WO2003048185A2

WO2003048185A2 - Angiopioetin related factors

Info

Publication number: WO2003048185A2
Application number: PCT/US2002/037660
Authority: WO
Inventors: Paul D. Kessler; Imre Kovesdi
Original assignee: Genvec, Inc.
Priority date: 2001-11-30
Filing date: 2002-11-21
Publication date: 2003-06-12
Also published as: AU2002359464A1; AU2002359464A8; WO2003048185A3

Abstract

The invention provides fusion proteins comprising a fibrinogen-like domain comprising an amino acid sequence identical or homologous to the amino acid sequence of the fibrinogen-like domain of a first wild-type angiopoietin-related factor and a coiled-coil domain comprising an amino acid sequence identical or homologous to the coiled-coil domain of a second wild-type angiopoietin-related factor, where the first and second wild-type angiopoietin related factors are different. The invention further provides polynucleotides encoding such fusion proteins, vectors comprising such polynucleotides, and methods of making and administering such fusion proteins, polynucleotides, and vectors.

Description

ANGIOPOIETIN RELATED FACTORS

FIELD OF THE INVENTION [0001] This invention pertains to novel angiopoietin related factors, nucleic acids encoding such factors, vectors comprising such nucleic acids, and methods of producing and administering such factors, nucleic acids, and vectors.

BACKGROUND OF THE INVENTION [0002] The angiopoietins are a recently discovered family of related proteins. Angiopoietins are characterized by their ability to bind Tie receptors (sometimes also referred to as "ork" or "tek" receptors) (described in, e.g., International Patent Application WO 98/05779, U.S. Patent 5,447,860, Jones et al., Nat. Rev. Mol. Cell Biol., 2(4), 251-61 (2001), and references cited therein), and in having a structure that includes a N-terminal coiled-coil domain and C-terminal fibrinogen-like domain. The angiopoietins are further functionally characterized by their ability to modulate angiogenesis and related biological activities in the endothelium. Valenzuela et al., Proc. Natl. Acad. Sci. USA, 96(5), 1904- 1909 (1999). For example, angiopoietin- 1 (Ang-1) has been linked to embryonic vascular stabilization, branching morphogenesis, and post-natal angiogenesis. However, not all angiopoietins induce angiogenesis. For example, angiopoietin-2 (Ang-2) acts as an antagonist for Ang-1, disrupting angiogenesis in vivo. Maisonpierre et al., Science, 277, 55- 60 (1997).

[0003] Several factors structurally and/or functionally related to the angiopoietins also have been identified. For example, U.S. Patents 5,972,338, 6,030,831, and 6,057,435 describe a number of protein factors homologous to the angiopoietins. These factors exhibit significant levels of amino acid sequence identity to Ang-1 and/or other angiopoietins and typically contain a fibrinogen-like domain, coiled-coil domain, or both. However, some of these factors may not bind a Tie receptor.

[0004] Fusion proteins containing portions of angiopoietins or angiopoietin homologs are known in the art. Most of these fusion proteins are limited to a protein comprising a single angiopoietin or angiopoietin homolog peptide portion and a heterologous portion typically added to promote protein stability, targeting, or purification. For example, U.S. Patent 6,074,873 discloses fusion proteins comprising an angiopoietin homolog portion where either the N-terminal region of the angiopoietin amino acid sequence is fused to a peptide portion that promotes secretion or the C-terminal portion of the angiopoietin amino acid sequence is fused to an immunogenic polypeptide. International Patent Application WO 00/37642 discloses fusion proteins comprising an angiopoietin peptide portion and a heterologous multimerizing portion derived from an immunoglobulin. International Patent Application WO 00/75329 discloses fusion proteins comprising an angiogenic factor linked to a molecule that specifically binds the vascular endothelium. [0005] International Patent Application WO 98/05779 discloses fusion proteins comprising two peptide portions derived from different angiopoietins. However, the fusion proteins of the '779 application are limited to generalized combinations of Ang-1, Ang-2, angiopoietin-3 (Ang-3), and angiopoietin-4 (Ang-4) peptide portions. Similarly, International Patent Applications WO 00/64946 and WO 01/03735 describe particular Ang- l/Ang-2 fusion proteins. International Patent Application WO 01/47951 describes angiopoietin homologs having at least two fibrinogen-like domains or coiled-coil domains. Thus, the fusion proteins of the '779, '946, '735, and '951 applications are limited in their characteristics.

[0006] There remains a need for novel angiopoietin-related factors that exhibit different biological properties than the angiopoietins, angiopoietin homologs, and angiopoietin fusion proteins previously known in the art. The present invention provides such angiopoietin- related factors, as well as polynucleotides encoding such angiopoietin-related factors, vectors comprising such polynucleotides and methods of producing and using such factors, polynucleotides, and vectors. These and other advantages of the invention, as well as additional inventive features, will be apparent from the description of the invention provided herein.

BRIEF SUMMARY OF THE INVENTION [0007] The invention provides a fusion protein comprising (a) a fibrinogen-like domain that exhibits at least about 75% amino acid sequence identity to the fibrinogen-like domain of a first wild-type protein, which binds a receptor that, when activated, promotes or inhibits angiogenesis, and (b) a coiled-coil domain that exhibits at least about 75% amino acid sequence identity to the coiled-coil domain of a second wild-type protein, which increases the affinity of the fusion protein for the receptor, promotes formation of a multimer comprising the fusion protein, or both, wherein the first protein and second protein are different, the first protein and second protein exhibit at least about 30% overall amino acid sequence identity with Ang-1, and the first protein, the second protein, or both proteins, are not Ang-1, Ang-2, Ang-3, or Ang-4.

[0008] The invention further provides a fusion protein comprising (a) a fibrinogen-like domain that exhibits at least about 30% amino acid sequence identity to the fibrinogen-like domain of Ang-1, and (b) a coiled-coil domain that comprises (1) an amino acid sequence that exhibits at least about 60% amino acid sequence identity to the coiled-coil domain of zapol, murine FDRG, or both, (2) an amino acid sequence that exhibits at least about 50% amino acid sequence identity to the coiled-coil domain of NL1, NL5, or both, (3) an amino acid sequence that exhibits at least about 40% amino acid sequence identity to the coiled- coiled domain of NL8, (4) an amino acid sequence that exhibits at least about 25% amino acid sequence identity to the coiled-coiled domain of FLS 139, or (5) an amino acid sequence that fulfills any combination of (l)-(4), wherein in vivo administration or production of the fusion protein promotes or inhibits angiogenesis in a mammalian host. [0009] The invention also provides a fusion protein comprising (a) a fibrinogen-like domain that comprises (1) an amino acid sequence that exhibits at least about 60% amino acid sequence identity to the fibrinogen-like domain of NLl, NL5, NL8, or any combination thereof, (2) an amino acid sequence that exhibits at least about 60% amino acid sequence identity to the fibrinogen-like domain of NL4, (3) an amino acid sequence that exhibits at least about 40% amino acid sequence identity to the fibrinogen-like domain of zapol, murine FDRG, or both, (4) an amino acid sequence that exhibits at least about 45% amino acid sequence identity to the fibrinogen-like domain of NL3, (5) an amino acid sequence that exhibits at least about 30% amino acid sequence identity to the fibrinogen-like domain of FLS 139, or (6) an amino acid sequence which fulfills any combination of (l)-(5), and (b) a coiled-coil domain that exhibits at least about 35% amino acid sequence identity to the coiled-coil domain of Ang-1, Ang-2, Ang-2X, Ang-3, Ang-4, or any combination thereof.

BRIEF DESCRIPTION OF THE DRAWINGS [0010] Figure 1 is a Kyte & Doolittle hydropathy profile of a first exemplary coiled-coil domain suitable for incorporation in a fusion protein of the invention. [0011] Figure 2 is a Kyte & Doolittle hydropathy profile of a second exemplary coiled- coil domain suitable for incorporation in a fusion protein of the invention. [0012] Figure 3 is a Kyte & Doolittle hydropathy profile of a third exemplary coiled- coil domain suitable for incorporation in a fusion protein of the invention. [0013] Figure 4 is a Kyte & Doolittle hydropathy profile of a fourth exemplary coiled- coil domain suitable for incorporation in a fusion protein of the invention. [0014] Figure 5 is a Kyte & Doolittle hydropathy profile of a fifth exemplary coiled-coil domain suitable for incorporation in a fusion protein of the invention. [0015] Figure 6 is a Kyte & Doolittle hydropathy profile of a sixth exemplary coiled- coil domain suitable for incorporation in a fusion protein of the invention. [0016] Figure 7 is a Kyte & Doolittle hydropathy profile of a seventh exemplary coiled- coil domain suitable for incorporation in a fusion protein of the invention. [0017] Figure 8 is a Kyte & Doolittle hydropathy profile of an eighth exemplary coiled- coil domain suitable for incorporation in a fusion protein of the invention. [0018] Figure 9 is a Kyte & Doolittle hydropathy profile of a ninth exemplary coiled- coil domain suitable for incorporation in a fusion protein of the invention. [0019] Figure 10 is a Kyte & Doolittle hydropathy profile of a tenth exemplary coiled- coil domain suitable for incorporation in a fusion protein of the invention. [0020] Figure 11 is a Kyte & Doolittle hydropathy profile of an eleventh exemplary coiled-coil domain suitable for incorporation in a fusion protein of the invention. [0021] Figure 12 is a Kyte & Doolittle hydropathy profile of a twelfth exemplary coiled-coil domain suitable for incorporation in a fusion protein of the invention.

[0022] Figure 13 is a Kyte & Doolittle hydropathy profile of a thirteenth exemplary coiled-coil domain suitable for incorporation in a fusion protein of the invention.

[0023] Figure 14 is a Kyte & Doolittle hydropathy profile of a fourteenth exemplary coiled-coil domain suitable for incorporation in a fusion protein of the invention.

[0024] Figure 15 is a Kyte & Doolittle hydropathy profile of a fifteenth exemplary coiled-coil domain suitable for incorporation in a fusion protein of the invention.

[0025] Figure 16 is a Kyte & Doolittle hydropathy profile of a sixteenth exemplary coiled-coil domain suitable for incorporation in a fusion protein of the invention.

[0026] Figure 17 is a Kyte & Doolittle hydropathy profile of a seventeenth exemplary coiled-coil domain suitable for incorporation in a fusion protein of the invention.

[0027] Figure 18 is a Kyte & Doolittle hydropathy profile of an eighteenth exemplary coiled-coil domain suitable for incorporation in a fusion protein of the invention.

[0028] Figure 19 is a Kyte & Doolittle hydropathy profile of a nineteenth exemplary coiled-coil domain suitable for incorporation in a fusion protein of the invention.

DETAILED DESCRIPTION OF THE INVENTION [0029] The invention provides novel angiopoietin-related factors, particularly novel fusion proteins, the novel ARF homolog Ang-2X, and synthetic homologs of known ARFs. The invention further provides polynucleotides encoding such ARFs, vectors comprising such polynucleotides, and methods of producing such fusion proteins, polynucleotides, and vectors. The invention additionally provides methods of using such novel ARFs, and new methods of using ARFs already known in the art.

[0030] An "angiopoietin related factor" (ARF) is any protein of at least about 100 amino acid residues, preferably at least about 200 amino acid residues, more preferably at least about 300 amino acid residues, and typically about 350-600 amino acid residues (most typically about 500 amino acid residues) that exhibits at least about 20% overall amino acid sequence identity, preferably at least about 25% amino acid sequence identity, and even more preferably at least about 30% amino acid sequence identity (e.g., at least about 35%, at least about 40%, or even at least about 50% amino acid sequence identity) to angiopoietin- 1 (SEQ ΓD NO: !). [0031] The fusion proteins of the invention comprise at least two peptide portions, each of which independently and separately (1) share an identical amino acid sequence with a naturally occurring ARF peptide portion, (2) comprise an amino acid sequence that exhibits a significant amount of sequence identity (at least about 25% amino acid sequence identity, preferably at least about 30% amino acid sequence identity) to the amino acid sequence of a wild-type ARF peptide portion, (3) comprise an amino acid sequence that exhibits a significant amount of sequence homology (at least about 30% sequence homology, preferably at least about 35% sequence homology) to a wild-type ARF peptide portion, (4) comprise an amino acid sequence that exhibits significant weight homology to and/or similar hydropathy profile with a wild-type ARF peptide portion, (5) are encoded by a polynucleotide that hybridizes, under at least moderate stringency conditions with the complement of a polynucleotide that, when transcribed and translated, produces (or otherwise results in the production of, e.g., after post-translational modification) a protein that is identical to a wild-type ARF peptide portion, or would so hybridize but for the degeneracy of the genetic code, (6) are bound by antibodies to a naturally occurring ARF peptide portion, or (7) exhibit any combination of characteristics (l)-(6). ARF fusion protein peptide portions that exhibit more than one of the six characteristics are preferred, and ARF fusion protein peptide portions that exhibit all of characteristics (4)-(6) and at least one of characteristics (l)-(3) (most preferably characteristic (1)) are particularly preferred. [0032] A "peptide portion" is any amino acid sequence contained within a larger peptide, polypeptide, or protein (which terms are used synonymously herein, unless otherwise indicated) of at least about 15, preferably at least about 20, more preferably at least about 25, and even more preferably at least about 30 amino acid residues (e.g., at least about 40, at least about 50, at least about 70, at least about 100, or even more amino acid residues).

[0033] "Identity" (sometimes referred to as "overall" identity) with respect to amino acid or polynucleotide sequences refers to the percentage of residues or bases that are identical in the two sequences when the sequences are optimally aligned. If, in the optimal alignment, a position in a first sequence is occupied by the same amino acid residue or nucleotide as the corresponding position in the second sequence, the sequences exhibit identity with respect to that position. The level of identity between two sequences (or "percent sequence identity") is measured as a ratio of the number of identical positions shared by the sequences with respect to the size of the sequences (i.e., percent sequence identity = (number of identical positions/total number of positions) x 100). [0034] The "optimal alignment" is the alignment that provides the highest identity between the aligned sequences. In obtaining the optimal alignment, gaps can be introduced, and some amount of non-identical sequences and/or ambiguous sequences can be ignored. Preferably, if a gap needs to be inserted into a first sequence to achieve the optimal alignment, the percent identity is calculated using only the residues that are paired with a corresponding amino acid residue (i.e., the calculation does not consider residues in the second sequences that are in the "gap" of the first sequence). However, it is often preferable that the introduction of gaps and/or the ignoring of non-homologous/ambiguous sequences are associated with a "gap penalty."

[0035] A number of mathematical algorithms for rapidly obtaining the optimal alignment and calculating identity between two or more sequences are known and incorporated into a number of available software programs. Examples of such programs include the MATCH-BOX, MULTAIN, GCG, FASTA, and ROBUST programs for amino acid sequence analysis, and the SIM, GAP, NAP, LAP2, GAP2, and PIPMAKER programs for nucleotide sequences. Preferred software analysis programs for both amino acid and polynucleotide sequence analysis include the ALIGN, CLUSTAL W (e.g., version 1.6 and later versions thereof), and BLAST programs (e.g., BLAST 2.1, BL2SEQ, and later versions thereof).

[0036] For amino acid sequence analysis, a weight matrix, such as the BLOSUM matrixes (e.g., the BLOSUM45, BLOSUM50, BLOSUM62, and BLOSUM80 matrixes), Gonnet matrixes (e.g., the Gonnet40, Gonnet80, Gonnetl20, GonnetlόO, Gonnet250, and Gonnet350 matrixes), or PAM matrixes (e.g., the PAM30, PAM70, PAM120, PAM160, PAM250, and PAM350 matrixes), are used in determining identity. BLOSUM matrixes are preferred. The BLOSUM50 and BLOSUM62 matrixes are typically most preferred. In the absence of availability of such weight matrixes (e.g., in nucleic acid sequence analysis and with some amino acid analysis programs), a scoring pattern for residue/nucleotide matches and mismatches can be used (e.g., a +5 for a match and -4 for a mismatch pattern). [0037] The ALIGN program produces an optimal global alignment of the two chosen protein or nucleic acid sequences using a modification of the dynamic programming algorithm described by Myers and Miller, CABIOS, 4, 11-17 (1988). Preferably, if available, the ALIGN program is used with weighted end-gaps. If gap opening and gap extension penalties are available, they are preferably set between about -5 to -15 and 0 to -3, respectively, more preferably about -12 and -0.5 to -2, respectively, for amino acid sequence alignments, and -10 to -20 and -3 to -5, respectively, more preferably about -16 and -4, respectively, for nucleic acid sequence alignments. The ALIGN program and principles underlying it are further described in, e.g., Pearson et al., Proc. Natl. Acad. Sci. USA, 85, 2444-48 (1988), and Pearson et al., Methods Enzymol, 183, 63-98 (1990). [0038] The BLAST programs provide analysis of at least two amino acid or nucleotide sequences, either by aligning a selected sequence against multiple sequences in a database (e.g., GenSeq), or, with BL2SEQ, between two selected sequences. BLAST programs are preferably modified by low complexity filtering programs such as the DUST or SEG programs, which are preferably integrated into the BLAST program operations (see, e.g., Wooton et al., Compu. Chem., 17, 149-63 (1993), Altschul et al., _Vαf. Genet., 6, 119-29 (1994), Hancock et al., Comput. Appl. Biosci., 10, 67-70 (1994), and Wootton et al., Meth. in Enzym., 266, 554-71 (1996)). If a lambda ratio is used, preferred settings for the ratio are between 0.75 and 0.95, more preferably between 0.8 and 0.9. If gap existence costs (or gap scores) are used, the gap existence cost preferably is set between about -5 and -15, more preferably about -10, and the per residue gap cost preferably is set between about 0 to -5, more preferably between 0 and -3 (e.g., -0.5). Similar gap parameters can be used with other programs as appropriate. The BLAST programs and principles underlying them are further described in, e.g., Altschul et al., J. Mol. Biol, 215, 403-10 (1990), Karlin and Altschul, Proc. Natl Acad. Sci. USA, 87, 2264-68 (1990) (as modified by Karlin and Altschul, Proc. Natl. Acad. Sci. USA, 90, 5873-77 (1993)), and Altschul et al., Nucl Acids Res., 25, 3389-3402 (1997).

[0039] For multiple sequence analysis, the CULSTAL W program can be used. The CLUSTAL W program desirably is run using "dynamic" (versus "fast") settings. Preferably, nucleotide sequences are compared using the BESTFIT matrix, whereas amino acid sequences are evaluated using a variable set of BLOSUM matrixes depending on the level of identity between the sequences (e.g., as used by the CLUSTAL W version 1.6 program available through the San Diego Supercomputer Center (SDSC)). Preferably, the CLUSTAL W settings are set to the SDSC CLUSTAL W default settings (e.g., with respect to special hydrophilic gap penalties in amino acid sequence analysis). The CLUSTAL W program and underlying principles of operation are further described in, e.g., Higgins et al., CABIOS, 8(2), 189-91 (1992), Thompson et al., Nucleic Acids Res., 22, 4673-80 (1994), and Jeanmougin et al., Trends Biochem. Sci., 23, 403-07 (1998).

[0040] Several commercially available software suites incorporate the ALIGN, BLAST, and CLUSTAL W programs and similar functions, and may include significant improvements in settings and analysis. Examples of such programs include the GCG suite of programs and those available through DNASTAR, Inc. (Madison, Wisconsin). Particular preferred programs include the Lasergene and Protean programs sold by DNASTAR. [0041] Because various algorithms, matrixes, and programs are commonly used to analyze sequences, amino acid and polynucleotide sequences are preferably characterized in terms of approximate identities by indicating a range of identity "about" a particular identity (e.g., +/- 10%), more preferably +/- 8%, and even more preferably +/- 5% of the particular identity). Alternatively, an exact identity can be measured by using only one of the aforementioned programs, preferably one of the BLAST programs, as described herein. An amino acid sequence of a fusion protein peptide portion can alternatively exhibit significant (at least about 30%, preferably at least about 35%) sequence "homology" to a wild-type ARF peptide portion, while failing to exhibit a significant level of amino acid sequence identity. Homology is a function of the number of corresponding conserved and identical amino acid residues in the optimal homology alignment. The "optimal homology alignment" is the alignment that provides the highest level of homology between two amino acid sequences, using the principles described above with respect to the "optimal alignment." Conservative amino acid residue substitutions involve exchanging a member within one class of amino acid residues for a residue that belongs to the same class. Fusion protein peptide portions containing conservative substitutions are expected to substantially retain the biological properties and functions associated with their wild-type counterpart or wild-type counterpart portions. The classes of amino acids and the members of those classes are presented in Table 1.

Table 1 —Amino Acid Residue Classes Amino Acid Class Amino Acid Residues

Acidic Residues ASP and GLU

Basic Residues LYS, ARG, and HIS

Hydrophilic Uncharged Residues SER, THR, ASN, and GLN

Aliphatic Uncharged Residues GLY, ALA, VAL, LEU, and ILE

Non-polar Uncharged Residues CYS, MET, and PRO

Aromatic Residues PHE, TYR, and TRP

[0042] The fusion protein peptide portion preferably will exhibit high weight homology to a naturally occurring ARF peptide portion. "High weight homology" means that at least about 40%), preferably at least about 60%, and more preferably at least about 70%> of the non-identical amino acid residues are members of the same weight-based "weak conservation group" or "strong conservation group" as the corresponding amino acid residue in the wild-type ARF peptide portion. Strong group conservation is preferred. Weight-based conservation is determined on the basis of whether the non-identical corresponding amino acid is associated with a positive score on one of the weight-based matrices described herein (e.g., the BLOSUM50 matrix and preferably the PAM250 matrix). Weight-based strong conservation groups include Ser Thr Ala, Asn Glu Gin Lys, Asn His Gin Lys, Asn Asp Glu Gin, Gin His Arg Lys, Met lie Leu Val, Met He Leu Phe, His Tyr, and Phe Tyr Trp. Weight-based weak conservation groups include Cys Ser Ala, Ala Thr Val, Ser Ala Gly, Ser Thr Asn Lys, Ser Thr Pro Ala, Ser Gly Asn Asp, Ser Asn Asp Glu Gin Lys, Asn Asp Glu Gin His Lys, Asn Glu Gin His Arg Lys, Phe Val Leu He Met, and His Phe Tyr. The CLUSTAL W sequence analysis program provides analysis of . weight-based strong conservation and weak conservation groups in its output, and offers the preferred technique for determining weight-based conservation, preferably using the CLUSTAL W default settings used by SDSC.

[0043] Additionally, the fusion protein peptide portion can be a peptide portion that exhibits a similar hydropathy profile (hydrophilicity) to a wild-type ARF protein peptide portion. A hydropathy profile can be determined using the Kyte & Doolittle index, the scores for each naturally occurring amino acid in the index being as follows: I (+4.5), V (+4.2), L (+3.8), F (+2.8), C (+2.5), M (+1.9); A (+1.8), G (-0.4), T (-0.7), S (-0.8), W (- 0.9), Y (-1.3), P (-1.6), H (-3.2); E (-3.5), Q (-3.5), D (-3.5), N (-3.5), K (-3.9), and R (-4.5) (see, e.g., U.S. Patent 4,554,101 and Kyte & Doolittle, J. Molec. Biol, 157, 105-32 (1982) for further discussion). Preferably, at least about 45%, preferably at least about 60%, and more preferably at least about 75% (e.g., at least about 85%, at least about 90%, or at least about 95%) of the amino acid residues which differ from the naturally occurring ARF peptide portions exhibit less than a +/-2 change in hydrophilicity, more preferably less than a +/-1 change in hydrophilicity, and even more preferably less than a +/-0.5 change in hydrophilicity. Overall, the fusion protein peptide portions of the invention preferably exhibit a total change in hydrophilicity of less than about 150, more preferably less than about 100, and even more preferably less than about 50 (e.g., less than about 30, less than about 20, or less than about 10) from their wild-type counterparts. Examples of typical amino acid substitutions that retain similar or identical hydrophilicity include arginine- lysine substitutions, glutamate-aspartate substitutions, serine-threonine substitutions, glutamine-asparagine substitutions, and valine-leucine-isoleucine substitutions. The GREASE program, available through the SDSC, provides a convenient way for quickly assessing the hydropathy profile of a peptide portion. For example, a coiled-coil domain fusion protein peptide portion will desirably exhibit a predominately negative hydropathy profile (i.e., at least 50% of the GREASE program line output is less than 0, representing that at least 50% of the amino acid residues exhibit a Kyte-Doolittle index score of less than 0). The fusion protein coiled-coil domain will typically exhibit a hydropathy profile where all residues exhibit a hydropathy score equal or less than leucine and more preferably equal of less than alanine. Desirably, a majority of the residues in the coiled-coil domain will exhibit a hydropathy profile equal to or less than glycine. Typically and preferably, the coiled-coil domain will be substantially formed (e.g., at least about 70%, preferably at least about 80%), and more preferably at least about 90%) of residues having a Kyte & Doolittle score of between -4 and 2, and more preferably between -3 and 1. Desirably, all amino acid residues in the coiled-coil domain will have a score between -3 and 1. A fusion protein coiled-coil domain will preferably exhibit a GREASE program graphical line output comprising a pattern of several (typically about 10-15) peaks and valleys of residues scoring between -3 to -2 and 0 to +1 respectively. GREASE program outputs of select examples of preferred coiled-coil domains are provided in Figs. 1-19. The fusion protein preferably comprises a coiled-coil domain having a hydropathy profile substantially identical (e.g., less than about a 10% difference from) or identical to one of the profiles shown in either Figs. 1- 3 or 4-9.

[0044] In yet another alternative, the fusion protein peptide portion can comprise or consist of a peptide of at least about 40 amino acid residues, preferably at least about 75 amino acid residues, and more preferably at least about 150 (e.g., at least about 200, at least about 250, or more) amino acid residues encoded by a polynucleotide that hybridizes to (1) the complement of a polynucleotide that, when expressed, results in a naturally occurring ARF peptide portion, under at least moderate, preferably high, stringency conditions, or (2) a polynucleotide which would hybridize to the complement of such a sequence under such conditions but for the degeneracy of the genetic code. Alternatively, the peptide portion can comprise a sequence encoded by a polynucleotide that selectively hybridizes to a wild-type ARF peptide portion-encoding polynucleotide of at least about 60 nucleotides (preferably at least about 120 nucleotides, and more preferably at least about 150 nucleotides, or more) with respect to other wild-type ARF-encoding polynucleotide sequences, and, more preferably selectively with respect to other wild-type proteins of the same organism, species, family, and/or kingdom.

[0045] Exemplary moderate stringency conditions include overnight incubation at 37°C in a solution comprising 20% formamide, 5x SSC (150 mM NaCl, 15 mM trisodium citrate), 50 mM sodium phosphate (pH 7.6), 5x Denhardt's solution, 10%) dextran sulfate, and 20 mg/mL denatured sheared salmon sperm DNA, followed by washing the filters in lx SSC at about 37-50°C, or substantially similar conditions, e.g., the moderately stringent conditions described in Sambrook et al., MOLECULAR CLONING: A LABORATORY MANUAL (Cold Spring Harbor Press 1989). High stringency conditions are conditions that use, for example, (1) low ionic strength and high temperature for washing, such as 0.015 M sodium chloride/0.0015 M sodium citrate/0.1% sodium dodecyl sulfate (SDS) at 50°C, (2) employ a denaturing agent during hybridization, such as formamide, for example, 50% (v/v) formamide with 0.1 % bovine serum albumin (BSA)/0.1% Ficoll/0.1% polyvinylpyrrolidone (PVP)/50 mM sodium phosphate buffer at pH 6.5 with 750 mM sodium chloride, 75 mM sodium citrate at 42°C, or (3) employ 50% formamide, 5x SSC (0.75 M NaCl, 0.075 M sodium citrate), 50 mM sodium phosphate (pH 6.8), 0.1 % sodium pyrophosphate, 5x Denhardt's solution, sonicated salmon sperm DNA (50μg/ml), 0.1 % SDS, and 10%> dextran sulfate at 42°C, with washes at (i) 42°C in 0.2x SSC, (ii) at 55°C in 50% formamide and (iii) at 55°C in O.lx SSC (preferably in combination with EDTA). Additional details and explanation of stringency of hybridization reactions are provided in, e.g., Ausubel et al., CURRENT PROTOCOLS IN MOLECULAR BIOLOGY (Wiley Interscience Publishers 1995). [0046] Another possibility is that the fusion protein peptide portion comprises or consists of an amino acid sequence that is bound by an antibody that also binds a wild-type ARF peptide portion. For example, the fusion protein peptide portion can be a peptide portion that is bound by the wild-type ARF antibodies described in, e.g., U.S. Patent 6,166,185, and/or International Patent Applications WO 00/18804, WO 00/18437, and WO 99/15653. Antibodies to Ang-1, Ang-2, Ang-3, and Ang-4 are commercially available through Research Diagnostics, Inc. (Flanders NJ). Other suitable wild-type ARF antibodies are described in the various patent documents and references directed to ARFs discussed elsewhere herein.

[0047] The production of new antibodies to the wild-type ARFs, and to the novel ARFs of the invention, also can be facilitated using any suitable technique known in the art. Examples of suitable techniques for obtaining such antibodies are provided in, e.g., Gavilodono et al., Biotechniques, 29(1), 128-32, 134-6, and 138 (passim) (2000), Nelson et al., Mol Pathol, 53(3), 111-7 (2000), Laurino et al, Ann. Clin. Lab. Sci., 29(3), 158-66 (1999), Rapley, Mol Biotechnol, 3(2), 139-54 (1995), Zaccolo et al., Int. J. Clin. Lab. Res., 23(4), 192-8 (1993), Morrison, Annu. Rev. Immunol, 10, 239-65 (1992), "Antibodies, Antigens, and Molecular Mimicry," Meth. Enzymol, 178 (John J. Langone, Ed., Academic Press, November 1989), Moore, Clin. Chem., 35(9), 1849-53 (1989), Rosalki et al., Clin. Chim. Acta, 183(1), 45-58 (1989), and Tami et al, Am. J. Hosp. Phαrm., 43(11), 2816-25 (1986), as well as U.S. Patents 4,022,878, 4,350,683, and 4,022,878. A preferred technique for producing antibodies is provided in Border et al., Proc. Nαtl Acαd. Sci., USA, 97(20), 10701-05 (2000).

[0048] The fusion protein peptide portion alternatively, or, typically and preferably, additionally, can comprise or consist of a peptide portion that exhibits structural similarity to a wild-type ARF peptide portion. It is further desired that the fusion protein exhibits overall structural similarity to a wild-type ARF. Structural similarity can be determined by any suitable technique, preferably using a suitable software program for making such assessments. Examples of such programs include the MAPS program and the TOP program (described in Lu, Protein Data Bank Quarterly Newsletter, #78, 10-11 (1996), and Lu, J. Appl Cryst., 33, 176-183 (2000)). The fusion protein peptide portion, and preferably the fusion protein overall, will desirably exhibit a low structural diversity, topological diversity (e.g., a topical diversity of less than about 20, preferably less than about 15, and more preferably less than about 10), or both, with respect to a wild-type ARF counterpart. Alternatively, the structure of the fusion protein or fusion protein peptide portion can be compared to the desired ARF or ARF peptide portion using the PROCHECK program (described in, e.g., Laskowski, J. Appl. Cryst., 26, 283-291 (1993)), the MODELLER program, or commercially available programs incorporating such features. Alternatively, a sequence comparison using a program such as the PredictProtein server (available at http://dodo.cpmc.columbia.edu/predictprotein/) can reveal similar structure for the fusion protein or fusion protein peptide portion and a wild-type ARF peptide portion counterpart. Additional techniques for analyzing protein structure that can be applied to determine structural similarity are described in, e.g., Yang and Honig, J. Mol Biol, 301(3), 665-78 (2000), Aronson et al., Protein Sci., 3(10), 1706-11 (1994), Marti-Remon et al., Annu. Rev. Biophys. Biomol Struct., 29, 291-325 (2000), Halaby et al., Protein Eng., 12(7), 563-71 (1999), Basham, Science, 283, 1132 (1999), Johnston et al., Crit. Rev. Biochem. Mol. Biol, 29(1), 1-68 (1994), Moult, Curr. Opin. Biotechnol, 10(6), 583-6 (1999), Benner et al., Science, 274, 1448-49 (1996), and Benner et al., Science, 273, 426-8 (1996). [0049] The fusion protein of the invention desirably has an overall structural conformation such that the overall structure of the fusion protein does not interfere with the function of the individual ARF peptide portions. In addition to the techniques described above, conformation of the peptide portions can be assessed using conformation-dependent antibodies. If necessary, linker sequences, such as the linker sequences described further herein, can be used to separate peptide portions in order to avoid steric hindrance and/or undesired interactions at the tertiary or quaternary levels of structure. The fusion protein will advantageously be designed such that multimerization of a heterologous peptide portion, e.g., a VEGF peptide portion, is not interfered with by the multimerization of a second peptide portion, e.g., the multimerization domain of an ARF, such as the Ang-1 CCD. The fusion protein also will desirably be designed such that post-translational modification of the fusion protein is of a manner suitable for retaining the function of the peptide portions, typically by expression of the fusion protein with appropriate signal sequences, examples of which are discussed further herein.

[0050] The fusion protein can include any combination of peptide portions having any of the aforementioned qualities. Thus, for example, the fusion protein can include a first peptide portion comprising an amino acid sequence identical to the amino acid sequence of a first wild-type ARF peptide portion and a second peptide portion comprising a sequence exhibiting 100% sequence homology, but less than 100%) sequence identity, but to a second ARF peptide portion. The fusion protein can include any suitable number of ARF peptide portions (peptide portions exhibiting one of the aforementioned seven qualities, preferably exhibiting significant sequence identity or sequence homology to a wild-type ARF peptide portion). For example, the fusion protein can include 3, 4, 5, or even more ARF peptide portions. Where the fusion protein comprises three or more ARF peptide portions, more than one of the peptide portions can be identical, homologous, and/or otherwise similar to the same wild-type ARF peptide portion.

[0051] Where the fusion protein comprises only two ARF peptide portions each ARF peptide portion will desirably exhibit higher sequence identity and/or homology to a different wild-type ARF than the other ARF peptide portion. Preferably, at least one of the ARF fusion protein peptide portions in such a fusion protein will not be an angiopoietin and, more preferably, is not selected from Ang-1, Ang-2, Ang-3, and Ang-4. [0052] The fusion protein exhibits less than 100% overall sequence identity to a wild- type ARF. For example, the fusion protein can exhibit about 95%) or less, about 90% or less, about 80% or less, about 70% or less, about 60%> or less, or even about 50% or less overall amino acid sequence identity to any wild-type ARF. The fusion protein also can lack any peptide portions that exhibit 100%> local sequence identity to a wild-type ARF. For example, the fusion protein can lack any peptide portions that exhibit about 90% or more, about 80%) or more, about 70%> or more, or even about 60%> or more local sequence identity to a wild-type ARF.

[0053] Local sequence identity can be determined using local sequence alignment software, e.g., the BLAST programs described above, the LFASTA program, or, more preferably, the LALIGN program. Preferably, the LALIGN program using a BLOSUM50 matrix analysis used for amino acid sequence analysis, and a +5 match/-4 mismatch analysis used for polynucleotide sequence analysis. Gap extension and opening penalties are preferably the same as those described above with respect to analysis with the ALIGN program. For LALIGN (or other program) analysis using k-tup value settings (also referred to as "k-tuple" or ktup values), a k-tup value of 0-3 for proteins, and 0-10 (e.g., about 6) for nucleotide sequences, is preferred.

[0054] The fusion protein will comprise a fibrinogen-like domain (FLD), a coiled-coil domain (CCD), or, typically and preferably, a FLD and a CCD. Desirably, the fusion protein fibrinogen-like domain exhibits at least about 25%> amino acid sequence identity, preferably at least about 30% amino acid sequence identity, and more preferably at least about 35%o amino acid sequence identity to the Ang-1 fibrinogen-like domain (SEQ ID NO: 2) and/or the KIA0003 associated protein (GenBank Accession No. NP 001137) (SEQ ID NO: 3) (hereinafter referred to as KAP). The fusion protein coiled-coil domain desirably will exhibit at least about 10%> amino acid sequence identity, preferably at least about 15% amino acid sequence identity, more preferably at least about 20% amino acid sequence identity, and even more preferably at least about 30% amino acid sequence identity to the CCD of Ang-1 (SEQ ID NO: 4). Desirably, the fusion protein FLD, CCD, or both domains will exhibit a higher level of local sequence identity to a non-angiopoietin ARF than to angiopoietin ARFs, and more particularly exhibits a higher level of local sequence identity to an ARF other than Ang-1, Ang-2, Ang-3, or Ang-4. While the fusion protein can comprise any suitable number of CCDs and FLDs (e.g., 2, 3, or more CCDs and/or FLDs), it will preferably comprise a single FLD and CCD.

[0055] A "fibrinogen-like domain" (FLD) is a peptide domain of at least about 75, more preferably at least about 150, and even more preferably at least about 210 (e.g., about 175- 250, typically about 210-230) amino acid residues in length, that exhibits at least about 25%>, preferably at least about 35%, and more preferably at least about 40% amino acid sequence identity to (e.g., 20-70% identity) or at least such levels of local sequence identity to the alpha, beta, or gamma chains (typically to the C-terminal globular domain thereof) of a naturally occurring mammalian fibrinogen (e.g., human fibrinogen alpha (1^st example (SEQ ID NO: 4) and 2^nd example (SEQ ID NO: 5)), human fibrinogen beta (1^st example (SEQ ID NO: 6) and 2^nd example (SEQ ID NO: 7)), and/or human fibrinogen gamma (1^st example (SEQ ID NO: 8) and 2^nd example (SEQ ID NO: 9))). As indicated above, the fusion protein FLD also desirably exhibits at least about 25% amino acid sequence identity to the Ang-1 FLD and/or KAP.

[0056] FLDs can be identified using any suitable technique, such as by applying any of the sequence alignment algorithms or programs discussed elsewhere herein to the subject sequence and the chain of a wild-type fibrinogen. Preferably, the subject fusion protein (or other ARF) will comprise an amino acid sequence, which, when submitted to Conserved Domain Database (CDD) analysis (e.g., by submission to the NCBI BLAST program with the CDD (sometimes indicated as CD) analysis option selected), will result in the program identifying the presence of a conserved "FBG" (fibrinogen related domain - CD=smart00186) and/or "Fibrinogen-C" domain (CD=pfam00147), or, most desirably, both types of domains, in the subject amino acid sequence.

[0057] The fusion protein FLD desirably will comprise at least two cysteine residues, preferably at least four cysteine residues, more preferably at least five cysteine residues, and even more preferably at least six cysteine residues, which desirably form internal disulfide bonds. Preferably, at least two, and more preferably at least four, of the FLD's cysteine residues are positioned in the C-terminal half of the FLD.

[0058] A "coiled-coil domain" (CCD) is a peptide domain comprising at least one alpha helix domain, wherein one side of the alpha helix domain comprises a predominantly hydrophobic portion (i.e., at least about 50%_> of the amino acid residues on a side of the alpha helix are hydrophobic, typically at least about 65 %> of the amino acid residues on a side of the alpha helix are hydrophobic (hydrophobic amino acid residues are listed in Table 1, supra)). The CCD can include any suitable number of alpha helix domains. Typically, the CCD will include 1-10 alpha helices (more typically 1-5), each of which are usually about 10 amino acid residues in length. The alpha helices can be antiparallel or parallel in orientation. The alpha helix domain is well known in the art and, as such, is only briefly discussed here. Further description of the features associated with the alpha helix domain can be found in, e.g., Stryer, BIOCHEMISTRY, W.H. Freeman & Co. (New York), 3d Edition (1988), Scholtz et al., Annu. Rev. Biophys. Biomol Struct, 21, 95-118 (1992), Sugeta et al., Biopolymers, 5, 673 - 679 (1967), and references cited therein. Analytical techniques, and computer programs incorporating such techniques for identifying alpha helixes, also are known in the art. See, e.g., Kumar et al., Biophys. J., 71, 1574-1586 (1996), Kumar et al., Biophys. J, 75, 1935-44 (1998), and Bansal et al., J. Biomolec. Struct. Dynamics, 17, 811- 19 (2000). Examples of suitable programs for identifying alpha helices using an amino acid sequence include the GOR4 program, which uses the algorithm of Gamier et al., Meth. Enzym., 266, 97-120 (1996), the CHOFAS program, which uses the algorithms of Chou et al., Adv. Enzymol, 47,45-148 (1978), and Pearson et al., Proc. Natl Acad. Sci. USA, 85, 2444-48 (1988), and the PELE program which combines multiple algorithms. All of these programs are accessible through the San Diego Supercomputer Center (SDSC). [0059] Typically and preferably, the coiled-coil domain will include at least one sequence of the heptad pattern h_a-x_b-x_c-h_d-x_e-xrXg, where "h" represents a hydrophobic amino acid residue, "x" represents any amino acid residue, and subscripted letters reference positions in the sequence. Typically, the amino acid residues at positions x_e and x_g will be polar residues. Often the CCD will include multiple heptad repeats of this pattern. The hydrophobic heptad repeat usually imparts greater stability to the CCD than over an otherwise identical multiple alpha helix domain, by adopting the Crick "knobs-into-holes" packing configuration, where the h_a and h_d hydrophobic residues interact with the internal (i.e., central) portions of a neighboring alpha helix (typically an alpha helix positioned in the CCD, but interactions with a peptide portion of a different polypeptide also are possible, particularly when the CCD is involved with forming multimers, which are further discussed below).

[0060] CCDs can be identified by any suitable technique. Conveniently, CCDs can usually be identified through direct amino acid sequence analysis. Several analytical techniques for identifying CCDs and computer programs incorporating these techniques are available. Any suitable sequence analysis technique or program can be used. Examples of suitable programs include the COILS program (available at http://www.ch.embnet.org/ software/COILS_form.html) and the PARCOILS program (available at http://nightingale.lcs.mit.edu/cgi-bin score). Other suitable programs include the PEPCOIL program, the COILSCAN program, and the coil analysis subprograms of commercially available sequence analysis software (e.g., the GCG program suite, or the PROTEAN program available through DNA Star Inc. (Madison, WI)). Alternatively, other protein structure techniques can be used to identify the presence of a CCD, such as X-ray diffraction analysis, crystallographic analysis, or three-dimensional imaging techniques. CCDs and the identification of CCDs are further described in Burkhard et al., Trends Cell Bio., 11, 82-88 (2001), Beck et al., J. Struct. Biol, 122, 17-29 (1998), Lupas, Curr. Opin. Struct. Biol, 7(3), 388-93 (1997), Kammerer, Matrix Biol, 15, 555-68 (1997), Nilges et al., Proteins, 15, 133-46 (1993), and Adamson et al., Curr. Opin. Biotechnol, 4(4), 428-37 (1993).

[0061] The CCD typically and desirably will comprise an amino acid sequence of at least about 10 amino acid residues (preferably at least about 20 amino acid residues) which exhibit a score of at least about 0.5, more preferably at least about 0.65, and even more preferably at least about 0.75, in the 28 residue window, when the sequence is analyzed by the COILS program. Alternatively, or preferably additionally, the CCD will comprise a sequence of at least about 10 amino acid residues that exhibit a score of at least about 0.3, more preferably at least about 0.4, and even more preferably at least about 0.45 when the sequence is analyzed by the PARCOILS program.

[0062] The fusion protein FLD can comprise or consist of a peptide portion that exhibits 100%) amino acid sequence identity to a wild-type ARF FLD. The fusion protein FLD can exhibit 100%> sequence identity to any suitable ARF FLD. Preferred wild- type FLDs include the Ang-1 FLD, KAP, the Ang-2 FLD (SEQ ID NO: 10), the Ang-3 FLD (SEQ ID NO: 11), the Ang-4 FLD (SEQ ID NO: 12), the NLl FLD (SEQ ID NO: 13), the NL3 FLD (SEQ ID NO: 14), the NL4 FLD (SEQ ID NO: 15), the NL5 FLD (SEQ ID NO: 16), the NL8 FLD (SEQ ID NO: 17), the FLS 139 FLD (SEQ ID NO: 18), the zapol FLD (SEQ ID NO: 19), and the murine FDRG FLD (SEQ ID NO: 20). Particularly preferred wild-type FLDs include the Ang-1 FLD and/or KAP, the NLl FLD, and the NL5 FLD. [0063] The fusion protein CCD can comprise or consist of a peptide portion that exhibits 100% amino acid sequence identity to a wild-type ARF CCD. The fusion protein CCD can exhibit 100% amino acid sequence identity to any suitable wild-type ARF CCD. Preferred wild-type CCDs include the Ang-1 CCD, the Ang-2 CCD, which is predicted to be sequence 1 (SEQ ID NO: 21), sequence 2 (SEQ ID NO: 22), sequence 3 (SEQ ID NO: 23), or sequence 4 (SEQ ID NO: 24), the Ang-3 CCD (SEQ ID NO: 25), the Ang-4 CCD (SEQ ID NO: 26), the NLl CCD (SEQ ID NO: 27), the NL2 CCD (SEQ ID NO: 28), the NL4 CCD (SEQ ID NO: 29), the NL5 CCD (SEQ ID NO: 30), the NL8 CCD (SEQ ID NO: 31), the FLS 139 CCD (SEQ ID NO: 32), the zapol CCD (SEQ ID NO: 33), and the murine FDRG CCD (SEQ ID NO: 34). Particularly preferred wild-type CCDs include the Ang-1 CCD, the Ang-2 CCD, the NLl CCD, and the NL5 CCD. Another preferred CCD (described further herein) is the CCD of the novel ARF, Ang-2X (predicted to be amino acid sequence 1 (SEQ ID NO: 35), sequence 2 (SEQ ID NO: 36), or sequence 3 (SEQ ID NO: 37)). The fusion protein also can comprise an Ang-2X FLD (SEQ ID NO: 38). [0064] The above described wild-type ARF FLDs and CCDs can be combined in any suitable manner in the fusion protein, so long as the FLD and CCD do not exhibit 100%> local sequence identity to the same wild-type ARF. In addition, the fusion protein can comprise a FLD or CCD that exhibits 100% identity to a wild-type ARF peptide portion and a FLD or CCD that is homologous to a wild-type ARF peptide portion, but does not exhibit 100%) local sequence identity to a wild-type ARF. A "homologous peptide portion" or "homolog" refers to any peptide portion that exhibits any of qualities (2)-(6) described above. Preferably, the homologous ARF peptide portion will exhibit at least about 30% amino acid sequence identity, or at least about 35%> amino acid sequence homology, to a wild-type ARF peptide portion.

[0065] A preferred fusion protein comprises a FLD that exhibits 100% amino acid sequence identity to KAP or the Ang-1 FLD and a CCD that exhibits 100%> amino acid sequence identity to a wild-type ARF other than Ang-1, Ang-2, Ang-3, or Ang-4. Preferably, the fusion protein comprises a CCD that exhibits 100% amino acid sequence identity to a wild-type ARF that is not an angiopoietin. More preferably, the fusion protein comprises a CCD that exhibits 100%> amino acid sequence identity to the NLl CCD, NL4 CCD, the NL5 CCD, the NL8 CCD, the FLS 139 CCD, the zapol CCD, or the murine FDRG CCD. Most preferably, the fusion protein comprises a CCD that exhibits 100% amino acid sequence identity to the NLl CCD or NL5 CCD.

[0066] Another preferred fusion protein comprises a FLD that exhibits 100%> amino acid sequence identity to the FLD of a wild-type ARF other than the Ang-1, Ang-2, Ang-3, or Ang-4 FLD, and more preferably exhibits 100% amino acid sequence identity to a non- angiopoietin ARF FLD. Such a fusion protein desirably further comprises a CCD that exhibits 100%> amino acid sequence to a wild-type ARF, most preferably to the CCD of Ang-1 or the CCD of Ang 2X.

[0067] Another preferred fusion protein comprises a FLD or CCD that exhibits 100%> amino acid sequence identity to the FLD or CCD of NLl. Preferably, the fusion protein will comprise a non-NLl CCD or FLD, which will exhibit at least about 45%, more preferably at least about 55%>, and optimally about 100% sequence identity to the CCD or FLD of NL5, as applicable.

[0068] In yet another aspect, the invention provides a fusion protein comprising a FLD or CCD that exhibits about 100% amino acid sequence identity to the FLD or CCD of NL5, as applicable. Such fusion proteins desirably comprise a non-NL5 FLD or CCD that exhibits at least about 45%), more preferably at least about 55%>, and optimally about 100%> sequence identity to the CCD or FLD of NLl.

[0069] The fusion protein can comprise one, two, or more ARF peptide portions that exhibit less than 100% sequence identity to a wild-type ARF peptide portion (an ARF wild- type CCD or FLD). In such fusion proteins, the biological properties of the ARF peptide portions can be changed in a therapeutically advantageous manner. For example, the removal, substitution, or addition of particular amino acids to the fusion protein FLD can reduce a host immune response to the fusion protein and/or increase the stability of the fusion protein, examples of which are discussed elsewhere herein.

[0070] A preferred fusion protein of the invention includes (1) a fibrinogen-like domain that exhibits at least about 65%>, preferably at least about 75%, and more preferably at least about 85%), amino acid sequence identity to the fibrinogen-like domain of a first wild-type protein, which binds a receptor that, when activated, promotes or inhibits angiogenesis, and (2) a coiled-coil domain that exhibits at least about 50%>, preferably at least about 75%>, and even more preferably at least about 85%> amino acid sequence identity to the coiled-coil domain of a second wild- type protein, which increases the affinity of the fusion protein for the receptor, promotes formation of a multimer comprising the fusion protein, or both, where the first protein and second protein are different, the first protein and second protein each exhibit at least about 30%> overall amino acid sequence identity to Ang-1, and the first protein, the second protein, or both proteins are not Ang-1, Ang-2, Ang-3, or Ang-4. [0071] The first wild-type protein binds a receptor, that, when activated, promotes or inhibits angiogenesis. The first wild-type protein is typically and preferably an ARF, and, most preferably, an ARF that binds and activates a receptor that promotes angiogenesis (e.g., Ang-1). A "receptor" is any protein that interacts with at least one ligand (e.g., a wild- type ARF) to initiate a biological response apart from the receptor-ligand interaction. Receptor activation occurs when the receptor-ligand interaction produces the biological response. Preferably, the receptor is a TIE receptor, and most preferable a TIE-2 receptor. [0072] Receptor binding can be measured using any suitable technique. Simple techniques for determining receptor binding include saturation and competition assays performed with labeled (typically radiolabeled) ligands or putative ligands, which are briefly discussed elsewhere herein. Alternatively, a labeled receptor, or material containing the receptor, can be associated with a filter or a chromatography matrix, and a solution comprising the ligand or putative ligand can be permitted to associate with the receptor or receptor composition, to determine whether receptor binding occurs. Scintillation techniques also are useful for evaluating radiolabeled ligand-receptor interactions. Alternatively, streptavidin/biotin labeling systems, or labeling with fluorescent dyes or peptide portions (e.g., a Green Fluorescent Protein portion), can be used to determine receptor binding. Phage display techniques (such as those described in Cwirla et al., Proc. Natl. Acad. Sci. USA, 87, 6378-6382 (1990) and McCafferty et al, Nature, 384, 552-554 (1990)) also can be used to identify ligands. Such techniques are known in the art, and, as such, are not discussed in detail herein. Principles and techniques related to ligand-receptor assays are further described in, e.g., Smisterova et al., J. Pharm. Biomed. Anal, 12(6),123- 45 (1994), Shaw et al., Curr. Opin. Biotechnol, 3(1), 55-58 (1992), Strosberg et al., Curr. Opin. Biotechnol, 2(1), 30-36 (1991), Zoon et al., J Pharm. Biomed. Anal, 7(2), 147-54 (1989), and Smith et al., Clin. Chem., 26(5), 543-50 (1980), and references cited therein, as well as U.S. Patents 5,480,792 and 5,939,272 (and references cited therein). Commercially available receptor binding assays include, e.g., the HitHunter® assay system (available through DiscoveRx (Fremont CA)). Related techniques for measuring receptor binding are known in the art and discussed further herein.

[0073] When the first wild-type protein binds the receptor it also activates the receptor, resulting in a biological response that modulates (either increases or decreases) angiogenesis. "Angiogenesis," in the context of the invention, encompasses promoting the formation of new blood vessels (also referred to in the art as neovascularization), e.g., by attracting endothelial cells to promote blood vessel sprouting, promoting blood vessel growth from or within existing blood vessels (such as by increasing the size of existing blood vessels or inducing collateral blood vessel growth from existing blood vessels (also known as vasculogenesis and often referred to as angiogenesis)), promoting blood vessel remodeling, promoting blood vessel maturation, and reducing blood vessel permeability (leakage). Thus, an angiogenic peptide or peptide portion can be any sequence of amino acids that induces the initiation of blood vessel growth at a location not otherwise undergoing angiogenesis, enhances or heightens collateral blood vessel growth in a location already undergoing angiogenesis, or both.

[0074] The angiogenic activity of a protein can be determined by any suitable technique. Examples of techniques for measuring angiogenesis include administering the angiogenic protein or DNA encoding the angiogenic protein (preferably in a suitable vector) in the rabbit or rat hind limb models (using a protocol as described in, e.g., Poliakova et al., J. Thorac. Cardiovasc. Surg., 118(2), 339-47 (1999), Rosengart et al., J. Vase. Surg., 26(2), 302-12 (1997), Walder et al., J. Cardiovasc. Pharmacol, 27, 91-98 (1996), Takeshita et al., J. Clin. Invest., 93(2), 662-70 (1994), and U.S. Patent 6,121,246), and/or the mouse ear model (using a protocol similar to that described in, e.g., Kjosleth et al., Microsurgery, 15(6), 390-98 (1994), or as discussed herein in Examples 1, 2, 4, and 5). Similar techniques are discussed in, e.g., Takeshita et al., J Clin. Invest., 93(20), 662-70 (1994). Other assays for assessing the angiogenic potential of an angiogenic factor include performing exercise tolerance tests (as described in, e.g., Fujita et al., Circulation, 77 (5), 1022-29 (1988), Komowski et al., Am. J. Cardiol, 81(7 A), 44E-48E (1998), and Rosengart et al., Circulation, 100(5), 468-74 (1999)), magnetic resonance imaging (MRI) testing for local perfusion, rest and stress (adenosine) ^99m Tc-sestamibi SPECT tests, rest and stress (dobutamine) echocardiograms, gradient echo tests, intravascular ultrasound (IVUS) (as described in, e.g., Oshima et al., Vase. Med., 3(4), 281-90 (1998)), angiography tests, or any combinations thereof, after administration of the putative angiogenic factor to a tissue (preferably a potentially ischemic or ischemic tissue in a mammalian host). Other quantitative angiogenesis activity assays include the corneal pocket assay, capillary number assay, the matrigel angiogenesis/endothelial cell assay, endothelial cell chemotaxis assays, umbilical artery outgrowth assay, choriollantoic membrane development assay, and related assays described in, e.g., Dellian et al., Am. J. Path., 149, 59-72 (1996), Folkman, Cell, 79, 315-28 (1994), O'Reilly et al., Cell, 88, 277-84 (1997), and Ribatti et al., J. Vase. Res., 34, 455-63 (1997). A more recent assay specifically designed for analytically comparing the angiogenic potential of different factors is described in Wang et al., Int. J. Mol. Med., 6(6), 645-53 (2000). The angiogenesis-inducing capability of a protein also can be determined by comparative measurement of the number of blood vessels, blood vessel density, total blood vessel volume, blood flow measurements, blood pressure ratios, or the like, in a particular tissue to which an angiogenic factor has been administered (as described in, e.g., Sands et al., Cancer Lett., 27(1), 15-21 (1985), Pu, et al., Circulation, 88, 208-15 (1993), Bauters et al., Am. J. Physiol, 267, H1263-71 (1994), Takeshita et al., supra, Bauters et al., Circulation, 91, 2802- 09 (1995), Bauters et al., J. Vase. Surg., 21, 314-25 (1995), and Witzenbichler et al., Am. J. Pathol, 153(2), 381-94 (1998)). Other useful techniques for measuring angiogenesis include those described in U.S. Patents 5,976,782, 5,972,639, and 5,919,759, as well as the angiogenesis assays described in Brown et al., Nat. Med., 7(7), 864-68 (2001), LaForge et al., Breast J., 6(2), 103-107 (2000), Tolsma et al., J. Cell Biol, 122, 497 (1993), and Vogel et al., J. Cell. Biochem., 53, 74 (1993). [0075] Although the first wild-type protein binds and activates a receptor, the fusion protein FLD does not have to activate the receptor. Thus, the fusion protein can comprise a FLD that binds the receptor that associates with the first wild-type protein, without activating the angiogenic or anti-angiogenic response upon binding. In this respect, the fusion protein can act as an antagonist for a wild-type ARF, competing for binding to at least one of its receptors. Preferably such fusion proteins comprise an FLD that exhibits at least about 90%> amino acid sequence identity to Ang-2, and preferably which comprises 100%) amino acid sequence identity to the Ang-2 FLD or the Ang-2X FLD. In some cases, the fusion protein can inhibit angiogenesis.

[0076] Where the fusion protein FLD does not exhibit 100% amino acid sequence identity to a wild-type ARF FLD, the FLD will desirably comprise a first amino acid sequence of the sequence pattern Glu Phe/Tyr/His Trp Leu Gly Leu/ Asn Glu/ Asp Xaa Val/Ile/Leu Xaa Xaa Ile/Leu Xaa₍ι_2-ι₄₎ Asp Trp (SEQ ID NO: 39) ("Xaa" represents any amino acid, unless otherwise noted herein, and a "/" between residues indicates that any of the residues divided by the slash is suitable at that position in the sequence). The FLD preferably comprises a second sequence of the sequence pattern Phe Ser Thr Xaa Asp Xaa Asp Asn/His Xaaø-io) Cys Ala/Ser Xaa₍₄₎ Gly Gly Trp Trp Xaa(_2-4) Cys (SEQ ID NO: 40) (subscripted numbers in parentheses herein refer to numbers of residues at a position, subscripted numbers without parentheses identify a particular residue for purposes of discussing preferred residues or residue types present at that position (e.g., "Xaa(_4-5)" represents that any four or five amino acids may occupy that position in the sequence)). Typically and preferably, the first FLD sequence is positioned N-terminal to the second FLD sequence.

[0077] The first FLD sequence and second FLD sequence are typically and preferably separated by at least about 20 amino acid residues, and more preferably at least about 30-65 amino acid residues. This sequence of amino acids between the first and second FLD sequences (FLD sequence 1 and FLD sequence 2, respectively) desirably comprises a sequence of the pattern Tyr Xaai Leu Arg Leu (SEQ ID NO: 41) or Tyr Xaa] Leu Gin (SEQ ID NO:42), which can be referred to as the first intervening FLD sequence. Typically, the residue at position Xaai in the first intervening FLD sequence will be an Arg, Ser, Lys, Ala, or Gin residue. The fusion protein FLD further preferably includes a second intervening sequence, located N-terminal to the first intervening sequence but C-terminal to FLD sequence 1 , of a sequence pattern such as Phe Xaa Xaa His Xaai Gly Xaa₂ Glu Xaa₃ Xaa4 (SEQ ID NO: 43), Tyr Xaa Xaa His Xaai Gly Xaa₂ Glu Xaa₃ Xaa₄ (SEQ ID NO: 44), Phe Xaa Xaa Phe Xaai Gly Xaa₂ Glu Xaa₃ Xaa₄ (SEQ ID NO: 45), Tyr Xaa Xaa Phe Xaai Gly Xaa₂ Glu Xaa₃ Xaa₄ (SEQ ID NO: 46), Phe Xaa Xaa His Xaai Glu Xaa₂ Glu Xaa₃ Xaa₄ (SEQ ID NO: 47), Phe Xaa Xaa Phe Xaai Glu Xaa₂ Glu Xaa₃ Xaa₄ (SEQ ID NO: 48), Phe Xaa Xaa Phe Xaa, Glu Xaa₂ Glu Xaa₃ Xaa* (SEQ ID NO: 49), Tyr Xaa Xaa His Xaai Glu Xaa₂ Glu Xaa₃ Xaa₄ (SEQ ID NO: 50), and Tyr Xaa Xaa Phe Xaai Gly Xaa₂ Glu Xaa₃ Xaa₄ (SEQ ID NO: 51), where Xaai, typically represents an He or Val residue, Xaa₂ typically represents either an Ala, Asp, or Ser residue, Xaa₃ represents a Gin, Glu, or Thr residue, and Xaa₄ represents an Ala, Gly, or Tyr residue. The Glu residue between Xaa₂ and Xaa₃ can be replaced by any other amino acid, although Glu is preferred at this position. The intervening amino acids further preferably comprise a third intervening sequence, positioned N-terminal to the second intervening sequence, but C-terminal to FLD sequence 1, of the pattern Asp/Glu/Ser Asp/Gly Arg/Asn Xaai Xaa₂ Leu/Phe/Tyr Ala/Ile/Leu Xaa₃, where Xaai typically represents an Ala or Lys residue, Xaa₂ typically represents an Ala, Glu, or Val residue, and Xaa₃ typically represents a Glu or Gin residue. The intervening amino acids typically comprise a fourth intervening sequence, positioned C-terminal to the first intervening sequence, but N-terminal to FLD sequence 2, of one of the following sequence patterns: Xaa, Xaa₂ Xaa₃ Thr Gly Xaa o_-3) Ala Gly (SEQ ID NO: 52), Xaai Xaa₂ Xaa₃ Thr Gly Xaa(_0-3) Leu Gly (SEQ ID NO: 53), Xaa, Xaa₂ Xaa₃ Thr Ala Xaa^ ₎ Ala Gly (SEQ ID NO: 54), Xaai Xaa₂ Xaa₃ Thr Ala Xaa(o_-3) Leu Gly (SEQ ID NO: 55), Xaai Xaa₂ Xaa₃ Val Gly Xaa(_0-3) Ala Gly (SEQ ID NO: 56), Xaa, Xaa₂ Xaa₃ Val Gly Xaa(₀.₃₎ Leu Gly (SEQ ID NO: 57), Xaa, Xaa₂ Xaa₃ Val Ala Xaa₍o_-3) Ala Gly (SEQ ID NO: 58), Xaa, Xaa₂ Xaa₃ Val Ala Xaa^₎ Leu Gly (SEQ ID NO: 59), Xaai Xaa₂ Xaa₃ His Gly Xaa(_0-3) Ala Gly (SEQ ID NO: 60), Xaa, Xaa₂ Xaa₃ His Gly Xaa_(0-3) Leu Gly (SEQ ID NO: 61), Xaa, Xaa₂ Xaa₃ His Ala Xaa<_0-3) Ala Gly (SEQ ID NO: 62), Xaai Xaa₂ Xaa₃ His Ala Xaa_(0-3) Leu Gly (SEQ ID NO: 63), Xaai Xaa₂ Xaa₃ Glu Gly Xaa_(0-3) Ala Gly (SEQ ID NO: 64), Xaa, Xaa₂ Xaa₃ Glu Gly Xaa_(0-3) Leu Gly (SEQ ID NO: 65), Xaai Xaa₂ Xaa₃ Glu Ala Xaa(o_-3) Ala Gly (SEQ ID NO: 66), Xaai Xaa₂ Xaa₃ Glu Ala Xaa_(0-3) Leu Gly (SEQ ID NO: 67), where Xaai typically represents a Gly, Lys, Ser, or Thr residue, Xaa₂ represents an Ala or Gly residue, and Xaa₃ represents a Phe or Tyr residue. The intervening amino acids will usually comprise a fifth intervening sequence of 9-22 residues of the pattern Xaa^-π) Asp/Xaa Glu/Ser/Val Phe/Pro/Ser/Tyr Leu/Pro/Ser Ser/Trp/Xaa Gly/His/Xaa Asn/Gln/Leu Asn/Gly/Lys/Met Gln/Pro/Thr positioned C-terminal to the fourth intervening sequence but N-terminal to FLD sequence 2. Desirably the intervening sequence maintains the hydropathy characteristics commonly associated with wild-type ARFs, discussed elsewhere herein.

[0078] Preferably, the fusion protein comprises sequences of more than one, and preferably all, of above-described preferred sequence patterns, in their preferred orientation to one another. As such, the fusion protein FLD desirably comprises a sequence of the above-described sequence patterns in the following order (N-terminal to C-terminal): FLD sequence 1 - third intervening sequence - second intervening sequence - first intervening sequence - fourth intervening sequence - fifth intervening sequence - FLD sequence 2. [0079] FLD sequence 1 desirably comprises a sequence of the more particular sequence pattern Phe Ser Thr Xaai Asp Xaa Asp Asn Asp Xaa Cys Xaa Cys Lys Cys Xaa₂ Xaa₃ Xaa» Xaa₅ Xaae Gly Gly Trp Trp Phe Asp Ala Cys Gly Xaa₇ Ser Asn Leu Gly (SEQ ID NO: 68), where Xaai preferably represents a Leu or Lys residue, Xaa₂ preferably represents an Ala or Ser residue, Xaa₃ preferably represents a Gin or Leu residue, Xaa preferably represents a Met or a Val residue, Xaa₅ preferably represents a Leu or a Met residue, Xaa^ preferably represents a Ser or a Tyr residue, and Xaa preferably represents a Leu or a Pro residue. Also preferably, FLD sequence 2 comprises a sequence of the more specific sequence pattern Glu Xaa, Trp Leu Gly Asn Glu Xaa Xaa₂ Xaa ₂₎ Xaa₃ Thr Xaa Xaa₄ Xaa₍₂₎ Tyr Xaa Leu Xaa₅ Xaa* Glu Leu Xaa Asp Trp Glu Gly (SEQ ID NO: 69), where Xaa, preferably represents a His or a Tyr residue, Xaa₂ preferably represents an He or a Val residue, Xaa₃ preferably represents an He or a Leu residue, Xaa preferably represents an Arg or a Gin residue, Xaa₅ preferably represents an Arg or a Lys residue, and Xaa₆ preferably represents an He or Val residue. [0080] In some situations, a fusion protein comprising an FLD homologous to an angiopoietin FLD is preferred. Such an "angiopoietin-like FLD" desirably exhibits at least about 40%), preferably at least about 50%>, and more preferably at least about 65%> (e.g., at least about 70%, at least about 80%>, at least about 90%>, or even at least about 95%>) amino acid sequence identity to at least one wild-type angiopoietin ARF. In such aspects, the fusion protein FLD desirably exhibits at least about 40%, preferably at least about 50%>, more preferably at least about 60%>, and even more preferably at least about 70% (e.g., at least about 80%>, at least about 90%>, or even at least about 95%>) amino acid sequence identity to the Ang-1 FLD.

[0081] Another set of preferred fusion proteins comprise a fibrinogen-like domain that exhibits at least about 30%> amino acid sequence identity to the fibrinogen-like domain of Ang-1, and a coiled-coil domain that comprises (1) an amino acid sequence that exhibits at least about 50%>, preferably at least about 60%, and more preferably at least about 75%> (e.g., at least about 85%>, about 90%, or even 100%) amino acid sequence identity to the coiled-coil domain of zapol, murine FDRG, or both, (2) an amino acid sequence that exhibits at least about 45%>, preferably at least about 50%>, more preferably at least about 60%, and even more preferably at least about 75% (e.g., at least about 80%, at least about 90%), or even about 100%>) amino acid sequence identity to the coiled-coil domain of NLl, NL5, or both, (3) an amino acid sequence that exhibits at least about 35%>, preferably at least about 40%, more preferably at least about 60%), and even more preferably at least about 15% (e.g., at least about 80%>, at least about 90%, or even about 100%) amino acid sequence identity to the coiled-coiled domain of NL8, (4) an amino acid sequence that exhibits at least about 25%>, preferably at least about 35%>, and more preferably at least about 75%o (e.g., at least about 80%>, at least about 90%, or even about 100%) amino acid sequence identity to the coiled-coiled domain of FLS 139, or (5) an amino acid sequence that fulfills any combination of (l)-(4), wherein in vivo administration or production (typically by transcription and translation of a fusion protein-encoding polynucleotide) of the fusion protein desirably promotes or inhibits (most preferably promotes) angiogenesis in a mammalian host. Preferably, in such fusion proteins, the fusion protein FLD will exhibit at least about 50%>, more preferably at least about 65%, and even more preferably at least about 75%o (e.g., at least about 80%>, at least about 90%>, or even about 100%) amino acid sequence identity to the Ang-1 FLD or KAP. Desirably, the CCD of such a fusion protein will exhibit less than about 40%, preferably less than about 30%>, and even more preferably less than about 20%> amino acid sequence identity to the Ang-1 CCD, Ang-2 CCD, Ang-3 CCD, or Ang-4 CCD (except in some instances where it is preferred that the CCD is the Ang-2X CCD), and can even exhibit less than about 40%_>, less than about 30%>, or even less than about 20%> amino acid sequence identity to the CCD of any angiopoietin (e.g., Ang-6). [0082] The angiopoietin-like FLD can be any suitable size, but preferably is about 200- 250 amino acids in length, and more preferably about 210-230 amino acids in length. The angiopoietin-like FLD desirably includes a first FLD sequence that is of the more particular sequence pattern Glu His/Tyr Trp Leu Gly Asn Glu Xaa Ile/Val Xaa₍₂₎ He/Leu Thr Xaa Arg/Gln Xaa(₂) (SEQ ID NO: 70) (ALFLD sequence 1) and a second FLD sequence is of the sequence pattern Phe Ser Thr Lys/Leu Asp Xaa Asp Asn Asp Xaa Cys Xaa Cys Lys Cys Ala/Ser Gin/Leu Met/Val Leu Met Ser/Thr Gly Gly Trp Trp Phe Asp Ala Cys Gly Leu/Pro Ser Asn Leu Gly (SEQ ID NO: 71) (ALFLD sequence 2). ALFLD sequence 2 preferably is positioned C-terminal to ALFLD sequence 1. Preferably, ALFLD sequence 1 is positioned in the N-terminal-most half of the FLD and ALFLD sequence 2 is positioned in the C- terminal-most half (typically in the C-terminal-most third) of the FLD. [0083] Usually, an intervening amino acid sequence of about 40-60 amino acid residues separate AFLD sequence 1 and ALFLD sequence 2. Also typically, the angiopoietin-like FLD will include a sequence of about 50-60 amino acid residues N-terminal to ALFLD sequence 1 and a sequence of about 30-50 amino acid residues C-terminal to ALFLD sequence 2. Preferably, the portion of the FLD N-terminal to ALFLD sequence 1 will include a sequence of the sequence pattern Gly Gly Xaai Tip Thr Xaa₂ He Gin Xaa₃ Arg Glu Xaa Gly Xaa₅ Xaaβ Xaa₇ Phe Gin Arg Xaa Tip Xaa₈ Xaa₉ Tyr Lys Xaa Gly Phe Xaaio Xaai, Xaa,₂ Xaa,₃ (SEQ ID NO: 72) (ALFLD N-terminus sequence 1), where Xaai preferably is an Arg or a Gly residue, Xaa₂ preferably is a Leu residue, Xaa₃ preferably is an Arg or His residue, Xa^ preferably is an Asn or Asp residue, Xaa₅ preferably is a Ser or Thr residue, Xaa_ό preferably is a Leu or Val residue, Xaa₇ preferably is an Asn or Asp residue, Xaa₈ preferably is a Glu or Lys residue, Xaa₉ preferably is an Asp or Glu residue, Xaaio preferably is an Asn or Asp residue, Xaan preferably is a Pro or Val residue, Xaaι₂ preferably is an Ala or Ser residue, and Xaa_]3 preferably is an Arg or Gly residue. An example of a more particularly preferred sequence pattern of the ALFLD N-terminus sequence 1 pattern is Gly Gly Gly Trp Thr Leu He Gin Xaa Arg Glu Asp Gly Ser Val Asp Phe Gin Arg Xaa Trp Lys Glu Tyr Lys Xaa Gly Phe Gly Asn Pro Ser Gly (SEQ ID NO: 73). The angiopoietin-like FLD alternatively, or preferably additionally, comprises a sequence N-terminal to ALFLD sequence 1, and more preferably N-terminal to ALFLD N-terminus sequence 1, of the pattern Phe Xaa, Asp Cys Ala Xaa Xaa₃ Xaa(₂) Xaa₄ Gly Xaa(₃₎ Xaa₅ Gly Xaa<s Tyr Thr Xaa Xaa(₃) Asn Xaa Xaa₈ Xaa₉ Xaa,₀ Xaa Lys Xaa, , Xaa,₂ Cys Xaa,₃ Xaa,₄ Xaa(₃) (SEQ ID NO: 74) (ALFLD N-terminus sequence 2), where Xaa, preferably is an Arg or a Gin residue, Xaa₂ preferably is an Asp or Glu residue, Xaa₃ preferably is an He or Val residue, Xaa preferably is an Ala or Ser residue, Xaa₅ preferably is an Asn or Ser residue, Xaa₆ preferably is an He or Val residue, Xaa₇ preferably is an He or Val residue, Xaa₈ preferably is a Pro or Thr residue, Xaa₉ preferably is a Lys residue, Xaa,₀ preferably is a Glu or Pro residue, Xaa, , preferably is an Ala or Val residue, Xaa,₂ preferably is a Phe or Tyr residue, Xaan preferably is an Asn or Asp residue, and Xaaι preferably is a Leu or Met residue. Examples of more particular preferred sequence patterns of the ALFLD N- terminus sequence 2 sequence pattern include Phe Arg Asp Cys Ala Gly Val Xaa Xaa Ser Gly Xaa Xaa Xaa Ser Gly He Tyr Thr He Xaa Xaa Xaa Asn Xaa Thr Lys Pro Xaa Lys Val Phe Cys Asp Met Xaa Xaa Xaa (SEQ ID NO: 75) and Phe Arg Asp Cys Ala Asp Val Xaa Xaa Ala Gly Xaa Xaa Xaa Ser Gly He Tyr Thr He Xaa Xaa Xaa Asn Xaa Pro Lys Pro Xaa Lys Val Phe Cys Asn Met Xaa Xaa Xaa (SEQ ID NO: 76).

[0084] The angiopoietin-like FLD desirably comprises a sequence positioned between ALFLD sequence 1 and ALFLD sequence 2 of the sequence pattern Tyr Xaa Leu Xaai Xaa₂ Glu Leu Xaa Asp Trp Glu Gly Xaa Xaa Xaa Xaa Xaa Xaa₅ Tyr Xa∑L Xaa Phe Xaa Xaa₇ Xaa₈ Xaa₉ Glu Xaa Xaaio Xaa Tyr Xaan Xaaι₂ Xaaι₃ Xaaj₄ Xaa Xaaι₅ Xaa Xaaι₆ Xaa, Xaa,₈ Ala Gly Xaa,₉ Xaa Xaa₂₀ Ser Xaa₂, Xaa Xaa Xaa₂₂ Xaa₂₃ Xaa Xaa (SEQ ID NO: 77) (ALFLD intervening sequence). Preferred residues for the selected variable residues in the ALFLD intervening sequence pattern are set forth in Table 2.

Table 2

Position Preferred Residues Position Preferred Residues

[0085] Preferred examples of more particular sequence patterns of the AFLD intervening sequence pattern include Tyr Xaa Leu Arg He Glu Leu Xaa Asp Trp Glu Gly Xaa Xaa Ala Tyr Ser Gin Tyr Asp Xaa Phe Xaa He Gly Asn Glu Xaa Gin Xaa Tyr Arg Leu Ser Leu Xaa Gly Xaa Thr Ala Gly Lys Xaa Ser Ser Leu Xaa Xaa Gin Gly Xaa Xaa (SEQ ID NO: 78), Tyr Xaa Leu Arg He Glu Leu Xaa Asp Trp Glu Gly Xaa Xaa Ala Tyr Ser Gin Tyr Glu Xaa Phe Xaa Leu Gly Ser Glu Xaa Gin Xaa Tyr Arg He Ser Leu Xaa Gly Xaa Thr Ala Gly Lys Xaa Ser Ser Leu Xaa Xaa Gin Gly Xaa Xaa (SEQ ID NO: 79), and Tyr Xaa Leu Arg Val Glu Leu Xaa Asp Trp Glu Gly Asn Xaa Ala Tyr Ser Gin Tyr Glu Xaa Phe Xaa Leu Gly Ser Glu Xaa Gin Asn Tyr Arg He Xaa Val Lys Gly Xaa Ser Gly Ser Ala Gly Arg Xaa Ser Ser Leu Xaa Xaa Xaa Gly Xaa Asp (SEQ ED NO: 80).

[0086] The angiopoietin-like FLD further desirably comprises a sequence positioned C- terminal to ALFLD sequence 2 of the sequence pattern Xaa Xaai Tyr Xaa(₃) Xaa₂ Xaa₃ Xaa₍₂₎ Lys Xaa Xaa₄ Gly He Xaa₅ Tip Xaa<₅ Tyr Xaa₇ Xaa₈ Xaa₉ Xaaio Xaai _I Xaa,₂ Leu Xaa Xaa,₃ Xaa Xaa,₄ Met Met Xaa_!5 Xaaι₆ Pro Xaa Xaaι Xaa (SEQ ID NO: 81) (ALFLD C- terminus sequence 1). Preferred residues for the selected variable residues in ALFLD C- terminus sequence 1 are set forth in Table 3.

Table 3

Position Preferred Residues Position Preferred Residues

[0087] Preferred examples of more particular ALFLD C-terminus sequence patterns of the broader ALFLD C-terminus sequence pattern include Xaa Phe Tyr Xaa Xaa Xaa Gin Asn Xaa Xaa Lys Xaa Asn Gly He Arg Tip His Tyr Phe Lys Phe Ser Ser Tyr Ala Leu Xaa Ala Xaa Arg Met Met He Glu Pro Xaa Asp Xaa (SEQ ID NO: 82), Xaa Tyr Tyr Xaa Xaa Xaa Gin Asn Xaa Xaa Lys Xaa Asn Gly He Arg Trp His Tyr Phe Lys Gly Pro Ser Tyr Ser Leu Xaa Ala Xaa Arg Met Met He Arg Pro Xaa Asp Xaa (SEQ ID NO: 83), and Met Tyr Tyr Xaa Xaa Xaa Gin Asn Xaa Xaa Lys Xaa Asn Gly He Xaa Trp His Tyr Phe Arg Gly Pro Xaa Tyr Ser Leu Xaa Ala Xaa Arg Met Met He Arg Pro Xaa Asp Phe (SEQ TD NO:84). [0088] The angiopoietin-like FLD can comprise any suitable combination of the above- described sequences, preferably within the pattern (from N-terminus to C-terminus) ALFLD N-terminus sequence 2, ALFLD N-terminus sequence 1, ALFLD sequence 1, ALFLD intervening sequence, ALFLD sequence 2, ALFLD C-terminus sequence. Thus, the angiopoietin-like FLD will desirably comprise a sequence of the overall pattern Phe Xaa Asp Cys Ala Xaa(₅₎ Gly Xaa(₄₎ Gly Xaa Tyr Thr Xaa₍₄₎ Asn Xaa(₅₎ Lys Xaa(₂₎ Cys Xaa₍₅₎ Gly Gly Xaa Xaa Trp Thr Xaa He Gin Xaa Arg Glu Xaa Gly Xaa(₃) Phe Gin Arg Xaa Trp Xaa Xaa Tyr Lys Xaa Gly Phe Gly Xaa(₄) Glu Xaa Trp Leu Gly Asn Glu Xaa<₅) Thr Xaa(₄) Tyr Xaa Leu Xaa(₂) Glu Leu Xaa Asp Trp Glu Gly Xaa(₆) Phe Xaa(₄) Glu Xaa₍₃₎ Tyr Xaa(io> Ala Gly Xaa ₃) Ser Xaa₍ ) Phe Ser Thr Xaa Asp Xaa Asp Asn Asp Xaa Cys Xaa Cys Lys Cys Xaa ₅₎ Gly Gly Trp Trp Phe Asp Ala Cys Gly Xaa Ser Asn Lys Asn Gly Xaa₍₂) Tyr Xaa ₎ Lys Xaa₍₂) Gly He Xaa Trp Xaa Tyr Xaa( ) Tyr Xaa Leu Xaa(₄) Met Met Xaa(₂) Pro Xaa₍₃) (SEQ ID NO: 85).

[0089] In other aspects, the fusion protein FLD exhibits desirably exhibits less than about 65% amino acid sequence identity to the Ang-1 FLD, less than about 60%> amino acid sequence identity to the Ang-1 FLD, or even less than about 45% amino acid sequence identity to the Ang-1 FLD (e.g., less than about 40% amino acid sequence identity to the Ang-1 FLD). For example, the fusion protein can comprise an FLD that exhibits at least about 60%) amino acid sequence identity (e.g., at least about 70%, at least about 80%, or even at least about 90% amino acid sequence identity) to the fibrinogen-like domain of NL4. The fusion protein FLD alternatively can comprise a sequence that exhibits at least about 45%) amino acid sequence identity (e.g., 50%> amino acid sequence identity), more preferably at least about 60% amino acid sequence identity, even more preferably at least about 75%) amino acid sequence identity (e.g., at least about 80%> or even at least about 90%> amino acid sequence identity) to the fibrinogen like domain of NLl, the fibrinogen-like domain of NL5, the fibrinogen-like domain of NL8, or any combination thereof. Alternatively, the fusion protein FLD can comprise an amino acid sequence that exhibits at least about 40%> amino acid sequence identity (e.g., at least about 50%, at least about 70%>, or even at least about 90% amino acid sequence identity) to the fibrinogen-like domain of zapol, the fibrinogen like domain of murine FDRG, or both. In another alternative, the fusion protein FLD can comprise a sequence that exhibits at least about 45%> amino acid sequence identity (e.g., at least about 50%), about 60%, about 70%>, about 80%, about 90%>, or even 100%) to the fibrinogen-like domain of NL3 (SEQ ED NO: 86) or the Hakata antigen (SEQ ID NO: 87) (as described in, e.g., Sugimoto et al., J. Biol. Chem. 273 (33), 20721-20727 (1998) and Yae, Biochim. Biophys. Acta 1078 (3), 369-376 (1991)). In yet another alternative, the fusion protein FLD can comprise an amino acid sequence that exhibits at least about 30%> amino acid sequence identity, preferably at least about 50% amino acid sequence identity, and more preferably at least about 70%> amino acid sequence identity (e.g., at least about 80%, about 90%, or about 95%> amino acid sequence identity) to the fibrinogen-like domain of FLS 139. The fusion protein FLD also can comprise a sequence that fulfills any combination of the above-described levels of identity to these non- angiopoietin ARF FLDS (e.g., in the case of a chimeric FLD peptide portion derived from two wild-type ARFs). [0090] A preferred set of fusion proteins, having an angiopoietin-like FLD, comprise (1) an amino acid sequence that exhibits at least about 50%>, preferably at least about 60%>, and even more preferably at least about 75%> (e.g., at least about 80%, about 90%o, or even 100%)) amino acid sequence identity to the fibrinogen-like domain of NLl, NL5, NL8, or any combination thereof, (2) an amino acid sequence that exhibits at least about 50%>, preferably at least about 60%>, and more preferably at least about 75%> (e.g., at least about 80%), about 90%), or even 100%>) amino acid sequence identity to the fibrinogen-like domain of NL4, (3) an amino acid sequence that exhibits at least about 35%, preferably at least about 40%, and more preferably at least about 75%> (e.g., at least about 80%>, at least about 90%), or even about 100%) amino acid sequence identity to the fibrinogen-like domain of zapol, murine FDRG, or both, (4) an amino acid sequence that exhibits at least about 40%>, preferably at least about 45%>, and more preferably at least about 75%> (e.g., at least about 80%, at least about 90%>, or even about 100%>) amino acid sequence identity to the fibrinogen-like domain of NL3 or the Hakata antigen, (5) an amino acid sequence that exhibits at least about 25%, preferably at least about 30%>, and more preferably at least about 75%o (e.g., at least about 80%>, at least about 90%>, or even about 100%) amino acid sequence identity to the fibrinogen-like domain of FLS 139, or (6) an amino acid sequence which fulfills any combination of (l)-(5), and a coiled-coil domain that exhibits at least about 30%), preferably at least about 35%>, and more preferably at least about 75%> (e.g., at least about 80%, at least about 90%>, or even about 100%ι) amino acid sequence identity to the Ang-1 CCD, Ang-2 CCD, Ang-2X CCD, Ang-3 CCD, Ang-4 CCD, or any combination thereof. A particularly preferred subset of this set of fusion proteins comprise a coiled coil domain that exhibits at least about 65%>, preferably at least about 75%>, and more preferably at least about 85%> amino acid sequence identity to the coiled-coil domain of at least one of Ang-1, Ang-2, or Ang-2X. An alternative preferred subset comprise a coiled-coil domain exhibits at least about 50%, more preferably at least about 75%>, and even more preferably at least about 85%> amino acid sequence identity to the coiled-coil domain of Ang-3, Ang-4, or both.

[0091] In aspects where the fusion protein includes an FLD that is homologous to the NLl FLD, NL5 FLD, or both wild-type FLDs, the fusion protein FLD desirably contains at least four cysteine residues, which preferably can form inter-FLD disulfide bonds. The NLl/NL5-like FLD further desirably comprises a first sequence of the pattern Gly Gly Tip Xaa₍₃₎ Gin Xaa Arg Xaa(₂) Gly Xaa Val Xaa Phe Xaa< ₎ Trp Xaa(₂) Tyr Xaa ₂) Gly Phe Gly Xaa<₆₎ Tip Leu Gly Xaa Glu Xaa(₄₎ Leu Xaa (SEQ ID NO: 88), which preferably is positioned in the N-terminal-most half of the FLD, and a sequence of the sequence pattern Tyr Xaa₍₂₎ Phe Xaa Leu Xaa(₂₎ Glu Xaa< ₎ Tyr Xaa Leu Xaa Leu Gly Xaa_(6-7) Asp Xaa(₄₎ His Xaa< ₎ Phe Xaa Thr Xaa Asp Xaa Asp Xaa₍₅₎ Cys Ala Xaa< ₎ Gly Xaa<₂) Trp Tyr Xaa(₂₎ Cys Xaa( > Ser Asn Leu Asn Gly (SEQ ID NO: 89), positioned C-terminal to first sequence. More preferably, the NLl/NL5-like FLD will comprise an amino acid sequence of the pattern Xaai Cys Xaa Gin Xaa₂ Xaa₃ His Xaa(₂) Ser Xaa Xaa₅ Tyr Xaa₍₂) Lys Pro Glu Asn Xaa₍o-i₎ Asn Xaa(_2-4) Met Gin Xaa₆ Xaa Cys Xaa₈ Xaa(₃₎ Asp Xaa₉ Gly Gly Tφ Xaaio Xaai ₁ Xaaι₂ Gin Xaan Arg Xaa Xaa_] Gly Xaaι₅ Val Asn Phe Xaai ₆ Xaan Asn Tφ Glu Xaa Tyr Xaaι₈ Xaa Gly Phe Gly Asn He Asp Xaaι Xaa₂o Xaa₂ι Tφ Leu Gly Xaa₂₂ Glu Asn He Tyr Xaa Leu Xaa₂₃ Asn Xaa₂₄ Xaa(o-i) Asn Tyr Lys Leu Xaa₂₅ Xaa₂₆ Xaa₂₇ Xaa₂₈ Glu Asp Xaa₂₉ Ser Xaa₃₀ Xaa Lys Val Xaa Ala Xaa_3! Tyr Xaa Ser Phe Arg Leu Glu Pro Glu Ser Glu Xaa Tyr Xaa₃₂ Leu Arg Leu Gly Xaa Tyr Xaa Tyr Xaa Xaa₃₃ Asn Ala Xaa₃₄

Asp Xaa₃₅ Xaa₍₂₎ Tφ His Xaa₃₆ Gly Lys Gin Phe Xaa₃₇ Thr Leu Asp Arg Asp Xaa Asp Xaa Tyr Xaa Gly Xaa₃₈ Cys Ala His Xaan.) Lys Gly Xaa₃₉ Xaa4o Tφ Tyr

Ala Cys Ala His Ser Asn Leu Asn Gly Xaa₄₂ Xaa ₃ Tyr Arg Gly Gly His Tyr Arg Ser Xaa Xaa₄₄ Gin Asp Gly Xaa_ι₅ Xaa Tφ Ala Glu Xaa Xaa₄₆ Gly Gly Ser Xaa₄ Xaa₄₈ Xaa.ι Xaa ₀ Xaa Xaa₅ι Xaa Met Met Xaa₅₂ (SEQ ID NO: 90). Preferred residues for the selected variable residues in this NL1/NL5 FLD are set forth in Table 4.

Table 4

Position Preferred Residues Position Preferred Residues

[0092] Where the fusion protein comprises a FLD homologous to the zapol FLD, the fusion protein FLD will desirably comprise a sequence of the sequence pattern Arg Asp Cys Gin Glu Leu Phe Gin Xaai Gly Glu Arg Xaa₂ Ser Gly Leu Phe Xaa₃ He Gin Pro Xaa₄ Gly Ser Pro Pro Phe Leu Val Asn Cys Xaa₅ Met Thr Ser Asp Gly Gly Tφ Thr Val He Gin Arg Arg Xaaβ Xaa₇ Tφ Glu Ala Tyr Lys Xaa₈ Gly Phe Gly Asp Pro Xaa₉ Gly Glu Phe Tφ Leu Gly Leu Glu Lys Xaaio His Ser He Xaai 1 Gly Xaan Arg Xaan Ser Xaan Leu Ala Val Gin Leu Xaaι₅ Asp Tφ Asp Gly Asn Ala Xaaι₆ Leu Leu Gin Phe Xaan Xaaι₈ His Leu Gly Gly Glu Asp Thr Ala Tyr Ser Leu Gin Leu Thr Xaaι₉ Pro Xaa₂₀ Ala Xaa₂ι Xaa₂₂ Leu Gly Ala Thr Thr Val Xaa₂₃ Pro Xaa₂₄ Gly Leu Ser Val Pro Phe Ser Thr Tφ Asp Gin Asp His Asp Leu Arg Xaa₂₅ Asp Xaa₂₆ Asn Cys Ala Lys Ser Leu Ser Gly Gly Tφ Tφ Phe Gly Thr Cys Ser His Ser Asn Leu Asn Gly Gin Tyr Phe Xaa₂₇ (SEQ ID NO: 91). Optionally, the zapol- like FLD sequence can include an additional sequence, N-terminal to this sequence, of the sequence pattern Ser He Pro Xaa₂₈ Gin Arg Gin Xaa₂₉ Xaa₃₀ Lys Lys Gly He Phe Tφ Lys Thr Tφ Xaa₃, Gly Arg Tyr Tyr Pro Leu Gin Ala Thr Thr Xaa₃₂ Leu He Gin (SEQ ID NO: 92). Preferred residues for the selected variable residues in these zapol -like FLD sequences are set forth in Table 5.

Table 5

Position Preferred Residues Position Preferred Residues

[0093] The fusion protein FLD can contain any combination of the above-described sequence portions, or peptide portions of such sequence portions. Any of the preferred selected amino acids in the above-described sequences also can desirably be replaced with a functionally homologous and/or weight-based homologous residue (e.g., where a preferred residue at a particular variable (i.e., Xaa) position is an Tφ or Phe residue, a Tyr residue at the indicated position, which also is aromatic, is more preferred than other residues, though less preferred than a Tφ or Phe residue). This principle applies to all amino acid sequence patterns discussed herein, unless otherwise stated. [0094] The fusion protein preferably also comprises a coiled-coil domain. The coiled- coil domain can be any suitable coiled-coil domain, preferred features of which, e.g., a substructure comprising an alpha helix comprising a primarily hydrophobic side, are discussed elsewhere herein. For example, the coiled-coil domain can be a parallel coiled- coil, a leucine zipper CCD, a TRAF2 CCD, an antiparallel coiled-coil, or an apoliproprotein A-I CCD. The CCD also can, for example, be a viral ecroprotein CCD or an epidermal growth factor CCD. ARF CCDs and homologs thereof are most preferred. [0095] Preferably, the coiled-coil domain increases the affinity of the fusion protein for the receptor that the first wild-type protein binds. "Affinity" is a measure of the degree of attractiveness between a ligand and a receptor. Affinity can be measured by any suitable technique. Typically, affinity is characterized by the receptor-ligand equilibrium dissociation constant, K_d, and/or the maximal specific binding, B_maχ, which can be empirically determined for a particular ligand-receptor pair by saturation analysis or other analytical techniques.

[0096] The K_d for a receptor-ligand interaction is the concentration at which 50%> of the receptor's ligand binding sites are occupied by a receptor-ligand interaction, and, as such, is a measure of affinity of receptor for a ligand. In general, the lower the K_d the higher the affinity, as only a small amount of ligand is required to fill half the receptor binding sites in such situations. The B_max is the maximum amount of ligand that can bind to the receptor, and can be used to indicate receptor concentration and/or number of receptors per cell. [0097] Saturation analysis techniques typically used to determine affinity are known in the art, and, as such, only briefly discussed herein. In a simple saturation analysis, cell membranes containing the receptor, or whole cells comprising the receptor, are incubated in a first set of assays with a range of concentrations of radioligand from a desired low point until an equilibrium or saturation point is reached. A second set of assays are performed under similar conditions except that a large excess of an unlabelled antagonist for the receptor (e.g., a wild-type ARF versus the fusion protein ligand) is added until an equilibrium is reached, such that the ligand is only permitted to bind to nonspecific sites (i.e., sites other than the receptor). The membranes or cells are then passed through a filter that retains the membranes and/or cells. The radioactivity (or other phenomena associated with the label, e.g., fluorescence intensity) is measured with a suitable detector. The bound radioactivity is then plotted as a function of concentration of the ligand to obtain two curves, representing total binding (obtained in the first set of assays), and nonspecific binding (obtained in the second set of assays), respectively. Nonspecific binding is subtracted from total binding to provide a saturation plot. The top of the saturation plot curve is experimentally equal to the number of binding sites in the system (the B_maχ), and the concentration at 50%> of this maximum is the K_d. While this technique is effective for determining the K and B_max for a ligand-receptor interaction, more typically a Scratchard plot (or Rosenthal plot), or even more typically a mathematical model consisting of a saturable hyperbolic component plus a linear non-saturable component, is used to provide an estimate of the K_d and B_max.

[0098] Alternatively or additionally, affinity can be measured by a competition assay. Briefly, in a competition assay, a single concentration of labeled ligand is used for every assay point (which typically are limited to concentrations below the ligand' s Ka value if known). The level of specific binding of the radioligand is then determined in the presence of a range of concentrations of other competing non-radioactive compounds to measure the potency with which they compete for the binding of the radioligand. The results of the competition assay provide a measure of the relative affinity of competing ligands for a receptor. The data for each competing ligand is usually fitted to a hyperbolic equation from which the IC₅o (i.e., the molar concentration of competing ligand (agonist or antagonist) that reduces the specific binding of a radioligand by 50%>) can be determined. The IC o can be correlated to K_d using the Cheng-Prusoff equation (Ki = IC₅₀ divided by 1+[L]/K_d, where L is the concentration of free ligand used in the assay) if the rate of ligand-receptor association (Ki) is known.

[0099] Other suitable techniques for measuring the affinity of receptor-ligand pairs are known in the art. Examples of such techniques and related principles are described in, e.g., Chang et al., Biochim Biophys Acta, 406(2), 294-303 (1976), Tallarida, Fed. Proc, 41(7), 2323-7 (1982), Levesque et al., Exp. Cell Res., 156(2), 558-62 (1985), Kenakin, Drugs, 40(5), 666-87 (1990), Smisterova et al., J. Pharm. Biomed. Anal, 12(6), 723-45 (1995), Van der Graaf et al., Int. J. Clin. Pharmacol. Ther., 35(10), 442-6 (1997), U.S. Patents 3,901,654, 4,251,360, and 5,639,5687, and references cited therein. Alternatively or additionally, the affinity of a ligand for a receptor can be measured by thermodynamic techniques, such as those described in Raffa, Life Sci., 44(4), 245-58 (1989), Pliska, J. Recept. Signal Transduct. Res., 17 (1-3), 495-510 (1999), and Pliska, J. Recept. Signal Transduct. Res., 17(4), 667 (1997). Additional related techniques are described in "Affinity Techniques Enzyme Purification," Meth. Enzymol, 34 (Jakoby and Wilcheck eds. 1974). [0100] Alternatively, or preferably additionally, the fusion protein CCD promotes formation of a multimer comprising the fusion protein. The multimer can be any collection of two or more non-covalently linked peptides. Thus, the multimer can be a dimer, although typically and preferably the multimer is a trimer or a higher ordered multimer (i.e., a collection of four or more non-covalently linked peptides).

[0101] Multimer formation can be determined by any suitable technique. Several suitable approaches to determining multimer formation are known in the art. A simple technique for assessing multimerization comprises subjecting a first portion of a composition comprising the putative multimer to size-exclusion chromatography, under conditions where the multimer will not be denatured, to determine the weight of the multimer. Another portion of the composition can be subjected to denaturing SDS-PAGE. If a multimer is formed the weights indicated in the two assays will be different, as the SDS-PAGE gel will exhibit a bond reflecting the weight of the monomeric fusion protein, rather than a multimer. Alternatively, two Western blots, one performed under denaturing conditions and the other under non-denaturing conditions can be performed on the multimer containing composition, if an antibody exhibits binding for both the multimer and the monomer. Recently, fluorescent microscopy, mass spectrometry, and light scattering techniques also have been used to determine multimerization. Alternatively, multimer- specific antibody binding assays can be used to assess multimerization. [0102] The fusion protein CCD can be any suitable size. The CCD will usually be at least about 20, typically about 25-500, more typically about 30-300, even more typically about 30-220 (e.g., at least about 190-210) amino acid residues in length. [0103] The fusion protein CCD desirably exhibits at least about 25% amino acid sequence identity, preferably at least about 50%> sequence identity, more preferably at least about 75%o amino acid sequence identity, and even more preferably at least about 90%> amino acid sequence identity to the coiled-coil domain of a second wild-type protein, where the second wild-type protein is typically and preferably an ARF, and most preferably is an ARF that promotes angiogenesis in vivo.

[0104] In instances where the fusion protein comprises a CCD that is homologous to an angiopoietin, the fusion protein CCD will desirably comprise the sequence Glu Val Lys Leu Leu Arg Lys Glu Ser Arg Asn Met Asn Ser Arg Val Thr Gin Leu Tyr Met Gin Leu Leu His Glu He He Arg Lys Arg Asp Asn (SEQ ID NO: 93). The angiopoietin-like CCD desirably further comprises an amino acid sequence of the sequence pattern Val Asp Gly Xaa He Val Xaai Glu Val Lys Leu Leu Arg Lys Glu Ser Arg Asn Met Asn Ser Arg Val Thr Gin Leu Tyr Met Gin Leu Leu His Glu He He Arg Lys Arg Asp Asn Xaa Leu Glu Leu Ser Gin Leu Glu Asn Xaa₂ He Leu Asn (SEQ TD NO: 94), where Xaai represents a hydrophobic residue and Xaa₂ represents a basic residue. More particularly still, the angiopoietin-like CCD desirably comprises a sequence of the sequence pattern Leu Glu Xaa Leu Xaa Xaa Xaa Leu Xaa Xaa Gin Lys Arg Xaa He Xaa Xaa Leu Gin Xaa Xaa Val Xaa Val Asp Gly Xaa He Val Xaa Glu Val Lys Leu Leu Arg Lys Glu Ser Arg Asn Met Asn Ser Arg Val Thr Gin Leu Tyr Met Gin Leu Leu His Glu He He Arg Lys Arg Asp Asn Xaa Leu Glu Leu Ser Gin Leu Glu Asn Xaa He Leu Asn Xaa Thr Xaa Xaa Met Leu Xaa Xaa Ala Xaa Xaa Tyr Xaa Xaa Leu Glu Xaa Lys Tyr Xaa Xaa Leu (SEQ ID NO: 95).

[0105] Alternatively, and desirably, where the fusion protein FLD exhibits higher identity to an ARF FLD other than the Ang-1 FLD, Ang-2 FLD, Ang-3 FLD, or Ang-4 FLD, and optionally higher identity to a non-angiopoietin ARF FLD than any angiopoietin FLD, the fusion protein CCD will desirably comprise a sequence of the sequence pattern Tyr Xaa Xaa Asn Ala Xaa Gin Arg Asp Ala Pro Xaa<ι_-2) Glu Xaa Asp Xaa Ser Xaa Gin Xaa Leu Gin Xaa Leu Glu Xaa Xaa Met Glu Asn Xaa Thr Gin Tφ Leu Xaa Lys Leu Glu Asn Tyr He Xaa Xaa Asn Met Lys Xaa Glu Met Xaa Xaa He Gin Gin Asn Ala Val Gin Asn Xaa Thr Ala Xaa Met Xaa Glu He Gly Thr Xaa Leu Leu Xaa Gin Thr Ala Glu Gin (SEQ ID NO: 96).

[0106] In another aspect, and desirably where the fusion protein FLD exhibits higher identity to a wild-type ARF FLD other than the Ang-3 or Ang-4 FLD, preferably other than the Ang-1 FLD, Ang-2 FLD, Ang-3 FLD, and Ang-4 FLD, and optionally higher identity to any ARF other than to any angiopoietin, the fusion protein CCD will desirably include a sequence of the sequence pattern Thr Asn Lys Leu Glu Xaa Gin Xaa Leu Xaa Gin Xaa Xaa Xaa Leu Gin Xaa Leu Gin Gly Xaa Asn Xaa Ala Leu Glu Xaa Arg Leu Gin Ala Leu Glu Xaa Xaa Xaa Gin Xaa Xaa Leu Xaa Ser Xaa Xaa Xaa Lys Xaa Xaa Xaa Leu Xaa Xaa Xaa Leu (SEQ ID NO: 97). More preferably, such a fusion protein CCD will comprise a sequence of the sequence pattern Thr Asn Lys Leu Glu Xaai Gin Xaa₂ Leu Xaa₃ Gin Xaa₄ Xaa₅ Xaa_ό Leu Gin Xaa Leu Gin Gly Xaa₈ Asn Xaa₉ Ala Leu Glu Xaaio Arg Leu Gin Ala Leu Glu (SEQ ID NO: 98), where Xaai preferably represents an Asn or Arg residue, Xaa₂ preferably represents a Leu or Met residue, Xaa₃ preferably represents a Leu or Met residue,

an Arg or a Ser residue, Xaa₅ preferably represents an Arg or a Gin residue, Xaa$ preferably represents a Glu or a Lys residue, Xaa₇ preferably represents an Arg or Gin residue, Xaa₈ preferably represents an Arg or a Gin residue, Xaa₉ preferably represents an Arg or a Ser residue, and Xaaio preferably represents a Lys or Thr residue. [0107] The first wild-type protein and second wild-type protein are different. Preferably, the first wild-type protein and second wild-type protein exhibit less than about 90% overall amino acid sequence identity, preferably less than about 70%> overall amino acid sequence identity, even more preferably less than about 50%> overall amino acid sequence identity, and optimally less than about 30%> overall amino acid sequence identity to each other (e.g., less than about 20%>, less than about 15%>, or even less than about 10%> amino acid sequence identity to each other). Advantageously, at least one, if not both, of the first wild-type protein and second wild-type protein exhibit a higher level of sequence identity to a wild-type ARF other than Ang-1, Ang-2, Ang-3, and Ang-4, or another angiopoietin. Thus, although the first protein and second protein typically exhibit at least about 30%) overall amino acid sequence identity to Ang-1, the first wild-type protein, the second wild-type protein, or both proteins, are not Ang-1, Ang-2, Ang-3, or Ang-4. [0108] The first wild-type protein can be any suitable wild-type ARF. For example, the first wild-type protein can be an angiopoietin (e.g., Ang-1, Ang-2 (SEQ ID NO: 99), Ang-3 (SEQ ID NO: 100), or Ang-4 (SEQ ID NO: 101)), or a non-angiopoietin ARF, such as a muscle ALGF (SEQ ID NO: 102) or liver ALGF (SEQ ID NO: 103) (as described in, e.g., International Patent Application WO 99/67382), NLl (SEQ ID NO: 104), NL2 (SEQ ID NO: 105), NL3 (SEQ ID NO: 106), NL4 (SEQ ID NO: 107) (which appears to be the same as CDT6, described in International Patent Application WO 01/29085 (SEQ ID NO: 108)), NL5 (SEQ ID NO: 109), NL6 (SEQ ID NO: 110), NL8 (SEQ ID NO: 111), zapol (SEQ ID NO: 112), FARP (SEQ ID NO: 113) and HFARP (SEQ ID NO: 114) (as described in, e.g., Lee et al., Mol. Cells, 11(1), 100-04 (2001), and Kim et al., Biochem. J., 346 (part 3), 603- 10 (2000)), the angiopoietin-like factors described by Peek et al., Invest. Opthalmol Vis. Sci., 39(10), 1782-88 (1998) (e.g., Genbank Accession No. NP 066969), angiopoietin- related protein 5 (SEQ ID NO: 115), angiopoietin related protein 2 (GenBank Accession No. NP 036053), angiopoietin-like 1 precursor (GenBank Accession No. NP 004664), the factors recorded at GenBank Accession Nos. BAB27951, XP_049490, BAB31076, AAK11499, P14448, NP_082609, AK018113, AF159049_1, AB056477, XP_041622, AF169312_1, AF278699_1, AAC97965, NP_065606, AAK83349, AF110520, AF278699, AF379604, and AK019860, fibrinogen/angiopoietin related protein (SEQ ID NO: 116) (as described in Kim et al., Biochem. J, 346, 603-10 (2000)), PPAR gamma (SEQ ID NO: 117) , Ang-2A (SEQ ID NO: 118), Ang-2B (SEQ ID NO: 119), and Ang-2C (SEQ ID NO: 120) (as described in, e.g., Mezquita et al., Biochem. Biophys. Res. Commun., 275(2), 643-51 (2000)), Ang2(443) (SEQ ID NO: 121) (as described in, e.g., Kim et al., J. Biol. Chem., 275(24), 18550-56 (2000)), Ang-6 (SEQ ID NO: 122) (as described in, e.g., International Patent Application WO 01/02429), Ang-7 (SEQ ID NO: 123) (as described in, e.g., International Patent Application WO 01/02434), the angiopoietin related proteins described in Yoon et al., Mol. Cell. Biol, 20(14), 5343-9 (2000), a FAIF (as described in, e.g., Kerten et al., J. Biol. Chem., 275(37), 28488-93 (2000)), the angiopoietin related factors described in, or encoded by nucleotide sequences described in, International Patent Applications WO 98/05779, WO 99/15653, WO 99/32515, WO 99/32639, WO 99/40193, WO 99/45135, WO 99/54353 (e.g., sequences 16 and 125), WO 99/62956, WO 99/62925, WO 00/05369 (e.g., the Ang-3 described therein (SEQ ID NO: 124)), WO 00/05241, WO 00/37642, WO 00/52167, WO 00/59938, WO 00/73452 (e.g., sequence 48), WO 01/04264 (e.g., sequence 27), WO 01/05825 (e.g., sequences 1, 3, 5, 11, 13, 14, 45, and 47), WO 01/05972 (e.g., PR0356 (SEQ ID NO: 125)), WO 01/14550 (e.g., PG-3 (SEQ ID NO: 126)), WO 01/36684 (e.g., sequence 40), and variants of such angiopoietins or ARFs (as described in, e.g., Kim et al., J. Biol Chem., 274, 26523-28 (1999), U.S. Patents 5,521,073, 5,643,755, 5,877,289, 5,879,672, 5,972,338, 6,030,831, 6,057,435, and 6,074,873, and International Patent Applications WO 96/11269, WO 96/31598, WO 99/15653, WO 99/32515, WO 99/45135, WO 99/67382, and WO 01/05825). Other suitable ARFs include the Hakata antigen, PR01346 (also known as NL7) (SEQ ID NO: 127), PR0179 (SEQ ID NO: 128), PR0196 (SEQ ID NO: 129), SECP3 (SEQ ID NO: 130) (as described in International Patent Application WO 01/05971), z-alpha5 (as described in, e.g., International Patent Application WO 99/55869), and murine FDRG (SEQ ID NO: 131). In some situations, it is preferred that the first wild-type protein is NLl or NL5. In other situations, it is preferred that the first wild-type protein is an angiopoietin, and, more preferably, Ang-1. Other preferred ARFs include the factors recorded under GenBank Accession Nos. AI972163 (SEQ ID NO: 132), AI341252 (SEQ NO: 133), and AW293408 (SEQ ID NO: 134). Factors encoded by the sequences recorded under GenBank Accession No. AA048880 also can be suitable, as can the factors recorded at GenBank Accession Nos. AA122061, AB056477, AC092300, AF153606, AF169312, M62290, T57280, T50719, T35448, T11442, and W77823. [0109] The first wild-type protein alternatively can be a non- ARF protein. For example, the first wild-type protein could be a human fibrinogen chain, a tenascin (e.g., a human tenascin (SEQ ID NO: 135), or a ficolin (e.g., human ficolin (SEQ ID NO: 136)). [0110] The second wild-type protein can be any suitable protein comprising a CCD (e.g., an epidermal growth factor receptor), but preferably is an ARF (such as any of the aforementioned first wild-type proteins), and most preferably an ARF that promotes angiogenesis in vivo. In some aspects, angiopoietin ARFs are preferred, and particularly Ang-1. In other aspects, the second wild-type protein is desirably NLl, FLS 139, NL3, the Hakata antigen, NL4, NL5, NL8, zapol, or murine FDRG. Where the first wild-type protein is KAP or an angiopoietin, preferred second wild-type proteins include NLl and NL5.

[0111] The first wild-type protein or second wild-type protein also can be the novel angiopoietin homolog Ang-2X (SEQ ID NO: 137), which is described further herein. Where Ang-2X is the first wild-type protein, the second wild-type protein preferably is Ang-1, Ang-2, Ang-3, or Ang-4, most preferably Ang-1. Alternatively, in such fusion proteins, the second wild-type protein desirably preferably is NLl or NL5. Where Ang-2X is the second wild-type protein, the first wild-type protein preferably is Ang-1, Ang-2, NLl, or NL5, and most preferably Ang-1.

[0112] A desirable quality of the fusion proteins of the invention is their ability to reduce blood vessel leakage (permeability) when administered or produced in vivo. The reduction in blood vessel permeability can be measured relative to the level of permeability in selected blood vessels prior to administration or production of the fusion protein, or, more typically, at a time prior to some selected time after administration or sufficient production of the fusion protein (e.g., at least about 6 hours, about 12 hours, about 1 day, about 2 days, about 3 days, about one week, or longer (e.g., about 2 weeks) after administration of the fusion protein). Alternatively or additionally, the reduction in permeability can be measured by comparing the level of blood vessel permeability exhibited in blood vessels receiving administration of a VEGF or a polynucleotide encoding a VEGF in the absence of the fusion protein, with blood vessels receiving a substantially identical amount of VEGF in the presence of an effective amount of fusion protein for a similar time period. Preferably, blood vessels affected by the administration or production of the fusion protein exhibit about 90% or less, more preferably about 75 % or less, and even more preferably about 50%> or less (e.g., about 25%> or less) of the vascular permeability exhibited by blood vessels induced or promoted by administration of a VEGF or VEGF-encoding polynucleotide without administration or production of the fusion protein, at about 1 day, about 2 days, about 3 days, about one week, about two weeks, or longer, after administration or production of the fusion protein. The fusion protein of the invention will desirably exhibit such levels of reduced permeability even when co-administered with, or administered closely after (e.g., within about 6 hours, 12 hours, 1 day, or 3 days) administration of a VEGF or VEGF-encoding polynucleotide.

[0113] Blood vessel permeability is the rate at which macromolecules move through a blood vessel wall and outside of the blood vessel. Blood vessel permeability can be determined by any suitable technique, and several techniques are known in the art. One measure of permeability is the lack of force that resists the movement through the blood vessel wall and out of the blood vessel. A simple technique for determining blood vessel permeability involves the application of a labeled, typically radiolabeled, protein that is normally carried in the blood (e.g., FITC labeled, or I labeled, human serum albumin (HSA)), to a blood vessel for a period of time followed by measuring the amount of the labeled protein that enters the tissue surrounding the blood vessel. More sophisticated techniques for assessing blood vessel formation involve MRI, Doppler imaging, and computed tomography. For example, first-pass perfusion imaging MRI can be suitable for assessing vascular perfusion in cardiac tissue in vivo. See, e.g., Dranga et al., J. Mag. Resonance Imaging, 7 664-68 (1997) and Wilke et al., Radiology, 204, 373-84 (1997). Scanning electron microscopy and/or light microscopy techniques also can be used to assess permeability. See, e.g., Grunt et al., Scanning Electron Microscopy, part 2, 575-89 (1986). Other techniques known in the art for assessing blood vessel permeability and related principles are discussed in, e.g., Thurston et al., Nat. Med., 6(4), 460-63 (2000), Bates et al., Microcirculation, 6, 83-96 (1999), Thurston et al., Science, 286, 2511-14 (1999), Cox et al., J. Surg. Res., 83(1), 19-26 (1999), Carter et al., Biophys. J, 74(4), 2121-28 (1998), Mori et al., Neurol Med. Chir., 38(11), 697-703 (1998), Kendall et al., Exp. Physiol., 80(3), 359-72 (1995), Adamson et al., Microcirculation, 1(4), 251-65 (1994), Yuan et al., Microvasc. Res., 45(3), 269-89 (1993), Olson et al., J. Appl Physiol, 70(3), 1085-96 (1991), Shibata et al., Jpn. J. Physiol, 41(5), 725-34 (1991), Tompkins, Am. J. Physiol, 260, HI 194-1204 (1991), and Kern et al., Am. J. Physiol, 245(2), H229-36 (1983).

[0114] Typically and preferably, in vivo production or administration of the fusion protein also or alternatively is blood vessel maturity (e.g., results in more blood vessels which exhibit greater density and/or structural similarity to mature mammalian blood vessels). The promotion of blood vessel maturity can be determined by comparing selected blood vessels before and after a suitable time after administration or production of the fusion protein (e.g., about 1 day, 2 days, 3 days, one week, or longer) and/or by comparing selected blood vessels that result from the administration of a VEGF or a polynucleotide encoding a VEGF in the absence of the fusion protein versus substantially identical blood vessels that also receive an effective amount of fusion protein (either by production or administration). Particular examples of maturation events include pericyte coating of forming blood vessels and arterialization of newly formed vessels. Generally, immature blood vessels are poorly associated with underlying basement membranes and/or other vessels (as described in, e.g., Dumont et al., Genes Dev. 8, 1897-1909 (1994) and Sato et al., Nature, 376, 10-14 (1995)). Blood vessel maturation can be assessed using any suitable standard. For example, maturity can be observationally assessed by assessing vessel shape, density, luminal regularity, and vessel opening size (as described in, e.g., Bloch et al., FASEB J, 14(5), 2373-76 (2000)). Maturity also can be assessed by evaluating signal intensity changes in response to hyperoxia (elevated oxygen) and hypercapnia (elevated carbon dioxide), for example by measuring physiological vasodilatory response to carbon dioxide (as described in, e.g., Gilead and Neeman, Neoplasia, 1(3), 226-30 (1999)), smooth muscle plasticity, the association of blood vessels with basement membrane or each other, or smooth muscle and non-muscle vascular associated myosin isoform distribution (as described in, e.g., Pauletto et al., Am. J. Hypertens., 7, 661-74 (1994)). Preferably, maturation is determined by assessing recruitment of pericytes to the vasculature, pericyte coating of new vessels, association between the vascular tube and the mural cells, or any combination thereof (as discussed in, e.g., Darland et al., J. Clin. Invest., 103(2), 157-58 (1999), which can be quantified, e.g., by using the microvessel maturation index (MMI) (see, e.g., Goede et al., Lab. Invest., 78(11), 1385-94 (1998)). The in vivo administration of the angiogenic fusion protein to a mammalian host typically will result in a higher number or greater concentration of smooth muscle cells, pericytes, mural cells, total endothelial cells, or any combination thereof, than blood vessels resulting from administration of a protein limited essentially to the first peptide portion. The increase in number of such cells can be detected using techniques known in the art. Other techniques and principles related to assessing blood vessel maturation are discussed in Koffet et al., FASEB J., 15(2), 447-57 (2001), as well as several other references cited elsewhere herein. [0115] In a related sense, administration or production of the fusion protein desirably promotes blood vessel remodeling. "Blood vessel remodeling" includes any type of vascular restructuring not associated with maturation, although the events often typically overlap and/or occur simultaneously in living systems. Typical types of blood vessel remodeling events include increase in vascular mass, vessel wall thickening, vessel enlargement or dilation, alteration in capillary density, vascular bed modification, change in vessel tone, or combinations thereof. Blood vessel remodeling can be assessed through stress state and pressure testing, MRI (e.g., as described in Nikol et al., Angiology, 49(4), 251-58 (1998)), in vivo ultrasound imaging or histologic analysis (as described in, e.g., Fung et al., J. Biomech. Eng., 115(4B), 453-59 (1993)), and techniques otherwise used to assess angiogenesis (e.g., gradient echo testing). The promotion of blood vessel remodeling can be measured by applying such techniques to selected blood vessels some time before and after a suitable period (e.g., about 1 day, about 2 days, about 3 days, about 1 week, or longer) after administration or production of the fusion protein. ARF-associated blood vessel remodeling is further described in e.g., Goldman- Wohl et al., Mol. Hum. Reprod., tfft), 81-87 (2000).

[0116] Administration or production of the fusion protein will preferably promote angiogenesis. Tests for assessing whether angiogenesis promoted are discussed above. The assessment of whether administration or production of the fusion protein promotes angiogenesis can be assessed with respect to the tissue, organ, or host of interest, by applying such techniques before and after a suitable time after administration or production of an effective amount of the fusion protein (e.g., about 1 day, about 2 days, about 3 days, about 1 week, or longer).

[0117] Preferably, the fusion protein, when administered or produced in a mammalian host, will exhibit an in vivo half-life which is at least about 110%>, preferably at least about 125%), and more preferably at least about 150% (e.g., at least about 200%, about 250%, about 300%), about 500%>, or more) as long as the half-life of Ang-1. Half-life marks a point where the total amount of fusion protein administered to, or produced in, the tissue or host exhibits a reduction in biological activity of about 50%. The biological activity can be any of the aforementioned biological activities associated with the fusion protein (e.g., promotion of angiogenesis, more particularly reduction of vessel leakage, promotion of vascular maturity, and/or promotion of vascular remodeling). Preferably, the fusion proteins will exhibit a half-life of at least three minutes, desirably at least about four minutes, more preferably at least five minutes, and even more preferably at least ten minutes (e.g., at least about 15, at least about 20, at least about 30, at least about 60, at least about 90, at least about 180, at least about 360, at least about 720 minutes, or even longer) in a mammalian host upon administration (or production). Extended half-life can be obtained by the removal of proteolytically susceptible residues from the ARF peptide portions. For example, the conversion of Cys_{2 5} in wild-type Ang-1 to Ser₂₆₅, can promote resistance to proteolytic degradation in a fusion protein comprising a modified Ang-1 peptide portion where the residue corresponding to Cys₂₆₅ is so modified. In Ang-1, this cysteine is of a 20-residue sequence positioned between the Ang-1 CCD and Ang-1 FLD. The exclusion of this sequence (or at least a portion containing Cys₂₆₅) from the fusion protein, or substitution of this sequence with an alternative "middle" domain sequence (e.g., Val Asn Leu Ser Thr Lys Glu Gly Val Leu Leu Lys Gly Lys Arg Glu Glu Glu Lys (SEQ ID NO: 138)) is expected to result in better half-life than that exhibited in Ang-1. Half-life also can be extended by the addition of a heterologous peptide portion to the fusion protein (e.g., an IgG domain) (as described in, e.g., International Patent Applications WO 98/22577 and WO 00/24782 and/or U.S. Patents 6,121,022 and 6,277,375), by the insertion of sequences that promote N-terminal glycosylation of the fusion protein (as described in, e.g., U.S. Patent 5,218,092), or by administering the fusion protein with a non-proteinaceous polymer, such as those described elsewhere herein.

[0118] The fusion protein peptide portions can be associated in any suitable manner. Typically and preferably, the first and second peptide portions will be covalently associated (e.g., by means of a peptide or disulfide bond). The peptide portions can be directly fused (e.g., the C-terminus of the fusion protein CCD can be fused to the N-terminus of the fusion protein FLD through a peptide bond between the two portions). The fusion protein can include any suitable number of modified bonds, e.g., isosteres, within or between the peptide portions. Alternatively, the fusion protein can include a peptide linker between the peptide portions that includes one or more amino acid sequences not forming part of the biologically active peptide portions. Any suitable peptide linker can be used. The linker can be any suitable size. Typically, the linker will be less than about 30 amino acid residues, preferably less than about 20 amino acid residues, and more preferably about 10 or less amino acid residues. Typically the linker will predominantly consist of neutral amino acid residues. Suitable linkers are generally described in, e.g., U.S. Patents 5,990,275, 6,010,883, 6,197,946, and European Patent Application 0 035 384. More particular examples of suitable linkers are provided further herein.

[0119] The linker can include one or more cleavage sites to promote separation of linked peptide portions under specific conditions (e.g., exposure to certain proteolytic enzymes) if desired. Examples of such cleavage sites include the He Glu Gly Arg linker sequence (SEQ ID NO: 139), which is cleaved by Factor X_a protease. Other sites can include sequences which are cleaved by, for example, trypsin, enterokinase, collagenase, and thrombin. Alternatively, the cleavage site in the linker sequence can be a site capable of being cleaved upon exposure to a selected chemical or chemical state, e.g., cyanogen bromide, hydroxylamine, or low pH. Additional examples of suitable cleavable linkers are provided in U.S. Patent 4,719,326.

[0120] Typically and preferably, the first and second peptide portions will be directly fused, or separated by a non-cleavable linker of less than about 10 amino acid residues (e.g., 1-5 amino acid residues).

[0121] Linkers that reduce the immunogenicity of the fusion protein in a mammalian host, preferably a human host, are preferred. Any linker that reduces the immune response of the intended recipient of the fusion protein is suitable in this respect. Preferably, a flexible linker, which does not interfere with the tertiary structure of the fusion protein ARF peptide portions, and preferably which permits the fusion protein to exhibit an overall tertiary structure similar to at least the CCD FLD, and any intervening portions of a wild- type ARF, is used. By not interfering with the tertiary structure of the peptide portions and/or the fusion protein overall, the flexible linker will not configure the fusion protein in a manner such that foreign epitopes are presented to the target's immune system. The flexible linker is desirably immuno logically inert in the host system, and addition of it to the fusion protein desirably does not produce epitopes resulting in a strong immunological host response against the fusion protein, and desirably eliminates any sequences that might result in such immune response from the otherwise direct fusion of the first and second peptide portions. Any flexible linker can be used. Typically and preferably the flexible Gly Ser₃ linker or derivative thereof (i.e., a linker comprising the sequence Gly Gly Gly Gly Ser Ser Ser (SEQ ID NO: 140) is used in such fusion proteins. The use of such flexible linkers is described in, e.g., McCafferty et al., Nature, 348, 552-554 (1990), Huston et al., Proc. Natl Acad. Sci. USA, 85, 5879-5883 (1988), Glockshuber et al., Biochemistry, 29, 1362-1367 (1990), and Cheadle et al., Molecular Immunol, 29, 21-30 (1992). Other glycine-rich flexible linkers also can be suitable, such as the Pro Gly He Ser Gly Gly Gly Gly Gly linker (SEQ ID NO: 141), described in Guan et al., Anal. Biochem., 192(2), 262-67 (1991), the Gly Gly Gly Gly Ser Gly Gly Gly Gly Ser Gly Gly Gly Gly Ser linker (SEQ ID NO: 142), described in Huston et al., Proc. Natl. Acad. Sci. USA, 85, 5879-5883 (1988), and the Glu Gly Lys Ser Ser Gly Ser Gly Ser Glu Lys Glu Phe linker (SEQ ID NO: 143), described in Bird et al., Science, 242, 423-26 (1988). Other suitable flexible linkers include the immunoglobulin hinge linkers (as described in, e.g., U.S. Patents 5,672,683 and 6,165,476), and helical peptide linkers (as described in, e.g., U.S. Patent 6,132,992). [0122] Alternatively, where the first and second peptide portions are directly fused, the fusion protein can be designed such that the intersection ("fusion point") of the FLD and CCD does not generate a sequence which results in a strong immune response against the fusion protein (e.g., as compared to the direct fusion of the corresponding (i.e., most homologous/identical) wild-type ARF peptide portions). Such determinations can be made by using algorithms that identify MHC class I and MHC class II epitope sequences (preferably through the use of bioinformatics software incoφorating such algorithms or through the use of databases which provide listings of such epitopes identified with such algorithms). Any suitable algorithm, database, or program can be used. Examples of such algorithms, programs, and databases include the EPIMER/EPIMAX algorithm developed at the Brown University School of Medicine, the BONSAI algorithm developed at Stanford University, the TEPITOPE algorithm, the Zycos, Inc. "EPIQUEST" database, the SYFPEITHI program (which applies the algorithm of Rammensee et al.), the MAPPP program (available at http://www.mpiib-berlin.mpg.de/MAPPP/addquery.html), and the BEV1AS program (available at http://bimas.dcrt.nih.gov/molbio/hla_bind/), which are variously described in, e.g., Altuvia et al., Mol Immunol, 31, 1-19 (1994), Brusic et al., Nuc Acids Res., 22, 3663-3665 (1994), Hammer et al., J. Exp. Med, 180, 2353-2358 (1994), Parker et al., J. Immunol, 152, 163-175 (1994), Sturniolo et al., Adv. Immunol, 66, 67-100 (1997), and Cunha-Neto, Braz. J. Med. Biol. Res., 32(2), 199-205 (1999). The amino acid sequence which would result upon the production or production of the fusion protein of interest, particularly the area where the first and second peptide portions intersect and surrounding region (typically about 15 residues or less, more typically about 10 amino acid residues or less, in both directions from the fusion point), can be inputted into such a program, referenced against such databases, or analyzed by similar technique, to determine whether the sequence would result in an undesired host immune response (e.g., formation of a complex with an MHC class I molecule, MHC class II molecule, or both). The polynucleotide sequences encoding the termini of the CCD and/or FLD can be deleted or modified accordingly to obtain a fused sequence that will not induce such an immune response. The insertion of a linker also can prevent such a sequence from being formed. [0123] Thus, the invention provides a fusion protein wherein the CCD, FLD, or both lack one or more amino acid residues corresponding to residues in their wild-type counteφarts near the fusion point (typically within about 20 amino acids or less, more typically within about 10 amino acids or less of the fusion point). Such that administration or production of the fusion protein results in a lower level of host immune response than occurs against an unmodified fusion protein (i.e., a fusion protein comprising a wild-type ARF FLD directly fused to a wild-type ARF CCD). The residues that would result in the immunologically-undesirable amino acid sequence can be removed either through deletion or through non-immunologically equivalent substitutions (which typically will be non- homologous in nature). Typically and preferably, about 15 or less, at the polynucleotide level and more typically about 5 or less of the residues will require deletion or substitution. In some fusion proteins, even a single deletion addition, or substitution will result in the desired reduction in the immunogenicity of the sequence formed by the fusion of the first peptide portion and second peptide portion. By "corresponding" in this context, it is meant that the deleted/substituted residue is homologous to, or more typically identical to, a sequence occurring in the wild-type peptide, and would align with the residue in the peptide portion's wild-type counteφart in an optimal alignment if present in the fusion protein. Similar techniques can be applied to fusion proteins that contain a linker, if necessary. [0124] By using the above-described techniques to modify a wild-type FLD-encoding and/or CCD-encoding sequence, a fusion protein comprising a FLD directly fused to a CCD, where the amino acid sequence within about 10 residues, preferably within about 20 residues, and more preferably within about 30 amino acid residues of the FLD/CCD junction will not form a complex with an MHC Class I molecule, an MHC Class II molecule, or both, in vivo, can be obtained. Alternatively, using the above-described techniques, a fusion protein comprising a linker positioned between the fibrinogen-like domain and the coiled-coil domain can be prepared, such that the amino acid sequence within about 15 amino acid residues, preferably about 20 amino acid residues, and more preferably within about 30 amino acid residues of the linker does not form a complex with an MHC Class I molecule, an MHC Class II molecule, or both, when the fusion protein is produced in, or administered to, a mammalian host. "Within about 10 amino acid residues of the linker" refers to the sequence about 5 residues N-terminal of the center of the linker sequence and about 5 amino acid residues C-terminal to the center of the linker sequence. [0125] Immunogenicity testing of the fusion protein or polynucleotides of the invention also can be assessed using any suitable immunogenicity model prior to administration to the target, particularly where the target of administration is a human, to determine whether the area of fusion will exhibit an acceptable level of immunogenicity upon in vivo administration or produced. Reducing exposure to epitope sequences identified by the above-described techniques also can reduce or eliminate the immune response to the fusion protein of the invention. For example, identified epitope sequences or adjacent sequences can be PEGylated (e.g., by insertion of lysine residues to promote PEG attachment) to shield identified epitopes from exposure. Other techniques for reducing immunogenicity of the fusion protein of the invention can be used in association with the administration of the fusion protein include the techniques provided in U.S. Patent 6,093,699. [0126] Where a linker is incoφorated into the fusion protein, the presence of the linker preferably does not impede the biological activity of the FLD, CCD, and more preferably of either domain, and more desirably enhances the biological activity of the separate peptide portions over a direct fusion of the peptide portions (e.g., the promotion of vascular maturation and/or reduction of plasma leakage). Examples of techniques used to assess the effect of linker sequences on the biological activities of fusion proteins are described in, e.g., Newton et al., Biochemistry, 35, 545-553 (1995), which can be modified as appropriate for the fusion proteins of the invention (e.g., using the biological assays described elsewhere herein). It will typically be advantageous for the linker to permit the first ARF peptide portion, second ARF peptide portion, or both ARF peptide portions, to exhibit a secondary and/or tertiary structure similar to that of their native peptide counteφarts, which can be assessed using techniques provided herein or similar techniques known in the art. [0127] The in vivo administration or production of fusion protein of the invention desirably promotes angiogenesis, or any aspect thereof (e.g., reduces plasma leakage, promotes vessel maturity, and/or promotes vessel remodeling) more than the administration or production of a wild-type ARF corresponding to one of the ARF peptide portions, and more preferably than the administration or production of either (most preferably both) of the two wild-type ARFs most closely related to the fusion protein's ARF peptide portions under substantially identical conditions and doses. The determination that the fusion protein promotes angiogenesis more than a wild- type ARF can be made by applying any of the techniques described herein for measuring angiogenesis to substantially identical systems (e.g., tissues) separately receiving the wild-type ARF(s) and the fusion protein of the invention.

[0128] The fusion protein can include any suitable number of non-ARF peptide portions. The fusion protein portion can include any number of other elements (e.g., peptide portions) or modifications (e.g., additional amino acid sequences or other peptide fragments), as long as the biological functions (e.g., reduction of vascular permeability) of the fusion protein are not substantially diminished (i.e., not diminished by more than about 20%), preferably not more than about 10%>, and even more preferably not at all) over a fusion protein lacking such additional elements. Examples of such elements include sequences encoding proteins for post-translational modification or for binding to a small molecule ligand.

[0129] The fusion protein can comprise any suitable post-translational modifications or modifications. Examples of common and suitable post-translational modifications include carboxylation, glycosylation, hydroxylation, lipid or lipid derivative-attachment, methylation, myristylation, phosphorylation, and sulfation. Other post-translational modifications include acetylation, acylation, ADP-ribosylation, amidation, covalent attachment of flavin, covalent attachment of a heme moiety, covalent attachment of a nucleotide or nucleotide derivative, covalent attachment of phosphotidylinositol, cross- linking, cyclization, disulfide bond formation, demethylation, formylation, GPI anchor formation, iodination, oxidation, proteolytic processing, prenylation, racemization, selenoylation, transfer-RNA mediated addition of amino acids to proteins such as arginylation, and ubiquitination. Similar modifications are described in, e.g., Creighton, supra, Seifteretal., Meth. Enzymol, 182, 626-646 (1990), and Rattan et al., Ann. NY. Acad. Sci., 663, 48-62 (1992). Moreover, the fusion protein can be either a methionine-containing or methionineless N-terminal variant of any of the fusion proteins described herein. The nature and extent of post-translational modifications is largely determined by the host cell's posttranslational modification capacity and the modification signals present in the polypeptide amino acid sequence. For instance, glycosylation often does not occur in bacterial hosts such as E. coli. Accordingly, when glycosylation is desired, a polypeptide should be expressed (produced) in a glycosylating host, generally a eukaryotic cell (e.g., a mammalian cell or an insect cell). Post-translational modifications can be verified by any suitable technique, including, e.g., x-ray diffraction, NMR imaging, mass spectrometry, and/or chromatography (e.g., reverse phase chromatography, affinity chromatography, or GLC). The fusion protein or portion thereof also or additionally can comprise one or more modified amino acids, non-naturally occurring amino acids (e.g., β amino acids), alternative amino acids (e.g., selenocysteine), or amino acid analogs, such as those listed in the Manual of Patent Examining Procedure § 2422 (7th Revision - 2000), which can be incoφorated by protein synthesis, such as through solid phase protein synthesis (described in, e.g., Merrifield, A dv. Enzymol, 32, 221-296 (1969)).

[0130] A common additional element present in the fusion protein is a signal sequence, which directs either organelle trafficking (e.g., an endoplasmic reticulum trafficking signal as described in, e.g., U.S. Patent 5,846,540) and/or cell secretion. Such sequences are typically present in the immature (i.e., not fully processed) form of the fusion protein, and are subsequently removed/degraded to arrive at the mature form of the protein. Both naturally occurring and heterologous signal sequences are suitable (e.g., a secretion sequence associated with the protein incoφorated in the second peptide portion as discussed herein). For example, a VEGF-A secretion signal sequence Met Asn Phe Leu Leu Ser Tφ Val His Tφ Ser Leu Ala Leu Leu Leu Val Leu His His Ala Lys Tφ Ser Gin Ala (SEQ ID NO: 144) or a portion thereof, can be used, preferably bound to the N-terminus of the fusion protein CCD, or a sequence N-terminal to the CCD. Such sequences will necessarily vary with the host in which the fusion protein is produced. Examples of heterologous secretion sequences include STII or Ipp for E. coli, alpha factor for yeast, and viral signals such as heφes gD for mammalian cells. Further examples of signal sequences are described in, e.g., U.S. Patents 4,690,898, 5,284,768, 5,580,758, 5,652,139, and 5,932,445. Additional signal sequences can be identified using skill known in the art. For example, sequences identified by screening a library can be analyzed using the SignalP program (see, e.g., Nielsen et al., Protein Engineering, 10, 1-6 (1997)), or similar sequence analysis software capable of identifying signal-sequence-like domains, or by otherwise analyzing the sequences for features associated with signal sequences, as described in, e.g., European Patent Application 0 621 337. [0131] Alternatively or, preferably additionally, the fusion protein can comprise a wild- type ARF variable domain or homolog thereof. An "ARF variable domain" is any amino acid sequence of at least about 10 residues, preferably at least about 20 residues, positioned N-terminal to the wild-type ARF CCD. For example, the fusion protein can comprise the Ang-1 variable region (SEQ ID NO: 145). Such variable domains typically will comprise wild-type ARF signal sequences where the fusion protein comprises a homologous variable domain peptide portion, the homologous variable domain will desirably exhibit at least about 70%), preferably at least about 80%>, and more preferably at least about 90%> amino acid sequence identity to a wild-type ARF variable domain. The fusion protein variable domain desirably exhibits the highest level of local sequence identity to the same wild-type ARF as the fusion protein CCD or FLD.

[0132] In view of the above, it should be clear that the fusion proteins of the invention include both mature (fully processed) and immature (nascent) peptide portions, particularly where such fusion proteins are produced through the expression of a polynucleotide of the invention. In this respect, a peptide portion of the fusion protein can comprise one or more "propeptide" regions, which are removed during processing. Accordingly, nucleotide sequences encoding such propeptide portions along with the "mature" amino acid sequence associated with the peptide portion are within the scope of the invention. [0133] Other non-ARF peptide portions that can be included in the fusion protein include binding regions, such as avidin or an epitope, which can be useful for purification and processing of the fusion protein. Examples of such sequences are described in, e.g., International Patent Application WO 00/15823. In addition, detectable markers can be attached to the fusion protein, so that the traffic of the fusion protein through a body or cell can be monitored conveniently. Such markers may include radionuclides, enzymes, fluorophores, small molecule ligands, and the like.

[0134] Recently, the production of fusion proteins comprising a prion-determining domain has been used to produce a protein vector capable of non-Mendelian transmission to progeny cells (see, e.g., Li et a , J. Mol. Biol, 301(3), 567-73 (2000)). The inclusion of such prion-determining sequences in the fusion protein is contemplated, ideally to provide a hereditable protein vector comprising the fusion protein that does not require a change in the host's genome.

[0135] The mature fusion protein also can include additional peptide portions which act to promote stability, purification, and/or detection of the fusion protein. For example, a reporter peptide portion (e.g., green fluorescent protein (GFP), β-galactosidase, or a detectable domain thereof) can be incoφorated in the fusion protein. Purification facilitating peptide portions include those derived or obtained from maltose binding protein (MBP), glutathione-S-transferase (GST), or thioredoxin (TRX). The fusion protein also or alternatively can be tagged with an epitope which can be antibody purified (e.g., the Flag epitope, which is commercially available from Kodak (New Haven, Connecticut)), a hexa- histidine peptide, such as the tag provided in a pQE vector available from QIAGEN, Inc. (Chatsworth, California), or an HA tag (as described in, e.g., Wilson et al., Cell, 37, 161 (1984)).

[0136] The fusion protein can be further modified or derivatized in any suitable manner (e.g., by reaction with organic derivatizing agents). For example, the fusion protein can be linked to one or more nonproteinaceous polymers, typically a hydrophilic synthetic polymer, e.g., polyethylene glycol (PEG), polypropylene glycol, or polyoxyalkylene, as described in, e.g., U.S. Patents 4,179,337, 4,301,144, 4,496,689, 4,640,835, 4,670,417, and 4,791,192, or a similar polymer such as polyvinylalcohol or polyvinylpyrrolidone (PVP). Mimetics of the fusion proteins are also contemplated. Suitable types of peptide mimetics are described in, e.g., U.S. Patent 5,668,110 and the references cited therein. Furthermore, the fusion protein can be modified by the addition of protecting groups to the side chains of one or more the amino acids of the fusion protein. Such protecting groups can facilitate transport of the fusion peptide through membranes, if desired, or through certain tissues, for example, by reducing the hydrophilicity and increasing the lipophilicity of the peptide. Examples of suitable protecting groups include ester protecting groups, amine protecting groups, acyl protecting groups, and carboxylic acid protecting groups, which are known in the art (see, e.g., U.S. Patent 6,121,236). Synthetic fusion proteins of the invention can take any suitable form. For example, the fusion protein can be structurally modified from its naturally occurring configuration to form a cyclic peptide or other structurally modified peptide.

[0137] The fusion protein also can include a non-CCD, or even a non-ARF multimerizing domain or multimerizing component (i.e., a peptide portion which facilitates multimer formation), which permits the fusion protein to form multimers (including dimers) in addition to, or more typically in place of a multimerizing CCD (e.g., in cases where a mutant ARF CCD is incoφorated in the fusion protein, a non-multimer forming CCD is incoφorated in the fusion protein, or no CCD is incoφorated in the fusion protein, such as where the fusion protein is limited to an ARF FLD and some other ARF peptide portion). In such aspects, the fusion protein can comprise any suitable non-ARF multimerization domain. Examples of suitable non-ARF multimerization domains include the human immunoglobulin (IgG) multimerization domains (see, e.g., European Patent Application 0464533) and IgG-derived domains (e.g., the Fc domain as described in, e.g., Johanson et al., J. Biol. Chem., 270, 9459-71 (1995)). Additional modified IgG multimerizing domains and other multimerizing domains are described in International Patent Application WO 00/37642 and the references cited therein. The fusion protein (or polynucleotide encoding the fusion protein) can be administered with a peptide or agent that facilitates multimer formation (e.g., through promoting formation of disulfide bonds), such as the peptides described in European Patent Application 1 074 563.

[0138] The fusion protein can include any suitable multimerization domain which results in the formation of any suitable multimer. The fusion protein multimer can be a heteromultimer (e.g., a heterodimer) or a homomultimer (e.g., a homodimer). Typically and preferably, the fusion protein will form multimers with other fusion proteins of the invention and/or with wild-type ARF peptides (e.g., Ang-1 peptides). Testing for the suitability of any particular multimer combination can be performed using the assays described herein as well as techniques described in, e.g., DiSalvo et al., J. Biol Chem., 270, 7717-23 (1995), Cao et al., J. Biol. Chem., 271, 3154-62 (1996), and Olofsson et al., Proc. Natl. Acad. Sci. USA, 93, 2567-81 (1996)).

[0139] In some aspects, the fusion protein contains a multimerization domain which permits dimer formation without permitting formation of higher order multimers. For example, fusion proteins that include the dimerization domains of VEGFm or TGF-β can limit multimerization to dimmer formation.

[01 0] The fusion protein can comprise one or more heterologous and/or artificial receptor sites, which preferably change the receptor binding profile of the peptide portion, and more preferably localize the fusion protein (or at least the peptide portion) to a specific cell, group of cells, tissue, or tissues. For example, the fusion protein can include an RGD domain or LDV domain of a hemostatic modifier (preferably from a decorsin or a homolog thereof), or other integrin binding domain, selectin binding domain, or similar binding domain (e.g., a lamanin, fibrinogen, and/or fibronectin binding domain), examples of which are described in U.S. Patent Application 09/832,355 and references cited therein. [0141] The invention further provides polynucleotides including at least one nucleotide sequence which, when expressed in a cell permissive for expression of the nucleotide sequence, result in the production of the fusion protein. The polynucleotide sequence can be any suitable sequence (e.g., single stranded or double stranded RNA, DNA, or combinations thereof) and can include any suitable nucleotide base, base analog, and/or backbone (e.g., a backbone formed by, or including, a phosphothioate, rather than phosphodiester, linkage). Examples of suitable modified nucleotides which can be incoφorated in the polynucleotide sequence are provided in the Manual of Patent Examining Procedure § 2422 (7th Revision - 2000). The polynucleotide sequence can be any suitable length (e.g., about 300 nt, about 600 nt, about 1200 nt, about 2000 nt or even larger). The polynucleotide sequence can be any sequence that results in the fusion protein being produced upon expression. As such, the polynucleotide sequence is not limited to sequences that directly code for production of the fusion protein. For example, the polynucleotide can comprise a sequence which results in the fusion protein through intein- like expression (as described in, e.g., Colson and Davis, Mol. Microbiol, 12(3), 959-63 (1994), Duan et al., Cell, 89(4), 555-64 (1997), Perler, Cell, 92(1), 1-4 (1998), Evans et a , Biopolymers 51(5), 333-42 (1999), and de Grey, Trends Biotechnol, 18(9), 394-99 (2000)), or a sequence which contains self-splicing introns (or other self-spliced RNA transcripts) which form the peptide portions and/or the fusion protein (as described in, e.g., U.S. Patent 6,010,884). The polynucleotides also can comprise sequences which result in other splice modifications at the RNA level to produce an mRNA transcript encoding the fusion protein and/or at the DNA level by way of trans-splicing mechanisms prior to transcription (as described in, e.g., Chabot, Trends Genet., 12(11), 472-78 (1996), Cooper, Am. J. Hum. Genet, 61(2), 259-66 (1997), and Hertel et al., Curr. Opin. Cell Biol, 9(3), 350-57 (1997)). RNA based vectors may include removal of regions which promote degradation in the absence of hypoxia, e.g., by removal of the mRNA 3' and/or 5' UTRs (see Dibbens et al., Mol. Biol. Cell, 10, 907-19 (1999)) or portion thereof, e.g., the AU rich haiφin structure region of the 3' UTR (see, e.g., Pages et al., J. Biol. Chem. (published on June 9, 2000 as manuscript M002104200 - American Society for Biochemistry and Molecular Biology, Inc.), and Levy, J. Biol. Chem., 271, 25492-25497 and 2746-2753 (1996)), particularly where RNA vectors are administered in the absence of hypoxic conditions. The polynucleotide can comprise one or more sequences encoding a fusion protein wherein the fusion protein-encoding sequence or a portion thereof is codon optimized, i.e., codon frequency optimized and/or codon pair (i.e., codon context) optimized for a particular species (e.g., humans, either by optimizing a non-human or human sequence by replacement of "rare" human codons based on codon frequency, such as by using techniques such as those described in Buckingham et al., Biochimie, 76(5), 351-54 (1994) and U.S. Patents 5,082,767, 5,786,464, and 6,114,148). For example, the polynucleotide encoding the fusion protein can comprise a codon optimized human Ang-1 FLD gene sequence (SEQ ID NO: 146) or a codon optimized human Ang-1 CCD gene sequence (SEQ ID NO: 147). Additionally, codon optimized sequences for particular peptides can be obtained by subjecting the amino acid sequences of such peptides to backtranslation using a suitable program, such as the Entelechon backtranslation tool (available at http://www.entelechon .com/eng/backtranslation.html). Partially codon optimized sequences also can be used, such as codon sequences where only some or all of the "rarest" sequences (for the particular organism of interest) are removed. For example, the polynucleotide can comprise an ARF peptide portion-encoding sequence (e.g., an Ang-1 CCD or FLD sequence) where at least one (preferably all) of the Ala-encoding GCA and/or GCT codons in the sequence are replaced with GCC (but GCG codons are retained), at least one (preferably all) of the Gly- encoding GGC codons in the sequence is replaced with GGG, at least one (preferably all) of the Asn-encoding AAT codons in the sequence is replaced with AAC, at least one (preferably all) of the Pro-encoding CCA codons in the sequence is replaced with CCG, at least one (preferably all) of the Arg-encoding AGA codons is replaced with CGG, and at least one (preferably all) of the Ser-encoding AGC codons is replaced with TCG. Replacement of such rare codons can be accomplished using any more frequently expressed codons, as desired and appropriate.

[0142] Because the fusion protein will desirably comprise a FLD that is homologous to a first ARF and a CCD that is homologous to a second ARF, it is preferred that the fusion protein's FLD, CCD, or both are encoded by a nucleotide sequence having a complement that hybridizes to the ARF peptide portion's nearest (i.e., most related based on overall polynucleotide sequence identity, amino acid sequence identity, or both) wild-type ARF- encoding counteφart. For example, in a polynucleotide encoding a fusion protein comprising an Ang-1 homolog FLD portion and a NL2 homolog CCD portion, the Ang-1 FLD portion desirably is encoded by a nucleotide sequence having a complement which hybridizes with a wild-type Ang-1 gene under at least moderate stringency conditions (preferably high stringency conditions), the NL2 CCD portion desirably is encoded by a polynucleotide having a complement which hybridizes with a wild-type NL2 gene under at least moderate stringency conditions (preferably high stringency conditions), or, most desirably, the polynucleotide hybridizes to both wild-type Ang-1 and wild- type NL2 gene sequences under such conditions.

[0143] Where the polynucleotide encodes a fusion protein comprising a peptide portion that exhibits 100%> amino acid sequence identity to a wild-type ARF peptide portion, the nucleic acid sequence encoding the peptide portion can comprise a portion of the naturally occurring gene encoding the wild-type ARF (a wild-type ARF gene fragment) encoding the wild-type peptide portion. The polynucleotide can comprise any suitable wild-type ARF gene fragment, or combinations of wild-type ARF gene fragments. Preferably, the polynucleotide comprises a wild-type gene fragment encoding a wild-type ARF FLD fused to a wild-type gene fragment encoding a wild-type ARF CCD. For example, the polynucleotide can comprise the Ang-1 FLD-encoding sequence (SEQ ID NO: 148), the Ang-1 CCD-encoding sequence (SEQ ID NO: 149), the Ang-2 FLD-encoding sequence, the Ang-2 CCD-encoding sequence (SEQ ID NO: 150), the NLl FLD-encoding sequence (SEQ ID NO: 151), or the NL5 FLD-encoding sequence (SEQ ID NO: 152). Sequences identical to those in the wild-type genes for Ang-3 (SEQ ID NO: 153), Ang-4 (SEQ ID NO: 154), Ang-6 (SEQ ID NO: 155), Ang-2A (SEQ ID NO: 156), Ang-2B (SEQ ED NO: 157), Ang- 2C (SEQ ED NO: 158), Ang-2X (SEQ ID NO: 159), NL2 (SEQ ID NO: 160), NL3 (SEQ ID NO: 161), the Hakata antigen, NL4 (SEQ ED NO: 162), NL6 (SEQ ED NO: 163), NL8 (SEQ ID NO: 164), FLS 139 (SEQ ID NO: 165), zapol (SEQ ED NO: 166), or murine FDRG, can be suitable. Other ARF gene fragments in the patent documents and publications known in the art, such as those incoφorated herein by reference, also can be suitable. [0144] Alternatively, the fusion protein-encoding polynucleotide can comprise a nucleic acid sequence that exhibits less than 100%) polynucleotide sequence identity to such a wild- type gene fragment, but which still encodes a peptide portion that exhibits 100% amino acid sequence identity to the wild-type ARF peptide portion. For example, the polynucleotide sequence can comprise (1) a nucleic acid sequence that exhibits at least about 35%>, desirably at least about 50%, preferably at least about 60%>, and more preferably at least about 70%) (e.g., at least about 80%, at least about 90%, or at least about 95%) polynucleotide sequence identity (either local or overall) to such a wild-type gene fragment, (2) a nucleic acid sequence which hybridizes under at least moderate stringency conditions (preferably high stringency conditions) to the wild-type gene fragment, or (3) a sequence that does not hybridize to and/or exhibit such levels of identity to the gene fragment due to the presence of other sequences that do not affect the produced protein product (e.g., degenerate sequences, sequences that are self-spliced out of the RNA transcript encoded by the nucleic acid sequence, or intein-like expression pattern sequences) but nonetheless when transcribed and translated results in the production of the protein product encoded by the wild-type gene fragment.

[0145] The polynucleotide also can comprise one or more nucleic acid sequences encoding any of the ARF peptide portion homologs described elsewhere herein. Polynucleotide sequences encoding such ARF peptide homologs can be obtained through constructing such sequences based on the desired ARF peptide homolog amino acid sequence.

[0146] Alternatively, ARF homologs or ARF peptide portion homologs (e.g., homologous CCDs and/or FLDs, even in non-ARF peptides) can be obtained by screening polynucleotide libraries (e.g., a genomic library, cDNA library, or sublibrary thereof). For example, the DNA of animal species can be screened to identify ARF homologs or ARF peptide portion homologs. Such screening can be performed by any suitable technique, including, e.g., screening the libraries with polynucleotide probes under conditions wherein hybridization to ARF homolog-encoding polynucleotides is likely to occur (e.g., under at least moderately stringent conditions). Such screening can be performed in a human DNA or cDNA library (e.g., to identify novel ARF splice variants or yet undiscovered ARF homologs), or in a polynucleotide library obtained from other species, preferably other mammalian species (e.g., Pan troglodytes, Gorilla gorilla, Pongo pygmaeus, Hvlobates concolor, Macaca mulatto, Papio papio, Papio hamadrvns, Cercopithecus aethiops, Cebus capucinus, Aotus trivirgatus, Sanguinus oedipus, Microcebus murinus, Mus musculus, Rattus norvegicus, Cricetulus griseus, Felis catus, Mustela vison, Canis familiar is, Orystolagus cuniculus, Bos taurus, Ovis aries, Sus scrofa, and Equus caballus). Fluorescence in situ hybridization (FISH) of a cDNA clone to a metaphase chromosomal spread can be used to perform chromosomal screening for ARF homolog-encoding genes. Pools of protein candidates can be similarly screened for ARF homologs using standard biochemical and proteomics-related techniques (e.g., the yeast two hybrid system as described in, e.g., Mendelsohn and Brat, Curr. Opin. Biotech., 5, 482-86 (1994), and/or affinity chromatography (e.g., using a TIE receptor or portion thereof)). [0147] The fusion protein ARF peptide portions can be identical to, or homologous to, wild-type ARF peptide portions of any animal. The fusion protein's ARF peptide portions can be derived from, or most closely related to, wild-type ARF peptide portions of different animals (e.g., the fusion protein can include a CCD that exhibits 100%» identity to a murine ARF CCD and a FLD that exhibits 100% identity to a human ARF FLD). The reference to any ARF (or peptide portion or related polynucleotide) herein encompasses both the human protein (or other particular animal, e.g., mice in the case of murine FDRG) and all orthologs thereof unless otherwise specified. Thus, for example, an "Ang-1 FLD" encompasses both the human Ang-1 FLD and the murine Ang-1 FLD.

[0148] Polynucleotides comprising sequences encoding novel ARF peptide portions, particularly sequences encoding peptide portions homologous to, but not identical with, wild-type ARF peptide portions also can be generated through inducing mutations in known ARF-encoding polynucleotides (e.g., a wild-type Ang-1, Ang-2, Ang-2X, NLl, or NL5 gene fragment). For example, ARF homolog peptide portion-encoding polynucleotides can be obtained through application of site-directed mutagenesis (as described in, e.g., Edelman et al., DNA, 2, 183 (1983), Zoller et al., Nucl Acids Res., 10, 6487-5400 (1982), and Veira et al., Meth. Enzymol, 153, 3 (1987)), alanine scanning, or random mutagenesis, such as iterated random point mutagenesis induced by error-prone PCR, chemical mutagen exposure, or polynucleotide expression in mutator cells (see, e.g., Bornscheueret al., Biotechnol. Bioeng., 58, 554-59 (1998), Cadwell and Joyce, PCR Methods Appl, 3(6), S136-40 (1994), Kunkel et al., Methods Enzymol, 204, 125-39 (1991), Low et al., J. Mol Biol, 260, 359-68 (1996), Taguchi et al, Appl. Environ. Microbiol, 64(2), 492-95 (1998), and Zhao et al., Nat. Biotech., 16, 258-61 (1998)). Suitable primers for PCR-based site- directed mutagenesis or related techniques can be prepared by the methods described in, e.g., Crea et al., Proc. Natl Acad. Sci. USA, 75, 5765 (1978). The application of site- directed mutagenesis to produce novel ARF peptide portion variants can be based on a modification of the techniques provided in, e.g., Shortle et al., Ann. Rev. Genet, 15, 288-94 (1981), Keyt et al., J. Biol Chem., 271, 5638-46 (1996), and Ki et al., J. Biol Chem., 275(38), 29823-28 (2000).

[0149] Other polynucleotide mutagenesis methods useful for producing novel ARF homolog peptide portions include PCR mutagenesis techniques (as described in, e.g., Kirsch et al., Nucl. Acids Res., 26(7), 1848-50 (1998), Seraphin et al., Nucl. Acids Res., 24(16), 3276-7 (1996), Caldwell et al., PCR Methods Appl, 2(1), 28-33 (1992), Rice et al., Proc. Natl Acad. Sci. USA. 89(12), 5467-71 (1992) and U.S. Patent 5,512,463), cassette mutagenesis techniques based on the methods described in Wells et al., Gene, 34, 315 (1985), phagemid display techniques (as described in, e.g., Soumillion et al., Appl. Biochem. Biotechnol, 47, 175-89 (1994), O'Neil et al., Curr. Opin. Struct Biol, 5(4), 443-49 (1995), Dunn, Curr. Opin. Biotechnol, 7(5), 547-53 (1996), and Koivunen et al., J. Nucl. Med., 40(5), 883-88 (1999)), reverse translation evolution (as described in, e.g., U.S. Patent 6,194,550), saturation mutagenesis described in, e.g., U.S. Patent 6,171,820), PCR-based synthesis shuffling (as described in, e.g., U.S. Patent 5,965,408) and recursive ensemble mutagenesis (REM) (as described in, e.g., Arkin and Yourvan, Proc. Natl. Acad. Sci. USA, 89, 7811-15 (1992), and Delgrave et al., Protein Eng., 6(3), 327-331 (1993)). Alternatively, ARF peptide portion homolog-encoding polynucleotides can be pre-designed and synthetically produced using techniques such as those described in, e.g., Itakura et al., Annu. Rev. Biochem., 53, 323 (1984), Itakura et al., Science, 198, 1056 (1984), and Ike et al., Nucl Acid Res., 11, 477 (1983). For example, sequence analysis of a number of related ARFs (e.g., a number of angiopoietins) can be subjected to sequence analysis (e.g., using CLUSTAL W) to identify an amino acid consensus sequence that can be used to design novel DNAs based on the genetic code (e.g., by subjecting the consensus sequence to reverse translation analysis). Further details regarding the above-described techniques are described in Sambrook et al, and Ausubel et al., supra.

[0150] Alternatively, ARF homolog peptide portion-encoding polynucleotides can be generated through directed evolution techniques (e.g., polynucleotide shuffling). Examples of such techniques are described in, e.g., Stemmer, Nature, 370, 389-91 (1994), Cherry et al., Nat. Biotechnol. 17, 379-84 (1999), and Schmidt-Dannert et al., Nat Biotechnol. , 18(7), 750-53 (2000). Preferably, shuffling of the ARF-encoding sequence is performed in combination with staggered extension (StEP), random primer shuffling, backcrossing of improved variants, or any combination thereof, e.g., as described in Zhao et al., supra, Cherry et al., supra, Arnold et al., Biophys. J., 73, 1147-59 (1997), Zhao and Arnold, Nucl. Acids Res., 25(6), 1307-08 (1997), and Shao et al., Nuc Acids Res., 26, 681-83 (1998). Alternatively, the incremental truncation for the creation of hybrid enzymes (ITCHY) method (see, e.g., Ostermeier et al., Nat. Biotechnol, 17(12), 1205-09 (1999)) can be applied to combinations of ARF-encoding genes or gene fragments to produce novel ARF- encoding or ARF peptide portion-encoding polynucleotides. Another set of techniques for introducing diversity into a library of homologs are provided in U.S. Patents 6,159,687 and 6,228,639. [0151] The biological activity of the products of molecular evolution are expected to vary, and, accordingly, some screening for biological activity of the directed evolution product can be required to ensure the novel ARF homolog peptide portion-encoding polynucleotide is suitable for incoφoration in the fusion protein and/or fusion protein- encoding polynucleotides of the present invention. Any suitable assay for measuring the desired biological activity of a molecule can be used, including the techniques described herein with respect to the fusion proteins of the invention (e.g., antibody reactivity, receptor binding, induction of angiogenesis or preferred aspects thereof (e.g., promoting vascular maturation and/or reducing plasma leakage)). In this manner, with a routine amount of experimentation, novel ARF homolog peptide portions can be generated which are suitable for incoφoration in the fusion protein-encoding polynucleotide. [0152] Preferably, in addition to the nucleic acid sequence which, when expressed, results in the fusion protein (the "fusion protein nucleic acid sequence"), the polynucleotide further includes one or more suitable "expression control sequences" operably linked to the sequence encoding the fusion protein. An expression control sequence is any nucleotide sequence that assists or modifies the expression (e.g., the transcription, translation, or both) of the nucleic acid encoding the angiogenic sequence. The expression control sequence can be naturally associated with a polynucleotide encoding any ARF peptide portion of the fusion protein (e.g., an Ang-1 promoter), or can comprise a heterologous element with respect to such polynucleotide sequences (e.g., a myocardin cardiac tissue regulator sequence or a desmin locus control region (LCR), both of which are associated with control of muscle gene expression). For example, the fusion protein nucleic acid sequence can be operably linked to a constitutive promoter (e.g., the Rous sarcoma virus long terminal repeat (RSV LTR) promoter/enhancer or the cytomegalo virus major immediate early gene (CMV IE) promoter, which is particularly preferred), an inducible promoter, (e.g., a growth hormone promoter, metallothionein promoter, heat shock protein promoter, E1B promoter, hypoxia induced promoter, radiation inducible promoter, or adenoviral MLP promoter and tripartite leader), an inducible-repressible promoter, a developmental stage-related promoter (e.g., a globin gene promoter), or a tissue specific promoter (e.g., a smooth muscle cell - actin promoter, VEGF receptor promoter, myosin light-chain 1 A promoter, or vascular endothelial cadherin promoter). In some instances, host-native promoters are preferred over non-native promoters (e.g., a human beta actin promoter or EFlα promoter driving expression of the fusion protein nucleic acid sequence can be preferred in a human host), particularly where strict avoidance of gene expression silencing due to host immunological reactions is desirable. The polynucleotide can include expression control sequences wherein one or more regulatory elements have been deleted, modified, or inactivated. The polynucleotide also or alternatively can include a bi-directional promoter system (as described in, e.g., U.S. Patent 5,017,478) linked to multiple genes of interest (e.g., multiple fusion protein encoding genes). Other suitable promoters and principles related to the selection, use, and construction of suitable promoters are provided in, e.g., Werner, Mamm. Genome, 10(2), 168-75 (1999), Walther et al., J. Mol. Med, 74(7), 379-92 (1996), Novina, Trends Genet, 12(9), 351-55 (1996), Hart, Semin. Oncol, 23(1), 154-58 (1996), Gralla, Curr. Opin. Genet Dev., 6(5), 526-30 (1996), Fassler et al., Methods Enzymol, 273, 3-29 (1996), Ayoubi et al., 10(4), FASEB J., 10(4), 453-60 (1996), Goldsteine et al., Biotechnol Annu. Rev., 1, 105-28 (1995), Azizkhan et al., Crit Rev. Eukaryot. Gene Expr., 3(4), 229-54 (1993), Dynan, Cell, 58(1), 1-4 (1989), Levine, Cell, 59(3), 405-8 (1989), and Berk et al., Annu. Rev. Genet, 20, 45-79 (1986), as well as U.S. Patent 6,194,191. In some aspects, radiation-inducible promoters such as those described in described in U.S. Patents 5,571,797, 5,612,318, 5,770,581, 5,817,636, and 6,156,736 can be suitable. In other instances, ecdysone and ecdysone-analog-inducible promoters (ecdysone-analog-inducible promoters are commercially available through Stratagene (LaJolla CA)). Other suitable commercially available inducible promoter systems include the inducible Tet-Off or Tet-on systems (Clontech, Palo Alto, CA). As an alternative to a promoter, particularly in RNA vectors and constructs, the nucleic acid sequence and/or vector can comprise one or more internal ribosome entry sites (ERESs), IRES-encoding sequences, or RNA sequence enhancers (Kozak analogs), such as the tobacco mosaic virus omega prime sequence. [0153] The polynucleotide sequence also or alternatively can comprise an upstream activator sequence (UAS), such as a Gal4 activator sequence (as described in, e.g., U.S. Patent 6,133,028) or other suitable upstream regulatory sequence (as described in, e.g., U.S. 6,204,060). The polynucleotide can include any other expression control sequences (e.g., enhancers, termination sequences, initiation sequences, splicing control sequences, etc.). Typically, the polynucleotide will typically include a Kozak consensus sequence, which can be a naturally occurring or modified sequence such as the modified Kozak consensus sequences described in U.S. Patent 6,107,477. The polynucleotide can further comprise site-specific recombination sites, which can be used to modulate transcription of the polynucleotide, as described in, e.g., U.S. Patents 4,959,317, 5,801,030 and 6,063,627, European Patent Application 0 987 326 and International Patent Application WO 97/09439. [0154] The polynucleotide can include or consist of any suitable fusion protein nucleic acid sequence. Preferred fusion protein nucleic acid sequences include nucleotide sequences which, when expressed, result in the production of the above-described fusion proteins, the polynucleotide can comprise any number of transcriptionally inert sequences (e.g., an intron). The polynucleotide can contain any suitable number of copies of the fusion protein-incoding nucleic acid sequence.

[0155] Preferably, the polynucleotide comprises a second nucleotide sequence that, when expressed, produces a second protein which modulates, and preferably promotes, angiogenesis, bone growth, wound healing, or any combination thereof, and most preferably promotes angiogenesis, particularly where the fusion protein also promotes angiogenesis. The second nucleotide sequence can encode, for example, a receptor that interacts, and preferably is activated by the fusion protein FLD (e.g., a TIE-2 receptor). In this respect, the polynucleotide can include any suitable number of protein-encoding sequences. Alternatively, the second nucleotide sequence can encode for a polynucleozyme (e.g., a ribozyme or DNAzyme) or for the production of an inhibitory (e.g., antisense) polynucleotide (suitable examples of which can be identified through standard gene walking procedures), which preferably facilitates one of the above-mentioned biological activities through inhibition of a biological activity inhibitor, examples of which are discussed further elsewhere herein. Polynucleotides, either encoding the ARFs of the invention or administered therewith, can take the form of aptamers (as described in, e.g., Famulok and Mayer - http://www.chemie.uni-bonn.de/oc/ak_fa/publications/CTMI-paper.pdf), capable of binding to suitable targets. Triplex-forming inhibitory nucleotides also can be co- administered with the ARF-encoding polynucleotide, as can "decoys" comprising DNA binding domains specific for particular genes which are sought to be silenced but which encode a different factor.

[0156] Examples of suitable angiogenic factors that can be expressed (produced) with the fusion proteins of the invention (and, as such, can be encoded by the second nucleotide sequence) include, but are not limited to, fibroblast growth factors (FGFs) (e.g., aFGF (FGF-1) (also known as heparin binding factor 1), bFGF (FGF-2), HST, int-2, FGF-4, FGF- 5, FGF-6, and KGF (as discussed in, e.g., Basilico and Moscatelli, "The FGF Family of Growth Factors and Oncogenes" in ADVANCES IN CANCER RESEARCH, 59, 115-65 (Woude and Klien eds., Academic Press 1992) and U.S. Patent 5,614,496) and their relatives (e.g., HDGFs, as described in, e.g., Klagsbrun et al., Proc. Natl. Acad. Sci. USA, 83, 2448 (1986)), angiogenins (e.g., angiogenin, angiogenin-2, and mAngiogenin-3, as described in, e.g., Strydom et al., Biochemistry, 24, 5486 (1985), Folkman et al., Science, 235, 442 (1987), Bond et al., Biochim. Biophys. Acta, 1162, \11 (1993), Hu et al., Biochem. Biophys. Res. Commun., 197, 682. (1993), Hu et al., Proc. Natl. Acad. Sci. USA, 91, 12096 (1994), and Moenner et al., Eur. J. Biochem., 226, 483 (1994)), pleiotrophin (PTN, also known as HBNF, HB-GAM, HBBM, pi 8, OSF-1, and HARP, among others, as described in, e.g., Kretschmer et al., Growth Factors, 5, 99 (1991), Kretschmer et al., Biochem. Biophys. Res. Commun., 192(2), 420-29 (1993), U.S. Patent 5,270,449, European Patent 0 441 763, and European Patent Application 0 474 979), midkine (MK) (as described in, e.g., Bόhlen and Kovesdi, Prog. Growth Factor Res., 3, 143-57 (1991), Inui et al., J Peptide Sci., 2, 28-39 (1996), Iwasaki et al., EMBOJ., 16, 6936-46 (1997), and U.S. Patent 5,210,026), transforming growth factors (TGFs - e.g., TGF-β), gelatinases (as described in, e.g., Nguyen et al., Int. J. Biochem. Cell Biol. 33(10), 960-970 (2001)), placental growth factors, Kallistatins (as described in, e.g., Chao et al., Biol. Chem., 382(1), 15-21 (2001), platelet- derived growth factors (e.g., platelet-derived endothelial cell growth factor and PDGF-BB (Regranex)), aniomotins (as described in, e.g., Troyanovsky et al., J. Cell Biol, 152(6), 1247-54 (2001)), ECGF (as described in, e.g., U.S. Patent 4,868,113), cyclooxylgenases (as described in, e.g., Dempke et al., J. Cancer Res. Clin. Oncol, 127(7), 411-7 (2001)), Del-1 (as described in, e.g., U.S. Patents 5,877,281 and 5,874,562), erythropoietin, follistatin, granulocyte colony-stimulating factor (G-CSF), GM-CSF, scatter factor/hepatocyte growth factor (HGF) (as described in, e.g., U.S. Patents 6,011,009 and 6,133,231), leptin, insulin like growth factors (IGFs, e.g., IGF-I and IGF-II), ORP150 (as described in, e.g., Ozawa et al., J. Clin. Invest, 108(1), 41-50 (2001)), endothelial growth factors (EGFs) (e.g., endothelial cell-derived growth factor (ECDGF) and PD-ECGF (as described in, e.g., Matsukawa et al., Biochim, Biophys, Acta, 1314(1-2), 71-82 (1996), Moghaddam et al., Biochemistry, 31, 12141-46 (1992), Miyazono et al., Biochemistry, 28, 1704-10 (1989), and Ishikawa, Nature, 338(6216), 557-62 (1989)), PP60 (C-Src) (as described in, e.g., Marx et al., Exp. Mol. Pathol, 70(3), 201-13 (2001)), HBEGFs (as described in, e.g., U.S. Patent 6,037,329), epidermal growth factors, connective tissue growth factors (CTGFs - as described in, e.g., U.S. Patent 6,149,916 and Moussad et al., Mol. Genet Metab., 71(1-2), 276-92 (2000), preferably CTGF-2), ganglosides (as described in, e.g., Lang et al., Biochem. Biophys. Res. Common., 282(4), 1031-37 (2001)), matrix metalloproteinases (MMPs) (as described in, e.g., Muφhy et al., Matrix Biol, 15(8-9), 511-8 (1997), Baramova et al., Cell Biol. Int., 19(3), 239-42 (1995), and Matrisian, Ann. NY. Acad. Sci., 732, 42-50 (1994)), tissue inhibitors of metalloproteinase (TEMPs, e.g., vasosten or TEVIP-4) (as described in, e.g., Vallamo et al., Human Pathol, 30(7), 795-802 (1999) and Dollery et al., Circ. Res., 84(5), 498-504 (1999)), Survivins, Stat-3(as described in, e.g., Yamauchi-Takihara et al., Trends Cardiovasc. Med., 10(7), 298-303 (2000)), Delta-3 (as described in, e.g., U.S. Patent 6,121,045), COUP-TFIT, eNOS, iNOS, MCP-1, proliferin, syndecan-4 protiens (as described in, e.g., International Patent Application WO 01/45751), RAF and RAS protein kinases (as described in, e.g., International Patent Application WO 01/12210), E-selectin, platelet-activating factors (PAFs) VCAMl, COX-2, UTV-tat, ephrins (e.g., EphBl, EphB2, or EphB4) (as described in, e.g., Yancopoulos et al., Cell, 93, 661-64 (1998) and references cited therein), TWEAK (as described in, e.g., Lynch et al., J. Biol. Chem., 273(13), 8455-49 (1999)), CYR 61 (as described in, e.g., Babic et al., Proc. Natl Acad. Sci. USA, 95, 6355 (1998)), VEGF- activating zinc finger protein (ZFP) transcription factors (as described in, e.g., International Patent Application WO 01/19981) Fibrin fragment E, PR39 (as described in, e.g., Li et al., Nat Med., 6(1), 49-55 (2000), and modified by Nat. Med., 6(3), 356 (2000)), Mekk3 (as described in, e.g., Yang et al., Nat Genet, 24(3), 309-13 (2000), MEF2C (as described in, e.g., Bi et al., Dev. Biol, 211(2), 255-67 (1999), prostacycins (such as PGI2 described in e.g., Spisniet et al., Exp. Cell Res., 266(1), 31-43 (2001)), SDF-1 (as described in, e.g., Moore et al., Ann: N.Y. Acad. Sci., 938, 36-47 (2001)) progestins (as described in e.g., U.S. Patent 6,245,757) tissue plasminogen activator (tPA), urokinase- plasminogen activator (uPA), coagulation/clotting factors (e.g., tissue factor, plasma coagulation factor Vila (FVIIa), cofactor FVa, Factor X (Xa), thrombin (prothrombin), and fibrinogen, as discussed in, e.g., Carmeliet, Science, 293, 1602-1604 (2001) and references cited therein), caveolin-1 and caveolin-1 -modifying agents (e.g., agonists) (as described in, e.g., European Patent Application 1 076 091), nicotine (as described in, e.g., International Patent Application WO 01/08684), angiogenic C-x-C chemokines (as described in, e.g., Colville-Nash et al., Mol. Med. Today, 13-23 (1997)), other angiogenic factors described in International Patent Applications WO 01/05825 and WO 01/32926, and the AHRs (as described in U.S. Patent 6,121,236). Other angiogenic peptides include cytokines such as tumor necrosis factor- alpha (TNF-α), interleukin-3 (IL-3), interleukin-18 (IL-18), and interleukin-8 (IL-8), and transcription factors such as HEF-1 (or HE?- lot and/or HE?-2α), chimeric HEF factors (e.g., the HIF-lα/VP16 factor described in Vincent et al., Circulation, 18, 2255-61 (2000)), homologs thereof (e.g., EPAS as described in, e.g., Maemura et al., J. Biol. Chem., 274(44), 3165-70 (1999)), fragments thereof, or heterodimers thereof (e.g., a HIF-lα/HIF-2α heterodimer). Preferably, the angiogenic factor produced by expression with the fusion protein is an angiogenic factor that functions in a manner other than as a transcription factor. Desirably, the second nucleic acid sequence encodes an angiogenic VEGF or VEGF homolog, most desirably a non-heparin-binding VEGF (e.g., human VEGFm), such as those VEGF and VEGF homologs described in U.S. Patent Application 09/832,355, and references cited therein. Other preferred angiogenic mediators include HBNFs, MKs, aFGFs, and Del-1. Non-peptide angiogenic mediators that can be associated with the fusion protein or co-administered therewith include hormones such as oestrogens and proliferin, alcohols such as glycerol, pyridine derivatives (e.g., nicotinamide), and oligosaccharides such as hyaluronan.

[0157] If the polynucleotide encodes multiple gene products, a combination of expression control sequences (e.g., promoters) can be used, preferably which correspond to a pre-planned pattern of activity with the desired pattern and level of expression of the encoded factors. Thus, nucleotide sequences in the polynucleotide can be under the control of separate promoters having different expression profiles, e.g., at least one nucleic acid sequence is operably linked to an RSV promoter and at least one other nucleic acid sequence is operably linked to a CMV promoter. Alternatively, a hybrid promoter can be constructed which combines the desirable aspects of multiple promoters. For example, a CMV-RSV hybrid promoter combining the CMV promoter's initial rush of activity with the RSV promoter's high maintenance level of activity is especially preferred for use in many embodiments of the inventive method. Thus, the invention provides polynucleotides where the fusion protein sequence is operably linked to a first promoter and a second nucleotide sequence is operably linked to a second promoter, such that the initiation of expression of the first nucleotide sequence and second nucleotide occurs at different times, in response to different factors, or both. Preferably, such promoter systems are designed to mimic expression patterns associated with normal biological activities, e.g., pathways or cascades. For example, a first promoter can drive the early expression (or separately inducible expression) of a vascular endothelial growth factor (VEGF) (e.g., human VEGFm, VEGF,₆₅, or VEGF, ₉), and a second promoter can be later induced or otherwise later cause expression of a nucleic acid sequence encoding an ARF fusion protein of the invention. [0158] The polynucleotide can include multiple fusion protein genes and/or related genes to be serially expressed and/or co-expressed. Thus, for example, the invention contemplates administration of polynucleotides that encode at least 3, at least 4, at least 5, or more, fusion protein genes or combinations of fusion protein genes, and other angiogenic factor genes, bone growth promoting factor genes, or wound healing promoting factor genes, which preferably mimic an expression pattern of a normal biological cascade. Suitable bone growth, wound healing, and angiogenic factors are described in U.S. Patent Application 09/832,355, and references cited therein. For example, the polynucleotide can include sequences encoding (1) a matrix metalloproteinase or other factor that promotes matrix degradation, (2) a VEGF or VEGF homolog (which preferably attracts endothelial cells and induces blood vessel formation), (3) a fusion protein of the invention, which preferably induces vascular maturation and/or reduces plasma leakage in the VEGF-induced vessels, and (4) an ephrin (which preferably induces stabilization and maintenance of the blood vessel), which are preferably operably linked to promoters such that they are expressed in a manner that mimics the normal biological cascade in blood vessel development in the host. Alternatively or additionally, multiple polynucleotides (e.g., within multiple vectors) can be administered, wherein the polynucleotides encode one or multiple genes, including at least one polynucleotide encoding one of the ARF fusion proteins of the invention, to provide such a cascade effect. However, the administration of a single polynucleotide under control of the above-described expression control sequence systems is preferred.

[0159] Production of the recombinant polynucleotide encoding the fusion protein can be accomplished by any suitable technique. Recombinant polynucleotide production is well understood, and methods of producing such molecules are provided in, e.g., Mulligan, Science 260, 926-932 (1993), Friedman, Therapy For Genetic Diseases (Oxford University Press, 1991), Ibanez et al., EMBOJ., 10, 2105-10 (1991), Ibanez et al., Cell, 69, 329-41 (1992), and U.S. Patents 4,440,859, 4,530,901, 4,582,800, 4,677,063, 4,678,751, 4,704,362, 4,710,463, 4,757,006, 4,766,075, and 4,810,648, and are more particularly described in Sambrook and Ausubel, supra.

[0160] The polynucleotide preferably is positioned in and/or administered in the form of a suitable delivery vehicle (i.e., a vector). The vector can be any suitable vector. For example, the nucleic acid can be administered as a naked DNA or RNA vector, including, for example, a linear expression element (as described in, e.g., Sykes and Johnston, Nat. Biotech., 17, 355-59 (1997)), a compacted nucleic acid vector (as described in, e.g., U.S. Patent 6,077,835 and/or International Patent Application WO 00/70087), a plasmid vector such as pBR322, pUC 19/18, or pUC 118/119, a "midge" minimal-sized vector (as described in, e.g., Schakowski et al., Mol. Ther., 3, 793-800 (2001)) or as a precipitated nucleic acid vector construct (e.g., a CaP0₄ precipitated construct). The vector also can be a shuttle vector, able to replicate and/or be expressed (desirably both) in both eukaryotic and prokaryotic hosts (e.g., a vector comprising an origin of replication recognized in both eukaryotes and prokaryotes). The nucleic acid vectors of the invention can be associated with salts, carriers (e.g., PEG), formulations which aid in transfection (e.g., sodium phosphate salts, Dextran carriers, iron oxide carriers, or gold bead carriers), and/or other pharmaceutically acceptable carriers, some of which are described herein. Alternatively or additionally, the polynucleotide vector can be associated with one or more transfection- facilitating molecules such as a liposome (preferably a cationic liposome), a transfection facilitating peptide or protein-complex (e.g., a poly(ethylenimine), polylysine, or viral protein-nucleic acid complex), a virosome, a modified cell or cell-like structure (e.g., a fusion cell), or a viral vector. The polynucleotide also can be administered with a vector which catalyzes disulfide bond formation (as described in, e.g., International Patent Application WO 98/37208).

[0161] More preferably, the polynucleotide is positioned in, and administered to the host via, a viral vector. The viral vector can be any suitable viral vector. A viral vector in the context of the invention includes any combination of nucleotides and proteins that are derived from, obtained from, or based upon proteins and or nucleic acids that are present in a wild-type virus. For example, the viral vector can comprise a virus-like particle (VLP) or a vector similar to those described in U.S. Patent 5,849,586 and International Patent Application WO 97/04748. The viral vector can be a vector which requires the presence of another vector or wild-type virus for replication and/or expression (i.e., a helper-dependent virus), such as an adenoviral vector amplicon. The viral vector preferably consists of an intact virus particle. Typically, such viral vectors consist essentially of a wild-type viral particle, or a viral particle modified in its protein and/or nucleic acid content to increase transgene capacity or aid in transfection and/or expression of the nucleic acid (examples of such vectors include the heφes virus/ AAV amplicons). Such vectors are typically named for the type of virus they are obtained from, derived from, or based upon, as applicable. Examples of preferred viral vectors include heφes viral vectors, adeno-associated viral vectors, and adenoviral vectors. Suitable examples of such vectors and other suitable viral vectors are provided in, e.g., Mackett et al., J. Gen. Virol, 67, 2067-82 (1986), Beaud et al., Dev. Biol. Stand., 66, 49-54 (1987), Levine, Microbiol Sci., 4(8), 245-50 (1987), Lebowski et al., Mol. Cell Biol, 8(10), 3988-96 (1988), Nicholas et al., Biotechnology, 10, 493-513 (1988), Moss et al., Curr. Top. Microbiol. Immunol, 158, 25-38 (1992), Berihoud et al., Curr. Opin. Biotechnol, 10(5), 440-47 (1999), Yonemitsu, Nat. Biotechnol, 18(9), 970-3 (2000), and Russell, J. Gen. Virol, 81, 2573-2604 (2000), as well as International Patent Application WO 00/32754. Desirably, the viral vector particle is a mammalian viral vector, a non-enveloped viral vector, and a vector that contains a DNA-based genome. [0162] The construction of recombinant viral vectors is well understood in the art. For example, adenoviral vectors can be constructed and/or purified using the methods set forth, for example, in Graham et al., Mol. Biotechol, 33(3), 207-220 (1995), U.S. Patents 5,922,576, 5,965,358 and 6,168,941 and International Patent Applications WO 98/22588, WO 98/56937, WO 99/15686, WO 99/54441, and WO 00/32754. Adeno-associated viral vectors can be constructed and/or purified using the methods set forth, for example, in U.S. Patent 4,797,368 and Laughlin et al., Gene, 23, 65-73 (1983). Similar techniques are known in the art with respect to other viral vectors, particularly with respect to heφes viral vectors (see e.g., Lachman et al., Curr. Opin. Mol. Ther., 1(5), 622-32 (1999)), lentiviral vectors, and other retroviral vectors.

[0163] The viral vector can be a chimeric viral vector. Examples of suitable chimeric viral vectors are described in, e.g., Reynolds et al., Mol. Med. Today, 5(1), 25-31 (1999) and Boursnell et al., Gene, 13, 311-317 (1991).

[0164] Desirably, the viral vector is capable of expressing the polynucleotide for a sustained period (e.g., for a period of at least about 1 day, preferably about 1 week), without expressing the polynucleotide so long that undesired effects associated with prolonged expression, e.g., promiscuous angiogenesis, occurs (e.g., for a period of less than about 2 weeks). Thus, the viral vector preferably is capable of therapeutic, and transient, self- terminating expression of the polynucleotide (e.g., expression for a period of about 1 week or less). Preferably, the viral vector achieves gene transfer in both dividing and non- dividing, as well as terminally differentiated, cells, with high levels of expression in cardiovascular relevant sites such as the myocardium, vascular endothelium, and skeletal muscle. The viral vector desirably is safe for administration to the host. Advantageously, the viral vector operates in an epichromosomal manner without insertion of genetic material to the host. Adenoviral vectors, which possess all of these aforementioned qualities, are particularly preferred delivery vectors for nucleic acid angiogenic mediators. [0165] Any suitable adenoviral vector can be used as a delivery vehicle for the polynucleotide. For instance, an adenovirus can be of subgroup A (e.g., serotypes 12, 18, and 31), subgroup B (e.g., serotypes 3, 7, 11, 14, 16, 21, 34, and 35), subgroup C (e.g., serotypes 1, 2, 5, and 6), subgroup D (e.g., serotypes 8, 9, 10, 13, 15, 17, 19, 20, 22-30, 32, 33, 36-39, and 42-47), subgroup E (serotype 4), subgroup F (serotypes 40 and 41), or any other adenoviral serotype. Preferably, the adenoviral vector is based on, derived from, or consists of a serotype-2 or serotype-5 adenovirus. In some aspects, type 35 adenoviral vectors (as described in, e.g., International Patent Application WO 01/41814) can be advantageous, particularly where immune response to the vector is a concern. [0166] Regions of the adenoviral genome (e.g., the E3 region) in the adenoviral vector can optionally and preferably will be deleted in order to provide space for insertion of the polynucleotide or other nucleic acid sequences. In addition, regions of the adenoviral genome can be deleted or altered in order to interfere with viral replication. The adenoviral vector used in the inventive method preferably is deficient in at least one gene function required for viral replication, thereby resulting in a "replication-deficient" adenoviral vector. Preferably the adenoviral vector will be deficient in at least one essential gene function of the El, E2, and/or E4 regions of the adenoviral genome. More preferably, the adenoviral vector is deficient in at least one essential gene function of the El region (e.g., deficient in at least part of the El a region and/or at least part of the Elb region) of the adenoviral genome. Other portions of the genome also can be deleted, e.g., typically the E3 region, which is non-essential for viral replication. Thus, the adenoviral vector can be lacking multiple adenoviral gene functions, e.g., at least one essential gene function of the El region and at least one essential gene function of the E4 region, in addition to at least part of the E3 region. Examples of El -deleted and other replication deficient adenoviral vectors are disclosed in, for example, U.S. Patents 5,851,806, 5,985,655, and 5,994,106 and International Patent Applications WO 95/34671 and WO 97/21826. The adenoviral vector desirably retains at least one adenovirus inverted terminal repeat (ITR) (preferably the 5' and 3' ITRs). The adenoviral vector also desirably retains the adenovirus packaging sequence. Preferably, the recombinant adenovirus also comprises a mutation in the major late promoter (MLP), as discussed in International Patent Application WO 00/00628. Alternatively or additionally, the adenoviral vector can retain an E3-gpl9K coding sequence, particularly where immune response to the adenoviral vector is a concern. [0167] The vector, including the viral vector, can integrate into the host cell's genome or be a non-integrative vector. Non-integrative vectors, e.g., naked DNA plasmids, and particularly non-integrative viral vectors (e.g., adenoviral vectors), are typically preferred. Alternatively, a lentiviral vector, naked DNA vector comprising integration-promoting sequences (as described in, e.g., International Patent Applications WO 98/41645 and WO 98/54345), or AAV viral vector that integrates into the host cell's genome can be used. Additionally, when using a non-integrating viral vector, control sequences that allow for retention of the delivered transgene in the host cell, either by integration into the target cell genome or by maintenance as an episomal nucleic acid can be utilized (as discussed in, e.g., International Patent Application WO 98/54345).

[0168] A particularly preferred adenoviral vector for use in the inventive method is deficient in the entire El a region, at least part of the Elb region, and at least part of the E3 region of the adenoviral genome and contains a DNA encoding a fusion protein comprising an Ang-1, KAP, or angiopoietin-like FLD directly fused to a sequence encoding a NLl, NL5, or Ang-2X CCD, under the control of the CMV IE promoter in the El region of the adenoviral genome. Such a vector supports in vivo production of the fusion protein that is maximized at one day following administration and is not detectable above baseline levels as little as one week after administration.

[0169] The adenoviral vector can be subject to any number of additional or alternative modifications. For example, a particularly preferred vector comprises a replication- deficient adenoviral vector which includes or produces (by expression) a modified adenoviral protein, non-adenoviral protein, or both, which increases the efficiency that the vector infects cells as compared to wild-type adenovirus, allows the vector to infect cells which are not normally infected by wild-type adenovirus, results in a reduced host immune response in a mammalian host as compared to wild-type adenovirus, or any combination thereof. Any suitable type of modification can be made to the vector, and several suitable modifications are known in the art. For example, the adenoviral vector coat protein can be modified. Examples of such modifications include modifying the adenoviral fiber, penton, pLX, pllla, or hexon proteins, and/or insertions of various native or non-native ligands into portions of such coat proteins. Manipulation of such coat proteins can broaden the range of cells infected by a viral vector or enable targeting of a viral vector to a specific cell type. One direct result of manipulation of the viral coat is that the adenovirus can bind to and enter a broader range of eukaryotic cells than a wild-type virus. Examples of adeno viruses including such modifications are described in International Patent Application WO 97/20051. Reduction of immune response against the adenoviral also or alternatively can be obtained through the methods described in U.S. Patents 6,093,699 and 6,211,160 and/or U.S. Patent Application 2001-0006947A1. In other embodiments, the viral coat is manipulated such that the virus is "targeted" to a particular cell type, e.g., those cells expressing unique receptors. Examples of such modified adenoviral vectors are described in Miller et a , FASEB J, 9, 190-99 (1995), Douglas et al., Nat. Biotechnol, 14(11), 1574- 78 (1996), Wickham, Gene Ther., 7(2), 110-14 (2000), U.S. Patents 5,559,099, 5,731,190, 5,712,136, 5,770,442, 5,846,782, 5,962,311, 5,965,541, 5,985,655, 6,030,954, 6,057,155 and International Patent Applications WO 96/07734, WO 96/26281, WO 97/20051, WO 98/07865, WO 98/07877, WO 98/40509, WO 98/54346, WO 00/15823 and WO 00/31285. Other adenoviral vector protein modifications that decrease the potential for immunological recognition by the host and resultant coat-protein directed neutralizing antibody production, as described in, e.g., International Patent Applications WO 98/40509 and WO 00/34496. In non-viral vector systems, the use of targeting through targeted proteins (e.g., an asialoorosomucoide protein conjugate which promotes liver targeting (such as is described in Wu and Wu, J. Biol Chem., 263 (29), 14621-24 (1988)) or the targeted cationic lipid compositions of U.S. Patent 6,120,799).

[0170] The adenoviral vector also can include a trans-acting factor, cw-acting factor, or both, which preferably increases the persistence of transgene expression from the adenoviral vector's genome. Any suitable trans-acting factor can be used, such as HSV ICP0, which prolongs transgene expression (e.g., expression of the fusion protein sequence). Such modifications are particularly preferred in E4-deleted adenoviral vectors. The use of transacting factors is further described in International Patent Application WO 00/34496. Additionally or alternatively, the adenoviral vector comprises a nucleic acid sequence encoding a cis-acting factor. For example, a matrix attachment region (MAR) sequence (e.g., an immunoglobulin heavy chain μ (as discussed in, e.g., Jenuwein et al., Nature, 355(16), 269 (1997)), locus control region (LCR) sequences, or apolipoprotein B sequence (as discussed in, e.g., Kalos et al., Molec. Cell Biol, 15(1) 198-207 (1995)) can be used to modify the persistence of expression from a transgene, such as a transgene inserted into an E4-deleted region of the adenoviral vector genome. LCR sequences are also believed to establish and or maintain transcription of transgenes in a cis manner. [0171] The polynucleotide can be positioned within any suitable location in the genome of the adenoviral vector. Typically, the polynucleotide will substitute for one or more of the aforementioned deleted regions of the adenoviral genome (e.g., the El, E2, E3, and/or E4 region, most preferably replacing at least a portion of the El region). Alternatively, several polynucleotides encoding multiple fusion proteins, or fusion proteins and other proteins (e.g., a second angiogenic, bone growth promoting, or wound healing promoting peptide- encoding sequence) can be inserted as expression cassettes into multiple deleted regions (e.g., the ARF fusion protein encoding sequence can be inserted in a portion of the El region, and the polynucleotide encoding the second factor (e.g., a VEGFm gene) can be inserted in the deleted E3 region, or vice versa).

[0172] Production of such deficient adenoviral vectors can be accomplished by use of a complementation cell line, which is capable of providing the deleted necessary adenoviral gene functions in trans. Several examples of suitable cells are known. Examples of suitable cells for producing such vectors include 293 cells (described in, e.g., Graham et al., J. Gen. Virol, 36, 59-72 (1977)), HER cells, such as 911 cells (as described in, e.g., Fallaux et al., Human Gene Therapy, 7, 215-222 (1996)) or PER.C6 cells (commercially available through Crucell (Leiden, Netherlands)), and 293-ORF6 cells (as described in, e.g., International Patent Application WO 95/34671 and Brough et al., J. Virol, 71, 9206-13 (1997)). The cell line can provide either no homologous overlapping regions with the adenoviral vector, ideally resulting in no replication competent adenovirus (RCA) or, alternatively, can partially overlap in one or more essential regions but lack homology in one or more essential regions (as exemplified by the cells in International Patent Application WO 95/34671). Desirably, the vector composition of the invention is formed from a purified stock of such vectors. A preferred method for purifying such vector stocks is provided in International Patent Application WO 99/54441. Methods for assessing the purity of such vector compositions are provided in International Patent Application WO 00/12765.

[0173] The polynucleotide encoding the fusion protein can be inserted in any of the above-described vectors in any suitable manner and in any suitable orientation. Whereas the polynucleotide can be inserted in any suitable orientation, preferably the orientation of the nucleic acid is from right to left. By the polynucleotide having an orientation "from right to left," it is meant that the direction of transcription of the nucleic acid is opposite that of the region of the vector into which the polynucleotide is inserted. [0174] The invention further provides methods of administering the fusion proteins, polynucleotides, and vectors of the invention to a cell, tissue, or host (preferably a human), to modulate, preferably to promote angiogenesis. Thus, for example, the invention provides a method of promoting angiogenesis in an individual comprising administering to the individual an amount of a fusion protein that is effective to promote angiogenesis. Administration of the fusion protein can be performed by any suitable method, and the fusion protein can be administered in any suitable form (including by way of the polynucleotide or vector described herein). Preferably, the fusion protein (or polynucleotide or vector encoding the fusion protein) is administered in a composition, with a carrier, preferably in a pharmaceutically acceptable composition, e.g., by combination with a pharmaceutically acceptable carrier.

[0175] The term "pharmaceutically acceptable" means that the composition is a non- toxic material that does not interfere with the effectiveness of the biological activity of the fusion protein or other effective ingredients. Any suitable carrier can be used, and several carriers for administration of therapeutic proteins are known in the art. The characteristics of the carrier will depend on the route of administration. [0176] The pharmaceutical composition and/or pharmaceutically acceptable carrier also can include diluents, fillers, salts (preferably monovalent salts in a concentration of about 100-200 mM), buffers (preferably which retain a pH at about 7-8.5 at room temperature for vector compositions), detergents (preferably nonionic detergents, such as Tween-80), stabilizers, solubilizers, and/or other materials suitable for inclusion in a pharmaceutically composition. The pharmaceutical composition of the invention also can contain preservatives, antioxidants, or other additives known to those of skill in the art. When the fusion protein (or polynucleotide or vector encoding is fusion protein) is administered with other agents or ingredients the combined amounts of the agents can be administered in combination, serially or simultaneously. Examples of suitable components of the pharmaceutical composition in this respect are described in, e.g., Berge et al., J. Pharm. Sci., 66(1), 1-19 (1977), Wang and Hanson, J. Parenteral Sci. Tech., 42, S4-S6 (1988), U.S. Patents 6,165,779 and 6,225,289, and elsewhere herein. [0177] The pharmaceutical composition of the invention can be in the form of a liposome in which the fusion protein (or polynucleotide or vector encoding the fusion protein) is combined, in addition to other pharmaceutically acceptable carriers, with amphipathic agents such as lipids which exist in aggregated form as micelles, insoluble monolayers, liquid crystals, or lamellar layers in aqueous solution. Suitable lipids for liposomal formulation include, without limitation, monoglycerides, diglycerides, sulfatides, lysolecithin, phospholipids, saponin, bile acids, and the like. Preparation of such liposomal formulations is described in, e.g., U.S. Patents 4,837,028 and 4,737,323. [0178] The pharmaceutical composition can be delivered to the individual by any suitable route of administration. Examples of suitable routes of administration include oral ingestion, inhalation, bucal application, rectal application, vaginal application, topical application, insufflation, implantation, transmucosal administration, or cutaneous, subcutaneous, intraperitoneal, parenteral, myocardial, pericardial (e.g., intrapericardial), or injection (e.g., intravenous injection). Intravenous administration and injection are preferred.

[0179] If the pharmaceutical composition is administered orally, the composition preferably is administered in the form of a tablet, capsule, powder, solution, elixir, or troches. Oral compositions can include any suitable carriers or other agents. For example, tablets will typically contain a solid carrier, such as a gelatin. Generally, oral compositions also can include binders (e.g., microcrystalline cellulose, gum tragacanth or gelatin), excipients (e.g., starch or lactose), disintegrating agents (e.g., alginic acid, Primogel, or corn starch), lubricants (e.g., magnesium stearate or Sterotes), glidants (e.g., colloidal silicon dioxide), and/or sweetening/flavoring agents. Oral compositions preferably contain about 5-95%, preferably about 25-90%>, fusion protein (or polynucleotide or vector encoding the fusion protein). [0180] To administer the fusion protein (or polynucleotide or vector encoding the fusion protein) in a liquid form, such as in delivery by injection, a liquid carrier such as water, petroleum, physiological saline, bacteriostatic water, Cremophor ELTM (BASF, Parsippany, NJ), phosphate buffered saline (PBS), or oils can be used as a carrier. Liquid pharmaceutical compositions can further contain physiological saline solution, dextrose (or other saccharide solution), polyols, or glycols, such as ethylene glycol, propylene glycol, PEG, coating agents which promote proper fluidity, such as lecithin, isotonic agents, such as manitol or sorbital, organic esters such as ethyoleate, and absoφtion-delaying agents, such as aluminum monostearate and gelatins. When administered in liquid form, the pharmaceutical composition preferably contains about 0.5-90% by weight (wt.%) of fusion protein (or polynucleotide or vector encoding the fusion protein), more preferably about 1- 50 wt.%) fusion protein (or polynucleotide or vector encoding the fusion protein). [0181] More particularly, when the pharmaceutical composition is administered by injection, the composition will preferably be in the form of a pyrogen-free, stable, parenterally acceptable aqueous solution. Preferably, the parenterally acceptable aqueous solution comprises an isotonic vehicle such as sodium chloride injection, Ringer's injection, dextrose injection, lactated Ringer's injection, or equivalent delivery vehicle (e.g., sodium chloride/dextrose injection).

[0182] In a particularly preferred aspect, the fusion protein polynucleotide, or vector is administered in or near the heart. Administration in or near the heart can be to any suitable heart-associated region or tissue, using any suitable technique. Examples of suitable types of administration include direct (needle or biolistic) intracoronary injection (e.g., of a vector composition) and/or intracoronary administration using implant devices (e.g., a fusion protein coated coronary stent). Pericardial, myocardial, and intracoronary administration are particularly preferred for angiogenic fusion proteins used to treat vascular occlusion in an individual's heart.

[0183] The fusion protein (and polynucleotide or vector, as applicable) can be desirably co-administered with another factor, preferably a factor that promotes angiogenesis, bone growth, or wound healing (most preferably angiogenesis). Where the fusion protein, polynucleotide, or vector is administered with a wound healing or bone growth-promoting factor, preferred methods of administration (e.g., topical administration systems, implants, and specialized compositions) preferably will be one set forth in U.S. Patent Application 09/832,355 and references cited therein. The second factor can be a separate protein, or a vector or polynucleotide that produces such a protein.

[0184] Preferably, the second factor is a non-heparin-binding VEGF, a polynucleotide comprising a nucleic acid sequence encoding a non-heparin-binding VEGF, or a vector comprising a polynucleotide encoding a non-heparin-binding VEGF. As discussed above, the fusion protein also can comprise non-ARF peptide portions that assist in promoting angiogenesis or other desired biological activities. Furthermore, the polynucleotide of the invention can comprise one or more nucleic acid sequences encoding additional factors that promote angiogenesis or other desired biological activity as well as the fusion protein of the invention.

[0185] Administration devices for the fusion protein, polynucleotide, or vector can be formed of any suitable material. Examples of suitable matrix materials for producing non- biodegradable administration devices include hydroxapatite, bioglass, aluminates, or other ceramics. In some applications, a sequestering agent, such as carboxymethylcellulose (CMC), methylcellulose, hydroxypropylmethylcellulose (HPMC), or autologous blood clot, can be used to prevent the fusion protein complex from disassociating from the device and/or matrix. Thus, such sequestering agents are preferably present in an amount which prevents desoφtion of the fusion protein from the matrix/device and/or provides better handling of the composition. Typically, such sequestering agents will make up about 0.5-20 wt.%), preferably about 1-10 wt.%, of the composition, based on total formulation weight. [0186] For administration by inhalation, the fusion protein (or polynucleotide or vector encoding the fusion protein) can be delivered in the form of an aerosol spray from pressured container or dispenser which contains a suitable propellant, e. g., a gas such as carbon dioxide, or a nebulizer.

[0187] For transmucosal or transdermal administration, penetrants appropriate to the barrier to be permeated are preferably included in the composition. Such penetrants are generally known in the art, and include, for example, for transmucosal administration, detergents, bile salts, and fusidic acid derivatives. Transmucosal administration can be facilitated through the use of nasal sprays or suppositories.

[0188] The invention further provides compositions comprising the fusion protein (or polynucleotide or vector encoding the fusion protein) for the preparation of pharmaceuticals or other medicaments, e.g., for the preparation of sterile injectable solutions (desirably any of the other compositions described herein also are sterile). Such powder compositions can be prepared by, e.g., vacuum drying and freeze-drying, which yields a powder of the active ingredient plus any additional desired ingredient from a previously sterile-filtered solution thereof. The compositions of the invention also or alternatively can be in provided in unit dose containers and devices, including ampoules, disposable syringes, or multiple dose vials.

[0189] Additional pharmaceutically acceptable carriers are known in the art. Examples of additional suitable carriers are described in, e.g., Urquhart et al., Lancet, 16, 361 (1980), Liebeπnan et al., PHARMACEUTICAL DOSAGE FORMS - DISPERSE SYSTEMS (2nd ed., vol. 3, 1998), Ansel et al., PHARMACEUTICAL DOSAGE FORMS & DRUG DELIVERY SYSTEMS (7th ed. 2000), REMINGTON'S PHARMACEUTICAL SCIENCES, and U.S. Patents 5,708,025 and 5,994,106.

[0190] The specific amount of fusion protein (or polynucleotide or vector encoding the fusion protein) in a dosage of the composition administered to the individual will depend upon the biological effect desired in the individual, condition to be treated, and/or the specific characteristics of the fusion protein (or polynucleotide or vector encoding the fusion protein) and individual. Preferably, the pharmaceutical composition is administered in a therapeutically effective amount. A "therapeutically effective amount" means an amount sufficient to show a meaningful benefit in an individual, i.e., promoting at least one aspect of angiogenesis other desired biological effect (e.g., receptor binding), or treatment, healing, prevention, or amelioration of other relevant medical condition(s). Therapeutically effective amounts may vary depending on factors such as those described above. Thus, the attending physician (or other medical professional responsible for administering the composition) will typically decide the amount of fusion protein with which to treat each individual patient. Generalized guidance in making such determinations can be found, for example, in Platt, Clin. Lab Med., 7, 289-99 (1987), and in "Drug Dosage," J. Kans. Med. Soc, 70(1), 30-32 (1969).

[0191] Proper dosage can be determined by any suitable technique. In a simple dosage testing technique, low doses of the composition are administered to a test subject or system (e.g., an animal model, cell-free system, or whole cell assay system). Larger doses of the composition then can be administered until the desired therapeutic effect is obtained. For example, Doppler imaging can be used to detect blood flow and/or microscopy can be used to detect changes in blood vessel number or quality. Preferably, the dosage is within a range that includes the ED₅₀, with low average toxicity. Such dosages are expected to typically contain about 0.01 mg-100 mg, preferably about 0.1-10 mg, more preferably about 0.1-1 mg, of fusion protein per kg body weight. Dosages with respect to vectors containing polynucleotides encoding the fusion protein are described elsewhere herein; however, it should be understood that the discussion provided with respect to dosage and administration of the fusion protein and of such vectors are considered interchangeable unless explicitly stated otherwise or clearly contradicted by the text.

[0192] The invention further provides a method of producing the fusion protein by introducing a vector containing a polynucleotide, which, when expressed, results in the production of a fusion protein of the invention, into a suitable cell, such that the nucleotide sequence is expressed and the fusion protein is produced. The vector can be introduced into a suitable host cell for puφose of producing the fusion protein, which is then substantially isolated, preferably purified, which can be administered to an individual as described above. Any cell permissive for the uptake and maintenance of the vector and expression of the polynucleotide can be suitable. Examples of suitable cells include bacterial cells, such as E. coli and mammalian cloned cells, such as HeLa cells, CHO cells, and VERO cells. Transformation and/or transfection of such cells can be accomplished using techniques described herein or in Sambrook and Ausubel, supra and U.S. Patent 5,371,017. The fusion protein produced in the host cell can be identified and substantially isolated (preferably completely isolated) using standard techniques, including genetic selection, cell surface display, phage and virus display, ribosome display, fluorescence-based cell sorting, and agar plate screening (preferably combined with automated colony picking). Where fewer candidates need to be screened, more sensitive and faster techniques such as HPLC, mass spectrometry, gas chromatography, or chromogenic techniques can be applied. [0193] The invention also provides a cell comprising a polynucleotide encoding any of the fusion proteins or other novel ARFs described herein. The cell can be any suitable cell, including any of the aforementioned cells or cells described elsewhere herein. The invention further provides in this respect a non-human transgenic animal comprising such a polynucleotide. The production of such transgenic animals are known in the art. [0194] Alternatively, and preferably, such a vector is administered to an individual (e.g., a mammalian host, such as a human), resulting in the in vivo production of the fusion protein. In vivo administration of the fusion protein by way of such vectors offers several advantages over direct protein administration, including, e.g., avoidance of the first pass effect and other metabolically-related processing problems, providing intracellular production and processing, and providing sustained administration over a period of time, thereby resulting in less need for repeated administration events. The vector containing the fusion protein-encoding polynucleotide will preferably be administered to an area of the individual's body such that it promotes angiogenesis.

[0195] The vector is desirably administered near one or more angiogenically functional locations (source locations) and at least one angiogenically dysfunctional location (target location). Desirably, the vector (or fusion protein) composition is administered in a gradient forming manner, as described in International Patent Application WO 01/34179, particularly where the vector comprises a polynucleotide sequence comprising a second nucleic acid sequence encoding an angiogenic factor (e.g., a VEGF). The source location can be any location in the individual (e.g., tissue or organ), which has physiologically normal levels of blood perfusion, such as an area near or imbued with existing blood vessels (e.g., a non-ischemic area). The target location preferably is an actual or potentially angiogenically dysfunctional location, e.g., a location in the host that is either undergoing or is at risk of undergoing ischemia or any other condition wherein the growth of new, or extension of existing, blood vessels is desirable. Thus, the target location typically will be suffering from or be at risk of suffering from ischemic damage, which results when the tissue is deprived of an adequate supply of oxygenated blood. The interruption of the supply of oxygenated blood is often caused by a vascular occlusion. Such vascular occlusion can be caused by arteriosclerosis, trauma, surgical procedures, disease, and/or other indications. There are many ways to determine if a tissue is at risk of suffering ischemic damage from undesirable vascular occlusion including, e.g., ^mTc-sestamibi scanning, x-ray imaging, Doppler imaging, and MRI scanning. The target location also can comprise a tissue in which blood flow is attenuated by trauma, surgery, or other events. The alleviation of such attenuated blood supply, regardless of its origin, is contemplated by the invention. Thus, prevention or alleviation of damage from indications such as myocardial ischemia (particularly in patients suffering from insulin dependent diabetes), delayed wound healing, Buerger's disease, and stroke are contemplated.

[0196] Additionally, the planning of a surgical procedure can be predictive of the interruption of blood supply through a particular portion of a patient's vasculature. Prior treatment according to the method of the invention can substantially improve the desired outcome of these surgeries. In that case, treatment preferably occurs about one day to about six weeks before the surgery, and more preferably about two to about fourteen days prior to surgery. Other prophylactic uses of the vector also are contemplated. [0197] The target and source locations can be in any suitable tissue susceptible to new blood vessel growth upon production of a therapeutic amount of the angiogenic fusion protein. For example, the target and source locations can be located in a discrete organ such as the brain, heart, pancreas, limbs, or generalized areas of the body, such as a leg or a foot. Preferably, the target location and source location comprise portions of an organ system that includes at least two arteries (e.g., a heart which comprises at least three major arteries). In such aspects, the target location typically comprises at least a portion of an angiogenically dysfunctional artery in the system (e.g., an artery suffering from vascular occlusion), and some, if not all, of the angiogenically functional arteries in the system serve as source locations. In such aspects, the angiogenic mediator preferably is administered in a distribution between the target artery and the source arteries. Where the target location is an artery suffering from vascular occlusion, the method can comprise administration of the vector upstream, downstream, or to the occluded region of the artery (i.e., with respect to normal blood flow), or any combination thereof, as desired, preferably such that induced collateral blood vessel development bypasses the occluded region. "Tissue" in this sense is thus meant to include interstitial spaces associated with solid tissue. The source and target locations also can comprise cavities or extracellular fluid next to a tissue. [0198] The polynucleotide or vector can be administered in the form of a composition, e.g., with or in any suitable acceptable carrier, preferably a pharmaceutically acceptable carrier, such as those described elsewhere herein. Additional pharmaceutically acceptable carriers particularly suitable for administration of vectors are described in, for example, International Patent Application WO 98/32859.

[0199] The desired dosage (i.e., total dosage to the host) of the vector composition is such that the amount of fusion protein produced by expression of the polynucleotide in the vector results in a therapeutic and/or prophylactic effect in the area where the vector is administered. The dosage will depend on the type of fusion protein to be produced. Because a wide range of suitable fusion proteins are provided by the invention, dosage is described generally, augmented by examples relating to specific vector compositions. It will be understood that this type of description is meant to further illustrate the invention without limiting it to any particular vector composition.

[0200] The dosage of an adenoviral vector containing a fusion protein-encoding polynucleotide will be at least about lxlO⁶ particle forming units (pfu) (e.g., lxlO⁶- lxlO¹³ pfu) to an area near, at, or between the target and source locations. The dose preferably is at least about lxlO⁷ pfu (e.g., about lxlO⁷ - lxlO¹³ pfu), more preferably at least about lxlO⁸ pfu (e.g., about lxlO⁸ - lxlO¹¹ pfu), and most preferably at least about lxlO⁹ pfu (e.g., about lxlO⁹ - lxlO¹⁰ pfu). The dose typically is for a volume of targeted tissue of about 0.5-15 cm , but can be for larger tissue volumes of up to 100 cm or even about 150 cm . The dose desirably is administered via multiple applications and, as such, is divided among the multiple applications. Thus, if the dose is administered via 10 administrations, each administration involves about 1x10 - 1x10 pfu. Preferably, each application involves about lxlO⁶ - lxlO¹² pfu, more preferably about lxlO⁷ - lxlO¹⁰ pfu, and most preferably about lxl 0⁸ - lxl 0⁹ pfu. For puφoses of considering the dose in terms of particle units (pu), also referred to as viral particles, it can be assumed that there are 100 particles/pfu (e.g., lxlO¹² pfu is equivalent to lxlO¹⁴ pu). In a single round of vector administration, using, for example, an adenoviral vector deleted of the entire Ela region, part of the Elb region, and part of the E3 region of the adenoviral genome, wherein the vector comprises a nucleic acid encoding, e.g., an NLl FLD/Ang-2X CCD fusion protein under the control of a standard CMV immediate early promoter, about 10⁷-10¹³ pfu, preferably about 10⁹-10^u pfu, are administered to the host (e.g., to a discrete organ containing the source and/or target locations) with an estimated volume of about 150 cm³. Under these conditions, a substantial level of fusion protein production is achieved in the tissue of interest without producing detectable levels of fusion protein production in distal tissues. [0201] The vector composition can be administered to the individual by any suitable technique, including those techniques described herein with respect to fusion protein-containing compositions or polynucleotides and vectors. Preferably, the vector is injected into the individual. Injection can be performed in any suitable tissue or body part (e.g., intravenously, myocardially, parenterally, intrathecally, intradermally, intraspinally, intraocularly, subdermally, or into the interstitial space of a tissue/organ (e.g., of a muscle tissue)). By the term "injecting," it is meant that the vector containing solution is forcefully introduced into the target tissue. The vector composition can be microinjected, injected directly by a needle, or injected by biolistic injection. Injection can be performed using any suitable device, such as the device described in U.S. Patent 5,846,225. Alternatively, the vector containing composition can be delivered by means of percutaneous administration, typically by use of a device, such as a catheter (e.g., inserted into the femoral artery) or by a stent coated with a suitable vector containing composition (e.g., which is placed in a suitable artery, such as a coronary artery).

[0202] The vector alternatively or additionally can be administered to any suitable surface, either internal or external, at or near the source and/or target locations. For example, with respect to directly injecting a vector containing a polynucleotide encoding an angiogenic fusion protein into cardiac tissue, it is contemplated that such an injection can be administered from any suitable surface of the heart (i.e., the angiogenic mediator can be administered endocardially, epicardially, and/or pericardially). Typically and preferably, cardiac administration will be to or in the left free ventricular wall of the heart that is easily accessible by minimally invasive thoracotomy. Alternatively, administration to other areas of the heart (e.g., the septum and/or right ventricle) can be accomplished by use of a catheter or other percutaneous delivery device. Such alternate techniques can be desired where the target location is positioned in the heart but away from the left free ventricular wall (e.g., where the target location is a vascular occlusion in the right coronary artery). For wounds at or near the skin surface, topical and/or transdermal administration of vectors containing polynucleotides encoding wound healing fusion proteins are often preferred routes of administration.

[0203] Adenoviral vectors can be used for short-term or intermediate term expression of the fusion protein in dosages such as those described above. Where longer expression (e.g., about three months, about six months, about nine months or longer) is desired, retroviral vectors (e.g., lentivirus vectors) or adeno-associated viral (AAV) vectors can be advantageously used (as described in, e.g., Buschacher et al., Blood, 5(8), 2499-504, Carter, Contrib. Microbiol, 4, 85-86 (2000), Smith-Arica, Curr. Cardiol Rep., 3(1), 41-49 (2001), Taj, J. Biomed. Set, 7(4), 279-91 (2000), Vigna et al., J. Gene Med., 2(5), 308-16 (2000), Klimatcheva et al., Front. Biosci., 4, D481-96 (1999), Lever et al., Biochem. Soc. Trans., 27(6), 841-47 (1999), Snyder, J. Gene Med., 1(3), 166-75 (1999), Gerich et al., Knee Surg. Sports Traumatol Arthrosc, 5(2), 118-23 (1998), and During, Adv. Drug Deliv. Review, 27(1), 83-94 (1997), and U.S. Patents 4,797,368, 5,139,941, 5,173,414, 5,614,404, 5,658,785, 5,858,775, and 5,994,136, as well as other references discussed elsewhere herein). Alternatively, polynucleotide vectors can be used, or host integrative techniques can be employed.

[0204] Non-polynucleotide and non-protein factors that are useful for co-administration with the fusion protein, polynucleotide, and vector of the invention, particularly where a polynucleotide encoding a second factor that promotes angiogenesis, bone growth, or wound healing is administered, also can be administered to the cell, tissue or host. Preferred examples of such factors are set forth in U.S. Patent Application 09/832,355. [0205] The fusion protein, polynucleotide, or vector can be administered by or in association with ex vivo delivery of cells, tissues, or organs. Therefore, for example, a target tissue can be removed, contacted with the vector composition, and then reimplanted into the host (e.g., using techniques described in or similar to those provided in Crystal et al., Cancer Chemother. Pharmacol, 43(Suppl), S90-S99 (1999)). Ex vivo administration of an angiogenic fusion protein, or preferably angiogenic vector composition, to the target tissue also helps to minimize undesirable induction of angiogenesis in non-targeted tissue. A specific example of such a technique is the administration of an angiogenic vector composition to a tissue flap in surgical procedures involving replacement and/or transfer of tissue flaps (e.g., in breast reconstruction). "Tissue flaps" thus can comprise portions of removed tissue from a living tissue, a tissue of the recently deceased, a tissue from a different species (e.g., a pig tissue, preferably a tissue that is modified to exhibit a reduced immune response upon application to a human), or a synthetically generated tissue. Examples of suitable tissues are described in, e.g., U.S. Patent 6,140,039. Cultures of cells, particularly three-dimensional cultures, which can be a suitable substitute, additive, or alternative to such tissues, also can be administered in association with the fusion protein, polynucleotide, or vector of the invention. Examples of suitable cultures in this respect are provided in U.S. Patents 6,039,760, 6,022,743, 5,902,741, 5,863,531, 5,858,721, 5,849,588, 5,843,766, 5,830,708, 5,785,964, 5,624,840, 5,580,781, 5,578,485, 5,541,107, 5,518,915, 5,516,681, 5,516,680, 5,512,475, and 5,510,254. Related methods and compositions are provided in, e.g., U.S. Patents 6,121,042, 6,060,306, 6,027,306, 6,008,049, 5,928,945, 5,842,477, 5,780,295, 5,714,588, and 5,559,022. Cells that are genetically transformed with the polynucleotides or a host genome integrative vector also can be administered in an ex vivo manner to the host (e.g., using the techniques described in, or similar to those described in, U.S. Patent 5,399,346). For example, keratinocytes or fibroblasts can be cultured in vitro, transformed so as to produce wound healing fusion protein at high levels, and subsequently administered to a wound site (typically re-administered), thereby effecting long term production of the wound healing fusion protein (via expression), which is particularly preferred in skin regeneration (e.g., in treating severe burns). In other aspects, the fusion protein, polynucleotide, or vector of the invention can be administered with a cell that is modified ex vivo, such as activated macrophages (particularly where the treatment of spinal injuries is sought).

[0206] As previously mentioned, the fusion protein, polynucleotide, vector, together or separately can be co-administered with any suitable factor, preferably a factor which promotes angiogenesis, wound healing, bone growth, related biological activity, or enhances the activity of the fusion protein, polynucleotide or vector. Thus, in some situations, combinations of fusion protein, polynucleotide, or vector and another factor (e.g., bone growth promoting, angiogenic, or wound healing promoting protein), or co- administration of the vector and fusion protein can be desirable. Such co-administration can facilitate a multi-dimensional approach to the treatment of diseases. For example, in the context of angiogenesis-related disorders, such as vascular ischemia, the administration of the fusion protein, fusion protein-encoding polynucleotide, or vector comprising such a polynucleotide can be associated with the administration of a smooth muscle tension modifier (e.g., a vasodilator, such as a direct vasodilator (e.g., hydralazine, minoxidil, reseφine, or combinations thereof), an atrial natriuretic peptide, a vasoactive intestinal peptide, a histamine, an epinephrine or modified epinephrine (e.g., a β-2 receptor targeted epinephrine homolog or a naturally occurring epinephrine administered in a β-2 receptor- targeting manner), a bradykinin, a paracrine which induces vasodilatation (e.g., adenosine, carbon dioxide, hydrogen ion, nitric oxide, or an endothelin), an ACE inhibitor (e.g., an ACE2 inhibitor), an adrenergic receptor blocker, a vascular-associated parasympathetic nervous system stimulator (e.g., acetylcholine), an angiotensin Il-receptor blocker (ARB - e.g., tasosartan), and/or a calcium channel blocker). Other suitable non-vasodilator compounds which lower vascular resistance can be administered, and/or the application of mechanical techniques for lowering resistance (and, thus, increasing blood flow) can be applied, near or at tissues associated with the administration of the angiogenic fusion protein, fusion protein-encoding polynucleotide, or vector, and/or at one or more distal/peripheral tissues. Additionally, one or more biologically active catecholamines can be co-administered in association with the fusion protein, polynucleotide, or vector, particularly in association with the administration of an angiogenic fusion protein, polynucleotide, or vector to or near the heart. When an angiogenic fusion protein, polynucleotide, or vector is administered as a prophylactic (e.g., to a tissue at risk of ischemia due to an imminent vascular occlusion), co-administration of a factor which reduces the risk of occlusion, e.g., an anti-coagulant (such as a heparin, antithrombin III, a plasminogen, a prostacyclin (e.g., prostaglandin I or PGI₂), Protein C, tissue plasminogen activator (t-PA), the anti-coagulants described in U.S. Patent 6,121,435, or homologs thereof), or an LDL cholesterol reducing factor (e.g., a bile acid sequestrant, such as cholestyramine, colestipol, and nicotinic acid (niacin), a statin (HMG CoA reductase inhibitor), such as, lovastatin, pravastatin, simvastatin, and atorvastatin (Lipitor), rosuvastatin calcium (Crestor), an endothelin agonist (e.g., tezosentan), a gemfibrozil, a probucol, or a clofibrate) also is contemplated. Certain statins (e.g., simvastatin, atorvastatin, mevastatin) have also been reported to stimulate endothelial progenitor cells which may lead to vasculogenesis (Llevadot et al., J Clin Invest 108, 399-405 (2001) and Dimmeler et al., J. Clin. Invest. 108, 391-397 (2001)). Administration of the fusion protein, polynucleotide, or vector can be in conjunction with a surgical method where an occlusion is removed, or where lipids (e.g., LDL cholesterol) are removed from cells which then are re-administered (i.e., an autotransplant). Administration of the fusion protein, polynucleotide or vector with or in association with a fast-twitch protein (e.g., a prealbumin) can be advantageous, particularly where the treatment of congestive heart failure is desired.

[0207] In certain situations, it can be desirable to co-administer a factor that induces or promotes hematopoiesis with the fusion protein, polynucleotide, or vector of the invention. Any suitable hematopoietic factor can be co-administered in any suitable form. The hematopoietic factor can be any suitable type of hematopoietic factor. Examples of such factors include red blood cell growth promoting factors (e.g., erythropoietin (EPO)), megakaryocyte growth promoting factors (e.g., granulocyte-macrophage colony stimulating factor (GM-CSF)), eosinophil growth promoting factors (e.g., GM-CSF), neutrophil growth promoting factors (e.g., granulocyte colony-stimulating factor (G-CSF)), and monocytes growth promoting factors (e.g., macrophage colony-stimulating factor (M-CSF)). Such factors can be administered in association with an administration of stem cells or, more particularly, haematopoietic precursor cells or angioblasts, such as bone maoow derived angioblasts (as described in, e.g., Kocher et al., Nat. Med., 7(4), 430-36 (2001)), or alternatively, an administration of developed cells, such as cardiac myocytes (using techniques described in or similar to those provided in Li et al., J. Mol. Cell Cardiol, 31, 513-22 (1999)). Such cells can be obtained from a heterologous source or from a patient to which they are to be re-administered (e.g., through obtaining such cells from removed (and possibly cultured) bone marrow, blood, or fatty tissues of the individual). Similar co- administration of relevant cells can be performed for wound healing and bone growth promoting aspects of the invention (e.g., co-administration of keratinocytes in wound healing or of osteoblasts for promotion of bone growth). Co-administration of hematopoietic factors is particularly preferred in association with the administration of a wound healing fusion protein, polynucleotide, or vector of the invention. [0208] Factors that block or enhance events in the angiogenic, wound healing, or bone growth promoting pathway also can be administered in association with the fusion protein, polynucleotide, or vector. Co-administration of factors that upregulate expression of a gene or genes encoding a desired angiogenic factor, bone growth promoting factor, or wound healing promoting factor, where such a factor does not correspond or related to a peptide portion of the fusion protein also is within the scope of the invention (e.g., administration of a factor which upregulates expression of a VEGF). For example, co-administration of PLC- γ, Ras, She, Nek, PKC and/or PI3-kinase can be co-administered with the polynucleotide, vector, or fusion protein, to induce downstream signal pathways associated with VEGFR-2, as can factors which block such downstream interactions. Factors which induce VEGF gene expression, such as PDGF, keratinocyte growth factor, EGF, TNF-α, IGF-1, thyroid- stimulating hormone, IL-lα, IL-4, IL-6, TGF-β, IL-lβ, prostaglandin E2 (PGE₂), ACTH, v- Ha-ras, v-raf, and v-myc, also can be co-administered with the fusion protein, polynucleotide, or vector, as can chemical agents which upregulate VEGF gene expression, such as phorbol myristate acetate (as described in, e.g., Han et al., J. Cell Sci., Ill, 3621-31 (1998)) or other phorbol esters. The fusion protein, polynucleotide, or vector can be advantageously administered after or during administration of such a phorbol ester compound, which may induce vascular tube formation in collagenous tissues, as administration of an angiogenic fusion protein may sustain the integrity of the newly formed vascular tube and prevent endothelial cell apoptosis thereafter which might otherwise result from phorbol ester-induced angiogenesis. Progesterone can similarly be co-administered with a fusion protein, polynucleotide, or vector to upregulate HBNF gene expression. [0209] The fusion protein, polynucleotide, or vector also or alternatively can be administered in association with a factor that affects a pathway associated with the biological activity of a wild-type ARF. For example, a Bmx tyrosine kinase (as described in, e.g., Rajantie et al., Mol. Cell Biol, 21(14), 4647-55 (2001)), or polynucleotide encoding a Bmx tyronsine kinase, which is a factor that acts "downstream" of Ang-1 and VEGF in arterial endothelial cell biological pathways, can be co-administered with the fusion protein, polynucleotide, or vector of the invention. Also or alternatively, the fusion protein, polynucleotide, or vector can be administered in association with a factor that induces expression of a wild-type ARF gene. For example, the fusion protein, polynucleotide, or vector of the invention can be co-administered with angiotensin II, Her 2, or leptin (which induce Ang-2 gene expression in adipose tissues), or thrombin (which induces Ang-1 release from platelets), or a polynucleotide encoding such factors or homologs thereof. [0210] In some situations, the fusion protein, polynucleotide, or vector is desirably administered with an integrin, polynucleotide encoding an integrin, or a protein or factor that blocks the activity of an integrin, preferably of an integrin that binds to a wild-type ARF. For example, the fusion protein, polynucleotide, or vector can be administered with an α5 integrin or a factor that blocks the expression/production of α5 integrin and/or its binding to Ang-1, Ang-2, and/or the fusion protein of the invention. Also or alternatively, the fusion protein, polynucleotide, or vector can be administered with a vitronectin, a polynucleotide encoding a vitronectin, or a factor that blocks the expression/production of vitronectin or its binding to Ang-1 , Ang-2, and/or the fusion protein of the invention. The fusion protein, polynucleotide, or vector also can be administered with a factor that upregulates the expression of a receptor for the fusion protein. For example, the fusion protein, polynucleotide, or vector can be co-administered with a thyroid stimulating hormone, TNF-α, or a combination of desferirioxamine and cobalt, which upregulate Tie-2 expression.

[0211] Factors that inhibit inflammation also or alternatively can be administered with the fusion protein, polynucleotide, or vector of the invention. The inflammation inhibitor can be any suitable inflammation inhibitor. Examples of suitable inflammation inhibitors are provided in, e.g., U.S. Patent 5,830,880. In some circumstances, co-administration of a suitable factor which inhibits thrombosis can be desirable, such as the factors described in U.S. Patent 5,955,576.

[0212] Factors which are co-administered with the fusion protein, polynucleotide, or vector of the invention can be co-administered in any suitable manner, and in any suitable order (i.e., concurrently or sequentially), such as administering a fusion protein, polynucleotide, or vector of the invention and separately administering a vector containing a polynucleotide encoding such a factor (or homolog thereof), or administering a vector containing a polynucleotide encoding such a factor which also encodes a fusion protein of the invention.

[0213] Factors which reduce naturally occurring anti-angiogenic factors (e.g., an endostatin (or fragment thereof, such as the collagen XVIII fragment), angiotensin (or fragment thereof, such as the plasminogen fragment), thrombospondins (e.g., thrombospondin-1), the 16 kDa fragment of prolactin, and vasostatin (or calreticulin)), Cartilage-derived inhibitor (CDI), CD59 complement fragment, Gro-beta, Heparinases, Heparin hexasaccharide fragment, human chorionic gonadotropin (hCG), IFNs, Interferon inducible protein (IP- 10), IL-12, Kringle 5 (plasminogen fragment), 2-Methoxyestradiol, Placental ribonuclease inhibitor, Plasminogen activator inhibitor, Platelet factor-4 (PF4), Proliferin-related protein (PRP), Retinoids, Tetrahydrocortisol-S, other anti-angiogenic C- X-C chemokines, Yes and Src (as described in International Patent Application WO 01/45751), and/or vasculostatin also can be suitable for co-administration with the fusion protein, polynucleotide, or vector. For example, one or more factors which block one or more anti-angiogenic factors from binding with receptors required for activation, or which prevent cleavage or other conformational changes required for immature anti-angiogenic proteins to develop anti-angiogenic activity (e.g., blocking cleavage required for development of a mature, anti-angiogenic, endostatin, or preventing conversion of plasminogen to angiostatin), can be administered with the angiogenic fusion protein, polynucleotide, or vector. Such factors can be administered in any suitable form (e.g., as a polynucleotide inserted into a separate vector or the same vector with a fusion protein- encoding polynucleotide). Alternatively, one or more antisense polynucleotides that prevent transcription and/or translation of an anti-angiogenic gene, or one or more monoclonal antibodies that deactivate the anti-angiogenic factor or block its activity. In some situations, the inventive method comprises blocking the expression or activity of native Ang-2, such that the angiogenic activity of the fusion protein is not impeded by it, particularly where the fusion protein contains an Ang-1 peptide portion. [0214] Administration of anti-angiogenic factors or angiogenic factor antagonists in association with the administration of angiogenic fusion proteins, polynucleotides, and/or vectors can be desirable in some conditions. For example, administration of such factors can provide control over the level of blood vessel growth and/or vascular permeability to be achieved by administration of the fusion protein, polynucleotide, or vector. In a preferred aspect, the fusion protein, fusion protein-encoding polynucleotide, or fusion protein- producing vector is administered with a PEDF anti-angiogenic protein, PEDF polynucleotide (or PEDF homolog polynucleotide), or vector comprising such a PEDF polynucleotide (such proteins, polynucleotides, and vectors are described in, e.g., U.S. Patent 5,840,686 and International Patent Applications WO 95/33480 and WO 01/58494). Alternatively or additionally, the second nucleic acid sequence can comprise a PEDF gene or homolog thereof. Such methods are particularly preferred wherein the fusion protein promotes blood vessel maturation. A particularly preferred combination of PEDF protein, PEDF polynucleotide (whether separate from the fusion protein or fusion protein-producing constructs of the invention or as the second nucleic acid sequence encoded by the fusion protein-encoding polynucleotide), or PEDF vector and a fusion protein, fusion protein polynucleotide, or fusion protein-encoding vector, comprises co-administration to the eye to treat or prevent macular degeneration.

[0215] The fusion protein of the invention can be used in any manner that is suitable for one of its constituent peptide portions (examples of such uses are described herein and/or in the references cited herein). Thus, a fusion protein comprising, for example, portions with significant homology to Ang-2, NL4, or CDT6 can be suitable for inhibition of angiogenesis. This includes the use of the fusion protein for treatment of ocular diseases dependent on angiogenesis, such as corneal graft neovascularization, retrolental fibroplasias, diabetic retinopathy, and neo vascular diseases of the eye (as described in, e.g., International Patent Application WO 01/29085). Additionally, the fusion protein can be used to treat other diseases which are associated with angiogenesis, including angiofibroma, Osier- Weber syndrome, psoriasis, solid tumors, trachoma, hemophilic joints, vascular adhesions, and hypertrophic scars (as described in, e.g., International Patent Application WO 01/29085). The fusion protein can also be used to treat immune related and inflammatory diseases (as described in, e.g., International Patent Application WO 00/73452). In contexts where the fusion protein, polynucleotide, or vector is used to treat cancer, for example by reducing angiogenesis, the fusion protein, polynucleotide, or vector can be co-administered with tumor reducing agents (e.g., chemotherapy, radiation therapy, or reoviruses).

[0216] In addition to the other administration techniques described herein, the vector composition or fusion protein composition can be administered by direct surgical implantation. Alternatively or additionally, the fusion protein and/or vector composition can be co-administered with a group of therapeutic cells, e.g., stem cells, macrophages, or neurophils. For example, an angiogenic vector composition of the invention can be co- administered with stem cells to an ischemic location in the heart. The use of the vector composition and fusion protein of the invention also can be useful in organ generation and organ transfer. Stem cell administration can be associated with administration of stem cell- determining factors, such as a Casanova gene homolog (as described in, e.g., Stainier, Genes Devel, 15, 1493-1505 (2001)) or other suitable determining factor. Cord blood also can be advantageously co-administered with the fusion protein, polynucleotide, or vector. The administration of the fusion protein, polynucleotide, or receptor also can be advantageously associated with percutaneous in situ coronary vernous arterialization (PICVA) (as described in e.g., Vesterle et al., Circulation, 103(21), 2539-43 (2001)). [0217] The angiogenic fusion protein and vector compositions of the invention can be used to treat a wide variety of ailments including, e.g., coronary artery disease, peripheral vascular disease, congestive heart failure (e.g., left ventricular dysfunction and left ventricular hypertrophy), neuropathy (peripheral or otherwise), avascular necrosis (e.g., bone or dental necrosis), mesenteric ischemia, impotence (or erectile dysfunction), incontinence, arterio-venous fistula, veno-venous fistula, stroke, cerebrovascular ischemia, muscle wasting, pulmonary hypertension, gastrointestinal ulcers, vasculitis, non-healing ischemic ulcers, retinopathies, restenosis, cancer, orthosclerosis, radiation-induced tissue injury (such as that common with cancer treatment), and other hypoxia-associated or low blood perfusion-associated disorders. In addition, the angiogenic fusion protein and vector compositions also find utility in the study and/or aid of wound healing (e.g., healing of ischemic ulcers), plastic surgery procedures (e.g., healing or reattachment of skin and/or muscle flaps), prosthetic implant healing, vascular graft patency, and transplant longevity. Thus, the invention provides methods of treating such ailments by administration of the fusion protein and/or vector compositions.

[0218] The fusion protein, fusion protein-encoding polynucleotide, and vector of the invention are believed to be useful in several medically related contexts. For example, in surgical contexts, the fusion protein, polynucleotide, and/or vector can be used to treat orthopedic surgery-associated avascular necrosis, treat mesenteric ischemia, provide prophylaxis against ischemia in association with ostomies, treat or provide prophylaxis for thoracic ischemia related spinal chord complications (aneurysm repairs), treat sexual dysfunction (e.g., urology-postprostatectomy associated sexual dysfunction - for example in association with radial prostatectomy), provide smooth muscle tone in tissues (e.g., treat incontinence), prevent radiation-induced vascular necrosis (e.g., prevent tooth loss associated with radiation use in dentistry), promote gum and/or tooth regeneration, create and/or promote veno-venous or arterio-venous anastamosis, and enhance cartilage, tendon, and/or ligament repair replacement (either through direct healing or by promoting angiogenesis in such tissues or tissues associated therewith).

[0219] The fusion protein, polynucleotide, and vector of the invention also can be useful in neurological applications, such as inducing angiogenesis in the treatment of cerebrovascular-associated vascular obstructive disease. Pulmonary and gastrointestinal applications of the fusion protein, polynucleotide, and vector include administration in association with liver regeneration, treatment of pulmonary hypertension, and providing/increasing blood supply to a transplanted lung. Rheumatological/renal applications of the fusion protein, polynucleotide, and vector include the treatment of vasculitis, modulation of renal permeability and function, modulation of peritoneal permeability and function, and promotion or prevention of growth factor delivery to such tissues through such permeability modulation.

[0220] The fusion protein, polynucleotide, or vector also can be used to reduce or prevent tumor growth. Fusion proteins comprising an Ang-2, Ang-2X, or Ang-2 homolog peptide portion (and corresponding polynucleotides and vectors) are preferred in this respect.

[0221] A preferred aspect of the invention comprises the administration of a vector comprising a nucleic acid encoding a fusion protein of the invention under one promoter and a second nucleic acid sequence encoding a second protein, desirably an angiogenic factor (e.g., a VEGF), under the control of a second promoter, where the first promoter and second promoter have different expression profiles and/or are differentially inducible. Desirably, after administration of the vector to a suitable cell, tissue, or mammal, the first nucleic acid sequence is permitted, induced, or upregulated to express the nucleic acid sequence encoding the fusion protein, and the second nucleic acid is permitted, induced, or upregulated to express the second factor, in a manner which mimics a biological cascade as described elsewhere herein. The method can be applied with more than two factors (e.g., 3, 4, or more factors) to mimic more complex cascades. Constructs for practicing such methods are described elsewhere herein. [0222] The methods of this invention are closely related in function. Thus, it is to be understood that the disclosure with respect to any aspect of the invention can be applied to any other suitable aspect. For example, forms of administration and delivery techniques for fusion protein compositions can be used for polynucleotide or, more particularly, vector compositions, and vice versa. Similarly, a reference to administration of the fusion protein, polynucleotide, or vector of the invention encompasses the administration of a pharmaceutically acceptable composition containing the fusion protein, polynucleotide, or vector, as applicable.

[0223] The invention further provides Ang-2X, and polynucleotides encoding Ang-2X, as described in, e.g., Example 1, and modified by the techniques and principles applicable to the fusion proteins and fusion protein-encoding polynucleotides of the invention. The Ang- 2X peptide provided by the invention can be administered and produced according to the methods described herein with respect to the fusion proteins of the invention. Ang-2X can be inserted into a vector using the techniques described herein with respect to fusion protein-encoding polynucleotides of the invention. As such, the invention provides vectors comprising Ang-2X-encoding polynucleotides, methods of producing Ang-2X peptides, and methods of using such Ang-2X peptides, polynucleotides, or vectors to treat disease, and, most preferably, to modulate angiogenesis.

[0224] The methods of using the fusion proteins of the invention, Ang-2X, and related polynucleotides and vectors described herein, can be applied to any suitable ARFs, such as Ang-1, Ang-2, Ang-3, Ang-4, NLl, NL2, NL3, NL4, NL5, FLS139, NL7, NL8, and murine FDRG. Thus, the invention provides novel methods of using such factors, and polynucleotides encoding them. For example, the invention provides a replication-deficient adenoviral vector comprising a gene encoding NLl, NL5, or a codon optimized angiopoietin, and a second gene encoding another angiogenic factor (e.g., a HBNF, MK, or VEGF (preferably a non-heparin-binding VEGF, such as VEGF,₂,)). Additionally, for example, the invention provides a method of promoting angiogenesis comprising infecting a cell with such a viral vector, preferably at the dosages described elsewhere herein. The invention also provides the application of the techniques described herein to obtain novel homologs of such wild-type ARFs and the novel homologs thereby obtained. For example, the invention provides codon optimized Ang-1 genes (SEQ ID NO: 167), and vectors comprising polynucleotides encoding such genes.

[0225] The following examples further illustrate the invention but, of course, should not be construed as in any way limiting its scope. [0226]

[0227] EXAMPLE 1 [0228] This example describes the generation of the novel Angiopoietin-2 homolog Ang-2X.

[0229] Ang-2X was derived from the results of a TNBLAST search of the high-through put sequence database for the human genome project for sequences exhibiting significant levels of identity to Ang- 1. Hits were identified on B AC clone RPl l-16gl2 (GenBank accession number AC018398). Nine contigs were identified and assembled by joining

(complement 335846...336136), (complement 265610...265440), (complement

133812...133693), (complement 302082...302315), (complement 52191...52060),

(complement 238562...238455), (complement 47913...47746), (complement

141153...141079), and (complement 18606...18544) to derive the following polynucleotide sequence:

[0230]

[0231] ATGTGGCAGATTGTTTTCTTTACTCTGAGCTGTGATCTTGTCTTGGCCG

CAGCCTATAACAACTTTCGGAAGAGCATGGACAGCATAGGAAAGAAGCAATATC

AGGTCCAGCATGGGTCCTGCAGCTACACTTTCCTCCTGCCAGAGATGGACAACTG

CCGCTCTTCCTCCAGCCCCTACGTGTCCAATGCTGTGCAGAGGGACGCGCCGCTC

GAATACGATGACTCGGTGCAGAGGCTGCAAGTGCTGGAGAACATCATGGAAAAC

AACACTCAGTGGCTAATGAAGGTAGAGAATATATCCCAGGACAACATGAAGAAA

GAAATGGTAGAGATACAGCAGAATGCAGTACAGAACCAGACGGCTGTGATGATA

GAAATAGGGACAAACCTGTTGAACCAAACAGCGGAGCAAACGCGGAAGTTAACT

GATGTGGAAGCCCAAGTATTAAATCAGACCACGAGACTTGAACTTCAGCTCTTGG

AACACTCCCTCTCGACAAACAAATTGGAAAAACAGATTTTGGACCAGACCAGTG

AAATAAACAAATTGCAAGATAAGAACAGTTTCCTAGAAAAGAAGGTGCTAGCTA

TGGAAGACAAGCACATCATCCAACTACAGTCAATAAAAGAAGAGAAAGATCAGC

TACAGGTGTTAGTATCCAAGCAGAATTCCATCATTGAAGAACTCGAAAAAAAAA

TAGTGACTGCCACGGTGAATAATTCAGTTCTTCAGAAGCAGCAACATGATCTCAT

GGAGACAGTTAATAACTTACTGACTATGATGTCCACATCAAACGCAGCTAAGGA

CCCCACTGTTGCTAAAGAAGAACAAATCAGCTTCAGAGACTGTGCTGAAGTATTC

AAATCAGGACACACCACGAATGGCATCTACACGTTAACATTCCCTAATTCTACAG

AAGAGATCAAGGCCTACTGTGACATGGAAGCTGGAGGAGGCGGGTGGACAATTA

TTCAGCGACGTGAGGATGGCAGCGTTGCATTTCAGAGGACTTGGAAAGAATATA

AAGTGGGATTTGGTAACCTCTCAGAAAAATATTGGCTGGGAAATGAGTTTGTTTC

GCAACTGACTAATCAGCAACGCTATGTGCTTAAAATACACCTTAAAGACTGGGA

AGGGAATGAGGCTTACTCATTGTATGAACATTTCTATCTCTCAAGTGAAGAACTC

AATTATAGGNN π^NNNNNNNNNNNNNNNGGCAATGATTTTAGCACAAGGGATG

GAGCCACCGNCANATGTATTTGCAAATGTTCACAAATGCTAACAGNAGGTNNNN

NNNNNNNNNNNNNNNNNNNNNNNNNNTACTGGAAAGGCTCAGGCTATTCGCTC

AAGGCCACAACCATGATGATCCGACCAGCAGATTTC [SEQ ID NO: 160] [0232]

[0233] where N represents any polynucleotide. A polynucleotide having this sequence is generated using any standard polynucleotide synthesis.

[0234] The Ang-2X polynucleotide is predicted to encode a polypeptide having the following amino acid sequence:

[0235]

[0236] Met Tφ Gin He Val Phe Phe Thr Leu Ser Cys Asp Leu Val Leu Ala Ala Ala Tyr

Asn Asn Phe Arg Lys Ser Met Asp Ser He Gly Lys Lys Gin Tyr Gin Val Gin His Gly Ser

Cys Ser Tyr Thr Phe Leu Leu Pro Glu Met Asp Asn Cys Arg Ser Ser Ser Ser Pro Tyr Val

Ser Asn Ala Val Gin Arg Asp Ala Pro Leu Glu Tyr Asp Asp Ser Val Gin Arg Leu Gin Val

Leu Glu Asn He Met Glu Asn Asn Thr Gin Tφ Leu Met Lys Leu Glu Asn He Ser Gin Asp

Asn Met Lys Lys Glu Met Val Glu He Gin Gin Asn Ala Val Gin Asn Gin Thr Ala Val Met

He Glu He Gly Thr Asn Leu Leu Asn Gin Thr Ala Glu Gin Thr Arg Lys Leu Thr Asp Val

Glu Ala Gin Val Ser Asn Ala Thr Thr Arg Leu Glu Leu Gin Leu Leu Glu His Ser Leu Ser

Thr Asn Lys Leu Glu Lys Gin He Leu Asp Gin Thr Ser Glu He Asn Lys Leu Gin Asp Lys

Asn Ser Phe Leu Glu Lys Lys Val Leu Ala Met Glu Asp Lys His He He Gin Leu Gin Ser He

Lys Glu Glu Lys Asp Gin Leu Gin Val Leu Val Ser Lys Gin Asn Ser He He Glu Glu Leu

Glu Lys Lys He Val Thr Ala Thr Val Asn Asn Ser Val Leu Gin Lys Gin Gin His Asp Leu

Met Glu Thr Val Asn Asn Leu Leu Thr Met Met Ser Thr Ser Asn Cys Lys Xaa Xaa Xaa

Xaa Val Ala Lys Glu Glu Gin He Ser Phe Arg Asp Cys Ala Glu Val Phe Lys Ser Gly His

Thr Thr Asn Gly He Tyr Thr Leu Met Tφ Gin He Val Phe Phe Thr Leu Ser Cys Asp Leu

Val Leu Ala Ala Ala Tyr Asn Asn Phe Arg Lys Ser Met Asp Ser He Gly Lys Lys Gin Tyr

Gin Val Gin His Gly Ser Cys Ser Tyr Thr Phe Leu Leu Pro Glu Met Asp Asn Cys Arg Ser

Ser Ser Ser Pro Tyr [SEQ ID NO: 174]

[0237]

[0238] where Xaa represents any amino acid residue. Ang-2X is believed to have angiogenesis modulating activities similar to angiopoietins, and particularly to Ang-2.

[0239]

[0240] EXAMPLE 2

[0241] This example demonstrates the production of a polynucleotide encoding a fusion protein of the invention, the insertion of the polynucleotide into a suitable vector, and the expression of the fusion protein-encoding polynucleotide resulting in production of the fusion protein.

[0242] A first polynucleotide comprising a sequence encoding Ang-2X CCD1 is generated by standard polynucleotide synthesis techniques. A second polynucleotide comprising a sequence encoding the Ang-1 FLD is produced by standard polynucleotide synthesis techniques. The first polynucleotide and second polynucleotide are fused, using standard techniques, to form a fused polynucleotide.

[0243] The oligonucleotide primers AGTAATGCAGTCCAAAGAGACG (SEQ ID NO: 168) and AAAGTCTAGTGGTTCGATCATCA (SEQ ID NO: 169) are used to amplify a PCR product encoding a fusion protein comprising a modified Ang-2X CCDl variant, which lacks the N-terminal valine residue of Ang-2X CCDl (SEQ ID NO: 170), directly fused to the Ang-1 FLD from the fused polynucleotide. The PCR product, first polynucleotide, or third polynucleotide optionally can be fused to an additional sequence encoding a wild-type ARF signal sequence and/or variable domain (such that the variable domain and/or signal sequence is positioned N-terminal to the Ang-2X CCDl variant sequence).

[0244] The Ang-2X CCDl variant/ Ang-1 FLD fusion protein-encoding PCR product is cut with Xho I to obtain a sequence encoding a fusion protein comprising an Ang-2X CCDl -derived CCD, which lacks the ten N-terminal residues of Ang-2X CCDl (hereinafter referred to as Ang-2X CCD4) (SEQ ID NO: 171) fused to the Ang-1 FLD. The fusion protein-encoding polynucleotide (SEQ ID NO: 172) is cloned into a pAd351 ICMV transfer vector, which comprises nucleotides 1-4511 of the adenoviral serotype 5 genome, except nucleotides 353-3511 (which encompass the adenoviral EIA and E1B coding regions), the CMV promoter, a multiple cloning site (including Xho I), an S V40 poly A site, and a splice donor/acceptor site between Ad5 nucleotides 353 and 3511.

[0245] After insertion of the Xho I fragment, the recombinant transfer vector is used to generate a transfection plasmid capable of producing an El -deleted adenoviral vector containing the Ang-2X CCD4/ Ang-1 FLD fusion protein-encoding sequence positioned in the El deletion upon transfection into a suitable host cell. The transfection plasmid is generated by any suitable technique. Examples of such techniques include homologous recombination, or ligation to, one or more additional plasmids comprising the remainder of the adenoviral genome except the desired deleted regions (i.e., El, E3, and optionally other regions, e.g., the E4 region). Any suitable homologous recombination technique can be used to generate the vector-producing plasmid. Examples of such techniques are provided in, e.g., Chinnadurai et al., J. Virol, 32, 623-28 (1979), Berkner et al., Biotechniques, 6, 616-28 (1998), Chartier et al., J. Virol, 70, 4805-10 (1996), and International Patent Application WO 96/25506. A preferred homologous recombination technique is described in International Patent Application WO 99/15686. Alternatively, any suitable ligation technique can be used, such as the techniques described in, e.g., Stow, J. Virol, 37(1), 171- 80 (1981), Stow, Nucl. Acids Res., 10(17), 5105-19 (1982), and Berkner et al., Nucl. Acids Res., 11 (17), 6003-20 (1983).

[0246] After a suitable transfection plasmid containing the Ang-2X CCD4/ Ang-1 FLD fusion protein-encoding sequence is generated, the transfection plasmid is transfected into a suitable El complementing cell line, such as a 293-ORF6 cell line (described in International Patent Application WO 95/34671), using standard techniques (e.g., calcium phosphate precipitated transfection), thereby resulting in the production of a stock of El- deleted, replication-deficient, adenoviral vectors. Preferably, the vector-cell line system selected is such that replication competent adenovirus (RCA) levels in the stock are confirmed to be less than about 1 x 10⁷ plaque forming units (pfu), preferably by using the techniques described in U.S. Patent 5,994,106. Levels of viral pfu can be determined using standard techniques (such as the techniques described in Chinnadurai et al., supra and Precious et al., "Purification and Titration of Adenoviruses" in VIROLOGY: A PRACTICAL APPROACH, 193-205 (Mahay et al., Eds., IRL Press 1985)).

[0247] The Ang-2X CCD4/ Ang-1 FLD vector is administered by needle injection in an appropriate carrier to at least one target location in a mammalian host. Resultant Ang-2X CCD4/ Ang-1 FLD fusion gene expression is confirmed by mRNA expression analysis, subsequent administration of an anti- Ang-1 and/or anti-Ang-2X antibodies to the site of vector administration after sufficient time for fusion protein expression, and/or observation of the angiogenic effects of administering the vector, for example, by using the mouse ear or rat hind limb models for testing the angiogenesis-inducing capacity of a molecule, as described in more detail here.

[0248] In the mouse ear model, 10⁹- 10 ' ° particles units (pu) of the Ang-2X CCD4/Ang- 1 FLD vector and a similar dosage of AdVEGF]₂ι, which contains the human VEGF]₂ι gene, are administered to Apo E^_/" mice. All injections are delivered subcutaneously at the base of the ears of anesthetized mice (12 mg/kg xylazine and 60 mg/kg ketamine, B?). Gross moφhological changes to the target tissue are observed at various days post-injection. Serial laser Doppler perfusion measurements are taken at various time points post-injection. Changes in blood vessel number are identified using an Olympus BX40F microscope at 400X to examine harvested ears that are perfusion fixed and embedded in paraffin. Control groups receiving similarly administered

and/or null vectors are used for comparative testing.

[0249] It is expected that at about four days post-injection, administration of the Ang- 2X CCD4/ Ang-1 FLD adenoviral vector with AdVEGF₎₂ι and resulting production of the Ang-2X CCD4/ Ang-1 FLD fusion protein and VEGFι₂ι will result in the formation of blood vessels in greater number and/or volume of blood vessels which exhibit a greater level of vessel maturation than vessels resulting from administration of only AdVEGFι₂ι. [0250] It is expected that at about 14-28 days post-injection, animals receiving the Ang- 2X CCD4/ Ang-1 FLD adenoviral vector and AdVEGFι_2ϊ in the hind limb model will exhibit tissue perfusion levels higher and/or less blood vessel leakage than in control groups receiving administration of only AdVEGF_]2ι and or a null vector. [0251] This Example provides a suitable vector comprising a polynucleotide encoding an ARF fusion protein, which, when produced by expression, promotes angiogenesis. The steps recited in the present Example (e.g., selection of appropriate primers for a wild-type ARF FLD and CCD, amplification thereof to form a fusion protein-encoding polynucleotide, insertion of the fusion protein-encoding polynucleotide into a suitable vector, and expression of the polynucleotide therein to produce the fusion protein) can be applied to any suitable combination of ARF-encoding polynucleotide sequences as a viable strategy for obtaining novel ARF fusion proteins by routine experimental techniques. [0252]

[0253] All amino acid or nucleotide sequences of one of the aforementioned sequence patterns are to be considered individually disclosed herein. Thus, for example, an amino acid sequence pattern of three residues, where a "Xaa" represents one of the amino acid positions in the pattern represents a disclosure of at least twenty different sequences (i.e., one sequence for each naturally occurring amino acid residue that could be present in the Xaa position).

[0254] All references, including publications, patent applications, and patents, cited herein are hereby incoφorated by reference to the same extent as if each reference were individually and specifically indicated to be incoφorated by reference and were set forth in its entirety herein.

[0255] The use of the terms "a" and "an" and "the" and similar referents in the context of describing the invention (especially in the context of the following claims) are to be construed to cover both the singular and the plural, unless otherwise indicated herein or clearly contradicted by context. Terms such as "including," "having," "comprising," "containing," and the like are to be construed as open-ended terms (i.e., meaning "including, but not limited to") unless otherwise indicated, and as encompassing the phrases "consisting of and "consisting essentially of." Recitation of ranges of values herein are merely intended to serve as a shorthand method of referring individually to each separate value of the range, unless otherwise indicated herein, and each separate value is incoφorated into the specification as if it were individually recited herein. All methods described herein can be performed in any suitable order unless otherwise indicated herein or otherwise clearly contradicted by context. The use of any and all examples, or exemplary language (e.g., "such as") provided herein, is intended merely to better illuminate the invention and does not pose a limitation on the scope of the invention unless otherwise claimed. No language in the specification should be construed as indicating any non- claimed element as essential to the practice of the invention.

[0256] Preferred embodiments of this invention are described herein, including the best mode known to the inventors for carrying out the invention. Variations of those preferred embodiments may become apparent to those of ordinary skill in the art upon reading the foregoing description. The inventors expect skilled artisans to employ such variations as appropriate, and the inventors intend for the invention to be practiced otherwise than as specifically described herein. Accordingly, this invention includes all modifications and equivalents of the subject matter recited in the claims appended hereto as permitted by applicable law. Moreover, any combination of the above-described elements in all possible variations thereof is encompassed by the invention unless otherwise indicated herein or otherwise clearly contradicted by context.

Claims

WHAT IS CLAIMED IS:

1. A fusion protein comprising:

(a) a fibrinogen-like domain that exhibits at least about 75% amino acid sequence identity to the fibrinogen-like domain of a first wild-type protein, which binds a receptor that, when activated, promotes or inhibits angiogenesis, and

(b) a coiled-coil domain that exhibits at least about 75% amino acid sequence identity to the coiled-coil domain of a second wild-type protein, which increases the affinity of the fusion protein for the receptor, promotes formation of a multimer comprising the fusion protein, or both, wherein the first protein and second protein are different, the first protein and second protein exhibit at least about 30%> overall amino acid sequence identity with Ang-1, and the first protein, the second protein, or both proteins, are not Ang-1, Ang-2, Ang-3, or Ang-4.

2. The fusion protein of claim 1 , wherein the fibrinogen-like domain comprises a first sequence of the sequence pattern Glu Phe/Tyr/His Tφ Leu Gly Leu/ Asn Glu/ Asp Xaa Val/Ile/Leu Xaa Xaa He/Leu Xaa₍i_2-ι₄) Asp Tφ and a second sequence of the sequence pattern Phe Ser Thr Xaa Asp Xaa Asp Asn/His Xaa_3-ιo Cys Ala/Ser Xaa₄ Gly Gly Tφ Tφ Xaa(_{2- )} Cys, wherein the first sequence is positioned N-terminal to the second sequence.

3. The fusion protein of claim 2, wherein the first sequence is of the sequence pattern Phe Ser Thr Lys/Leu Asp Xaa Asp Asn Asp Xaa Cys Xaa Cys Lys Cys Ala/Ser Gin Leu Met/Val Leu/Met Tyr/Ser Gly Gly Tφ Tφ Phe Asp Ala Cys Gly Pro/Leu Ser Asn Leu Gly and the second sequence is of the sequence pattern Glu Tyr/His Tφ Leu Gly Asn Glu Xaa Val/Ile Xaa₍₂₎ He/Leu Thr Xaa Gln Arg Xaa Xaa Tyr Xaa Leu Arg/Lys Ile/Val Glu Leu Xaa Asp Tφ Glu Gly.

4. The fusion protein of claim 1, wherein the second protein is NLl, FLS 139, NL3, NL4, NL5, NL8, zapol, or murine FDRG.

5. The fusion protein of claim 4, wherein the first protein is the KIA003- associated protein (KAP) or an angiopoietin.

6. The fusion protein of claim 1, wherein the first protein is NLl, FLS 139, NL3, NL4, NL5, NL8, zapol, or murine FDRG.

7. The fusion protein of claim 6, wherein the second protein is an angiopoietin.

8. The fusion protein of claim 1, wherein the first protein or the second protein is Ang-2X.

9. The fusion protein of claim 1 , wherein

(a) the fibrinogen-like domain and coiled-coil domain are directly fused and the amino acid sequence within about 20 amino acid residues of the junction of the fibrinogen- like domain and coiled-coil domain does not form a complex with an MHC Class I molecule, an MHC Class II molecule, or both, when the fusion protein is produced in, or administered to, a mammalian host, or

(b) the fusion protein comprises a linker positioned between the fibrinogen-like domain and the coiled-coil domain, such that the amino acid sequence within about 20 amino acid residues of the linker does not form a complex with an MHC Class I molecule, an MHC Class II molecule, or both, when the fusion protein is produced in, or administered to, a mammalian host.

10. The fusion protein of any of claims 1 -9, wherein the fusion protein reduces blood vessel leakage when administered or produced in vivo, promotes blood vessel maturation when administered or produced in vivo, or both.

11. The fusion protein of any of claims 1 -9, wherein the fusion protein, when administered or produced in a mammalian host, exhibits a half-life which is at least 125% as long as the half- life of Ang-1.

12. A fusion protein comprising:

(a) a fibrinogen-like domain that exhibits at least about 30% amino acid sequence identity to the fibrinogen-like domain of Ang-1, and

(b) a coiled-coil domain that comprises

(1) an amino acid sequence that exhibits at least about 60% amino acid sequence identity to the coiled-coil domain of zapol, murine FDRG, or both,

(2) an amino acid sequence that exhibits at least about 50% amino acid sequence identity to the coiled-coil domain of NLl, NL5, or both,

(3) an amino acid sequence that exhibits at least about 40% amino acid sequence identity to the coiled-coiled domain of NL8,

(4) an amino acid sequence that exhibits at least about 25%> amino acid sequence identity to the coiled-coiled domain of FLS 139, or

(5) an amino acid sequence that fulfills any combination of (l)-(4), wherein in vivo administration or production of the fusion protein promotes or inhibits angiogenesis in a mammalian host.

13. The fusion protein of claim 12, wherein

(a) the fibrinogen-like domain exhibits at least about 50%> amino acid sequence identity to the fibrinogen-like domain of Ang-1,

(b) the coiled-coil domain exhibits less than about 30% amino acid sequence identity to the coiled-coil domains of Ang-1, Ang-2, Ang-3, and Ang-4, or

(c) both (a) and (b).

14. A fusion protein comprising:

(a) a fibrinogen-like domain that comprises

(1) an amino acid sequence that exhibits at least about 60% amino acid sequence identity to the fibrinogen-like domain of NLl, NL5, NL8, or any combination thereof,

(2) an amino acid sequence that exhibits at least about 60%_> amino acid sequence identity to the fibrinogen-like domain of NL4,

(3) an amino acid sequence that exhibits at least about 40% amino acid sequence identity to the fibrinogen-like domain of zapol, murine FDRG, or both,

(4) an amino acid sequence that exhibits at least about 45 %> amino acid sequence identity to the fibrinogen-like domain of NL3,

(5) an amino acid sequence that exhibits at least about 30%> amino acid sequence identity to the fibrinogen-like domain of FLS 139, or

(6) an amino acid sequence which fulfills any combination of (l)-(5), and

(b) a coiled-coil domain that exhibits at least about 35% amino acid sequence identity to the coiled-coil domain of Ang-1, Ang-2, Ang-2X, Ang-3, Ang-4, or any combination thereof.

15. The fusion protein of claim 14, wherein

(a) the coiled-coil domain exhibits at least about 65%> amino acid sequence identity to the coiled-coil domain of at least one of Ang-1, Ang-2, or Ang-2X, or

(b) the coiled-coil domain exhibits at least about 50% amino acid sequence identity to the coiled-coil domain of Ang-3, Ang-4, or both.

16. The fusion protein of claim 15 or claim 16, wherein the fibrinogen-like domain exhibits less than about 45% amino acid sequence identity to the fibrinogen-like domains of Ang-1, Ang-2, Ang-3, and Ang-4.

17. A polynucleotide comprising a nucleic acid sequence encoding the fusion protein of any of claims 1-16.

18. A replication-deficient and non-integrating vector comprising the polynucleotide of claim 17.

19. The vector of claim 18, wherein the polynucleotide comprises a first nucleic acid sequence encoding the fusion protein and a second nucleic acid sequence encoding an angiogenic factor.

20. The vector of claim 19, wherein the first nucleic acid sequence is operably linked to a first expression control sequence, and the second nucleic acid sequence is operably linked to a second expression control sequence, such that the expression of the first and second nucleic acid sequences is initiated at different times, upregulated at different times, initiated in response to different factors, upregulated in response to different factors, or any combination thereof.

21. The vector of claim 19 or claim 20, wherein the second nucleic acid sequence encodes a non-heparin-binding VEGF.

22. A method of promoting angiogenesis in an individual comprising administering to the individual an amount of the fusion protein of any of claims 1-16 effective to promote angiogenesis.

23. The method of claim 22, wherein the method comprises co-administering at least one additional angiogenic factor.

24. The method of claim 22 or claim 23, wherein the method comprises co- administering a non-heparin-binding VEGF, a polynucleotide comprising a nucleic acid sequence encoding a non-heparin-binding VEGF, or a vector comprising a polynucleotide encoding a non-heparin-binding VEGF.

25. A method of producing a fusion protein comprising introducing the polynucleotide of claim 17 into a cell such that the nucleic acid sequence is expressed to produce a fusion protein.

26. The method of claim 25, wherein the cell is in a mammal.

27. The method of claim 25 or claim 26, wherein the polynucleotide is introduced into the cell via the vector of any of claims 18-21.

28. A method of promoting angiogenesis in an individual comprising introducing the vector of claim 20 into a cell and permitting, inducing, or upregulating expression of the first nucleic acid sequence, the second nucleic acid sequence, or both.