Title of Invention

A FUSION PROTEIN FOR DELIVERY OF A COMPOUND OF INTEREST INTO A CELL

Abstract

The present invention is directed to membrane penetrating peptides useful as in vivo, ex vivo and in vitro intracellular delivery devices for compound of interest. More particularly, the invention involves identification of membrane penetrating peptides which may be used as protein carriers for delivery of acompound of interest to cells, to methods of delivering a compound of interest attached to membrane penetrating peptides to cells.

Full Text

MEMBRANE PENETRATING PEPTIDES AND USES THEREOF This application claims the benefit of U.S. Provisional Application No. 60/227,647, filed August 25, 2000 and GB Application 0103110.3, filed February 7, 2001. FIELD OF THE INVENTION The invention relates to membrane penetrating peptides useful as in vitro, ex vivo and in vivo delivery devices for intracellular delivery of a compound of interest to cells in vitro, ex vivo and in vivo, compositions comprising the same and methods of using the same. The invention also includes identification of additional membrane penetrating peptides useful as delivery devices for intracellular delivery of a compound of interest to cells in vitro, ex vivo and in vivo. BACKGROUND OF THE INVENTION The delivery of small molecules, oligonucleotides, and proteins through biological membranes is a major challenge facing therapy and validation paradigms. It has recently been established that transducing peptides derived from Antennapedia, TAT-HIV, and VP22 can * penetrate biological membranes, act as cargo vehicles, and target to specific subcellular compartments. Here we show the identification of a nuclear localization sequence (NLS) within human Period 1 (hPERl) circadian protein that functions as a transducing peptide. More importantly, using database mining, we have uncovered additional transducing peptides embedded within the NLS's of other proteins and extend the number of gene-encoded transducing peptides from 3 to 14. Our data suggest that transducing peptides are found within NLS's and are prevalent, diverse, and distributed widely throughout the genome. It is well established that certain extracellular and intracellular proteins are targeted to specific organelles within a cell transmembrane or secreted from the cell The biological mechanisms by which intracellular protein targeting occurs continues to be characterized, but is well recognized that one mechanism for localization occurs by virtue of specific leader sequence contained within the protein of interest, or intraprotein sequence. Localization of proteins within selected cellular organelles is aided by specific targeting sequences. A number of nuclear localization sequences (NLSs) have been identified in proteins that permit the protein to be tranported or otherwise pass from the cytoplasm into the nuclear membrane. Fusion proteins containing the targeting sequence and another, otherwise non-targeted protein, are localized in the selected cellular organelle depending on the targeting sequence selected. For example, Ferullo, J. M. and Paget, E. FR 279695, disclose selective compartmentalization of an hydroxyphenylpyruvate dioxygenase (HPPD) fused to a signal sequence directing the enzyme to a cellular compartment other than the cytosol, e.g., a vacuole. Similarly, WO 0147950 (Wehrle-Haller, Bemhard M.; Imhof, Beat A) identify a new determinant responsible for basolateral targeting and prolonged exposure of cell-surface-anchored growth factors at cell surfaces. The signal is a mono-leucine dependent basolateral sorting signal consisting of the amino acid sequence Xlh2X3h4Lp5p6, wherein: XI represents a polar amino acid residue or alanine, h2 represents any hydrophobic amino acid residue, X3 represents any amino acid residue, h4 represents any hydrophobic amino acid residue, except leucine and isoleucine, L represents a leucine residue, p5 represents any polar amino acid residue, and p6 represents any polar amino acid. Richardson, A. E., et al. Plant J. (2001), 25(6), 641-649 describe manipulation of the enzyme aspergillus phytase to include the signal peptide sequence from the carrot extensin gene. The resulting fusion protein was only effective when secreted as an extracellular enzyme into the adjacent soil, and resulted in a 20-fold increase in total root phytase activity in transgenic lines and subsequent improved * phosphorus nutrition, such that the growth and phosphorus content of the plants was equivalent to control plants supphed with inorganic phosphate. WO 0132894 (Lok, S.) disclose use of the signal anchor domain sequences of type II cell surface proteins to anchor recombinant proteins into surface of transfected cells. A characteristic feature of type II cell surface proteins is that they are held within the cellular membrane by a single hydrophobic transmembrane domain and are oriented with their C-terminus outside the cell. More recently, a few proteins have been identified which are capable of passing through the cellular membrane without requiring active transport mechanisms or 'pores'. It is recently established that membrane penetrating peptides (MPPs, also known as protein transduction domain, "PTD") derived from Antennapedia, TAT, and VP22 can penetrate biological membranes and target to specific subcellular compartments. None of these previously disclosed proteins are derived from mammalian proteins. The present invention is directed to the discovery that polypeptides derived from mammalian or yeast proteins nuclear localization sequences (NLSs) or overlapping with NLS's are capable of acting as MPPs, and identification of a specific polypeptide sequences capable of penetrating cellular membranes, even when conjugated to large proteins, such as biologically active proteins, or other organic compounds. Nuclear transport is essential to a number of biological processes including gene expression and .cell division, as well as to viral replication, tumorigenesis and tumor cell proliferation. The mechanism of nuclear transport has only recently been characterized in detail and has been shown to involve a number of discrete steps. Proteins that are destined to be transported into the nucleus contain within their amino acid sequence a short stretch of amino acids termed a nuclear localization sequence ("NLS"). These sequences may occur anywhere within the amino acid sequence and are typically four to about eight amino acids. These sequences are generally basic (i.e., positively charged) in nature, however, there has been no consensus sequence identified. Thus, there is a wide variety of these sequences that appear to be specific for particular proteins. Within the cell, these NLSs may be either masked or unmasked by accessory proteins or by conformational changes within the NLS-containing protein. An NLS may be masked because it is buried in the core of the protein and not exposed on the surface of the protein. Unmasking of NLSs, and nuclear translocation of cytoplasmic proteins may be triggered by phosphorylation, dephosphorylation, proteolytic digestion, subunit association or dissociation of an inhibitory subunit, or the like. Accordingly, the masking and unmasking of NLSs provides a mechanism by which the transport of these cytoplasmic proteins into the nucleus may be regulated. For example, the transcription factor NF-AT contains nuclear localization sequences which allow NF-AT to translocate to the nucleus in the presence of intracellular calcium, but which are shielded by fonning intramolecular associations with other domains in the NF-AT polypeptide in the absence of calcium-Lee, H, C. and Bernstein, H. D. Proc, Natl Acad. Sci. U. S. A. (2001), 98(6), 3471-3476 studied the mechanism involved for presecretory proteins such as maltose binding protein (MBP) and outer membrane protein A (OmpA) that are targeted to the E. coli inner membrane by the molecular chaperone SecB, in contrast to the targeting of integral membrane proteins by the signal recognition particle (SRP). The authors found that replacement of the MBP or OmpA signal peptide with the first transmembrane segment of AcrB abolished the dependence on SecB for transport and rerouted both proteins into the SRP targeting pathway. Some proteins contain cytoplasmic localization sequences (CLS), or nuclear export sequences, which ensure the protein remains predominantly in the cytoplasm. For example, Hamilton, M, H. et al., J. Biol, Chem. (2001), 276(28), 26324-26331 demonstrate that the ubiquitin-protein ligase (E3), hRPFl/Nedd4, a component of the ubiquitin-proteasome pathway responsible for substrate recognition and specificity, is capable of entering the nucleus, but the presence of a functional Rev-like nuclear export sequence in hRPFl/Nedd4 ensures a predominant cytoplasmic localization. The cytoplasmic domains of many membrane proteins contain sorting signals that mediate their endocytosis from the plasma membrane. Heineman, T. C, and Hall S, L. Virology (2001), 285(1), 42-49 studied three consensus internalization motifs within the cytoplasmic domain of VZV gB and determined that internalization of VZV gB, and its subsequent localization to the Golgi, is mediated by two tyrosine-based sequence motifs in its cytoplasmic domain. In mammalian cells and yeasts, amino acid motifs in the cytoplasmic tails of transmembrane proteins play a prominent role in protein targeting in the early secretory pathway by mediating localization to or rapid export from the endoplasmic reticulum (ER). Hoppe, H. C. and Joiner, K. A. Cell Microbiol. (2000), 2(6), 569-57S. The mammalian endopeptidase, furin, is predominantly localized to the trans-Golgi network (TGN) at steady state. The localization of furin to this compartment seems to be the result of a dynamic process in which the protein undergoes cycling between the TGN and the plasma membrane. Both TGN localization and internalization from the plasma membrane are mediated by targeting information contained within the cytoplasmic domain of furin. Voorhees. P.. et al, EMBO J. (1995), 14(20), 4961-75 report that there are at least two cytoplasmic determinants that contribute to the steady-state localization and trafficking of furin. The first determinant corresponds to a canonical tyrosine-based motif, YKGL (residues 758-761), that functions mainly as an internalization signal The second determinant consists of a strongly hydrophilic sequence (residues 766-783) that contains a large cluster of acidic residues (E and D) and is devoid of any tyrosine-based or di-leucine-based motifs. This second determinant is capable of conferring localization to the TGN as well as mediating internalization from the plasma membrane. The trans-Golgi network (TGN) plays a central role in protein sorting/targeting and the sequence SXYQRL can by itself confer significant TGN localization. Wong, S.H., and Hong, W, J. Biol. Chem. (1993), 268(30), 22853-62 report detailed mutagenesis of the 32-residue sequence of TGN3S, an integral membrane protein confined mainly to the TGN, and determined that the Ser, Tyr, and Leu residues at positions 23, 25, and 28, respectively, are essential for TGN localization. When the cytoplasmic 32-residue sequence of TGN38 was fused to the ecto- and transmembrane domains of glycophorin A (a surface protein), the resulting chimeric protein was localized to the TGN. It is well recognized that certain proteins are either only active in a specific organelle, or are capable of different functions depending on their localization. For example, appropriate subcellular localization is crucial for regulation of NF-KB function. Huang, T. T., et al., Proc, Natl. Acad, Sci, U. S. A, (2000), 97(3), 1014-1019, show that latent NF-KB complexes can enter and exit the nucleus in preinduction states and identified a previously uncharacterized nuclear export sequence in residues 45-54 of IKBa that was required for cytoplasmic localization of inactive complexes. It appears that NF-KB/lKBa complexes shuttle between the cytoplasm and nucleus by a nuclear localization signal-dependent nuclear import and a CRMl "dependent nuclear export and that the dominant nuclear export over nuclear import contributes to the largely cytoplasmic localization of the inactive complexes to achieve efficient NF-KB activation by extracellular signals. Nuclear import of classical nuclear localization sequence-containing proteins involves the assembly of an import complex at the cytoplasmic face of the nuclear pore complex (NPC) followed by movement of this complex through the NPC and release of the import substrate into the nuclear interior. In combination with Ran, two other soluble factors are thought to be _ absolutely required to mediate the nuclear import of a protein containing a classical or basic NLS into the nucleus. The first is karyopherin/importin a (Kap a), which binds a classical NLS and then forms a complex with karyopherin/importin pi (Kappl). Adam, S. A., and Gerace, L. (1991) Cell 66, 837-S47; Goriich, D„ et al. (1994) Cell 79, 761-71^; Moroianu, J., et al. (1995) Proc. Natl. Acad ScL U. S. A. 92, 2008^2011; Radu, A., et ah (1995) Proc. Natl. Acad ScL U. S. A. 92,1769-1773; Goriich, D.,., et al (1995) Cwr, BioL 5, 383-392; Chi, N, C., et al. (1995)/. Cell BioL 130,265-274. Kap pi interacts with nuclear pore complex (NPC) proteins and appears to mediate movement of the import complex through the NPC via these interactions. Rexach, M., and Blobel, G. (1995) Cell 83, 6S3-692; Radu, A., Blobel, G., and Moore, M. S. (1995) Proc. Natl Acad ScL U. S A. 92,1769-1773; lovine, M. K., Watkins, L L„ and Wente, S. R, (1995) J. Cell BioL 131, 1699-1713; Radu, A„ Moore, M. S„ and Blobel, G. (1995) Cell 81, 215-222. Another protein, plO/NTF2, has also been implicated in nuclear import, but its function may only be to take Ran into the nucleus, where it is subsequently needed to disassemble an incoming import complex. Moore, M, S,, and Blobel, G. (1994) Proc, NatL Acad ScL a S A. 91,10212-10216; Paschal, B. M., and Gerace, L. (1995) J. Cell BioL 129, 925-937; Ribbeck, K., Lipowsky, G., Kent, H, M., Stewart, M., and Gorlich, D. (1998) EMBOJ. 17, 6587-6598; Smith, A., Brownawell, A., and Macara, I. G. (1998) Curr. Biol. 8, 1403-1406. Although there is only one Kap a homologue in yeast (SRPl or Kap60), vertebrate cells contain a number of proteins that can bind a classical NLS and share sequence homology (see Ref. Nachury, M. V., Ryder, U. W., Lamond, A. L, and Weis, K. (1998) Proc. Natl. Acad. Sci. U. S. A. 95, 582-587, and references therein). These proteins have been given a variety of names but can be grouped into three major families. The Kap al family contains the human protein NPI-1/importin al/karyopherin al/Rch2/liSRPl and a second related protein importin a6, in addition to the mouse S2 protein. Moroianu, J., et al., (1995) Proc. Natl. Acad. Sci. U. S. A. 92, 2008-2011; Cortes, P., et al., (1994) Proc. Natl Acad. Sci. U. S A. 91, 7633-7637; O'Neill, R. E., et al, (1995) J. Biol Chem. 270, 22701-22704; Kohler, M., et al., (1997) FEES Lett. 417, 104-lOS; Tsuji, L., et al., (1997) FEES Lett. 416, 30-34. The second family, Kapa2, contains human Rchl/hSRPl/importin a2/karyopherin a2 and the mouse protein pendulin/PTAC 58. Gorlich, D., Prelin, S., Laskey, R. A., and Hartmann, E. (1994) Cell 19,161-11%; Cuomo, C. A., Kirch, S. A., Gyuris, J., Brent, R., and Oettinger, M. A. (1994) Proc. Natl Acad. Sci. U. S A. 91, 6156-6160; Kussel, P., and Frasch, M. (1995) Mo/. Gen. Genet. 248, 351-363; Imamoto, N., Shimamoto, T., Takao, T., Tachibana, T., Kose, S., Matsubae, M., Sekimoto, T., Shimonishi, Y., and Yoneda, Y. (1995) EMBOJ. 14, 3617-3626;, K., Mattaj, I. W., and Lamond, A. I. (1995) Science 268, 1049-53. The third family, Kapa3, consists of the two human proteins, QEP-l/importin a3 and KPNA3/hSPRl Y/1ISRP4, and the mouse proteins Ql and Q2. Nachury, M. V., et al., (1998) Proc. Natl. Acad. Sci. U. S. A. 95, 582-587; Kohler, M., et al., (1997) FEES Lett. 417, 104-108; Tsuji, L., et al, (1997) FEES Lett. 416, 30-34; Takeda, S., et al, (1997) Cytogenet. Cell Genet. 16, 87-93; Seki, T., et al., (1991) Eiochem. Biophys. Res. Commun. 234, 48-53; Miyamoto, Y., et al, (1997) J. Biol Chem. 272, 26375-26381. Each of these classes share about 50% homology with each other and to the yeast SRPl, and each of these mammalian proteins has been shown to be capable of mediating the import of one or more classical NLS-containing proteins. Nachury, M. V., et al., (1998) Proc. Natl. Acad. Sci. U. S. A. 95, 582-587; Sekimoto, T., etal., (1991) EMBOJ. 16, 7067-7077; Nadler, S. G., et al., (1997) J Biol Chem. 272, 4310-4315; Prieve, M. G., et al, (1998) Mol. Cell. Biol 18, 4819-4832. Stat-1 import is mediated by Kapαl/NPM but not Kapa2/Rchl, but activated Stat-1 appears to bind to a COOH-terminal region of Kapαl distinct from the NLS binding Armadillo repeats. The binding differences of the different Kapαs to RCCl observed appear to be due solely to the NLS on RCCl and thereforeprobably due to the NLS binding region of Kapa3. Sekimoto, T., et al., (1997) EMBOJ. 16, 7067-7077, Kamei, Y., et al., (1999) J. Histocheiiv Cytocheuu 47, 363-372 showed that, in mice, the Kapa3 homologue is expressed in many tissues and theorized that Kapa3 may play a role in importing "a limited number of unique karyophilic proteins, such as helicase Ql." The results provided by Talcott, B. and Moore, M.S., 2000 J Biol Chem, 275(14) 10099-10104 suggest that RCCl should be included in the group of proteins that use Kapα3 to mediate their nuclear import, USP 6,191,269 teaches the existence of a nuclear localization sequence contained within the cDNA sequence of the N-terminal IL-1 alpha propiece, T76-NGKVLKKEIRL, which had characteristics of a nuclear localization sequence (NLS) and could mediate nuclear localization of the propiece (Stevenson et al (1997) Proc. Natl. Acad. Sci. USA 94:508-13). Introduction of the cDNA encoding the N-terminal IL-.alpha, propiece into cultured mesangial cells resulted in nuclear accumulation (Stevenson et al. id). USP 5,877,282 teaches that the antennapedia homeodomain signal sequence peptide is the amino acid sequence RQIKIWFQNRRJVIKWKK; the fibroblast growth factor signal sequence peptide is AAVALLPAVLLALLA; the HIV Tat signal sequence peptide is the amino acid sequence CFITKALGISYGRKKRRQRRRPPQGSQTH. « Schwartze, S.R., et al., Science 285:1569-1572 (1999) report delivery of an ip injected reporter protein, 116 kD beta-galacatosidase, as a TAT fusion protein into tissues and across the blood-brain barrier. Schwartze used an 11 amino acid protein transduction domain (PTD) derived from HIV tat protein with an N-terminal fluorescein isothiocyanate (FITC)-Gly-Gly-Gly-Gly motif The authors report that earlier attempts to transduce beta-Gal chemically cross-hnked to the TAT PTD resulted in sporadic and weak beta-Gal activity in a limited number of tissues. They speculate that the improved transduction was due to the in-frame fiision and purification strategy used. Nuclear locahzation of IFNy is mediated by a polybasic NLS in its C terminus, which is required for the full expression of biological activity of IFNy, both extracellularly and intracellularly. Subramaniam, Prem S., et al, J. Cell Sci. (2000), 113(15), 2771-2781. This NLS is thought to play an integral intracellular role in the nuclear translocation of the transcription factor STAT la activated by IFNy because treatment of IFNy with antibodies to the C-terminal region (95-133) containing the NLS blocked the induction of STATlα nuclear translocation, but these antibodies had no effect on nuclear translocation of STAT la in IFNa treated cells. A deletion mutant of human IFNy, IFN7(1-123), which is devoid of the C-terminal NLS region was biologically inactive, but was still able to bind to the IFNy receptor complex on cells with a Kd similar to that of the wild-type protein. Deletion of the NLS specifically abolished the ability of IFNY(1-123) to initiate the nuclear translocation of STATlα, which is required for the biological activities of IFNy following binding to the IFNy receptor complex. A C-terminal peptide of murine IFNy, IFNy(95-133), that contains the NLS motif, induced nuclear translocation of STATlα when taken up intracellularly by a murine macrophage cell line. Deletion of the NLS motif specifically abrogated the ability of this intracellular peptide to cause STATlα nuclear translocation. In cells activated with IFNy, IFNy was found to as part of a complex that contained STATlα and the importin-a analog Npi-1, which mediates STATlα nuclear import. The tyrosine phosphorylation of STATlα, the formation of the complex EFNy/Npi-l/STATlα complex and the subsequent nuclear translocation of STATlα were all dependent on the presence of the IFNy NLS. The peptide representing amino acids 95-132 of EFN-y (IFN-y(95-132)), containing the 11^ ITT polybasic sequence " RKRKRSR , was capable of specifying nuclear uptalce of the auto fluorescent protein, APC, in an energy-dependent fashion that required both ATP and GTP. Nuclear import was aboHshed when the above polybasic sequence was deleted. Subramaniam, P., et al., 1999 J Biol Chem 274(1) 403-407. A peptide containing the prototypical polybasic NLS sequence of the SV40 large T-antigen was also able to inhibit the nuclear import mediated by IFN-y(95-132), suggesting that the NLS in IFN-y may function through the components of the Ran/importin pathway utilized by the SV40 T-NLS. Intact EFN-y, when coupled to APC, was also able to mediate its nuclear import, and this nuclear import was blocked by the peptide IFN-y (95-132) and the SV40 T-NLS peptide, suggesting that intact IFN-y was also transported into the nucleus through the Ran/importin pathway. Nuclear proteins are imported into the nucleus through aqueous channels that span the nuclear envelope called nuclear pore complexes (NPCs), Although ions and molecules less than -20-40 Da can diffuse passively through the nuclear pore complexes, larger proteins are transported by saturable pathways that are energy- and signal-dependent. The signals that specify nuclear protein import (NLSs)l are commonly short stretches of amino acids rich in basic amino acid residues, although other classes of NLSs have been described recently. The initial step in the import of proteins containing basic amino acid-type NLSs occurs in the cytosol, where the NLS-containmg proteins are bound to a receptor (variously called the MLS receptor, importin a, and karyopherin (13). The substrate-receptor complex then associates with the cytoplasmic face of the nuclear pore complexes, and with the participation of other cytosolic factors, is transported through a gated channel in the nuclear pore complexes to the nuclear interior. The in vivo events of NLS-mediated nuclear import can be dupHcated in an in vitro system using digitonin-perraeabilized cells supplemented with cytosolic extracts and ATP (14). Transport in this in vitro assay is blocked by the same inhibitors that block in vivo import, is rapid, and is easily quantified. The NLS the sequence NYKKPKL in the N-terminus of fibroblast growth factor (FGF)-l, the precursor for acidic FGF, has been proposed to affect the long term activities of FGF-1 through its function as a nuclear translocation signal or its role in stabilization of the structure required to sustain binding and activation of the transmembrane receptor kinase. Luo. Y., et aL. J. Biol Chem. (1996), 271(43), 26S76-26883. For example, concurrent with a marked increase in dependence on exogenous heparin for optimal activity, sequential deletion of residues in the NYKKPKL sequence in FGF-1 resulted in a progressive loss of thermal stability, resistance to protease, mitogenic activity, and affinity for the transmembrane receptor. The largest change resulted from deletion of the entire sequence through the lysine- ■ leucine residues. In the presence of sufficiently high concentrations of heparin, the deletion mutants exhibited mitogenic activity equal to wild-type FGF-1. Although FGF-1 contains an NTS, nuclear translocation requires an exogenous and not an endogenous pathway. The NTS of FGF-1, NYKKPKL, is able to direct the expression of the bacterial P-galactosidase (pgal) gene to the nucleus of transfected NTH 3T3 cells, but this NTS is unable to target either FGF-1 itself of a FGF-l-(3gal fusion protein into the nucleus, suggesting that FGF-1 may contain an additional sequence which prevents endogenously expressed FGF-1 from being translocated into the nucleus. Zhan, X., et aL, Biochem. Biophvs. Res. Commun. (1992), 188(3), 982-91. Interferon-y (IFN-y), a protein that uses the Jak-Stat pathway for signal transduction, translocates rapidly to the nucleus in cells treated extracellularly with the cytokine. An NLS has been identified and characterized in the C-terminus of human and murine IFN-y. Larkin, J., et al.. J. Interferon Cytokine Res. (2001), 21(6), 341-348 report that human IFN-y (HuIFN-y) contains a second NLS at an upstream site. The primary sequence, analogous with the NLS sequence identified in murine EFN-y, representing amino acids 122-132 of HuIFN-y was capable of mediating the nuclear import of the autofluorescent protein allophycocyanin (APC) in an energy-dependent manner The second sequence, representing amino acids 78-92 of HuIFN-γ, was also capable of mediating the nuclear import of APC in an energy-dependent manner but to a greatly reduced extent. The nuclear import of both sequences conjugated to APC was strongly blocked by competition with unconjugated HuIFN-'y(122-132). Competition by the sequence HUIFN-Γ(78-92) effectively blocked the import of APCr conjugated HuIFN-γ(78-92) but, at the same concentration, was not capable of inhibiting the nuclear import of APC-conjugated HUIFN-Γ(122-I32), suggesting that HUIFN-Γ(78-92) was a less efficient NLS than HUIFN-Γ(122-132). This is consistent with >90% loss of antiviral activity of HuIFN-γ lacking the downstream NLS in 122-132. The nuclear import of APC-conjugated HUIFN-Γ(122-'132) was inhibited by a peptide containing the prototypical polybasic NLS of the SV40 T NLS, which suggests that the same Ran/importin cellular machinery is used in both cases. There appears to be strong conservation of the NLS motif as a mechanism for nuclear localization. Evolution seemed to have used part of the existing DNA-binding mechanism when compartmentalizing DNA-binding proteins into the nucleus. Cokol, M., et al., EMBO Rep. (2000), 1(5), 411-415 estimate that greater than 17% of all eukaryotic proteins maybe imported into the nucleus, and after analyzing a set of 91 experimentally verified NLSs from the literature and expanding this set to 214 potential NLSs through iterated "in silico mutagenesis". This final set matched in 43% of all known nuclear proteins and in no known non-nuclear protein. Cokel et al found an overlap between the NLS and DNA-binding region for 90% of the proteins for which both the NLS and DNA-binding regions were known, but only 56 of the 214 NLS motifs overlapped with DNA-binding regions. These 56 NLSs enabled a de novo prediction of partial DNA-binding regions for approximately 800 proteins in human, fly, worm and yeast. More recently, it has been reported that NLS signal peptide can induce structural changes of DNA. The plant enzyme, glutaminyl-tRNA synthetase (GlnRS) from Lupinus luteus, contains an NLS at the N-terminal, a lysine rich polypeptide, KPKKKKEK. Krzyzaniak, A,, et al.. Mol. Biol, Rep. (2000), 27(1), 51-54. Two synthetic peptides (20 and 8 amino acids long), derived from the NLS sequence of lupin GluRS interact with DNA. In addition, the shorter 8 amino acid peptide caused the DNA to change its conformation from the B to the Z form. This observation clearly suggests that the presence of the NLS polypeptide in a leader sequence of GlnRS is required not only for protein transport into nucleus but also for regulation of a gene expression. This is the first report suggesting a role of the NLS signal peptide in structural changes of DNA. Typically there is strong conservation of the NLS sequence within species. For example, the NLS in the N-tenninal region of Smad 3 protein, the major Smad protein involved in TGF-(3 signal transduction, has a basic motif Lys40- Lys-Leu-Lys-Lys44 which is conserved among all the pathway-specific Smad proteins, and is required for Smad 3 nuclear import in response to ligand. Smad proteins are intracellular mediators of transforming growth factor-β (TGF'P) and related cytokines. Xiao. Z.. et aL, J. Biol. Chem. (2000), 275(31), 23425-23428 identified the role the NLS plays in nuclear localization. The authors demonstrated that the isolated Smad 3 MHl domain displays significant specific binding to importin β, which is diminished or eliminated by mutations in the NLS. Full-size Smad 3 exhibits weak but specific binding to importin β, which is enhanced after phosphorylation by the type I TGF-p receptor. In contrast, no interaction was observed between importin a and Smad 3 or its MHl domain, indicating that nuclear translocation of Smad proteins may occur through direct binding to importin β. The authors conclude that activation of all of the pathway-specific Smad proteins (Smads 1, 2, 3, 5, 8, and 9) exposes the conserved NLS motif, which then binds directly to importin β and triggers nuclear translocation. In all cells, the lipid bilayer of cell membranes serves as a selective barrier for the passage of charged molecules, with the intemalization of hydrophilic macromolecules being achieved through classical transport pathways (Hawiger, J., Ciin- 0pm Chem BioL 3, 89-94 (1999), Schwarze, S.R., et al, Trends in Cell Biology 10, 290-295 (2000)). These classical mechanisms of internalization involve receptor-mediated endocytosis or transporter dependent uptake (Cleves, A.E., Current Biology 7, R318-R320 (1997)). In contrast, an increasing number of molecules have been discovered that lack classical import and/or export signals (Cleves, A.E., Current Biology 7, R318-R320 (1997)), These molecules gain direct access to either cytoplasmic or nuclear compartments using unconventional processes of which the mechanisms remain largely unknown. These novel mechanisms are generally termed "nonclassical" and refer to transport pathways being used that are atypical. Relevant examples of this latter type are found in the gene-encoded proteins of HIV-1 TAT (Frankel, A.D, and Pabo, CO. Cell 55,1189-1193 (1988)), herpes virus VP22 (Elliott, G. and CHare, P. Cell 88, 223-233 (1997)), and Antennapedia, Antp (Derossi, D., et al, J. Biol Chenu 269,10444-10450 (1994)), It is now well established that the full-length proteins of HIV-1 TAT (Helland D.E,, et al., J Virol 65, 4547-4549 (1991)), and VP22 (Pomeranz L.E. and Blalio J.A,, / Virol 73, 6769-6781 (1999)) rapidly translocate into and out of cellular membranes. In fact, distinct peptide regions have been identified within both of these proteins that are capable of translocating into cellular compartments either alone or in combination with chimeric cargo peptides, and proteins (Lindgren, M., et aL, Trends Pharmacol Set 3, 99-103 (2000), Derossi, D„ et al, Trends Cell Biol, 8, 84-87 (199S), Prochiantz A., Current Opinion in Cell Biology 12, 400-406 (2000), Steven R. Schwarze, S.R., et al., Trends in Cell Biology 10, 290-295 (2000)). In contrast, fulWength Antp protein has not been shown to traverse biological membranes; however, a 16 amino acid synthetic peptide derived from within its coding region does possess potent membrane penetrating abilities (Derossi, D,, et al. Trends Cell Biol, 8, 84-87 (1998)). The accepted view of atypical transport used by these molecules has been tenned "transduction" (Schwarze, S.R., et al. Trends in Cell Biology 10, 290-295 (2000)), and is cun-ently defined as an extremely rapid membrane transport pathway that is receptor and energy independent, and can occur at 4 C in all cell types (Schwarze, S.R. and Dowdy, S.F. Trends Pharmacol Sci. 21, 45-48 (2000)). Interestingly, these three proteins are all nuclear proteins involved in transcriptional regulation, and their respective transducing peptides consist of strings of amino acids rich in arginine and lysine (Lindgren, M., et al,, Trends Pharmacol Sci, 3, 99-103 (2000), Schwarze, S.R. and Dowdy, S.F. Trends Pharmacol Sci. 21, 45-48 (2000)). Hov/ever, irrespective of these similarities, these transducing peptides possess many different characteristics such as amino acid sequence, length of the sequence, cellular localization, and potency of membrane penetration. Thus, though each transducing sequence can penetrate cells and tissues, it has not been established whether they use the identical atypical transport mechanisms. Finally, USP 6,022,950 teaches the use of a hybrid molecule of a portion of the binding domain of a cell-binding polypeptide ligand effective to cause said hybrid protein to bind to a cell of an animal, a translocation domain of naturally occurring protein which translocates said third part across the cytoplasmic membrane into the cytosol of the cell; and a chemical entity to be introduced into the cell. However, the patent teaches translocation domains of toxins. Naturally-occurring proteins which are known to have a translocation domain include diphtheria toxin and Pseudomonas exotoxin A, and may include other toxins and non-toxin molecules, as well. The translocation domains of diphtheria toxin and Pseudomonas exotoxin A are well characterized (see, e.g., Hoch et al, Proc. Natl Acad. Sci. USA 82:1692-1696, 1985; Colombatti et al, J. Biol. Chem, 261:3030-3035,1986; and Deleers et al, FEBS 160:82-86, 1983), and the existence and location of such a domain in other molecules may be determined by methods such as those employed by Hwang et al.. Cell 48:129-136, 1987; and Gray et al., Proc. Natl. Acad. Sci. USA 81:2645-2649, 1984. Given the considerable body of literature teaching control mechanisms of cellular localization, the proteins involved in regulation of intracellular transport, the different properties and control mechanisms for plasma membrane and the nuclear envelope, it is unexpected that polypeptides derived from mammalian proteins could transduce through the plasma membrane using nonclassical mechanisms and thus could be useful as membrane penetrating peptides useful as in vitro, ex vivo and in vivo delivery devices of a compound of interest. There is also considerable literature teaching non-protein derived methods for delivering a compound of interest into cells, for example electroporation, membrane fusion with liposomes, high velocity bombardment with DNA-coated microprojectiles, incubation with calcium-phosphate-DNA precipitate, DEAE-dextran mediated transfection, infection with modified viral nucleic acids, and direct microinjection into single cells, usually ova and the like. Each of these methods is relatively inefficient, resulting in relatively low percentage of the cells containing the delivered compound of interest and most of the methods are clearly not capable of realistic in vivo delivery. Many of the methods are toxic to the cells, resulting in relatively high apoptosis. Therefore, there is a considerable need for simple and more efficient delivery of compounds of interest into cells, SUMMARY OF THE INVENTION The present invention is directed to polypeptides derived from mammalian and yeast proteins useful as a carrier for m vitro, ex vivo and in vivo delivery a compound of interest. The invention also provides compositions containing the same, and methods of delivering a compound of interest in vitro, ex vivo and in vivo. BRIEF DESCRIPTION OF THE DRAWINGS Figure 1. (A)- Schematic diagram of hPERl fusion constructs showing the locations of the PAS, cytoplasmic localization, and nuclear localization sequence (NLS, but indicated as nuclear localization domain (NLD) in Figure). The name and the position of the fusion constructs are listed on the left. The number indicates the first and last amino acid residues in the hPERl protein. The principal sites of accumulation of each fusion protein are summarized on the right, (n) nuclear, (no) nucleoli, (c) cytoplasmic, (diff) diffuse. All constructs were N- terminally tagged with EYFP, The alignment human and mouse PERl-NLS is shown at the bottom. Figure 1, (B). Cellular localization of hPERl fusion proteins as described in Figure 1 A, above, in living cells. CHO cells were transient transfected with the fusion constructs indicated on the top of each panel and the subcellular localization of EYFP reporters (green) was directly visualized using fluorescent microscopy 10 h post-transfection. EYFP vector alone is used as control (see 5. E\TP-VECTOR) Figure 2, (A). Membrane penetration assay in CHO cells. N-tenninal biotinylated synthetic peptides hPERl-PTD, Flag-hPERl-PTD, Flag-TAT-PTD (positive control), and Flag-Flag (negative control) were assayed for their ability to penetrate cellular membranes in living CHO cells in culture. The subcellular localization of internalized peptides was determined using a two color staining method, either Streptavidin-Alexa 594 (red) or anti-flag mAb (green). The third column is an overlay (yellow), Confocal microscopy was employed to further confirm intracellular and intranuclear localization. Single section of confocal imaging is shown. Figure 2, (B). Nuclear targeting of biotinylated peptides hPERl-NLD (also known as hPERl-PTD) compared with TAT-PTD and Flag-Flag (negative control) using Streptavidin Alexa -594 fluorescence (green). Hoechst 33258 at 5ng/ml was used to stain the nucleus (blue, middle column). The third column is an overlay of confocal imaging. Figure 3. Alanine scanning of hPERl-PDTs. Biotinylated hPERl-NPDs were synthesized with a single amino acid residue substitution at the indicated position with an alanine and assayed for membrane penetration in CHO cells. Cells were incubated for 10 minutes at 37 C at a peptide concentration of 10 µM followed by washing, fixation, permeablization, and then detected with labeled Streptavidin Alexa-594 (red, 2ug/ml) for 15 minutes at the RT. Control peptide was from hPERl N-terminal amino acids residues 486-500. Figure 4. Activation of serotonin 5HT2A receptor with hPERl-MPP fusion peptide. (A). hPERl-MPP and TAT-PTD peptides were synthesized alone or in fusion with either the first intracellular loop II (SLEKKLQNATN), or the C-terminal Transmembrane 7 domain, TM7 (KTYRSAFSRYIQYKENKKPLQLI) derived from the 5HT2A receptor, genebank accession numbr, MS6841). Receptor activities was assayed using standard FLIPR analysis and measuring endogenous and exogenous Ca+2 levels. Peptide designations are as follows: T (TAT-PTD), P (hPERl-MPP), H (intracellular loop 1), T-Il (TAT-PTD-Il), P-Il (hPERl- MPP-Il), TM7 (C-terminal domain), TTM7 (TAT-PTD-TM7), PTM7 (hPERl-MPP-TM7), and S (Serotonin). Figure 4 (B), Dose response of PTM7 (closed circles) and TTM7 (closed diamonds) peptides. Serotonin (control, open triangle) was used at the maximum receptor stimulatory concentration of 10 µM, Figure 5, Identification of additional PTDs. Putative PTD sequences were searched using a combined bioinformatics method that included SwissPro, PRF, PIR-Protein info Resource, PDB with peptides sequences translated from the annotated protein coding region in GenBank with "transcription factor" as the key word. We initially searched for all known or putative NLS's. Secondly, we employed the PHI-BLAST (Pattern-Hit Initiate BLAST) to search for the degenerative pattern occurrence [R/H/K]-[R/H/K]-[R/H/K]'[R/H/K], (X)n where n is an integer of 4 or larger and X each time is independently selected to be either arginine, histidine, or lysine. 7374 putative PTD sequences were identified. From the two searches we synthesized (A) biotinylated peptides to these sequences or (B) created in frame fusion proteins with GFP and transfected CHO cells. 9 of the 12 peptides were found to transduce, and all sequences localize to the nucleus in transfected cells. hPERl-PTD, hPER3- PTD, and TAT-PTD peptides were used as positive controls. Six positive sequences and 2 * negative sequences are shown. Numbers represent the ammo acid residues within the parental protein sequence and Gene bank accession numbers for these proteins are indicated as follows: (M24S99, human Thyroid hormone apha-I; L12699, human Homeobox protein Engrailed 1 HMEl; X16416, human Proto-oncogene tyrosine protein kinase ABLl;; Q025755 human HENl/NSLCl; Q02577, human HEN2/NSLC2; AAA74561, rat HNF-3; CAB65887, Drosophila cAfvIP dependent transcription factor). Three negative peptides are (V01512, c-Fos; AAD53184, human cyclin L ania-6a; CAB66914, Arabidopsis P-zip transcription factor). Figure 6. hPER-PTD cargo's P-Galactosidase into cells: At least one feature of HIV TAT transducing peptide is its ability to cargo proteins into cells and tissues. We therefore sought to determine if hPERl- transducing peptide could cargo beta galactosidase into cells. To perform this experiment, we followed a protocol by Frankel et aL 19S9 (19):7397-401, whereby, we chemically linked hPERl-PTD or hPER-PTD R7A to full length β-galactosidase and assayed for the ability of these conjugates and beta-galactosidase protein alone to transduce into CHO cells. As shown in the figure 6, left, cells incubated with hPER- PTD P-galactosidase fusion showed positive enzymatic activity for p-galactosidase as indicated by the blue color in the cells after the addition of X-gal However, neither hPER- MPP R7A (3-galactosidase (center) nor p-galactosidase protein (right) alone was able to enter the cells as indicated by a no blue staining reactivity after the addition of X-gal. These data indicate that like TAT peptide, hPERl-PTD can cargo a large (120 kD) protein into cells. DETAILED DESCRIPTION OF THE INVENTION The present invention is based on discovery that human Period! (hPERl) protein contains an NLS which has now also been identified as an MPP and is useful as a delivery device for intracellular delivery of a compound of interest. hPERl is involved in regulation of the circadian rhythm and the capacity of hPERl to translocate to adjacent cells maybe critical to its overall biological function of regulating circadian rhythm. The NLS identified within hPERl does not fit within previously identified NLS sequences, and its identification has resulted in identification of an algorithm for searching for other NLS sequences which may also function as MPPs, Period 1 (hPERl) is a nuclear protein involved with transcriptional regulation. It is an essential component in the "gears" of the biological clock (Brown, S.A., and Scliibler, U., Current Opinion in Genetics & Development 9, 5SS-594 (1999), Dunlap, J.C., Cell 96, 271-290 (1999)), and studies in mice have shown that nuclear entry of PERI is essential for the down regulation of CLOCK/BMAL transcriptional complexes (Gekakis N, et al.. Science 280, 1564-1569. (1998), Yagita, K., et al., Genes Dev 14,1353-1363 (2000), Lowrey, P.L., et ah, Science 288, 4S3-492 (2000)). However, to date, the functional NLS for human PERI has not been elucidated. The present inventors identified the NLS within hPERl, and demonstrate that the 16 amino acid and 13 amino acid sequence, see Figure 3.hPERl-NLS peptide, hPERl-MPP, has potent membrane penetrating ability. This work results in the identification of four additional MPPs also derived from nuclear proteins. PERI is a central component in the circadian clock, and its nuclear entry plays an important role in the regulation of daily oscillations (Jin, X., et al, Cell 96, 57-68 (1999), Sangoram, AM,, et al., Neuron 21,1101-13 (1998 )). Using deletion and fusion protein analysis, we identified a NLS that is necessary and sufficient for hPERl nuclear localization, Tliis functional analysis was necessary because the NLS of hPERl does not conform to classical nuclear localizing consensus motifs; and therefore, was not identified using standard NLS search procedures. We show that a single copy of hPERl-NLS is sufficient for inducing nuclear localization of a reporter protein and of tagged hPERl fragments (P1-F2 to P1-F7) in transfected cells. The PERl-NLS is located between amino acids (S30-S45) of hPERl, is embedded within a string of 13 amino acids rich in arginine, histidine, and lysine (see Table 1) that is not found in other PERs or other nuclear proteins in available databases. Therefore, tliough PERs 2 and 3 are nuclear proteins (Jin, X., et al, Cell 96, 57-68 (1999)), they apparently use alternative sequences and or mechanisms for their nuclear import. Peptide fragments of a limited number of nuclear proteins that are rich in basic residues have been shown to penetrate into cellular membranes in a receptorless, energy-independent fasliion. Sequences from three such proteins, TAT, Antp, and VP22 have been demonstrated to possess the ability to penetrate and cargo fusion molecules into cells and tissues by an as yet undefined mechanism. See, for example, USP 5,804,604, 5,747,641, 5,674,980, 5,670,617 and 5,652,122 issued to Frankel et al., which teach the use of a nine-amino acid HIV TAT-derived polypeptide (Arg-Lys-Lys-Arg-Arg-Gln-Arg-Arg-Arg) for intracellular delivery of cargo molecules. The similarities between hPERl, the hPERl-NLS, and other MPPs prompted us to investigate whether or not hPERl-MPP could have membrane penetrating capability. The immunohistochemical and cytological data presented herein indicates that the hPERl-MPP functions as a MPP in a variety of cell types. hPERl-MPP demonstrated intense focal staining in the nuclear plasma as well as in the nucleolus, suggesting that the subnuclear address of hPERl-MPP is different from the hPERl (Pl-FL) protein that was diffused in the nucleus but not concentrated in the nucleolusThe cellular penetration of hPERl-MPPs is not blocked even under the conditions of reversing the sequence (reversed hPERl-MPP), adding negatively charged residues or pre-fixing cells with 4% PFA, unpublished observation, the latter supports the idea that penetration is receptor and membrane independent. These results are in contrast to other peptide classes that have been described that are derived from signal peptide sequences (Hawiger, J., Curr Opin Immunol 9,189-94 (1997)), DNA antibodies (Deng, S.X., et al., Int Immunol. 12,415-423 (2000)), and other protein domains (Lindgren, M,, et al., Trends Pharmacol Sci. 3, 99-103 (2000)) that bind and cross the cell membranes using slow, temperature, energy, and receptor dependent mechanisms. The identification of other MPPs, has been limited by our lack of understanding the mechanisms and structural requirements necessary for membrane peptide penetration. The likelihood that a specific peptide structure and/or charge is important for membrane penetration is demonstrated in the alanine scanning experiments whereby a single amino acid change at arginine 7 appears to be critical for MPP potential. By comparing wild-type hPER-MPP to modified PI- R7A, in live cells or pre-fixed and permeabilized cells (data not show). P1-R7A is only defective in penetration but not in nuclear targeting once the cells have been permeabilized. This finding suggests that arginine 7 has a major role in structure based penetration, and thus provides a useful model for the future structure-function studies. No structural determinants for TAT peptide have been described, but in the case of Antp, replacing the two tryptophan residues with two phenylalanines abolishes penetration (Le Roux, L, et al Proc Natl Acad Sci USA. 90, 9120-9124 (1993)), Since hPERUMPP does not contain any tryptophan residues, membrane penetration between these two peptides may occur by different mechanisms. Full-length HIV TAT and VP22, both of which lack classical secretary signal sequences and are therefore exported by non-classical mechanisms, can also be imported "by transduction", into cells in a non-classical manner (Procliiantz A,, Current Opinion in Cell Biology 12, 400-406 (2000)). Therefore, it is interesting to speculate that perhaps hPERl distributes circadian clock information to adjacent SCN neurons or to circadian output pathways by "transduction" mechanisms similar to foil-length TAT and VP22 proteins. However, simply having membrane penetrating sequences within the body of a protein does not necessarily confer membrane penetrating capability, as full-length Antp protein is neither exported from nor imported into cells. Thus, the non-classical penetration of the Antp peptides into the cells is unlikely to have physiological relevance, and like Antp, there is no evidence to suggest that full-length hPERl is a cell membrane penetrating protein. However, these findings did encourage us to search for other MPP-containing proteins. By searching protein databases with an algorithm designed to identify strings of basic residues within nuclear proteins, we uncovered hundreds of proteins that contained potential membrane penetrating peptide regions and found 4 additional MPPs from several species (see Fig. 5). These and additional database mining searches suggest that MPP-like sequences are common, and present within a wade variety of proteins. However, like many putative NLSs that do not always confer nuclear localization when fused to reporter sequences (Moroianu, 1,, J Cell Biochem. 32-33, 76-83 (1999)), any potential MPPs must be functionally determined experimentally. Though it seems clear that either transducing or non-transducing proteins can encode MPP regions, the interesting question that remains is whether or not proteins containing MPP-like sequences use these domains to rapidly translocate intracellularly into cellular domains to activate normal physiological processes. The efficiency associated with the transduction phenomena might be particularly useful where the rapid delivery of intercellular information is critical, as may be the case in cell synchronization, development, ' and differentiation paradigms. The ability for MPPs to cargo molecules to intracellular compartments is becoming well-established (Lindgren, M., et al. Trends Pharmacol ScL 3, 99-103 (2000), Derossi, D., et al. Trends Cell Biol, 8, 84-87 (1998)), Similar to other MPPs, hPERl-MPP and other MPPs identified herein can deliver compounds of interest, such as large molecules, i.e., peptides and proteins, lipids, polysaccharides, other organic molecules, rapidly and efficiendy into cells. The data presented herein demonstrates that hPERl-MPP in fusion with either serotonergic and/or adrenergic 7TM-receptor derived peptides mimic the effects of ligand activated receptors (see Fig. 4, and data not shown), confirming that hPERl-MPP translocates compounds of interest to intracellular compartments, and supports the idea that physiologically relevant signaling can be initiated by MPPs linked to compounds of interest. Using the methods described herein, the present invention may be expanded to provide target validation using MPPs linked to targets, and/or therapeutic strategies using MPPs linked to specific enzymes or receptors as a method of altering, correcting or compensating for dysfunctional enzyme performance or within pathways. In addition, therapeutic strategies using MPPs linked to specific receptors may be used as a method of altering, correcting or compensating for dysfunctional receptor, low expression of normal or abnormal receptors. Taken together, the results provided herein demonstrate an MPP encoded by a mammalian protein and more specifically, a human nuclear protein, whose cellular penetration is membrane independent and likely depends on the peptide structure. hPERl-MPP targets to specific subnuclear sites, but has the potential to efficiently deliver other macromolecules to intracellular targets. More importantly, this invention also provides the first example for mapping a novel MPP based on a NLS domain, and suggests that many MPP-like regions are contained within a wide variety of proteins. The data provided herein demonstrate that an MPP may be based on part of an NLS, or overlap with part of the NLS, or alternatively, may be a novel peptide. Methods of identifying NLS sequences are well known in the art, and include NLSs previously identified as conferring the ability of the native protein to enter the nucleus, or is a putative NLS based on substantial sequence homology with a previously identified NLS. Alternatively, the NLS may be identified by sequence deletion experiments. See for example, Luo JC. Shibuya M A variant of nuclear localization signal of bipartite-type is required for the nuclear translocation of hypoxia inducible factors (1 alpha, 2alpha and 3 alpha). Oncogene, 2001 Mar 22;20(12):1435-44.or Hodel MR. Corbett AH. Hodel AE, Dissection of a nuclear localization signal. J Biol Chem. 2001 Jan 12;276(2):1317-25. Preferred membrane penetrating peptides (MPPs, also known as peptide transduction domain or TTD') of the present invention are small polypeptides, and may be derived from an NLS, or overlapping with an NLS, of a mammalian or yeast protein. Preferred mammalian proteins are those of human, primate, murine or rat species. It is generally preferred to use the same species for the NLS-derived protein as the cell to be treated. Human species are especially preferred as the NLS-derived protein when being used to treat human cells, NLSs may be found within a broad class of enzymes, and is not limited to nuclear proteins, transcription factors, cytokines and kinases. Preferred MPPs are those derived from nuclear proteins or transcription factors. Alternatively, MPPs of the present invention are small polypeptides comprising a sequence -(X-X-X-X)n- where n is an integer 1 to 7, and X each time is independently selected from the group consisting of arginine, histidine or lysine. It is preferred that small MPPs are used, and therefore, it is preferred that n is an integer 1 to 5, and more preferred that n is an integer 1 to 3. Selected embodiments of suitable MPPs are provided in Table 1 and Example 5. The MPP and/or compound of interest maybe chemically synthesized separately, for example, by chemical synthetic routes and using commercially available reagents. Alternatively, if the IVIPP and/or compound of interest is a polypeptide, it may be synthesized by recombinant technology and purified according to known methods. Host cells, cloning vectors, promoters and oligonucleotide linkers are well known and commercially available. Methodolgies for using recombinant technology and purification methods are also well known, see Current Protocols in Molecular Biology, 4 Vols. Wiley. Generally, recombinant technology is preferred, as it is more amenable to larger scale production and is more economical for mass production. Alternatively, MMPs may be obtained by specific protease degradation of a precursor proteins. The compound of interest may be attached or linked to the MPP via chemical crossliniking at the N- or C-terminus of the MPP to create a conjugated (also referred to a a fusion) MPP and compound of interest, for example, via disulfide or ester linkages. In an alternative embodiment, if the compound of interest is a peptide, the peptide may be synthesized by recombinant technology with a host cell with an expression vector encoding a fusion of the MPP sequence and the compound of interest under conditions to permit expression of the vector and obtaining the fusion MPP and compound of interest. In another embodiment, the MPP and the compound of interest may be attached or linked via a chemical linker. Chemical linkers are well known in the art, and include but are not limited to dicyclohexylcarbodiimide (DCC), N-hydroxysuccinimide (NHS), maIeiimidoben2oyl-N-hydroxysuccinimide ester (MBS), N-ethyloxycarbonyl-2-ethyIoxy"I,2-dihydroquinoline (EEDQ), N-isobutyloxy- carbonyl-2-isobutyloxy-l32-dihydroquinoline (IIDQ), Preferred linkers may also be monomeric entities such as a single amino acid, especially preferred are those amino acids with small side chains, or a small polypeptide chain, or polymeric entities of several amino acids. Preferred polypeptide linkers are fifteen amino acids or less, more preferred are polypeptide linkers of ten or less amino acids. Even more preferred are polypeptide linkers of five or less amino acids. In an alternative embodiment, the linker may be a nucleic acid encoding a small polypeptide chain; preferred linkers encode a polypeptide of fifteen or less amino acids. More preferred linkers are nucleic acids encoding a small polypeptide chains often or less amino acids. Even more preferred linkers are nucleic acid encoding a small polypeptide of five or less amino acids, such as Gly-Phe-Leu-Gly, Gly-Gly, Gly-Leu or Gly, and the like. Recombinant technology may be used to express a fusion MPP, linker and compound of interest, as described above and is well known in the art. In another embodiment, the linker may be a cleavable linker, resulting in cleavage of the MPP and compound of interest once delivered to the tissue or cell of choice. In such an embodiment, the cell or tissue would have endogenous (either naturally occuring enzyme or be recombinantly engineered to express the enzyme) or have exogenous (e.g., by injection, absorption or the like) enzyme capable of cleaving the cleavable linker. Suitable enzymes for cleavage include, for example, use of a KEX2 protease recognition site (Lys, Arg) inserted between glucoamylase and the desired polypeptide to allow in vivo release of the desired polypeptide from the fusion protein as a result of the action of a native Aspergillus KEX2-like protease. (Contreras et al, 1991; Broekhuijsen et al, 1993; Ward et al., 1995). Another example of a cleavable linker peptide comprises the recognition sequence Asp-Asp-Asp-Asp-Lys, and wherein said fusion protein is cleavable by enterokinase. Alternatively, the linker may be biodegradable such that the compound of interest is detached from the fusion MPP and compound of interest by hydrolysis and/or enzymatic cleavage inside cells. For example, tumors often express specific proteases, and be used in the delivery of prodrugs of cytotoxic agents. The linker may be selective for lysosomal proteases, such as cathepsin B, C, or D. Delivery of prodrugs and their subsequent activation is well recognized, and such an approach provides significantly less systemic toxicity due to premature linker hydrolysis in the blood, consequently a greater amount of compound of interest, i.e., drug or cytotoxic agent, is delivered to the tumor site. See for example, T, Higuchi and V. .Stella provide a thorough discussion of the prodrug concept in Pro-drugs as Novel Delivery Systems, Vol 14 of the A.C.S. Symposium Series, American Chemical Society (1975). Examples of readily-cleavable groups include acetyl, trimethylacetyl, butanoyl, methyl succinoyi, t-butyl succinoyl, ethoxycarbonyl, methoxycarbonyl, benzoyl, 3- aminocyclohexylidenyl, and the like. The compound of interest is any organic molecule, and includes small organic molecules, peptides, lipoproteins, and other modifed proteins, polysaccharides, oligonucleotides, antisense oligonucleotides, and any other compound thought to have pharmaceutical, prophylactic, diagnostic properties and/or research interest. The compound of interest may be a small organic molecule already known to have pharmaceutical properties, and thus the present invention may be used as a method of treating a patient with the compound of interest. Alternatively, the compound of interest may be a novel protein of unknown function, and thus the present invention maybe used as a method of identifying the function of the compound of interest. In another embodiment, the compound of interest may * be an antisense molecule, and thus the present invention may be used as a method of altering transcription. In yet another embodiment, the compound of interest may be a prodrug, e.g. in an inactive form but capable of being activated once within the cell. In another embodiment, the compound of interest may be a cytotoxic agent, and thus the invention may be used as a method of delivering a cytotoxic agent to a cell. The compound of interest also includes detectable proteins which are useful to generate conjugated MMP and the detectable protein for identification of new MIvIPs. Detectable proteins include GFP, beta galactosidase, radiolabeled proteins and biotinylated proteins, proteins capable of conferring a detectable phenotype in the cell. The present invention may be used to deliver the compound of interest into a cell in \itro, ex vivo or in vivo. For example, deUvery may be carried out in vitro by adding the conjugated MPP and compound of interest extracellularly to cultured cells. Delivery may be carried out ex vivo by adding the conjugated MPP and compound of interest extracellularly or exogenously to a cultured sample removed from a patient, for example, blood, tissue or bone marrow, and returning the treated sample to the patient. Delivery may be carried out in vivo by adminstering the conjugated MPP and compound of interest by transdermal administration, inhilation, or injection to a patient. Any type of cell may used in the present invention. The cell may be of mammalian, bacterial, viral-or yeast origin. The cell may be a cultured cell such as commonly used for oncology screening. Examples of cultured cells include CHO, HEK293T, HeLa, and NIH3T3. The cell may be a cultured cell from a patient suitable for ex vivo treatment with an MPP conjugate and reintroduction into a patient. The cell may be from the same or different patient than the patient to be treated. Compositions of the invention comprising the conjugated MPP and compound of interest may be used for therapeutic, prophylactic, diagnostic or research purposes. Compositions may further comprise adjuvants, stabilizers and the like to improve the handling, stability and storage properties of the compositions. Methods to identify novel MPPs are also part of the present invention. One method for identification of a membrane penetrating peptide is to generate a conjugate peptide comprising the sequence -(X-X-X-X)n- where n is an integer 1 to 7, and X each time is independently selected from the group consisting of arginine, histidine or lysine, with a detectable protein such as GFP, beta galactosidase and the like, adding the conjugate peptide to a.cell and determining if the conjugated peptide is located within the cytoplasm and/or nucleus of the cell. Another method for identification of a membrane penetrating peptide is to generate a conjugate peptide comprising a peptide derived from or overlapping with a nuclear localization sequence of a mammalian or yeast protein and a detectable protein such as GFP, beta galactosidase and the like, adding the conjugate peptide to a cell and determining if the conjugated peptide is located within the cytoplasm and/or nucleus of the cell. The following abbreviations are used for amino acids: A refers to Ala, or alanine; C refers to Cys or cysteine; D refers to Asp or aspartic acid; E refers to Glu or glutamic acid; F refers to Phe or phenylalanine; G refers to Gly or glycine; H refers to His or histidine; I refers to He or isoleucine; K refers to Lys or lysine; L refers to Leu or luecine; M refers to Met or methionine; N refers to Asn or asparagine; P refers to Pro or proline; Q refers to Gin or glutamine; R refers to Arg or arginine; S refers to Ser or serine; T refers to Thr or threonine; V refers to Val or valine; W refers to Trp or tryptophan; Y refers to Tyr or tyrosine. Proteins are written with the N-terminus to the left. The following abbreviations are used: 'v/v' refers to volume to volume; 'EYFP' refers to a peptide iragment of the sequence Glu-Tyr-Phe-Pro; 'ORF' refers to Open Reading Frame; TCR' refers to polymerase chain reaction; 'CHO' refers to Chinese Hamster Ovary cells; 'HEK293T' refers to Human Embroyonic Kidney cells, 'HeLa' refers to epithelial adenocarcinoma cells; 'NIH3T3' refers to Swiss mouse embryo fibroblast cells;' DMSO' refers to dimethyl sulfoxide; 'PCS' refers to fetal calf serum; 'DMEM' refers to Dulbecco's Modified Eagle's Medium; TBS' refers to Phosphate buffered saline;' BSA' refers to bovine serum albumin; 'C-terminus' refers to the carboxy-terminus; TSf-terminus* refers to the amino-terminus; 'PTD' refers to Peptide transduction domain; 'GPCR' refers to G-protein coupled receptor; TM' refers to a transmembrane domain of a GPCR; T refers to an intracellular loop of a GPCR; '5HT2A' refers to serotonin receptor 2 A; and 'mAb' refers to monoclonal antibody. EXAMPLES Example 1 Identification of an NLS within hPERl Plasviid Construction All hPerl fragments described here are cloned as in-frame C-temainal fusion to EYFP. EYFP-hPerl ORE, Pl-N and PI-NX (fig.lA) is generated by insertion of EcoRI and Xhol digested fragments into EYFP-Cl vector (Clontech). The other fragments are PCR amplified from the full-length hPerl cDNA and subcloned into EYFP-Cl vector. The first and the last residue present in each of fragment is indicated in Fig.lA. All constructs are verified by automated DNA sequencing. Cell culture and DNA Transfection CHO, H^La and 293T cells are maintained in Dulbecco's Modified Eagle's Medium (DMEM) supplemented with 10% fetal calf serum (FCS), 50 units/ml penicillin, 50 p,g streptomycin, and 4 mM L-glutamine (hereafter referred to as complete DMEM) at 37°C with 5% CO2. Transfection of the cells is carried in two-well Lab-Tek coverslips (Nunc Inc.) with LIPOFECT-AMINETM Reagent (Life Technologies) according to the manufacturer's instructions. Peptides and Peptide Internalization Peptides are synthesized by a commercial vendor (Bio Synthesis). For peptides internalization, cells are plated into two-well Lab-Tek coverslips (Nunc Inc) at a density of 2X105 cells/well and cultured overnight. The peptides are dissolved in DMSO diluted to indicated concentration with PBS, The cell monolayers were incubated with the appropriate peptide/PBS solution at 1µM standard concentration for 10 min at room temperature (RT) unless otherwise specified. For experiments at 4°C, the protocol was the same except that all incubations were performed at 4°C until the end of the fixation procedure. Immunofluorescence and Microscopy For direct detection of expression and subcellular localization of EYFP fusion protein, transfected cells were examined directly without fixation or afler fixation with 4%(v/v) fonnaldehyde in PBS for 20 min at A°C and washed with PBS. For indirect immunodetection of biotinylated peptides, fixed cell were washed twice with PBS and permeabilized with 0,3% Triton X-100 in PBS for 20 min at 4°C and blocked with 2% BSA in PBS for 30 min at RT. Cells were then washed with PBS and incubated with Streptavidin-FITCTM (Sigma) or -Alex499 (Molecular Probe), 1:400 diluted in 0.2% Tween 20, 2% BSA in PBS for 1 h at RT. Following 2 X 5 min washes with PBS and once with 0.3% Triton X400 in PBS for 20 min RT. In some experiment, the nucleus was stained with 50 ng/ml Hoechst 3325S (Sigma) or 3 j-Lg/ml propidium iodide in PBS. The subcellular localization of the fluorescence was analyzed on an Olympus microscope. Confocal images were taken on a Zeiss confocal laser scan microscope (CLSM phoibos lOOO). Though it is known that nuclear entry of PERI is important for its function, no putative NLS was identified using a standard Profile Scanning program (Shearman, L.P., et al„ Neuron 19, 1261-1269 (1997), Yagita, K., et al.. Genes Dev. 14, 1353-1363 (2000)). To determine theNLS of hPERl experimentally, three full-length hPERl (Pl-FL) were constructed and denoted as Pl-N, Pl-NM and Pl-C (Fig. lA). The ability of these constructs to localize to the nucleus in CHO cells then analyzed. An EYFP-tag was used to facilitate the detection of hPERl in living cells; however, the EYEP-tag had no apparent contribution on hPERl fusion protein localization since hPERl constructs made with an N-terminal Flag-tag presented an identical cytological distribution pattern (data not shown). After transient transfection, both Pl-FL and Pl-NM proteins were expressed in the nucleus of transfected cells as early as 10 hours post-transfection, while both Pl-N and PI-C accumulated only in the c>loplasm (Fig.IB). The EYFP vector control was diffuse in both the nucleus and cytoplasm. These results demonstrate that a fiinctional NLS in hPERl is located between Pl-N and Pl-C in what we designated as region M (see Fig.l A). To further localize the NLS in region M (amino acids 481-890), a series of 8 deletion constructs, Pl-Fl to P1-F8, were generated and the subcellular distribution of each mutant was assayed as indicated in Fig. lA and B. Sequential deletion from amino acid 581 (P1-E2) to position 82I(P1-F7) of region M resulted in nuclear localization. Further deletion of amino acids 821 to 841 (P1-F8) resulted in a diffused fluorescent pattern within transfected cells with a localization pattern similar to that of the EYFP vector control. These data indicate that a NLS exists between amino acids 821 and 890, and is located at the C-terminus of region M. This observation was confirmed by the construction of an additional EYFP fusion protein, Pl-NLS, which contained hPERl amino acids 830-845. This region contains a string of basic residues that might function as a NLS (Weis, K., Trends Biochem. ScL 23, 1S5-189 (1998), Truant, R. and CuUen, B.R. Mol Cell BioL 19, 1210-1217 (1999)). As expected, Pl-NLS exhibited nuclear localization in 100% of transfected cells (Fig. IB). Other regions of PERI in additional fusion constructs failed to localize to the nucleus (data not shown). Therefore, we conclude that the NLS of hPERl (hPERl-NLS) is localized to within amino acids 830-845. Interestingly, construct Pl-Fl has a strictly cytoplasmic localization pattern irrespective of the fact that it contains the NLS, supporting published observations that this region also contains and as yet unidentified cytoplasmic localization domain (Vielhaver, E., et al., Mol Cell BioL 20, 4SSS-4899 (2000)). Sequence alignment shows that the hPERl-NLS is conserved between human and mouse PERI proteins (Fig. 1 A), but not with other putative NLSs, or with other human, mouse or Drosophila PERs. After completion of our studies, Vielhaber et al (2000), identified a longer mouse PERI-NLS that contains our identified 16 The peptides are assayed for their ability to penetrate cellular membranes. Intracellular localization is assayed by direct staining with labeled Streptavidin ALEXA reagents or by indirect staining with anti-Flag niAb followed by the addition of labeled secondary antibodies. When added to the cells in culture at a concentration of 10 jiM, hPERl-MPP, Flag-hPERl-MPP and Flag TAT-PTD peptides are found to penetrate rapidly into 100% cells (Fig, 2A and Fig. 5). By both detection methods, hPERl-MPP, Flag-tagged hPERl-MPP, and Flag-tagged TAT-PTD are observed to be diffusely distributed throughout the cytoplasm, but concentrated within subnuclear domains that appear as distinct foci within the nucleoplasm and the nucleolus. In contrast, biotinylated negative control peptides,. Flag-Flag and several additional peptides derived from other hPERl regions are only barely discernible background staining, with no staining in the nucleus or nucleoli, even at high concentrations (data not shown). Confocal microscopy is used to confirm the intracellular and intranuclear staining of Flag-tagged hPERl-MPP, and that the negative control peptides are not internalized (Fig. 2A). hPERl-MPP rapidly penetrated the cellular membranes and localized in nuclear regions with efficiencies similar to the TAT-PTD peptide (Fig. 2B). Identical results are obtained using CHO, HEK293T, HeLa, NIH3T3 and cultured rat primary cortical neurons (data not shown), indicating cell type-independent penetration. hPERl-MPP internalization occurrs rapidly (within 5 min), with similar potencies at 4C and 37C and even after cell membrane fixation (data not shown). Thus, the amino acid sequence 830-845 of hPERl functions as both as a protein nuclear/nucleolar localization signal in the fusion protein and as a MPP, and that membrane penetration is independent of traditional receptor-mediated endocytic mechanisms. Example 3 Arginine 7 is essential for hPERl-MPP activity To date, the mechanisms as well as the structural basis whereby MPPs transverse cellular membranes have not been elucidated. We therefore sought to determine if there were key residues within hPERl-MPP that were important for maintaining those properties essential for its membrane penetrating potential We separately replaced each amino acid in hPERl-MPP to alanine (Table 2), and assayed for the ability of these mutated peptides to penetrate living cells relative to the wild-type hPERl-lVlPP- Alanine scaning: As shown in Fig. 3, most of the single alanine substitutions had very little effect on membrane penetrating capabilities as compared with wild-type peptide. However, changing arginine 7 to an alanine (R7A) reduced the detectable cytological signal to that observed for the negative control peptides. Thus, the arginine 7 to alanine mutation significantly reduced * the membrane penetrating properties of hPERl-MPP. Identical observations were observed after changing arginine 7 to glutamic acid (R7E) (data not shown). Furthermorej the simultaneous deletion of the N-terminal serine and of the two C-terminal histidines from hPERl-MPP (hPERl-MPP13) had little overall effect on the positive membrane penetrating potential of the peptide (Fig. 3). The arginine 7 residue plays a critical role in the cell penetrating ability of the hPERl-MPP. We therefore sought to determine if the R7A mutation affected nuclear translocation of a fusion protein Pl-NLS. CHO cells transfected with fusion protein PI-R7A (arginine 7 mutated to alanine) have intense nuclear staining similar to the wild-type, Pl-NLS (data not shown). Nuclear translocation appears to be normal in the P1-R7A mutant fusion protein, but subnuclear targeting to the nucleoli is disrupted (data not shown). These data indicate that membrane penetration and nucleoli targeting are affected by the single R7 amino acid residue and indicate that nuclear translocation of hPERl-NLS has separate and distinct detemiinants. Example 4 hPERl-MPP delivery of functioning molecules One of the features of MPPs is their ability to cargo proteins and peptides into cells. We were successful in coupling hPERl-MPP to B-galactosidase and in delivering the fusion protein into cells in culture (data not shown), as has been described by Fawells et al., 1994 (Fawell, S., et al., Proc Natl Acad Sci USA. 91, 664-668 (1994)), However, to further extend the functional utility of MPPs, we tested hPERl-MPP in fusion with a physiologically relevant and biologically active peptide. Wess and collegues (1993) have shown a functional role for the conserved transmembrane segment 7 (TM7) of the G-protein coupled receptor (GPCR) superfamily. Along with TM7, the third intracellular loop (13) plays a significant role in GPCR calcium signaling (Wess, JM.,et al., M50/ 12,331-338 (1993)) while intracellular loops 1 and 2 (II and 12) appear not to be important. Using the serotonin receptor, 5HT2A, we experimentally tested the ability of hPERl-MPP and TAT-PTD in fusion with peptides designed from II and the TM7 domains to activate the receptor. Biotinylated peptides hPERl-MPP TM7, TAT-PTD TM7, hPERl-MPP II, TAT-PTD II, hPER-MPP, TAT-PTD, TM7 or II were incubated at a concentration of 10 fiM with a 5HT2A receptor CHO stable cell line. Peptide membrane penetration was assayed using Streptavidin-Alexa 594 as described above. As shown in Fig. 4A, receptor signaling is activated by the addition of exogenous serotonin, hPERl-MPP TM7, and TAT-PTD TM7 as measured by the level of the calcium response. However, TM7 alone nor any of the other peptides were able to generate a calcium response. Furthermore, the activation of the receptor by hPERl-MPP TM7 m and TAT-PTD TM7 is peptide concentration dependent, Fig. 4B. The addition of increasing concentrations of the activating peptide, TM7, in fusion with hPERl-MPP or TAT-PTD results in a calcium response in a dose dependent manner. TAT-PTD TM7 appears to be a more potent 5HT2A receptor activator than is hPERl-MPP TM7. A simple explanation for this result is that TAT-PTD TM7 is more cytoplasmically locahzed or has greater cell penetrating capablilites than hPERl-MPP TM7, although we have not observed that to be the case. Similar results were also obtained in this laboratory using hPERl-MPP in fusion with P-adrenergic activating peptides (unpublished data). These data support previous results that hPERl-MPP not only penetrates cell membranes, but also demonstrates that it is capable of cargoing peptides to intracellular compartments to initiate biologically relevant signal transduction events. Example 5 Identification of other gene encoded MPPs Since hPERl is a nuclear protein proposed to be involved in transcriptional regulation, and since, to date, all PTDs derived from naturally occurring proteins are transcription factors (TAT,, Antp, and VP22), we sought to determine if other PTD sequences existed within the genome. To this end, we used two approaches; first, we searched the NCBI non-redundant protein database for all known and putative NLS's (table 1,10-17), We synthesized peptides corresponding to the NLS amino acid sequences and assayed for peptide transduction. As shown in table 1 and Fig. 5, 6 of the 7 peptides synthesized had membrane penetrating characteristics similar to hPER-PTD and TAT-PTD. These proteins included human proteins of the thyroid honnone receptor alpha-1, homeobox protein HMEl, and proto-oncogene protein ABL-L Furthennore, (table 1 and Fig, 5) when we create in frame fusion proteins between these peptide sequences and GFP then transfected into CHO or HEK 293T cells, all of the sequences conferred nuclear localization of the fusion protein. Our second approach to identifying PTDs involved searching the NCBI non-redundant protein database collection with a degenerative algorithm (see Fig. 5, legend). Using these search parameters, we found 533,291 sequences of which the conditions for the algorithm were satisfied 129,169 times (24%). By limiting our search to include either "transcription factors, cytokines or tyrosine kinases", we identified 8280 transcription factor protein sequences of which the algorithm pattern occurred 7374 times (89%); within 2333 cytokine protein sequences the pattern occurred 450 times (19%); and within 2513 tyrosine kinase protein sequences the pattern occurred 843 times (36%)). Because the algoritlim occurred most frequently in nuclear proteins, we synthesized peptides to putative PTDs for 6 of the "transcription factor" sequences and assayed for their ability to penetrate into the cells. As shown in table 1, results in lines 18-23 and fig 3 A, 4 of the 6 peptides tested had membrane penetrating properties similar to hPERl-PTD and TAT-PTD, These proteins included two human proteins HENl/NSLC-1 and HEN2/NSLC-2 which are reported to be involved in neuronal differentiation and development (Uittenbogaard, M., Peavy, D.R, and Chiaramello, A. 1999. Expression of the bHLH gene NSCL-1 suggests a role in regulation of cerebellar granule cell growth and differentiation. J, Neurosci Res. 51:710-7Sl, Lipkowitz, S. et al. 1992. A comparative structural characterization of the human NSCL-1 and NSCL-2 genes. Two basic helix-loop-helix genes expressed in the developing nervous system. 1 Biol, Chem. 267:21065-21071), rat HNF-3 (17), and a Drosophila cAMP dependent transcription factor (IS). Furthermore, (table 1 and Fig 5) when we create in frame fusion proteins between these peptides and GFP and transfected into CHO or HEK 293T cells, all of the sequences conferred nuclear localization of the fusion protein. These results indicate that PTD sequences can be found within or overlapping with NLSs. However not all NLSs are PTDs as is apparent in SV40, hPER2, C-FOS, Cyclin L ania-6 and beta Zip transcription factor NLSs (table 1), These results also suggest that PTDs sequences are prevalent throughout the genome and in particular within nuclear proteins. Example 6. hPER-PTD with β-Galactosidase At least one feature of HIV TAT transducing peptide is its ability to cargo proteins into cells and tissues. We therefore sought to determine if hPERl transducing peptide could cargo beta galactosidase into cells. To perform this experiment, we followed a protocol by Frankel et al PNAS 19S9 (19):7397-401, whereby, we chemically linked hPERl-PTD or hPER-PTD R7A(with Ala replacing Arg7) to full length β-galactosidase and assayed for the ability of these conjugates and beta-galactosidase protein alone to transduce into CHO cells. As shown in the figure 65 panel 1, cells incubated with hPER-PTD β-galactosidase fusion showed positive enzymatic activity for β-Galactosidase as indicated by the blue color in the cells after the addition of X-gal However, neither hPER-MPP R7A β-Galactosidase nor β-Galactosidase protein alone was able to enter the cells as indicated by a no blue staining reactivity after the addition of X-gal, panels 2 and 3. These data indicate that like TAT peptide, hPERl -PTD can cargo a large (120 kD) protein into cells. We claim: 1. A fusion protein for delivery of a compound of interest into a cell, comprising a membrane penetrating peptide attached to a compound of interest. 2. The fusion protein according to claim 1, wherein the membrane penetrating peptide is derived from a nuclear localization sequence, overlaps with a nuclear localization sequence of a mammalian or yeast protein or comprises a sequence -(X-X-X-X)n- where n is an integer 1 to 7, and X each time is independently selected from the group, consisting of arginine, histidine or lysine. 3. The fusion protein according to claim 2, wherein the nuclear localization sequence is derived from a nuclear protein or transcription factor. 4. The fusion protein according to claim 3, wherein the transcription factor is a Period protein. 5. The fusion protein according to claim 4, wherein the Period protein is a human Period protein. 6. The fusion protein according to claim 5, wherein the mammalian Period protein is human Period protein. 7. The fusion protein according to claim 2 wherein the membrane penetrating peptide comprises the sequence -(X-X-X-X)n- where n is an integer 1 to 7, and X each time is independently selected from the group consisting of arginine, histidine or lysine, 8. The fusion protein according to claim 7, wherein n is an integer 1 to 4. 9. The fusion protein according to claim 8, wherein n is an integer 1 to 2, 10. The fusion protein according to claim 1, wherein the compound of interest is a peptide, protein, chemical entity, nucleic acid, or any modified form thereof 11. A method of delivering a compound of interest into a cell, comprising contacting a cell with a fusion protein according to claim L 12. The method of delivering a compound of interest into a cell in vitro, comprising contacting a cultured cell with a fusion protein according to claim 1. 13. The method of delivering a compound of interest into a cell ex vivo, comprising contacting a cell with a fusion protein according to claim 1 and introducing the cell into the body of a patient. 14. The method of delivering a compound of interest into a cell in vivo, comprising administering to a patient a fusion protein according to claim 1. 15. A method for identifying a membrane penetrating peptide, wherein a peptide comprises a sequence -(X-X-X-X)n- where n is an integer 1 to 7, and X each time is independently selected from the group consisting of arginine, histidine or lysine, by generating a conjugate peptide comprising the sequence -(X-X-X-X)n- where n is an integer 1 to 7, and X each time is independently selected from the group consisting of arginine, histidine or lysine, with a detectable protein, adding the conjugate peptide exogenously to a cell and determining if the conjugated peptide is located within the cytoplasm and/or nucleus of the cell. 16. A method for identifying a membrane penetrating peptide, wherein a peptide comprises a sequence derived from or overlapping with a nuclear localization sequence of a mammalian or yeast protein, by generating a conjugate peptide comprising the part or all of the nuclear localization sequence with a detectable protein, adding the conjugate peptide exogenously to a cell and determining if the conjugated peptide is located within the cytoplasm and/or nucleus of the cell. 17. The method of delivering a compound of interest into a cell, comprising administering to a cell a fusion protein according to claim 1, wherein the membrane penetrating peptide comprises a sequence -(X-X-X-X)n- where n is an integer 1 to 7, and X each time is independently selected from the group consisting of arginine, histidine or lysine. IS. A fusion protein for delivering a compound of interest into a cell, wherein the fusion protein comprises a membrane penetrating peptide comprising a sequence -(X-X-X-X)n- where n is an integer 1 to 7, and X each time is independently selected from the group consisting of arginine, histidine or lysine, and a compound of interest. 19. The fusion protein of claim 18, wherein the compound of interest is directly chemically attached to the membrane penetrating peptide or by a linker. 20. The fusion protein of claim 19, wherein the linker is an amino acid linker or a polypeptide linker. 21. The fusion protein of claim 18, wherein the membrane penetrating protein is produced by recombinant technology, chemical synthesis or degradation of a precursor protein. 22. A fusion protein substantially as herein described with reference to the accompanying drawings. 23. A method for identifying a membrane penetrating peptide substantially as herein described with reference to the accompanying drawings.

Documents:

314-chenp-2003-abstract.pdf

314-chenp-2003-assignement.pdf

314-chenp-2003-claims filed.pdf

314-chenp-2003-claims granted.pdf

314-chenp-2003-correspondnece-others.pdf

314-chenp-2003-correspondnece-po.pdf

314-chenp-2003-description(complete) filed.pdf

314-chenp-2003-description(complete) granted.pdf

314-chenp-2003-drawings.pdf

314-chenp-2003-form 1.pdf

314-chenp-2003-form 26.pdf

314-chenp-2003-form 3.pdf

314-chenp-2003-form 5.pdf

314-chenp-2003-other documents.pdf

314-chenp-2003-pct.pdf