 3.85 Å) to be maintained with the Ser83 mutation.

Cys83→Thr mutation (ǻǻG = +4.1 kJ/mol)

[0114] The Cys83ĺ Thr mutation adopts a rotamer that juxtaposes its CȖ2 atom (and not OȖ1) with the wild-type Cys83 SȖ. The r.m.s.d. value of main chain CĮ groups within 4 Å is about 0.18 Å--indicating accommodation with only minimal structural change. The mutant OȖ1 group is oriented towards the side chain of adjacent Tyr55, and this ring swings away from position 83 by 0.54 Å to avoid a close contact. The mutant Thr83 OȖ1 group consequently establishes two novel H-bond interactions: the first is with the main chain amide of position Asn80 and effectively regenerates the former H-bond provided by the wild- type Cys83 SȖ; the second novel H-bond is with the main chain carbonyl of Glu81 (FIG. 5, Table 4). Thus, with minimal structural perturbation the mutant Thr83 maintains the exclusive H-bond with the main chain amide of Asn80, consequently, the local structure can provide two H-bond interactions to satisfy the H-bond requirements of the introduced Thr83 OȖ1. The former H-bond between the Thr78 main chain carbonyl and wild-type Cys83 SȖ is not maintained by the Thr83 mutation; however, Thr78 O has an additional H-bond partner with Ser76 OȖ that is conserved.

Cys83→Val mutation (ǻǻG = +11.1 kJ/mol)

[0115] The Cys83ĺ Val mutation adopts the same rotamer as the Cys83ĺ Thr mutant. The r.m.s.d. value of main chain CĮ groups within 4 Å is about 0.25 Å--indicating minor structural changes. The mutant Val83 is unable to preserve any of the H-bond interactions seen with the wild-type Cys83 SȖ atom, and no novel solvent structure is observed. Thus, while there is limited structural perturbation (i.e., the most significant is a >1 Å movement of the side chain aromatic ring of Y55 away from the introduced Val83 CȖ2 to avoid a steric clash) there is effective elimination of three H-bond interactions—including the unique H- bond partner to the main chain amide of Asn80 (FIG. 5, Table 4).

Position 117

Wild-type FGF-1

[0116] The wild-type FGF-1 Cys117 residue exhibits two alternative conformations– a solvent excluded (i.e., buried) gauche+ (X ¹ = -60°) rotamer, and a partially solvent accessible gauche- (X ¹ = +60°) rotamer (FIG. 6, Table 5). The buried gauche+ rotamer of Cys117 participates in a single H-bond interaction with the main chain carbonyl of Val31. Cys16 is not the only H-bond partner for this Val31 carbonyl as it also H-bonds with Sol1258 and Sol1273. The solvent accessible gauche- rotamer is able to maintain an H-bond with the main chain carbonyl of Val31, and also has several additional H-bond interactions with solvent molecules, including Sol1273, Sol1258, Sol1232, and Sol1234.

Cys117→Ala mutation (ǻǻG = -1.8 kJ/mol)

[0117] Elimination of the Cys117 SȖ by mutation to Ala results in minimal perturbation of the local structure (r.m.s.d. value for CĮ within a 4Å sphere is about 0.19 Å). Loss of the Cys117 SȖ results in alteration of local solvent structure, and Sol1234 in wild-type FGF-1 is replaced by two solvents (Sol315 and Sol108) in mutant FGF-1 having substitution of Cys117o Ala (FIG. 6, Table 5). The equivalent of Sol1258 of the wild-type structure, which H-bonds with Cys117 SȖ, has no counterpart in the Ala mutant, suggesting that the presence of this Sol principally satisfies the H-bond requirement of the wild-type Cys residue. Thus, all essential H-bond interactions appear to be effectively maintained with the Ala mutation. Cys117→Ser mutation (ǻǻG = +1.0 kJ/mol)

[0118] The mutant Ser residue adopts the solvent exposed gauche- rotamer and with minor alteration of the local structure (r.m.s.d. value for CĮ within a 4Å sphere is about 0.25 Å). The mutant Ser OȖ overlays that of the wild-type SȖ within about 0.38Å; consequently, minor positional shifts of local solvent result in all H-bonds observed with the Cys117 gauche- rotamer being effectively conserved with the Ser mutation (FIG. 6, Table 5). The mutant Ser in the gauche- rotamer is able to H-bond with the main chain carbonyl of Val31– effectively substituting for the H-bond interaction observed for the wild-type Cys in the gauche+ rotamer. Thus, all essential H-bond interactions also appear to be effectively maintained with the Ser mutation. Cys117→Thr mutation (ǻǻG = +1.3 kJ/mol)

[0119] The mutant Thr residue adopts a rotamer orientation that juxtaposes the OȖ1 group with the solvent exposed wild-type Cys gauche- rotamer, and the CȖ2 group with the buried wild-type Cys gauche+ rotamer. The Thr mutant side chain is accommodated with minimal alteration of the local structure (r.m.s.d. value for CĮ within a 4Å sphere is about 0.16 Å; within the error of the crystallographic data). Minor positional shifts of local solvent result in all H-bonds observed with the Cys117 gauche-rotamer being effectively conserved with the Thr mutation (FIG. 6, Table 5).

Cys117→Val mutation (ǻǻG = 0.0 kJ/mol)

[0120] The mutant Val residue does not adopt a rotamer isosteric with Thr. While the mutant Val CȖ2 atom juxtaposes with the wild-type buried Cys SȖ gauche+ rotamer, the mutant Val CȖ1 adopts a unique trans orientation. However, the Val mutant side chain is accommodated with minimal alteration of the local structure (r.m.s.d. value for CĮ within a 4Å sphere is about 0.14 Å). The local solvent structure is essentially unperturbed, and all local H-bonds are effectively conserved with the Val117 mutation (FIG. 6, Table 5).

Discussion

[0121] Each of the three buried free Cys residues in this study participate in H-bond interactions; each acting as a donor, and in two cases (Cys16, Cys83) also participating as an acceptor. In no case does the buried Cys act as an exclusively hydrophobic (i.e., aliphatic) amino acid. Evaluation of the structural details of Ser and Thr mutations serves to support the hydrophilic property of the buried Cys local environments. The Cys16o Ser and Cys83o Ser mutations each overlay the mutant Ser OȖ with the wild-type Cys SȖ, and the Cys16o Thr mutation overlays the mutant Thr OȖ1 with the wild-type Cys SȖ. Deviation from this general polar behavior was observed for the Cys83o Thr mutation (which overlays the mutant Thr CȖ2 with the wild-type Cys SȖ); however, in this conformation the mutant Thr OȖ1 was able to maintain critical local H-bonds that compensated for deleted Cys SȖ H-bonds. Since position Cys117 adopts two alternative rotamers (one buried, one exposed) the hydrophobic/hydrophilic environment of Cys at this position is potentially ambiguous; however, mutant Ser and Thr side chains position their OȖ and OȖ1 groups, respectively, towards solvent; and both the Thr CȖ2 and Val CȖ2 mutant methyl groups overlay with the buried Cys117 rotamer. Thus, position 117 may be a more hydrophobic environment for the buried Cys117; however, in this rotamer the buried Cys does make one local H-bond interaction. Thus, the results indicate that each buried Cys is acting as a buried polar residue, typically participating in several H-bond interactions.

[0122] In several instances, the donor or acceptor role of the buried Cys SȖ is observed to be the sole H-bond partner for an adjacent polar group (e.g., main chain amide or carbonyl); lack of a secondary H-bond partner identifies such interactions as likely to be energetically significant. H-bonds involving sulfur are characteristically about 0.5 Å longer than those involving oxygen (i.e., measuring the S/O---donor/acceptor distance) due to the larger atomic radius of sulfur. ³⁰ Thus, in the absence of a localized structural collapse, substitution of the wild-type Cys SȖ by mutant Ser OȖ or Thr OȖ1 cannot effectively maintain local H-bond interactions. Such collapse can also obviate formation of cavity space upon mutational loss of the large SȖ atom. Such a local collapse was observed for the Cys83o Ser mutation; however, this collapse permitted only two of the three wild-type H-bond interactions to be maintained upon Ser substitution, and was thus not entirely successful. In the absence of ability to structurally collapse (as might be the case with a well-packed local environment), to maintain H-bonds with Cyso Ser mutation, another potential solution is the introduction of a novel buried solvent. However, such introduction may require expansion of the local structure to provide requisite space. Such a situation is observed with position Cys16o Ala and Cys16o Ser mutations.

[0123] Of these two types of general structural responses (collapse or expansion) to mutation of H-bonded buried Cys residues, which (if any) is more destabilizing? On average, position 16 mutations (expansion) were observed to be more destabilizing (ǻ ǻ G = +11.7 ± 2.1 kJ/mol) than position 83 mutations (collapse) (ǻ ǻ G = +6.3 ± 3.2 kJ/mol). However, the Cys83o Val mutation is unique in presenting with structural expansion to accommodate the mutant Val; thus, grouping Cys83o Val with the expansion type mutations the stability effects are ΔΔG = +11.6 ± 1.8 kJ/mol, while the collapse type stability effects are +4.7 ± 0.8 kJ/mol. Position 117 mutations (essentially neutral in energetics) appear to make no critical buried H-bonds and have neither expansion nor collapse structural effects.

[0124] Although a limited set, the current data indicate that in predicting, or designing, mutations to eliminate buried free Cys residues it is important to determine the number and type of H-bond interactions the Cys SȖ participates in, the potential for the local structure to collapse to maintain such H-bonds, or to expand and accommodate a novel solvent to participate in maintaining critical buried H-bonds. Considering Ser as a simple isosteric substitution, and to consider Ala as an appropriate hydrophobic substitution at a buried site, are over-simplified models that do not capture the complexity of the structural role of buried Cys residues.

[0125] The present results suggest that proteins are structurally optimized at buried Cys positions, and destabilization upon mutation can be expected. In this regard, collapse-type structural responses appear to be the least deleterious. Curiously, while Ala may seem to be the mutation that would permit the greatest collapse, it is also the mutation most likely to permit space for the introduction of a novel solvent– leading to structural expansion. In other words, collapse appears most useful in the case of Ser or Thr mutation where the mutant hydroxyl can potentially substitute the Cys SȖ in maintaining H-bond interactions (Ala having no H-bonding group).

[0126] Three different amino acids, Ser, Thr, and Ala, are identified as the best choice from a stability standpoint to substitute for Cys at positions 16, 83, and 117, respectively. However, the best average substitution for the set of three positions is Ala, while the worst choice is Val (Ala, Ser, Thr, Val ΔΔG = 4.6, 5.2, 6.4, 8.0 kJ/mol, respectively). This observation appears to reflect the collapse vs. expansion effects associated with such mutations and maintenance of local H-bond interactions.

Example 2

Determination of the biological activity of internal disulfide mutants on human corneal endothelial cells (HCECs)

[0127] In this example, FGF-1 mutant proteins having the combination of substitutions of C16S/A66C/C117V and of K12V/C16S/A66C/C117/P134V are tested for their activities in the stimulation of cell proliferation, compared with the activity of FGF-1 K12V/C117/P134V (M1) mutant protein.

Materials and Methods

Protein mutagenesis and expression

[0128] FGF-1 mutant proteins having mutations of C16S/A66C/C117V, K12V/C16S/A66C/C117/P134V, and K12V/C117/P134V are respectively constructed, expressed, and purified according to the method described in Example 1. For purposes of convenience, FGF-1 mutant protein having mutations of K12V/C117/P134V is referred to as mutant protein M1; FGF-1 mutant protein having mutations of 16S/A66C/C117V with a His- tag at the N-terminus of the mutant protein is referred to as mutant protein C3; and FGF-1 mutant protein having mutations of K12V/C16S/A66C/C117/P134V with a His-tag at the N- terminus of the mutant protein is referred to as mutant protein C2V3.

[0129] Mutant proteins C3 and C2V3 each retains a His-tag used for purification after expression. The His-tag does not alter the biological activities of mutant proteins C3 and C2V3. As described above, in each of these two mutants, a Cys is introduced at position 66 to pair with the Cys83. Biophysical characterization shows that a disulfide bond between the Cys66 and Cys 83 is formed in each of these two mutant proteins and stabilizes the structure of each mutant protein. Cys at positions 16 and 117 have been removed from the mutant proteins C3 and C2V3 so the mutant proteins C3 and C2V3 have no free Cys and thus are minimally reactive to oxidation. Since the mutant protein C2V3 also includes the substitutions of K12V/C117/V/P134V as that in the mutant protein M1, the mutant protein C2V3 not only has all the properties of the mutant protein C3 but also has the stabilizing property of the mutant protein M1. Proliferation assay on human corneal endothelial cells

[0130] Primary cultures of human corneal endothelial cells (HCECs) are thawed and expanded in a FNC-coated T75 in growth media (OptiMEM with 8% fetal bovine serum (FBS), Insulin/transferrin/selenium, 20 g/ml ascorbic acid, 200mg/ml calcium chloride, and antibiotic/antimycotic) supplemented with 10ng/ml mutant protein FGF M1.

[0131] For the proliferation assay, HCEC cells are passaged using accutase, harvested by centrifugation (200g x 12 min), resuspended in growth media without FGF and plated into FNC-coated 24-well plates (3 plates) at a seeding of 25,000 cells per well in 0.5ml of growth medium.

[0132] At 24hr post plating, the media are removed and replaced with base media (OptiMEM with 0.4% fetal bovine serum (FBS), Insulin/transferrin/selenium, 20μg/ml ascorbic acid, 200mg/ml calcium chloride, and antibiotic/antimycotic) with indicated additions. Each mutant proteins M1, C3, and C2V3 is diluted d to 1ng/μl in base media as a stock solution and then used to make up the cell media. Mutant proteins M1, C3, and C2V3 are added respectively into base media as additions at four concentrations (0.3ng/ml, 1.0ng/ml, 3.0ng/ml, and 10ng/ml). Cells base media with no addition are used as negative control media. Growth media are used as positive control media. All negative control media, positive control media, and base media with additions of different mutant proteins of FGF-1at four concentrations are tested in quadruplicate. Cell numbers are counted manually as the number of cells per 20x field. The same area of the plate is counted each time, with location marked at day 1 using an ink dot on the bottom surface of the plate.

Results

[0133] The adhesion and regular distribution of the cells are checked 24 hours later after plating. The cells are counted per 20x field. At passage 4, cells are examined under microscope for their appearance. These cells have fibroblastic-like appearance. (FIG. 18) The cell numbers per 20x field for the plates are about 49 ±11, about 53 ± 14, and about 41 ± 8 for plates 1, 2 and 3, respectively.

[0134] Mutant protein M1 has been tested previously on rabbit and human CECs and shown to have an effect in stimulating cell proliferation. [0135] On day 3, cells are examined and cell numbers are counted. The chart of FIG. 19 shows the numbers of cells cultured in base media, in growth media and in base media with mutant proteins M1, C3, and C2V3 at concentrations of about 0.3 ng/ml, about 1.0 ng/ml, about 3.0 ng/ml, and about 10 ng/ml, respectively. Bars labeled with M1, C3, and C2V3 represent groups of data for testing the effect of the mutant proteins M1, C3 and C2V3, respectively, which include cell numbers counted from cell cultures in negative control media (no addition), in positive control media (high serum), and in base media with the addition of different amount of mutant protein M (FIG. 19.)

[0136] The number of cells cultured in negative control media in each plate approximately doubles over the three-day culture.

[0137] Cells cultured in base media with the addition of the mutant protein M1 significantly increase and show a dose-response to the addition of the mutant protein M1 (FIG. 19). On day 3, comparing cells growing in base media with the addition of the mutant protein M1 at concentrations of about 0.3 ng/ml, about 1.0 ng/ml, and about 10 ng/ml, cells growing in base media with about 3.0 ng/ml of the mutant protein M1 reach the highest number. The result indicates that the mutant protein M1 stimulates the HCEC to grow with a dose-response and has an EC ₅₀ of about 3.0 ng/ml.

[0138] Cells cultured in base media with the addition of the mutant protein C3 significantly increase and also show a dose-response to the mutant protein C3 (FIG. 19). On day 3, comparing to cells growing in base media with the addition of the mutant protein C3 at concentrations of about 1.0 ng/ml, about 3.0 ng/ml, and about 10 ng/ml, cells growing in base media with about 0.3 ng/ml of the mutant protein C3 reach the highest number. Unexpectedly, mutant protein C3 appears to stimulate maximally at the lowest dose (0.3ng/ml) of the four concentrations tested with a reverse dose-response curve. This result indicates that the mutant protein C3 has mitogenic activity at a low dose.

[0139] The morphology of cells cultured in base media with the addition of about 1.0 ng/ml of the mutant protein C3 is further examined. On day 3, cells cultured in positive control media have fibroblast-like appearance. (Panel A of FIG. 20) Cells cultured in base media with the addition of about 1.0 ng/ml of the mutant protein C3 appear flatter and less fibroblast-like. (Panel B of FIG. 20). This result suggests that the mutant protein C3 has a morphological effect on the cells.

[0140] On day 3, cells cultured in base media with the addition of the mutant protein C2V3 are not statistically different from the number of cells cultured in negative control media

(FIG. 19), similar to the result of cells cultured in the positive control.

[0141] On day 5, the numbers of cells cultured in positive control media and in base media with the addition of the mutant protein C2V3 are statistically different from the number of cells cultured in negative control media. (FIG. 21). Bars labeled with C2V3 in the chart of FIG. 21 represent the data for testing the effect of the mutant protein C2V3. Each bar represents the number of cells cultured in base media, in growth media, and in base media with the addition of mutant protein C2V3 at concentrations of about 0.3 ng/ml, about 1.0 ng/ml, about 3.0 ng/ml, and about 10 ng/ml, respectively. As shown in FIG. 21, cells cultured in base media with the addition of 3.0 ng/ml of the mutant protein C2V3 reach the highest number. This result indicates that the mutant protein C2V3 has a mitogenic effect on HCECs and such effect is dose-responsive.

Discussion

[0142] The previous result with the mutant protein M1 has been replicated and is consistent with an EC ₅₀ of 3ng/ml for the stimulation of cell proliferation in HCECs.

[0143] The mutant protein C3 appears to potently stimulate the proliferation of HCECs. The inverse dose response curve suggests that the peak of the effect may be at lower concentrations. More tests can be done with lower concentrations of the mutant protein C3. Notably, this data suggests that the EC ₅₀ of this mutant protein C3 may be at least 10-fold lower than the EC ₅₀ of wild type FGF or other tested FGFs. The morphological effect suggests that C3 may reverse the epithelial-mesenchymal transition (EMT) of these cells.

[0144] According to the this example, the mutant protein C2V3 may be no more potent than M1 or wild-type FGF, but the mutant protein C2V3 still has mitogenic potency and may be used to stimulate the proliferation of the cells. [0145] The many features and advantages of the invention are apparent from the detailed specification, and thus, it is intended by the appended claims to cover all such features and advantages of the invention which fall within the true spirit and scope of the invention. Further, since numerous modifications and variations will readily occur to those skilled in the art, it is not desired to limit the invention to the exact construction and operation illustrated and described, and accordingly, all suitable modifications and equivalents may be resorted to, falling within the scope of the invention.

[0146] It is intended that the invention not be limited to the particular embodiment disclosed herein contemplated for carrying out this invention, but that the invention will include all embodiments falling within the scope of the claims.

[0147] All documents, patents, journal articles and other materials cited in the present application are incorporated herein by reference.

References

[0149] The following references are referred to above and are incorporated herein by reference:

1. Fomenko, D. E., Marino, S. M. & Gladyshev, V. N. (2008). Functional diversity of cysteine residues in proteins and unique features of catalytic redox-active cysteines in thiol oxidoreductases. Mol Cells 26, 228-35.

2. Marino, S. M. & Gladyshev, V. N. (2012). Analysis and Functional Prediction of Reactive Cysteine Residues. Journal of Biological Chemistry 287, 4419-4425.

3. Nagano, N., Ota, M. & Nishikawa, K. (1999). Strong hydrophobic nature of cysteine residues in proteins. FEBS Letters 458, 69-71.

4. Ortega, S., Schaeffer, M.-T., Soderman, D., DiSalvo, J., Linemeyer, D. L., Gimenez- Gallego, G. & Thomas, K. A (1991). Conversion of cysteine to serine residues alters the activity, stability, and heparin dependence of acidic fibroblast growth factor. Journal of Biological Chemistry 266, 5842-5846.

5. Lee, J. & Blaber, M. (2009). The interaction between thermodynamic stability and buried free cysteines in regulating the functional half-life of fibroblast growth factor- 1. Journal of Molecular Biology 393, 113-127.

6. Fremaux, I., Mazeres, S., Brisson-Lougarre, A., Arnaud, M., Ladurantie, C. & Fournier, D. (2002). Improvement of Drosophila acetylcholinesterase stability by elimination of a free cysteine. BMC Biochem 3, 21.

7. Nakaniwa, T., Fukada, H., Inoue, T., Gouda, M., Nakai, R., Kirii, Y., Adachi, M., Tamada, T., Segawa, S.-i., Kuroki, R., Tada, T. & Kinoshita, T. (2012). Seven Cysteine-Deficient Mutants Depict the Interplay between Thermal and Chemical Stabilities of Individual Cysteine Residues in Mitogen-Activated Protein Kinase c-Jun N-Terminal Kinase 1. Biochemistry 51, 8410-8421.

8. Culajay, J. F., Blaber, S. I., Khurana, A & Blaber, M. (2000). Thermodynamic characterization of mutants of human fibroblast growth factor 1 with an increased physiological half-life. Biochemistry 39, 7153-7158. 9. Brych, S. R., Gokarn, Y. R., Hultgen, H., Stevenson, R. J., Rajan, R. & Matsumura, M. (2010). Characterization of antibody aggregation: Role of buried, unpaired cysteines in particle formation. Journal of Pharmaceutical Sciences 99, 764-781. 10. Franey, H., Brych, S. R., Kolvenbach, C. G. & Rajan, R. S. (2010). Increased aggregation propensity of lgG2 subclass over lgG1: Role of conformational changes and covalent character in isolated aggregates. Protein Science 19, 1601-1615.

11. Gevondyan, N. M., Volynskaia, A. M. & Gevondyan, V. S. (2006). Four free cysteine residues found in human lgG1 of healthy donors. Biochemistry (Moscow) 71, 279- 284.

12. Huh, J. H., White, A J., Brych, S. R., Franey, H. & Matsumura, M. (2013). The identification of free cysteine residues within antibodies and a potential role for free cysteine residues in covalent aggregation because of agitation stress. Journal of Pharmaceutical Sciences 102, 1701-1711.

13. Marshall, S. A, Lazar, G. A., Chirino, A J. & Desjarlais, J. R. (2003). Rational design and engineering of therapeutic proteins. Drug Discovery Today 8, 212- 221.

14. Copeland, R. A , Ji, H., Halfpenny, A J., Williams, R. W., Thompson, K. C., Herber, W. K., Thomas, K. A., Bruner, M. W., Ryan, J. A., Marquis-Omer, D., Sanyal, G., Sitrin, R. D., Yamazaki, S. & Middaugh, C. R. (1991). The structure of human acidic fibroblast growth factor and its interaction with heparin. Archives of Biochemistry and Biophysics 289, 53-61.

15. Engleka, K. A & Maciag, T. (1992). Inactivation of human fibroblast growth factor-1 (FGF-1) activity by interaction with copper ions involves FGF-1 dimer formation induced by copper-catalyzed oxidation. Journal of Biological Chemistry 267, 11307- 11315.

16. Mellin, T. N., Cashen, D. E., Ronan, J. J., Murphy, B. S., DiSalvo, J. & Thomas, K. A (1995). Acidic fibroblast growth factor accelerates dermal wound healing in diabetic mice. Journal of Investigative Dermatology 104, 850-855. 17. Blaber, S. I., Culajay, J. F., Khurana, A & Blaber, M. (1999). Reversible thermal denaturation of human FGF-1 induced by low concentrations of guanidine hydrochloride. Biophysical Journal 77, 470-477.

18. Wang, W., Lin, S., Xiao, Y., Huang, Y., Tan, Y., Cai, L. & Li, X. (2008).

Acceleration of diabetic wound healing with chitosan-crosslinked collagen sponge containing recombinant human acidic fibroblast growth factor in healing- impaired STZ diabetic rats. Life Sciences 82, 190-204.

19. Blaber, M., DiSalvo, J. & Thomas, K. A (1996). X-ray crystal structure of human acidic fibroblast growth factor. Biochemistry 35, 2086-2094.

20. Bernett, M. J., Somasundaram, T. & Blaber, M. (2004). An atomic resolution structure for human fibroblast growth factor 1. Proteins 57, 626-34.

21. Lee, J. & Blaber, M. (2009). Structural basis for conserved cysteine in the fibroblast growth factor family: evidence for a vestigial half-cystine. Journal of Molecular Biology 393, 128-139.

22. Brych, S. R., Blaber, S. I., Logan, T. M. & Blaber, M. (2001). Structure and stability effects of mutations designed to increase the primary sequence symmetry within the core region of a E-trefoil. Protein Science 10, 2587-2599.

23. Eftink, M. R. (1994). The use of fluorescence methods to monitor unfolding transitions in proteins. Biophysical Journal 66, 482-501.

24. Pace, C. N. & Scholtz, J. M. (1997). Measuring the conformational stability of a protein. In Protein structure: a practical approach (Creighton, T. E., ed.), pp. 299- 321. Oxford University Press, Oxford.

25. Terwilliger, T. C., Grosse-Kunstleve, R. W., Afonine, P. V., Moriarity, N. W., Zwart, P. H., Hung, L. W., Read, R. J. & Adams, P. D. (2008). Iterative model building, structure refinement and density modification with the PHENIX AutoBuild wizard. Acta Crystallographica D Biol Crystallogr 64, 61-69.

26. Emsley, P. & Cowtan, K. (2004). Coot: model-building tools for molecular graphics. Acta Crystallographica D Biol Crystallogr 60, 2126-2132. 27. Dubey, V. K., Lee, J., Somasundaram, T., Blaber, S. & Blaber, M. (2007). Spackling the crack: stabilizing human fibroblast growth factor-1 by targeting the N and C terminus beta-strand interactions. J Mol Biol 371, 256-68.

28. Lee, J. & Blaber, M. (2009). The interaction between thermostability and buried free cysteines in regulating the functional half-life of fibroblast growth factor-1. Journal of Molecular Biology 393, 113-127.

29. Brych, S. R., Kim, J., Logan, T. M. & Blaber, M. (2003). Accommodation of a highly symmetric core within a symmetric protein superfold. Protein Science 12, 2704 - 2718.

30. Zhou, P., Feifei, T., Fenglin, L. & Shang, Z. (2008). Geometric characteristics of hydrogen bonds involving sulfur atoms in proteins. Proteins, published online 18 Nov 2008.