AU742018B2 - Alternatively targeted adenovirus - Google Patents

Description

WO 98/54346 PCT/US98/11024 1 ALTERNATIVELY TARGETED ADENOVIRUS TECHNICAL FIELD OF THE INVENTION The present invention relates to an alternately targeted adenovirus and includes methods for producing and purifying such viruses as well as protein modifications mediating alternate targeting.

BACKGROUND OF THE INVENTION Adenoviral infection begins with the attachment of the virion to the target cell.

The adenovirus attaches to two cellular surface proteins, both of which must be present for the virus to infect the target cell (Wickham et al., Cell, 73, 309-19 (1993)). Wild-type adenovirus first binds the cell surface by means of a cellular adenoviral receptor (AR).

One such AR is the recently-identified coxsackievirus and adenovirus receptor (CAR) (Bergelson et al., Science, 275, 1320-23 (1997); Tanako et al., Proc. Nat. Acad. Sci., 94, 3352-56 (1997)); the MHC class I receptor also is an AR (Hong et al., EMBO 16(9), 2294-06 (1997)). After attachment to an AR, the virus attaches to av integrins, a family of a heterodimeric cell-surface receptors mediating interaction with the extracellular matrix and playing important roles in cell signaling (Hynes, Cell, 69, 11-25 (1992)).

Following attachment to the cell surface, infection proceeds by receptor-mediated internalization of the virus into endocytotic vesicles (Svensson et al., J. Virol., 51, 687-94 (1984); Chardonnet et al., Virology, 40, 462-77 (1970)). Within the cell, virions are disassembled (Greber et al., Cell, 75, 477-86 (1993)), the endosome disrupted (Fitzgerald et al., Cell, 32, 607-17 (1983)), and the viral particles transported to the nucleus via the nuclear pore complex (Dales et al., Virology, 56, 465-83 (1973)).

The adenoviral virion is a non-enveloped icosahedron about 65-80 nm in diameter (Home et al., J. Mol. Biol., 1, 84-86 (1959)). The adenoviral capsid comprises 252 capsomeres 240 hexons and 12 pentons (Ginsberg et al., Virology, 28, 782-83 (1966)).

The hexons and pentons are derived from three viral proteins (Maizel et al., Virology, 36, 115-25 (1968); Weber et al., Virology, 76, 709-24 (1977)). The hexon comprises three identical proteins of 967 amino acids each, namely polypeptide II (Roberts et al., Science, 232, 1148-51 (1986)). The penton contains a base, which is bound to the capsid, and a fiber, which is non-covalently bound to and projects from, the penton base. Proteins IX, VI, and IIIa also are present in the adenoviral coat and are thought to stabilize the viral capsid (Stewart et al., Cell, 67, 145-54 (1991); Stewart et al., EMBO 12(7), 2589-99 (1993)).

The penton base is highly conserved among serotypes of adenovirus and (except for the enteric adenovirus Ad40) has five RGD tripeptide motifs (Neumann et al., Gene, 69, 153-57 (1988)). In adenovirus, the RGD tripeptides apparently mediate adenoviral WO 98/54346 PCT/US98/11024 2 binding to a, integrins and endocytosis of the virion (Wickham et al. (1993), supra; Bai et al., J. Virol., 67, 5198-3205 (1993)).

The adenoviral fiber is a homotrimer of the adenoviral polypeptide IV (Devaux et al., J. Molec. Biol., 215, 567-88 (1990)). Structurally, the fiber has three discrete domains. The amino-terminal tail domain attaches non-covalently to the penton base. A relatively long shaft domain comprising a variable number of repeating 15 amino acid residues forming P-sheets extends outward from the vertices of the viral particle (Yeh et al., Virus Res., 33, 179-98 (1991)). Lastly, roughly 200 amino-acid residues at the carboxy-terminal form the knob domain. Functionally, the knob mediates primary viral binding to the cellular AR and fiber trimerization (Henry et al., J. Virol., 68(8), 5239-46 (1994)). Hence, the trimerization domain of a fiber is a ligand for a cell-surface receptor native for the adenoviral serotype. The trimerization domain also appears necessary for the tail of the fiber to properly associate with the penton base (Novelli et al., Virology, 185, 365-76 (1991)). In addition to recognizing cell ARs and binding the penton base, the fiber protein contributes to serotype integrity and mediates nuclear localization.

Fiber proteins from different adenoviral serotypes differ considerably. For example, the number of 15 amino-acid P-sheet repeats differs between adenoviral serotypes (Green et al., EMBO 2, 1357-65 (1983)). Moreover, the knob regions from the closely related Ad2 and Ad5 serotypes are only 63% similar at the amino acid level (Chroboczek et al., Virology, 186, 280-85 (1992)), and Ad2 and Ad3 fiber knobs are only identical (Signas et al., J. Virol., 53, 672-78 (1985)). In contrast, the penton base sequences are 99% identical. Despite these differences in the knob region, a sequence comparison of even the Ad2 and Ad3 fiber genes demonstrates distinct regions of conservation, most of which are also conserved among the other human adenoviral fiber genes.

A number of factors present the adenovirus as an attractive vector choice for use in a variety of gene transfer applications cellular protein production, therapy, academic study, etc.). For example, the adenovirus is a superior expression vector. Recombinant adenovirus can be produced in high titers about 10 1 3 viral particles/ml), and adenoviral vectors can transfer genetic material to non-replicating, as well as replicating, cells (in contrast with retroviral vectors). The adenoviral genome can be manipulated to carry a large amount of exogenous DNA (up to about 7.5 kb), and the adenoviral capsid can potentiate the transfer of even longer sequences (Curiel et al., Hum. Gene Ther., 3, 147-54 (1992)). Additionally, several features suggest that adenoviruses represent a safe choice for gene transfer, a particular concern for therapeutic applications. For example, adenoviruses do not integrate into the host cell chromosome, thus minimizing the likelihood that an adenoviral vector will interfere with normal cell function. Moreover, adenoviral infection does not correlate with human malignancy, and recombination of the WO 98/54346 PCT/US98/11024 3 adenoviral genome is rare. Due to these advantages, clinicians have employed adenoviral vectors safely as a human vaccine and for gene therapy for many years.

Based on the popularity of adenoviral vectors, efforts have been made to increase the ability of adenovirus to enter certain cells, those few cells it does not infect, an approach referred to as "targeting" (see, International Patent Application WO 95/26412 (Curiel et International Patent Application WO 94/10323 (Spooner et al.), U.S. Patent 5,543,328 (McClelland et International Patent Application WO 94/24299 (Cotten et Of course, while the ability to target adenoviruses to certain cell types is an important goal, far more desirable is an adenovirus which infects only a desired cell type, an approach referred to as "exclusive targeting." However, to exclusively target a virus, its native affinity for host cell ARs must first be abrogated, producing a recombinant adenovirus incapable of productively infecting the full set of natural adenoviral target cells. Efforts aimed at abrogating native adenoviral cell affinity have focused logically on changing the fiber knob. These efforts have proven disappointing, largely because they fail to preserve the important fiber protein functions of stable trimerization and penton base binding (Spooner et al., supra). Moreover, replacement of the fiber knob with a cell-surface ligand (McClelland et al., supra) produces a virus only suitable for infecting a cell type having that ligand. Such a strategy produces a virus having many of the same targeting problems associated with wild-type adenoviruses (in which fiber trimerization and cellular tropism are mediated by the same protein domain), thus decreasing the flexibility of the vector. Moreover, due to the necessity of having a host cell, and the integral connection between the fiber trimization and targeting functions, obtaining a mutant virus with substituted targeting is difficult. For example, removing the fiber knob and replacing it with a non-trimerizing ligand McClelland et al., supra) results in a virus lacking appreciable fiber protein. As such, there is currently a need for an adenoviral fiber having reduced affinity for natural ARs but retaining fiber trimerization and penton base-binding function.

While exclusive adenoviral targeting requires reducing native cellular tropism, the abrogation of natural targeting also reduces the ability of the virus to infect cell lines normally employed for its propagation 293 cells) (see Curiel et al., supra). One published attempt at surmounting this barrier fortuitously employed a cell line expressing the relevant cell surface binding site (McClelland et al., supra), and thus did not address this central concern. However, many cell lines do not express important cellular receptors. Moreover, many available cell lines expressing potentially useful cell surface binding sites are inadequate for production of recombinant adenoviruses, especially viruses useful for clinical application cell lines harboring and expressing the essential adenoviral immediate early genes from the El, E2, and/or E4 regions of the genome). There is thus a need for a cell line, and a means of producing a cell line, which WO 98/54346 PCT/US98/11024 4 can propagate and package a recombinant adenovirus substantially incapable of productively infecting cells via native ARs.

Typical protocols for purifying viral vectors from packaging cell lysates involve centrifuging the viruses through a CsCl 2 gradient one or more times. While such methods adequately isolate viruses, they generally require considerable material (CsCI 2 and are therefore relatively inefficient. Moreover, such protocols are not readily amenable to high throughput application, presenting a significant barrier to economic development of viral vectors on a commercial scale. Other methods involving column purification do not bind the viruses specifically (Shabram et al., Hum. Gene Ther., 8, 453 (1997); Huyghe et al., Hum. Gene Ther., 6, 1403 (1995)), often resulting in an unacceptable amount of contaminants compared to the purity obtainable in affinity purification of other materials.

Thus, there is a need for an efficient method of purifying and isolating recombinant viral vectors.

In many applications involving in vivo delivery of viral vectors, it is desirable to contain infection (and gene delivery) to the tissue of interest. For example, the threat of systemic infection and delivery of a biologically active gene represents a significant concern to gene therapy applications. Moreover, ectopic expression of a transgene would spoil many experimental applications. While, in theory, host blood cells can express proteins mediating the clearing of foreign substances, such as adenoviruses (News and Comment, Science, 275, 744-45 (1997)), engineering such cells and producing them in the host are difficult and intrusive. Moreover, while antibodies directed against the adenoviral hexon can inactivate the virus (Toogood et al., J. Gen. Virol., 73, 1429-35 (1992)), efficient protocols for delivering a sufficient quantity of anti-hexon antisera to the gene transfer recipient in time to reduce or prevent ectopic viral infection have not been forthcoming, and such a strategy can actually interfere with gene transfer protocols by blocking infection in desired tissues. Thus, there is a need for a method of inactivating recombinant viral vectors leaving the desired locus of delivery within a host animal.

BRIEF SUMMARY OF THE INVENTION The present invention provides a trimer comprising three monomers, each having an amino terminus of an adenoviral fiber protein and each having a trimerization domain.

The trimer exhibits reduced affinity for a native substrate than a native adenoviral fiber trimer. The present invention further provides an adenovirus incorporating the trimer of the present invention. The present invention also provides a cell line expressing a nonnative cell-surface receptor to which an adenovirus having a ligand for the receptor binds, and a method of propagating an adenovirus using the cell line.

The present invention also provides a method of purifying an adenovirus having a ligand for a substrate from a composition comprising the adenovirus. The method WO 98/54346 PCT/US98/11024 to selectively bind the substrate. Subsequently, the composition not bound to the substrate is separated from the substrate, after which the bound adenovirus is eluted from the substrate.

The present invention further provides a method of inactivating an adenovirus having a ligand recognizing a blood- or lymph-borne substrate by exposing the virus to the substrate. Within the blood or lymph, the ligand binds its substrate, thereby adsorbing the free virus from the blood or lymph.

Additionally, the present invention provides a chimeric blocking protein comprising a substrate for an adenovirus fiber, and a method of interfering with adenoviral receptor binding by incubating an adenovirus with such chimeric blocking protein in a solution such that the chimeric blocking protein binds the fiber.

The present invention is useful in a variety of applications, in vitro and in vivo, such as therapy, for example, as a vector for delivering a therapeutic gene to a cell with minimal ectopic infection. Specifically, the present invention permits more efficient production and construction of safer vectors for gene therapy applications. The present invention is also useful as a research tool by providing methods and reagents for the study of adenoviral attachment and infection of cells and in a method of assaying receptorligand interaction. Similarly, the recombinant fiber protein trimers can be used in receptor-ligand assays and as adhesion proteins in vitro or in vivo. Additionally, the present invention provides reagents and methods permitting biologists to investigate the cell biology of viral growth and infection. Thus, the vectors of the present invention are highly useful in biological research.

BRIEF DESCRIPTION OF THE FIGURES Figures 1A and 1B depict the three-dimensional structure of an adenoviral knob protein (serotype Figure 1A is a ribbon diagram representing p-sheets and the loops interconnecting the sheets. Figure 1B is a filled-in diagram taking into account the relative sizes of the amino acid residues.

Figure 2 is a sequence comparison between adenoviral serotypes.

Figures 3A-3C depict vectors for creating recombinant adenoviral fiber trimers having non-native trimerization domains. Figure 3A depicts pAcT5S7GCNTS.PS.LS.X.

Figure 3B depicts pAcT5sigDel.TS.PS.LS. Figure 3C depicts pAcT5S7sigDel.TS.PS.LS.

Figure 4 depicts pAcT5sigDel.GFP.TS.PS.LS, a vector containing a gene encoding a fiber-sigDel-GFP chimera.

Figures 5A-5D depict vectors useful for the construction of recombinant adenovirus vectors containing fiber trimers having non-native trimerization domains.

Figure 5A depicts pAS pGS HAAV. Figure 5B depicts pAS pGS pK7. Figure 5C depicts pAS T5S7sigDelpGS.HAAV. Figure 5D depicts pASTSS7sigDel.GFP.pGS.pK7.

WO 98/54346 PCT/US98/11024 6 Figures 6A-6D represent vectors used in the construction of fiber trimers having non-native trimerization domains. Figure 6A represents pAcPig4KN. Figure 6B represents pAcPigKN D363E. Figure 6C depicts pAcPigKN N437D. Figure 6D depicts pAcPig4KN(FLAG).

Figures 7A-7B represent vectors employed in creating a fiber trimer having a nonnative trimerization domain. Figure 7A depicts PNS F5F2K. Figure 7B depicts pNS Pig4.SS.

Figures 8A-8C represent vectors useful for creating an adenoviral vector having a chimeric fiber trimer comprising a mutant NADC-1 knob lacking native receptor-binding ability and containing a functional non-native ligand. Figure 8A depicts pAcPig4KN D363E N437D. Figure 8B depicts pAcPig4KN D363E N437D HAAV. Figure 8C depicts pNS Pig4 D363E N437D HAAV SS.

Figures 9A-9B represent vectors useful for creating chimeric blocking proteins of the present invention able to interfere with native adenoviral receptor binding. Figure 9A depicts pACSG2-sCAR. Figure 9B depicts pACSG2-sCAR-HAAV.

Figures 10A-10B represent vectors useful for creating chimeric blocking proteins able to form trimers interfering with native adenoviral receptor binding. Figure depicts pAcSG2sCAR.sigDel. Figure 10B depicts pAcSG2-sCARsigDel (HAAV).

Figures 11 A-1 1E depict vectors useful for creating construction of adenovirus vectors having specific non-native ligands. Figure 11A depicts pBSSpGS. Figure 11B depicts pBSS pGS (RKKK)2. Figure 11C depicts pNSF5F2K(RKKK)2. Figure 11D depicts pBSSpGS (FLAG). Figure 11E depicts pNS F5F2K(FLAG).

Figures 12A-12E represent vectors useful for creating a cell line expressing a nonnative cell surface binding site substrate. Figure 12A depicts pHOOK3. Figure 12B depicts pRC/CMVp-Puro. Figure 12C pScHAHK. Figure 12D depicts pNSE4GLP.

Figures 13A-13D represent vectors useful for creating a fiber-expressing cell line for the production of targeted adenovirus particles. Figure 13A depicts pCR2.1- TOPO+fiber. Figure 13B depicts pKSII Fiber. Figure 13C depicts pSMTZeo-DBP.

Figure 13D depicts pSMTZeo-Fiber.

Figure 14 depicts pAdE1(Z)E3/E4(B), a plasmid useful for the construction of targeted adenovirus particles having genomes encoding chimeric fibers.

Figures 15A-15E illustrate the locations of mutations within adenoviral knobs which interfere with ligand binding. Figure 15A is a top view, Figure 15B a side view, and Figure 15C a bottom view of the knob illustrating the location of the 3D9 mutation.

Figure 15D is a top view, Figure 15E a side view, and Figure 15F a bottom view of the knob illustrating the locations of the CD loop mutation, the FG loop mutation, and the IJ mutation.

WO 98/54346 PCT/US98/11024 7 Figure 16 depicts a vector useful for the construction of a recombinant adenovirus containing a short-shafted fiber and a mutant fiber knob exhibiting reduced affinity for its native receptor.

Figures 17A-17B depict vectors useful for constructing a cell line able to replicate adenoviruses lacking native cell-binding function (but targeted for a pseudo-receptor).

Figure 17A depicts pCANTAB5E(HA). Figure 17B depicts pScFGHA.

DETAILED DESCRIPTION OF THE INVENTION Definitions An adenovirus is any virus of the genera Mastadenoviridae or Aviaadenoviridae, and can be of any serotype within those genera. Adenoviral stocks that can be employed as a source of adenovirus or adenovirus coat protein such as penton base and/or fiber protein can be amplified from the adenovirus serotypes currently available from American Type Culture Collection (ATCC, Rockville, MD), or from any other source.

A ligand is any species selectively binding an identifiable substrate.

Native refers to a protein or property of an unmodified virus or cell. Thus, a nonnative protein can be a modified or mutated protein differing from its native homologue within the virus or cell. Alternatively, a non-native protein can be a protein having no native homologue within the virus or cell.

An AR refers to an adenoviral receptor. In particular, an AR is a ligand binding the mastadenoviral knob.

A first species is selectively bound to a substrate if it binds the substrate with greater affinity than a second species. The first species is not selectively bound if binds the substrate with the same or lesser affinity than the second species, even if the first species binds with some affinity.

Trimers The present invention provides a trimer comprising three monomers at least a portion of each of three adenoviral fiber monomers), each having an amino terminus derived from an adenoviral fiber protein and each having a trimerization domain. The inventive trimer exhibits reduced affinity for a native substrate, such as an antibody, a cellular binding cite, etc. native to the serotype from which the shaft, and particularly the amino-terminus, is drawn) as compared to a native adenoviral fiber trimer. The trimer can be a homotrimer or a heterotrimer of different fiber monomers. Any modification of the monomeric units reducing the affinity of the resulting trimer for its native cell surface binding site a native AR) is within the scope of the invention. Preferably, the reduction in affinity is a substantial reduction in affinity (such as at least an order of magnitude, and preferably more) relative to the unmodified corresponding fiber.

WO 98/54346 PCT/US98/11024 8 As mentioned, where a trimerization domain is itself a ligand for a native cell surface binding site, trimers possessing such trimerization domains present some of the same problems for targeting as native adenoviral fiber trimerization domains. Therefore, the trimerization domain of a monomer incorporated into the trimer of the invention preferably is not a ligand for the CAR or MHC-1 cell surface domains, or antibodies recognizing the fiber. Most preferably, the non-native trimerization domain is not a ligand for any native mammalian cell-surface binding site, whether the site is an AR or other cell surface binding site. As is discussed herein, adenoviruses incorporating such trimers exhibit reduced ability to appreciably infect their native host cells, and can serve as efficient source vectors for engineering selectively targeted vectors. Therefore, while the trimerization domain preferably is not a ligand for a cell surface binding site, the entire trimer can be such a ligand (by virtue of a non-native ligand as discussed herein).

Moreover, the trimerization domain can be a ligand for a substrate other than a native cell surface binding site, as such trimerization-ligands do not present the same concern for cell targeting as do trimerization domains which are ligands for cell surface binding sites.

Thus, for example, the non-native trimerization domain can be a ligand for a substrate on an affinity column, on a blood-borne molecule, or even on a cell surface when it is not a native cell-surface binding site on a cell engineered to express a substrate cell surface protein not native to the unmodified cell type).

A monomer for inclusion into a trimer can be all or a part of a native adenoviral fiber monomeric protein. For example, a modified monomer can lack a sizable number of residues, or even identifiable domains, as herein described. For example, a monomer can lack the native knob domain; it can lack one or more native shaft p-sheet repeats, or it can be otherwise truncated. Thus, a monomer can have any desired modification so long as it trimerizes. Furthermore, a monomer preferably is not modified appreciably at the amino terminus the amino-terminus of a monomer preferably consists essentially of the native fiber amino-terminus) to ensure that the resultant trimer interacts properly with the penton base. Hence, the present invention also provides a composition of matter comprising a trimer of the present invention and an adenoviral penton base. Preferably, the trimer and the penton base associate much in the same manner as wild-type fibers and penton bases. Of course, while the trimer comprises modified fiber monomers, the penton base can also be modified, for example, to include a non-native ligand, for example as is described in U.S. Patent 5,559,099.

Mutant Knobs A fiber monomer for incorporation into the trimer of the present invention has a trimerization domain which binds a native mammalian AR an AR native for the adenoviral serotype of interest) with less affinity than a native adenoviral fiber. Trimers WO 98/54346 PCT/US98/11024 9 incorporating such monomers preferably are not ligands for their native cellular binding sites. The monomers can be modified in any manner suitable for reducing the affinity of the fiber for native AR while permitting the monomers to trimerize. For example, in one embodiment, the trimerization domain is a modified adenoviral fiber knob domain lacking a native receptor-binding amino acid. Any native amino-acid residue mediating or assisting in the interaction between the knob and a native cellular AR is a suitable amino acid for mutation or deletion from the monomer. Moreover, the knob domain can lack any number of such native receptor-binding amino acids, so long as, in the aggregate, the monomers associate to form a trimer of the present invention.

Native amino acid residues for modification or deletion can be selected by any method. For example, the sequences from different adenoviral serotypes can be compared to deduce conserved residues likely to mediate AR-binding. Alternatively or in combination, the sequence can be mapped onto a three dimensional representation ofthe protein (such as the crystal structure) to deduce those residues most likely responsible for AR binding. These analyses can be aided by resorting to any common algorithm or program for deducing protein structural functional interaction. Alternatively, random mutations can be introduced into a cloned adenoviral fiber expression cassette. One method of introducing random mutations into a protein is via the Taq polymerase. For example, a clone encoding the fiber knob (see, SEQ ID NO:9; Roelvink et al., J.

Virol., 70, 7614-21 (1996)) can serve as a template for PCR amplification of the adenoviral fiber knob, or a portion thereof. By varying the concentration of divalent cations in the PCR reaction, the error rate of the transcripts can be largely predetermined (see, Weiss et al., J. Virol., 71, 4385-94 (1997); Zhou et al., Nucl. Acid. Res., 19, 6052 (1991)). The PCR products then can be subcloned back into the template vector to replace the sequence within the fiber coding sequence employed as a source for the PCR reaction, thus generating a library of fibers, some of which will harbor mutations which diminish native AR binding while retaining the ability to trimerize.

A monomer lacking one or more amino acids, as herein described, can optionally comprise a non-native residue several non-native amino acids) in addition to or in place of the missing native amino acid(s); of course, alternatively, the native amino acid(s) can simply be deleted from the knob. Preferably, the amino-acid is substituted with another non-native amino acid to preserve topology and, especially, trimerization.

Moreover, if substituted, the replacement amino acid preferably confers novel qualities to the monomer. For example, to maximally ablate binding to the native AR, a native amino acid can be substituted with a residue (or a plurality of residues) having a different charge.

Such a substitution maximally interferes with the electrostatic interaction between native adenoviral knob domains and cellular ARs. Similarly, a native amino acid can be substituted with a heavier residue (or a plurality of residues) where possible. Heavier WO 98/54346 PCT/US98/11024 residues have longer side-chains; hence, such a substitution maximally interferes with the steric interaction between native adenoviral knob domains and cellular ARs.

Non-native Trimerization Domains In another embodiment, the trimer includes modified monomers which are chimeric adenoviral fiber polypeptides. A suitable chimeric monomer lacks all or a portion of the trimerization domain native to the source adenoviral serotype. The trimerization domain of such a monomer can be deleted from the virus, or the trimerization domain can be ablated by inserting or substituting non-native amino acids into the domain. Of course, a monomer lacking the native trimerization domain can also lack the entire native knob. Because the native trimerization domain is a ligand for a native AR, a trimer of chimeric adenoviral fiber monomers lacking the native trimerization domain binds its native AR with less affinity than the native adenoviral fiber.

For the chimeric monomers to form a trimer of the present invention, they must incorporate a replacement non-native) trimerization domain. To maximally promote the targeting of the virus, preferably the non-native trimerization domain is not a ligand for a mammalian cell-surface receptor, or any cell-surface receptor. Any domain able to form homotrimers is a suitable trimerization domain for inclusion into the trimers of the present invention, and several are known in the art. For example, a chimeric monomer can include the trimerization domain from the heat shock factor (HSF) protein ofK. lactis (Sorger and Nelson, Cell, 59, 807 (1989)), trout axonal dynein (Garber et al., EMBO 8, 1727 (1989)), parainfluenza virus hemagglutanin protein (Coelingh et al., Virology, 162, 137 (1988)), the sigma 1 protein of reovirus type 1 (Strong et al., Virology, 184, 12 (1991)), or other suitable trimer. Alternatively, a chimeric monomer can include a modified leucine-zipper motif. Leucine zippers comprise heptad repeats of leucines, which mediate dimerization. However, replacement of one or more leucine with isoleucine results in stable trimerization of the domains. An example of such a modified leucine zipper motif is the 32 amino acid GCN4p-II trimer (Harbury et al., Science, 262, 1401 (1993)).

Of these trimerization domains, the reovirus sigma 1 trimerization domain is preferred. This protein contains 17 alpha helical heptad repeats, reminiscent of the coiled-coil trimer structure of the aforementioned mutant isoleucine zipper domains (Harbury et al., Nature, 371, 80-83 (1994)). Fiber chimeras containing the sigma 1 domain can thus protrude farther from the virus than corresponding chimeras containing shorter trimerization domains. An advantage of the reovirus sigma 1 trimerization domain over a mutant leucine-zipper GCN4) domain is that the sigma 1 domain is 22 nm long (Fraser et al., J. Virol., 64, 2990-3000 (1990)) whereas GCN4 domain is only 5 nm WO 98/54346 PCT/US98/11024 11 long (Harbury et al., supra). An additional advantage to employing the reovirus sigma 1 attachment protein is that, unlike the adenoviral shaft protein, it exhibits intrinsic trimerization propensity (Leone et al., Virology, 182, 336-45 (1991)). As fiber length appears to increase the efficiency and specificity of adenoviral-cell attachment (Roelvink et al., J. Virol., 70, 7614-21 (1996)), longer fibers possible with the sigma 1 domain are preferred to other chimeric fibers.

A chimeric monomer can alternatively include a knob domain from another adenoviral serotype. For example, the trimerization domain can be replaced with a mutated knob from an adenoviral serotype capable of productive infection within the host species a mutant knob of Ad3 containing a mutation in the HI loop). Alternatively, it can be replaced with a knob from a serotype not capable of productive infection within the host species. For example, the fiber knob of a mammalian adenoviral serotype can be replaced with a knob from an avian serotype. While the avian knob mediates trimerization of the fiber proteins, it is likely unable to recognize a mammalian AR; hence, such chimeric fibers lack the native ability to bind the native host AR. Similarly, the fiber knob of one mammalian adenoviral serotype can be replaced with a knob from another mammalian serotype. In this regard, a modified or unmodified knob from the porcine adenovirus NADC-1 fiber is a preferred domain, as the NADC-1 is well characterized. The NADC-1 knob has identifiable ligands, galectin (which binds galactose), and LDZ and RGD peptides, (which bind integrins) (see, Hirabayashi et al., J. Biol. Chem., 266, 13648-53 (1991)). Thus, chimeric human adenoviral fibers having NADC-1 knobs with such mutations can form trimers and associate with the penton base, but they bind native cell-surface receptors with reduced affinity.

The non-native trimerization domain can be ligated to the native fiber monomer at any suitable site, so long as the monomers can trimerize properly be capable of interacting with an adenoviral penton base). For example, the domain can be inserted into the native knob to disrupt knob topology. Alternatively, the trimerization domain can be inserted after any of the 15 amino acid shaft repeats, preferably after the 7 1 5 t h or 22 d repeats to mimic native adenoviral shaft size. Where the non-native trimerization domain is inserted into the adenoviral shaft, it can form the carboxy-terminus of the chimeric protein, or it can be inserted into the middle of the amino acid sequence. Moreover, any number of trimerization domains can be so inserted into the fiber monomer, so long as the resulting trimer properly associates with the penton base.

Blocking Domain Another suitable chimeric monomer has a novel domain blocking the ligand for the native host AR. The blocking domain is any peptide which can be tightly bound to the native ligand. (See, Hong et al., EMBO 16, 2294-2306 (1997)). In other WO 98/54346 PCT/US98/11024 12 words, the blocking domain is a substrate to which the (native or modified) fiber monomer ligand selectively binds. Desirably, the ligand-substrate interaction occurs at least immediately upon viral production and effectively continues until the fiber trimer is destroyed. Because the native ligand binds the blocking domain, the ligand is incapable of binding its native substrate on cell surfaces. Because the native trimerization domain is a ligand for a native AR, trimers of chimeric adenoviral fiber monomers having such blocking domains bind the native AR with less affinity than a native adenoviral fiber.

The blocking domain can be at any position on the adenovirus to bind the native ligand without appreciably affecting trimerization or penton base interaction. For example, the blocking domain can be appended to the above-referenced P-sheets or loops, either by fusion within the reading frame, by covalent post-translational modification, etc.

Alternatively, the blocking domain can be appended to another portion of the monomer, such as the shaft. The blocking domain can also include a linker or spacer polypeptide to afford an opportunity for the blocking domain to interact with the native ligand. If the blocking domain is attached via such a spacer, the spacer can include a protease recognition site for subsequent cleavage, as described herein.

Preparation The monomers for inclusion into the trimers of the present invention can be produced by any suitable method. For example, the mutant fiber protein can be synthesized using standard direct peptide synthesizing techniques as summarized in Bodanszky, Principles of Peptide Synthesis (Springer-Verlag, Heidelberg: 1984)), such as via solid-phase synthesis (see, Merrifield, J. Am. Chem. Soc., 85, 2149-54 (1963); Barany et al., Int. J. Peptide Protein Res., 30, 705-739 (1987); and U.S. Patent 5,424,398). Alternatively, site-specific mutations (such as replacing the knob with a nonnative trimerization domain, removing, replacing, or mutating the AR-binding residues, or adding a blocking domain, as herein described) can be introduced into the monomer by ligating into an expression vector a synthesized oligonucleotide comprising the modified site. Alternatively, a plasmid, oligonucleotide, or other vector encoding the desired mutation can be recombined with the adenoviral genome or with an expression vector encoding the monomer to introduce the desired mutation. Oligonucleotide-directed sitespecific mutagenesis procedures also are appropriate(e.g., Walder et al., Gene, 42, 133 (1986); Bauer et al., Gene, 37, 73 (1985); Craik, Biotechniques, 12-19 (1995); U.S.

Patents 4,518,584 and 4,737,462). However engineered, the DNA fragment encoding the modified monomer can be subcloned into an appropriate vector using well known molecular genetic techniques. The fragment is then transcribed and the peptide subsequently translated in vitro within a host cell. Any appropriate expression vector Pouwels et al., Cloning Vectors: A Laboratory Manual (Elsevior, NY: 1985)) and WO 98/54346 PCT/US98/11024 13 corresponding suitable host cells can be employed for production of recombinant peptides. Expression hosts include, but are not limited to, bacterial species, mammalian or insect host cell systems including baculovirus systems Luckow et al., Bio/Technology, 6, 47 (1988)), and established cell lines such 293, COS-7, C127, 3T3, CHO, HeLa, BHK, etc. An especially preferred expression system for preparing modified fibers of the invention is a baculovirus expression system (Wickham et al., J. Virol., 6831-38 (1995)) as it allows the production of high levels of recombinant proteins. Of course, the choice of expression host has ramifications for the type of peptide produced, primarily due to post-translational modification.

Once produced, the monomers are assayed for fiber protein activity. Specifically, the ability of the monomers to form trimers, interact with the penton base, and interact with native ARs is assayed. Any suitable assay can be employed to measure these parameters. For example, as improperly folded monomers are generally insoluble (Scopes, "Protein Purification" (3d Ed., 1994), Chapter 9, p. 270-82 (Springer-Verlag, New York)), one assay for trimerization is whether the recombinant fiber is soluble.

Determining solubility of the fiber is aided if an amount of radioactive amino-acid is incorporated into the protein during synthesis. Lysate from the host cell expressing the recombinant fiber protein can be centrifuged, and the supematant and pellet can be assayed via a scintillation counter or by Western analysis. Subsequently, the proteins within the pellet and the supernatant are separated on an SDS-PAGE gel) to isolate the fiber protein for further assay. Comparison of the amount of radioactivity in the fiber protein isolated from the pellet vis-A-vis the fiber protein isolated from the supernatant indicates whether the mutant protein is soluble. Alternatively, trimerization can be assayed by using a monoclonal antibody recognizing only the amino portion of the trimeric form of the fiber via immunoprecipitation, Western blotting, etc.). One such antibody is described in International Patent Application WO 95/26412, and others are known in the art. A third measure of trimerization is the ability of the recombinant fiber to form a complex with the penton base (Novelli and Boulanger, Virology, 185, 1189 (1995)), as only fiber trimers can so interact. This propensity can be assayed by coimmunoprecipitation, gel mobility-shift assays, SDS-PAGE (boiled samples run as monomers, otherwise, they run as larger proteins), etc. A fourth measure of trimerization is to detect the difference in molecular weight of a trimer as opposed to a monomer. For example, a boiled and denatured trimer will run as a lower molecular weight than a nondenatured stable trimer (Hong and Angler, J. Virol., 70, 7071-78 (1996)).

A trimeric recombinant fiber must also be assayed for its ability to bind native ARs. Any suitable assay that can detect this is sufficient for use in the present invention.

A preferred assay involves exposing cells expressing a native AR 293 cells) to the recombinant fiber trimers under standard conditions of infection. Subsequently, the cells WO 98/54346 PCT/US98/11024 14 are exposed to native adenoviruses, and the ability of the viruses to bind the cells is monitored. Monitoring can be by autoradiography employing radioactive viruses), immunocytochemistry, or by measuring the level of infection or gene delivery using a reporter gene). In contrast with native trimers which reduce or substantially eliminate subsequent viral binding to the 293 cells, those trimers not substantially reducing the ability of native adenoviruses to subsequently bind the cells are trimers of the present invention. The reduction of interference with subsequent viral binding indicates that the trimer is itself not a ligand for its native mammalian AR, or at least binds with reduced affinity.

Alternatively, a vector including a sequence encoding a mutated fiber (or a library of putative mutated fibers, such as described herein) can be introduced into a suitable host cell strain to express the fiber protein. For high-efficiency screening, preferably the host cells are bacteria. Where bacteria are employed as host cells, mutants can be identified by assaying the ability to bind the soluble CAR protein. For example, a replica of the bacterial plate on a nitrocellulose filter lift) can be cultured in a suitable medium to induce expression from the vector. Subsequently, the filter is exposed to a solution suitable for lysing the bacteria adhering to it, and the probed with a radiolabled CAR protein. Preferably, the filter is first "blocked" with a high protein solution to minimize nonspecific adherence of the CAR probe to the filter. After the hybridization, the filter is exposed to film to identify colonies expressing fiber proteins that bind the CAR. Those colonies not hybridizing to the radiolabeled CAR probe (or binding with reduced affinity as indicated by weaker signal) potentially express fiber monomers of the present invention. Because a reduction in CAR-binding could be due to either selective ablation of the ligand or structural modification affecting trimerization, mutant fibers identified as non-CAR binding by such a bacterial library screen must be assayed for the ability to trimerize, as described above.

Blocking Proteins As an alternate means for reducing native viral tropism, the present invention provides a chimeric blocking protein comprising a substrate for an adenovirus fiber. The chimeric blocking protein can include any suitable domain having a substrate recognized by the ligand on the adenoviral fiber. For example, for interfering with the receptorbinding of a wild-type adenovirus, the chimeric blocking protein can comprise the extracellular domain of the CAR cell-surface protein (Bergelson et al., Science, 275, 1320-23 (1997); Tomko et al., Proc. Nat. Acad. Sci. 94, 3352-56 (1997)), the extracellular domain for the MHC class I receptor (Hong et al., EMBO 16(9), 2294-06 (1997)), or other similar extracellular substrate domain for an AR. Moreover, for interfering with the substrate-binding of recombinant adenoviruses, such as adenoviruses WO 98/54346 PCT/US98/11024 having chimeric fiber trimers as described herein, the blocking protein can comprise a substrate recognized by a ligand present on the trimer. While, as mentioned, the chimeric blocking protein can comprise domains from cell-surface proteins, typically it is not itself a cell-surface protein. Instead, the chimeric blocking protein is preferably a free soluble protein able to interact with an adenovirus in solution.

A chimeric blocking protein of the present invention affords a method of interfering with adenoviral receptor-binding by incubating an adenovirus with the chimeric blocking protein in a solution such that the chimeric blocking protein binds the ligand present on the adenoviral fiber. The virus and the chimeric blocking protein can be incubated for any length of time, and under any suitable conditions, to promote the ligand on the fiber to bind the substrate on the chimeric blocking protein. The parameters of time, temperature, and solution chemistry suitable for promoting selective binding between the fiber ligand and the chimeric blocking protein substrate can vary according to the affinity with which the ligand selectively binds the substrate. Generally, where known ligand-substrate systems are employed, these parameters are also known. Where novel ligand-substrate systems are employed, however, the binding conditions can, in large measure, be predetermined as discussed herein by employing such conditions when screening the protein library for the novel ligand-substrate interaction). However, preferably the concentration of the chimeric blocking proteins is sufficient to saturate the cell-surface ligands present on the fibers of the adenovirus during the incubation.

In addition to including a domain having a substrate recognized by the ligand on an adenoviral fiber, a chimeric blocking protein also can have other domains. For example, the protein can include domains to promote secretion (see, Suter et al., EMBO, J, 10, 2395-2400 (1991); Beutler et al., J. Neurochem., 64, 475-81 (1995)), thus aiding in the collection of free chimeric blocking proteins from cells producing the protein. Additionally, the chimeric blocking protein preferably further includes a ligand domain a ligand in addition to the substrate for the viral knob), such as those ligands described herein. The presence of a ligand on the chimeric blocking protein, notably peptide tags and other similar sequences, facilitates purification and identification of the chimeric blocking protein after production. A more preferred ligand is one recognizing a cell surface binding site or other substrate, as discussed herein. Such blocking proteins function as "bi-specific" molecules for altering adenoviral receptor binding. For example, where a chimeric blocking protein includes a ligand for a cell-surface binding site, the blocking protein is able to effect selective targeting of the adenovirus by interfering with fiber-mediated receptor binding while directing novel targeting through the ligand present on the chimeric blocking protein. Thus, the present invention provides a method of directing adenoviral targeting by incubating an adenovirus with a chimeric blocking protein having a ligand recognizing a substrate present on a cell surface binding site in a WO 98/54346 PCT/US98/11024 16 solution such that the chimeric blocking protein binds the adenoviral fiber to form a complex, and thereafter exposing the complex to a cell having a substrate for the ligand.

In addition to including a domain having a substrate recognized by the ligand on an adenoviral fiber (and possibly a non-adenoviral ligand domain), the chimeric blocking protein also can include a trimerization domain, such as those trimerization domains discussed herein. The presence of such trimerization domains permits the chimeric blocking protein monomers to trimerize. While, as monomers, the chimeric blocking proteins can saturate the ligands present on the fibers, such bonds are, of course, subject to dissociation at a certain rate depending on the kinetics of the ligand-substrate interaction. However, because the probability that all three ligand/substrate bonds between a trimeric fiber and the trimeric blocking protein will be severed at the same time is significantly less than the probability that any one such bond will be broken, a trimeric blocking protein more easily saturates the available ligands present on the fiber. In effect, the trimeric structure effectively holds each substrate against the fiber knob ligand, thereby increasing the likelihood that each ligand is blocked.

The chimeric blocking proteins can be produced by any suitable method, such as by direct protein synthesis, cellular production, in vitro translation or other method known in the art. Many suitable methods for producing proteins are described elsewhere herein and are otherwise known in the art.

Viruses The present invention provides an adenovirus incorporating the recombinant fiber trimers of the present invention. The adenovirus of the present invention does not infect its native host cell via the native AR as readily as the wild-type serotype, due to the above-mentioned reduction in affinity of the fiber trimers present in the viral coat via replacement of the trimerization domain with a non-ligand trimerization domain, selective mutation of the responsible residues, or incorporation of a blocking domain, as herein described). Thus, the adenovirus preferably incorporates a non-adenoviral ligand to facilitate its propagation, isolation and/or targeting.

The virus can include any suitable ligand a peptide specifically binding to a substrate). For example, for targeting the adenovirus to a cell type other than that naturally infected (or a group of cell types other than the natural range or set of host cells), the ligand can bind a cell surface binding site any site present on the surface of a cell with which the adenovirus can interact to bind the cell and thereby promote cell entry) other than its native AR or even any native AR. A cell surface binding site can be any suitable type of molecule, but typically is a protein (including a modified protein), a carbohydrate, a glycoprotein, a proteoglycan, a lipid, a mucin molecule or mucoprotein, or other similar molecule. Examples of potential cell surface binding sites include, but WO 98/54346 PCT/US98/11024 17 are not limited to: heparin and chondroitin sulfate moieties found on glycosaminoglycans; sialic acid moieties found on mucins, glycoproteins, and gangliosides; common carbohydrate molecules found in membrane glycoproteins, including mannose, N-acetyl-galactosamine, N-acetyl-glucosamine, fucose, and galactose; glycoproteins such as ICAM-1, VCAM, selectins E-selectin, P-selectin, L-selectin, etc.), and integrin molecules; and tumor-specific antigens present on cancerous cells, such as, for instance, MUC-1 tumor-specific epitopes. The protein can thus be expressed in a narrow class of cell types cardiac muscle, skeletal muscle, smooth muscle, etc.) or expressed within a broader group encompassing several cell types.

In other embodiments to facilitate purification or propagation within a specific engineered cell type), the non-native ligand can bind a compound other than a natural cell-surface protein. Thus, the ligand can bind blood- and/or lymph-borne proteins albumin), synthetic peptide sequences such as polyamino acids polylisine, polyhistadine, etc.), artificial peptide sequences FLAG SEQ ID NO: 16), and RGD peptide fragments (Pasqualini et al., J. Cell. Biol., 130, 1189 (1995)).

Alternatively, the ligand can bind non-peptide substrates, such as plastic Adey et al., Gene, 156, 27 (1995)), biotin (Saggio et al., Biochem. 293, 613 (1993)), a DNA sequence (Cheng et al., Gene, 171, 1, (1996); Krook et al., Biochem. Biophys., Res.

Commun., 204, 849 (1994)), streptavidin (Geibel et al., Biochemistry, 34, 15430 (1995), Katz, Biochemistry, 34, 15421 (1995)), nitrostreptavidin (Balass et al., Anal. Biochem., 243, 264 (1996)), heparin (Wickham et al., Nature Biotechnol., 14, 1570-73 (1996)), cationic supports, metals such as nickel and zinc Rebar et al., Science, 263, 671 (1994); Qui et al., Biochemistry, 33, 8319 (1994)), or other potential substrates.

Examples of suitable ligands and their substrates for use in the method of the invention include, but are not limited to: CR2 receptor binding the amino acid residue attachment sequences, CD4 receptor recognizing the V3 loop of HIV gpl20, transferrin receptor and its ligand (transferrin), low density lipoprotein receptor and its ligand, the ICAM-1 receptor on epithelial and endothelial cells in lung and its ligand, linear or cyclic peptide ligands for streptavidin or nitrostreptavidin (Katz, Biochemistry, 34, 15421 (1995)), galactin sequences that bind lactose, galactose and other galactose-containing compounds, and asialoglycoproteins that recognize deglycosylated protein ligands. Moreover, additional ligands and their binding sites preferably include (but are not limited to) short 6 amino acid or less) linear stretches of amino acids recognized by integrins, as well as polyamino acid sequences such as polylysine, polyarginine, etc. Inserting multiple lysines and/or arginines provides for recognition of heparin and DNA. Also, a ligand can comprise a commonly employed peptide tag short amino acid sequences known to be recognized by available antisera) such as sequences from glutathione-S-transferase (GST) from Shistosoma manosi, thioredoxin p-galactosidase, or maltose binding protein WO 98/54346 PCT/US98/11024 18 (MPB) from E. coli., human alkaline phosphatase, the FLAG octapeptide (SEQ ID NO:16), hemagluttinin (HA) (Wickham et al., 1996, supra), polyoma virus peptides, the large T antigen peptide, BPV peptides, the hepatitis C virus core and envelope E2 peptides and single chain antibodies recognizing them (Chan, J. Gen. Virol., 77, 2531 (1996)), the c-myc peptide, adenoviral penton base epitopes (Stuart et al., EMBO 16, 1189-98 (1997)), epitopes present in the E2 envelope of the hepatitis C virus SEQ ID NO:17, SEQ ID NO:18 (see, Chan et al., 1996, supra), and other commonly employed tags. A preferred substrate for a tag ligand is an antibody directed against it, a derivative of such an antibody a FAB fragment, Single Chain antibody (ScAb)), or other suitable substrate.

As mentioned, a suitable ligand can be specific for any desired substrate, such as those recited herein or otherwise known in the art. However, adenoviral vectors can also be engineered to include novel ligands by first assaying for the ability of a peptide to interact with a given substrate. Generally, a random or semirandom peptide library containing potential ligands can be produced, which is essentially a library within an expression vector system. Such a library can be screened by exposing the expressed proteins the putative ligands) to a desired substrate. Positive selective binding of a species within the library to the substrate indicates a ligand for that substrate, at least under the conditions of the assay. For screening such a peptide library, any assay able to detect interactions between proteins and substrates is appropriate, and many are known in the art. However, one preferred assay for screening a protein library is the phage display system, which employs bacteriophage expressing the library Koivunen et al., Bio/Technology, 13, 265-70 (1995); Yanofsky et al., Proc. Nat. Acad. Sci. 93, 7381-86 (1996); Barry et al., Nature Med., 299-305 (1996)). Binding of the phage to the substrate is assayed by exposing the phage to the substrate, rinsing the substrate, and selecting for phage remaining bound to the substrate. Subsequently, limiting dilution of the phage can identify individual clones expressing the putative ligand. Of course, the insert present in such clones can be sequenced to determine the identity of the ligand.

Phage display is preferred for identifying potential ligands because it best mimics viral interaction with the microenvironment. Notably, phage display is an extracellular system (as is the initial stage of viral infection); moreover, phage display incorporates an actual virus (phage) presenting the actual potential ligand. Phage display also offers significantly more flexibility than other protein binding assays (especially intracellular assays). Notably, phage display not only identifies proteins (ligands) binding to a particular substrate, but it identifies those which bind under predefined conditions. Thus, the use of phage display can identify ligands useful for incorporation into an adenovirus to facilitate purification under largely predefined conditions. For example, the phage display library can be screened by exposure to a particular plastic, resin, or other desired WO 98/54346 PCT/US98/11024 19 substrate used in an affinity column. Phage expressing peptides that either bind the substrate or that are eluted from the substrate under a specific condition or range of conditions high or low salt, pH, temperature, etc.), but do not so bind or elute under other conditions, can be readily identified. Thereafter, adenovirus incorporating the ligand can be purified by expositing it to the substrate under like conditions, as discussed herein.

Once a given ligand is identified, it can be incorporated into any location of the virus capable of interacting with a substrate the viral surface). For example, the ligand can be incorporated into the fiber, the penton base, the hexon, or other suitable location. Where the ligand is attached to the fiber protein, preferably it does not disturb the interaction between viral proteins or monomers. Thus, the ligand preferably is not itself an oligomerization domain, as such can adversely interact with the trimerization domain as discussed above. Moreover, the ligand preferably does not replace a portion of the fiber protein, as such perturbance can adversely affect trimerization and interaction with the penton. Rather, the ligand preferably is added to the fiber protein, and is incorporated in such a manner as to be readily exposed to the substrate at the carboxy-terminus of the protein, attached to a residue facing the substrate, positioned on a peptide spacer to contact the substrate, etc.) to maximally present the ligand to the substrate. Where the ligand is attached to or replaces a portion of the penton, preferably it is within the hypervariable regions to ensure that it contacts the substrate. Furthermore, where the ligand is attached to the penton, preferably, the recombinant fiber is truncated or short from 0 to about 10 shaft repeats) to maximally present the ligand to the substrate (see, U.S. Patent 5,559,099 (Wickham et Where the ligand is attached to the hexon, preferably it is within a hypervariable region (Miksza et al., J. Virol., 70(3), 1836-44 (1996)).

When engineered into an adenoviral protein (or blocking protein), the ligand can comprise a portion of the native sequence in part and a portion of the non-native sequence in part. Similarly, the sequences (either native and/or nonnative) that comprise the ligand in the protein need not necessarily be contiguous in the chain of amino acids that comprise the protein. In other words, the ligand can be generated by the particular conformation of the protein, through folding of the protein in such a way as to bring contiguous and/or noncontiguous sequences into mutual proximity. Of course an adenovirus of the present invention (or a blocking protein) can comprise multiple ligands, each binding to a different substrate. For example, a virus can comprise a first ligand permitting affinity purification as described herein, a second ligand that selectively binds a cell-surface site as described herein, and/or a third ligand for inactivating the virus, also as described herein.

WO 98/54346 PCT/US98/11024 The protein including the ligand can include other non-native elements as well.

For example, a non-native, unique protease site also can be inserted into the amino acid sequence. The protease site preferably does not affect fiber trimerization or substrate specificity of the fiber ligand. Many such protease sites are known in the art. For example, thrombin recognizes and cleaves at a known amino acid sequence (Stenflo et al., J. Biol. Chem., 257, 12280-90 (1982)). The presence of such a protease recognition sequence facilitates purification of the virus in some protocols, as discussed herein. The protein can be engineered to include the ligand by any suitable method, such as those methods described above for introducing mutations into proteins.

In addition to the trimer and the ligand, a virus of the present invention can include one or more non-native passenger genes as well. A "passenger gene"" can be any suitable gene, and desirably is either a therapeutic gene a nucleic acid sequence encoding a product that effects a biological, preferably a therapeutic, response either at the cellular level or systemically), or a reporter gene a nucleic acid sequence which encodes a product that, in some fashion, can be detected in a cell). Preferably a passenger gene is capable of being expressed in a cell into which the vector has been internalized.

Preferably the passenger gene exerts its effect at the level of RNA or protein. For instance, a protein encoded by a transferred therapeutic gene can be employed in the treatment of an inherited disease, such as, the cystic fibrosis transmembrane conductance regulator cDNA for the treatment of cystic fibrosis. Alternatively, the protein encoded by the therapeutic gene can exert its therapeutic effect by effecting cell death. For instance, expression of the gene in itself can lead to cell killing, as with expression of the diphtheria toxin. Alternatively, a gene, or the expression of the gene, can render cells selectively sensitive to the killing action of certain drugs, expression of the HSV thymidine kinase gene renders cells sensitive to antiviral compounds including aciclovir, ganciclovir, and FIAU (1-(2-deoxy-2-fluoro-3-D-arabinofuranosil)-5iodouracil). Moreover, the therapeutic gene can exert its effect at the level of RNA, for instance, by encoding an antisense message or ribozyme, a protein which affects splicing or 3' processing polyadenylation), or a protein affecting the level of expression of another gene within the cell where gene expression is broadly considered to include all steps from initiation of transcription through production of a processed protein), perhaps, among other things, by mediating an altered rate of mRNA accumulation, an alteration of mRNA transport, and/or a change in post-transcriptional regulation. Of course, where it is desired to employ gene transfer technology to deliver a given passenger gene, its sequence will be known in the art.

The altered protein the trimer or the coat protein having the ligand) and the passenger gene (where present) can be incorporated into the adenovirus by any suitable method, many of which are known in the art. As mentioned herein, the protein is WO 98/54346 PCT/US98/11024 21 preferably identified by assaying products produced in high volume from genes within expression vectors baculovirus vectors). The genes from the vectors harboring the desired mutation can be readily subcloned into plasmids, which are then transfected into suitable packaging cells 293 cells). Transfected cells are then incubated with adenoviruses under conditions suitable for infection. At some frequency within the cells, homologous recombination between the vector and the virus will produce an adenoviral genome harboring the desired mutation.

Adenoviruses of the present invention can be either replication competent or replication deficient. Preferably, the adenoviral vector comprises a genome with at least one modification therein, rendering the virus replication deficient (see, International Patent Application WO 95/34671). The modification to the adenoviral genome includes, but is not limited to, addition of a DNA segment, rearrangement of a DNA segment, deletion of a DNA segment, replacement of a DNA segment, or introduction of a DNA lesion. A DNA segment can be as small as one nucleotide and as large as the adenoviral genome about 36 kb) or, alternately, can equal the maximum amount which can be packaged into an adenoviral virion about 38 kb). Preferred modifications to the adenoviral genome include modifications in the El, E2, E3, and/or E4 regions. An adenovirus also preferably can be a cointegrate, a ligation of adenoviral genomic sequences with other sequences, such as other virus, phage, or plasmid sequences.

The adenovirus of the present invention has many qualities which render it an attractive choice for use in gene transfer, as well as other, applications. For example, the adenovirus does not infect its native host cells as readily as does wild-type adenovirus, due to the mutant fiber trimers selective mutation of residues responsible for AR binding, replacement of the trimerization domain, or addition of a blocking domain, as herein described). Furthermore, the adenovirus has at least one non-native ligand specific for a substrate which facilitates viral propagation, targeting, purification, and/or inactivation as discussed herein. For ease in cloning, the ligands and the trimerization domains preferably are separate domains, thus permitting the virus to be easily be reengineered to incorporate different ligands without perturbing fiber trimerization.

Alternatively, if the fiber trimer incorporates a mutated fiber knob, the ligand can be incorporated into the knob, as herein described.

Of course, for delivery into a host (such as an animal), a virus of the present invention can be incorporated into a suitable carrier. As such, the present invention provides a composition comprising an adenovirus of the present invention and a pharmacologically acceptable carrier. Any suitable preparation is within the scope of the invention, the exact formulation, of course, depends on the nature of the desired application cell type, mode of administration, etc.), many suitable preparations are set forth in U.S. Patent 5,559,099 (Wickham et al.).

WO 98/54346 PCTIUS98/11024 22 Cell Line As mentioned herein, an adenovirus of the present invention does not readily infect its native host cell via the native AR because its ability to bind ARs is significantly attenuated (due to the incorporation of the chimeric trimers of the present invention).

Therefore, the invention provides a cell line able to propagate the inventive adenovirus.

Preferably, the cell line can support viral growth for at least about 10 passages about passages), and more preferably for at least about 20 passages about 25 passages), or even 30 or more passages.

For example, the adenoviruses can be first grown in a packaging cell line which expresses a native fiber protein gene. The resultant viral particles are therefore likely to contain both native fibers encoded by the complementing cell line and non-native fibers encoded by the adenoviral genome (such as those fibers described herein); hence a population of such resultant viruses will contain both fiber types. Such particles will be able to bind and enter packaging cell lines via the native fiber more efficiently than particles which lack native fiber molecules. Thus, the employment of such a fiberencoding cell line permits adenovirus genomes encoding chimeric, targeted adenovirus fibers to be grown and amplified to suitably high titers. The resultant "mixed" stocks of adenovirus produced from the cell lines encoding the native fiber molecule will contain both native and chimeric adenovirus fiber molecules; however, the particles contain genomes encoding only the chimeric adenovirus fiber. Thus, to produce a pure stock of adenoviruses having only the chimeric adenovirus fiber molecules, the "mixed" stock is used to infect a packaging cell line which does not produce native fiber (such as 293 for El-deleted viruses). The resultant adenoviruses contain only the fiber molecules encoded by the genomes the chimeric fiber molecules).

Similar fiber-complementing cell lines have been produced and used to grow mutant adenovirus lacking the fiber gene. However, the production rates of these cell lines have generally not been great enough to produce adenovirus titers of the fiberdeleted adenovirus comparable to those of fiber-expressing adenovirus particles. The lower titers produced by such mutants can be improved by temporally regulating the expression of the native fiber to more fully complement the mutant adenovirus genome.

One strategy to produce such an improved cell line is to use of an inducible promoter, the metallothionine promoter), to permit fiber production to be controlled and activated once the cells are infected with adenovirus. Alternatively, an efficient mRNA splice site introduced into the fiber gene in the complementing cell line improves the level of fiber protein production in the cell line.

When the adenovirus is engineered to contain a ligand specific for a given cell surface binding site, any cell line expressing that receptor and capable of supporting WO 98/54346 PCT/US98/11024 23 adenoviral growth is a suitable host cell line. However, because many ligands do not bind cell surface binding sites (especially the novel ligands discussed herein), a cell line can be engineered to express the substrate for the ligand.

The present invention provides a cell line expressing a non-native cell-surface biding site to which an adenovirus (or a bi-specific blocking protein) having a ligand for the receptor binds. Any cell line capable of supporting adenoviral growth is a suitable cell line for use in the present invention. Where the adenovirus lacks genes essential for viral replication, preferably the cell line expresses complementing levels of the gene products. As 293 cells are superior for supporting adenoviral growth, preferably the cell line of the present invention is derived from 293 cells.

The non-native cell surface binding site is a substrate molecule, such as those described herein, to which an adenovirus (or a bi-specific blocking protein) having a ligand selectively binding that substrate can bind the cell and thereby promote cell entry.

Where the ligand is on the adenovirus, the binding site can recognize a non-native ligand incorporated into the adenoviral coat or a ligand native to a virus. For example, where the non-native viral ligand is a tag peptide, the binding site can be a single chain antibody (ScAb) receptor recognizing the tag. Alternatively, the ScAb can recognize an epitope present in a region of a mutated fiber knob (where present), or even an epitope present on a native adenoviral coat protein, on the fiber, penton, hexon, etc.). Alternatively, where the non-native ligand recognizes a cell-surface substrate membrane-bound protein), the binding site can comprise that substrate. Where the substrate binding side is native to a cell-surface receptor, the cell line can express a mutant receptor with decreased ability to interact with the cellular signal transduction pathway a truncated receptor, such as NMDA, (Li, et al., Nat. Biotech., 14, 989 (1996)), attenuated ability to act as an ion channel, or other modification. Infection via such modified proteins minimizes the secondary effects of viral infection on host-cell metabolism by reducing the activation of intracellular messaging pathways and their various response elements. In short, the choice of binding site depends to a large extent on the nature of the adenovirus in question. However, to promote specificity of the cell type for the virus, the binding site preferably is not a native mammalian AR. Moreover, the binding site must be expressed on the surface of the cell to be accessible to the virus. Hence, where the binding site is a protein, it preferably has leader sequence and a membrane tethering sequence (see, e.g., Davitz et al., J. Exp. Med. 163, 1150 (1986)). to promote proper integration into the membrane.

The cell line can be produced by any standard method. For example, a vector an oligonucleotide, plasmid, viral, or other vector) containing a gene encoding the non-native receptor can be introduced into source cell line by standard means. Preferably, the vector also encodes an agent permitting the cells harboring it to be selected the WO 98/54346 PCT/US98/11024 24 vector can encode resistance to antibiotics which kill cells not harboring the plasmid). At some frequency, the vector will recombine with the cell genome to produce a transformed cell line expressing the binding site.

Method of Propagation In connection with the cell line expressing a non-native adenoviral cell-surface binding site, the present invention provides a method of propagating the inventive adenovirus. The inventive method involves infecting the cell with an adenovirus having a non-native ligand selectively binding to the receptor, incubating the cells, and recovering the adenoviruses produced within the cells. Adenoviruses recovered from the cells can be propagated again amplified) to produce viral stocks of very high titer. The ligand on the adenovirus can be any ligand, such as those discussed herein. The cells of the present invention are infected by the virus at any suitable m.o.i. to promote efficient infection of the cell line from about 1 m.o.i. to about 10 The conditions of cell culture largely depend on the nature of the host cell. However, it is within the skill of the art to select culture conditions suitable for a given cell type. Viruses are recovered from the cells by standard means, such as by cell lysis. Thereafter they can be purified by standard methods or the method of the present invention.

Method of Purifying As mentioned, the substrate for the ligand engineered into the adenovirus need not be present on the surface of a cell. For example, the substrate can be located on a support, an inanimate support such as plastic, glass, metal, resin, or other material commonly employed in chromatographic or affinity separation. Examples of such supports include metals, natural polymeric carbohydrates and their synthetically modified, cross-linked or substituted derivatives, such as agar, agarose, cross-linked alginic acid, substituted and cross-linked guar gums, cellulose esters, especially with nitric acid and carboxylic acids, mixed cellulose esters, and cellulose ethers; natural polymers containing nitrogen, such as proteins and derivatives, including cross-linked or modified gelatins; natural hydrocarbon polymers, such as latex and rubber; synthetic polymers which may be prepared with suitably porous structures, such as vinyl polymers, including polyethylene, polypropylene, polystyrene, polyvinylchloride, polyvinylacetate and its partially hydrolyzed derivatives, polyacrylamides, polymethacrylates, copolymers and terpolymers of the above polycondensates, such as polyesters, polyamides, and other polymers, such as polyurethanes or polyepoxides; porous inorganic materials such as sulfates or carbonates of alkaline earth metals and magnesium, including barium sulfate, calcium sulfate, calcium carbonate, silicates of alkali and alkaline earth metals, aluminum and magnesium; and aluminum or silicon oxides or hydrates, such as clays, alumina, talc, kaolin, zeolite, WO 98/54346 PCT/US98/11024 silica gel, or glass (these materials may be used as filters with the above polymeric materials); and mixtures or copolymers of the above classes, such as graft copolymers other material commonly employed in chromatographic or affinity separation. Such supports can be fashioned into beads, films, sheets, plates, etc., or coated onto, bonded, laminated, or otherwise joined to appropriate inert carriers, such as paper, glass, polymeric films, fabrics, etc.

The presence of a substrate for a ligand on the surface of an adenovirus of the present invention permits adenoviruses to be readily purified with high affinity and fidelity. Accordingly, the present invention provides a method of purifying an adenovirus having a ligand for a substrate from a composition comprising the adenovirus. The method involves exposing the composition to the substrate under conditions to promote the ligand present on the adenovirus to selectively bind the substrate. Subsequently, the composition at least a significant portion of the composition) not selectively binding the substrate is removed from the substrate, after which the adenovirus bound to the substrate is eluted from the substrate. Using this method, an adenovirus having a ligand can be purified from a variety of compositions solutions, dispersions, suspensions, gels, etc.). While adenoviruses can be present in a variety of compositions, a common composition containing adenoviruses is a cell lysate, such as produced from a packaging cell during adenoviral propagation.

Generally, the substrate is bound to a support, as previously described. Fusing desired ligand-substrates to a suitable support material is known in the art, and the present invention contemplates any suitable method for engineering a support having the substrate. Indeed, as mentioned, the substrate can itself be such a plastic, glass, metal, resin, etc. Any method of exposing the composition containing the adenovirus to the substrate is suitable for use in the present inventive method. For example, the composition can be passed through a column comprising the support onto which the substrate is bound. Of course, the composition also can be mixed with a slurry of such a support beads or other preparation comprising the support-bound substrate), placed into a container a tube, the well of a dish, etc.) which has been coated with the substrate, or otherwise exposed to the substrate.

The parameters of time, temperature, and solution chemistry necessary to promote selective binding can vary according to the affinity with which the ligand selectively binds the substrate. Generally, where known ligand-substrate systems are employed, these parameters are also known. Where novel ligand-substrate systems are employed, however, the binding conditions can, in large measure, be predetermined as discussed herein by employing such conditions when screening the protein library for the novel ligand-substrate interaction). Preferably, the conditions for selective binding do not permit selective binding of other constituents of the composition to the substrate. Where WO 98/54346 PCT/US98/11024 26 other constituents do not selectively bind the substrate, a significant amount of the adenovirus can be removed from the composition by association with the substrate.

After the selective binding step, the adenoviral-deprived composition is removed from the presence of the substrate selectively eluted). Any suitable method for so removing the adenoviral-deprived composition from the substrate can be employed, provided the adenovirus remains selectively bound to the substrate. In other words, the conditions employed for removing the adenoviral-deprived composition from the substrate generally are insufficient to elute the adenovirus from the substrate. The method of removing the adenoviral-deprived composition is largely a function of the type of substrate and support. For example, the adenoviral-deprived composition can be removed from a column comprising the substrate by rinsing the column with several volumes of a suitable solution. Moreover, the adenoviral-deprived composition can be removed from a slurry of the support containing the substrate by repeated centrifugation, resuspension in a suitable solution, and recentrifugation. Alternatively, where the support is a magnetic material, it can be physically removed from the solution by exposing the vessel containing the solution to a magnet and rinsing the magnetic support. Moreover, where the substrate is bound to a dish or a well, the dish can simply be rinsed with several volumes of a suitable solution.

After the adenoviral-deprived composition has been removed from the substrate, the adenovirus is eluted from the substrate. Any method for separating the adenovirus from the substrate is suitable for use in the present inventive method. In many applications, the adenovirus can be liberated by exposing the support-adenovirus complex to an elution solution incompatible with the ligand-substrate bond. The parameters of time, temperature, and solution chemistry necessary to promote selective elution of the virus from the support can vary according to the affinity with which the ligand selectively binds the substrate. Generally, where known ligand-substrate systems are employed, these parameters are also known. Where novel ligand-substrate systems are employed, however, the elution conditions can, in large measure, be predetermined, for example, by adjusting the conditions when screening a protein library, as discussed herein.

Additionally, where the ligand is incorporated into the adenovirus on a spacer or other peptide, as described, the spacer can include a peptidase recognition sequence or other specific cleavage motif. Adenoviruses containing such a cleavage sequence can be liberated from the support by exposing the support to an agent effecting the cleavage, such as an endoprotease or other agent. While the cleavage method severs the ligand from the adenovirus, in many applications this is preferred. For example, the ligand for purifying the virus might interfere with a second ligand for targeting the virus to a particular cell type. Removal of the purifying ligand thus permits the isolated adenovirus to more readily infect the cell type of interest.

WO 98/54346 PCT/US98/ 1024 27 While any suitable binding or elution conditions can be employed, a practical limit is set by the ability of the adenovirus to survive the conditions. However, as adenoviruses are able to withstand a wide variety of environmental variation, such as high salt, high osmolality, and basic conditions, the present method can be employed under a wide range of conditions. In any event, such conditions are known to those of skill in the art.

The inventive method for purifying adenoviruses need not remove all of the virus from the solution, or even a majority of the virus. Indeed, in many applications, the amount of virus present in the initial composition can saturate the amount of substrate present on the support. Moreover, while the ligand on the adenovirus selectively binds the substrate, such selective binding can be of any affinity. As such, a substantial amount of substrate can not bind available ligands in the separation step. Therefore, to obtain as much adenovirus from the initial composition as possible, the adenoviral-depleted composition removed from the support, as herein described, can be subjected to successive rounds of purification, and the viruses obtained from each round can be combined into a single stock. Similarly, while other constituents of the initial composition preferably do not selectively bind the resin, the complete absence of erroneous binding is not common, at least in early rounds of purification. The presence of background levels of erroneous binding necessarily results in some contamination of the initial viral stock obtained. To reduce or substantially eliminate such background contamination, the viral stock can be subjected to successive rounds of purification until the background level of contaminants approaches zero. As such, the present inventive method provides an economical, efficient, and reliable means of purifying adenoviruses having known ligands. Moreover, the use of slurries and columns is common in industrial applications, rendering the present method amenable to high throughput, or commercialscale application.

Method of Infecting a Cell As mentioned, the non-native ligand present on the virus of the present invention (or on the virus/blocking protein complex) can recognize a substrate present within a cell surface binding site. Therefore, the present invention provides a method of infecting a cell having a cell surface binding site including a substrate for the non-native ligand. The method involves contacting the cell with the adenovirus such that the non-native ligand of the adenovirus (or on the virus/blocking protein complex) binds the particular cell surface binding site and thereby effects entry of the adenovirus. Because the viruses of the present invention incorporate fiber trimers having reduced ability to bind native mammalian ARs, the adenovirus is internalized into the cell primarily due to the nonnative ligand. As such, the present inventive method effects selective targeting of the virus comprising the ligand to a cell type expressing a binding site comprising the WO 98/54346 PCT/US98/ 1024 28 substrate for that ligand without significant infection of cells via native mammalian ARs.

In the case where the ligand is on the penton base (such as a modified or unmodified penton base), the virus is internalized via the ligand on the penton.

Any cell expressing a cell surface binding site including a substrate for the ligand can be selectively targeted in accordance with the present invention. A cell can be present as a single entity, or can be part of a larger collection of cells, such as a cell culture (either mixed or pure), a tumor, a tissue epithelial, muscle, or other tissue), an organ, an organ system circulatory system, respiratory system, gastrointestinal system, urinary system, nervous system, integumentary system or other organ system), or even an entire organism a human). Preferably, the cells being targeted are selected from the group consisting of heart, blood vessel, smooth muscle, skeletal muscle, lung, liver, gallbladder, urinary bladder, and eye cells.

The method for infecting a cell ideally is carried out wherein the adenovirus includes a passenger gene, such as those vectors herein described. Where the adenovirus of the present invention includes a passenger gene, the method permits the adenovirus to serve as a vector for introducing that gene into a targeted cell. Once internalized, the passenger gene is expressed within the cell. Thus, the vectors and methods of the present invention provide useful tools for introducing a passenger gene into a selected class of cells without significantly providing the gene to cells ubiquitously or ectopicly.

Method of Inactivating a Virus As mentioned, the non-native ligand present on the virus of the present invention can recognize substrate present within blood or lymphatic fluid (such as a ligand present on a free blood-borne protein, a protein present on erythrocytes, etc.). Therefore, the present invention provides a method of inactivating an adenovirus having a ligand recognizing a blood- or lymph-borne substrate by exposing the virus to the substrate.

Within the blood or lymph, the ligand selectively binds its substrate, thereby adsorbing the free virus from the fluid. Preferably, the substrate is present within a large macromolecule albumin) or on the surface of erythrocytes (which lack transcription machinery required to propagate viruses). Of course, a ligand for inactivating the virus can be present at any location on the viral coat (Fender et al., Virology, 214, 110 (1995)).

However, as antibodies recognizing and/or neutralizing adenoviruses primarily bind epitopes on the hexon (Gahery-Segard et al., Eur. J. Immunol., 27, 653 (1997)), nonnative ligands for inactivation of the virus preferably are incorporated into the hexon, as herein described.

By providing a means of effectively inactivating adenoviruses, the method assists in confining the viral infection to a desired locus (tissue, cell type, etc.). Specifically, the method effectively inactivates an individual virus by tethering it to the substrate, thereby WO 98/54346 PCT/US98/11024 29 reducing its ability to contact (and therefore enter) a cell. Even where a virus so adsorbed does contact a cell, it is significantly less likely to be internalized due to the presence of the particle having the substrate. Due to the aggregation of these effects, the inventive method effectively inactivates a viral stock (outside of the desired locus of infection) by dramatically reducing its effective free titer.

The inventive method for inactivating the virus complements the other embodiments of the present invention. For example, as stated, the viruses of the present invention incorporate fiber trimers having reduced affinity for native mammalian ARs, thereby substantially reducing the likelihood that the virus will infect cell types other than the desired cell type. Moreover, the viruses of the present invention can include ligands specific for a substrate present on a cell surface binding site, permitting the virus to be targeted to a predetermined cell type. While those two qualities effect selective targeting, and thereby significantly attenuate ectopic infection, viruses also having a ligand recognizing a blood- or lymph- borne substrate are much less likely to even contact an ectopic tissue by reason of the effective reduction of viral titer.

While it is believed that one of skill in the art is fully able to practice the invention after reading the foregoing description, the following examples further illustrate some of its features. As these examples are included for purely illustrative purposes, they should not be construed to limit the scope of the invention in any respect. The procedures employed in these examples, such as affinity chromatography, Southern blots, PCR, DNA sequencing, vector construction (including DNA extraction, isolation, restriction digestion, ligation, etc.), cell culture (including antibiotic selection), transfection of cells, protein assays (Western blotting, immunoprecipitation, immunofluorescence), etc., are techniques routinely performed by those of skill in the art (see generally Sambrook et al., Molecular Cloning, A Laboratory Manual, Cold Spring Harbor Laboratory, Cold Spring Harbor, NY (1989)). Accordingly, in the interest of brevity, experimental protocols are not discussed in detail.

EXAMPLE 1 This example describes two different fiber trimers having non-native trimerization domains, each of which interacts properly with the adenoviral penton base. Specifically, the fiber chimeras incorporate the reovirus sigma 1 trimerization domain.

Two chimeras were constructed, T5S7sigDel and T5sigDel. T5sigDel contained only the Ad5 fiber tail (T5) fused to sigDel without any Ad fiber shaft sequence.

T5S7sigDel contained the tail plus the first 7 3-sheet repeats of the Ad shaft (S7) fused to sigDel. The DNA and respective amino acid sequences of these two clones are set forth at SEQ ID NO: 1 and SEQ ID NO:2.

WO 98/54346 PCT/US98/11024 The sigDel region of the reovirus sigma 1 gene was amplified via PCR and cloned into the vector, pAcT5S7GCNTS.PS.LS.X (Fig. 3A), to create the baculovirus transfer vector, pAcT5sigDel.TS.PS.LS (Fig. 3B). This vector encodes the Ad5 fiber tail fused to the N-terminal trimerization domain of reovirus type 3 sigma 1 protein followed by a FLAG epitope near the C-terminus. At the C-terminus of the gene, the vector also contains multiple restriction sites to facilitate the cloning of targeting and purification sequences into the gene.

The second vector, pAcT5S7sigDel.TS.PS.LS (Fig. 3C), was created by cutting the above PCR product with the restriction enzymes NheI and BamHI and cloning this fragment into the vector, pAcT5S7GCNTS.PS.LS.X (Fig. 3A), also cut with NheI and BamHI. The resultant vector encodes a protein containing the tail and first seven P-sheet shaft repeats of Ad5 fiber fused to sigDel, followed by a FLAG epitope.

Recombinant baculovirus clones encoding each of the fiber chimeras were then generated by standard means using each of the above plasmids. The resultant baculovirus clones were used to produce recombinant proteins in Tn5 insect cells. To compare the sigDel trimerization domain with the GCN domain, another baculovirus was constructed from the initial plasmid, pAcT5S7GCNTS.PS.LS.X (Fig. 3A), which contained the GCN trimerization domain in place of the sigDel trimerization domain.

The baculovirus-infected cells were pelleted at 3 days post infection. The cell pellet was resuspended in PBS plus protease inhibitors and freeze-thawed three times to release the soluble intracellular proteins. The cell debris were then pelleted by centrifugation at high speed and the cleared cell lysate was removed. The pellet was then resuspended in the same volume of PBS as previously.

Pellet and lysate samples were then run on an 0.1% SDS, 12.5% polyacrylamide gel and transferred to nitrocellulose for Western analysis using anti-FLAG M2 MAb (Kodak). These results demonstrated that over 90% of each of the proteins, T5S7GCN.TS.PS.LS, T5sigDel.TS.PS.LS and T5S7sigDel.TS.PS.LS were soluble in the lysate.

The proteins were further assayed for their ability to form trimers. To test for chimera trimerization, the lysates from each sample were either boiled or not boiled prior to running the samples on a 0.1% SDS, 12.5% polyacrylamide gel. Western analysis of the boiled samples showed that the boiled samples migrated at molecular weights corresponding to the size of the monomeric protein, whereas the unboiled proteins containing the sigDel trimerization domains migrated at molecular weights commensurate with a trimer. The unboiled T5S7GCN.TS.PS.LS protein also migrated as a trimer; however, a significant portion (over half) of the unboiled sample migrated as a monomer.

Similar analyses of wild type fiber and sigma 1 protein have shown that these proteins migrate completely as trimers when not boiled and as monomers when boiled.

WO 98/54346 PCT/US98/11024 31 That the vast majority of the proteins were soluble in the lysate (as opposed to the pellet) strongly suggests that they were correctly folded. Moreover, the migration of the unboiled samples demonstrates that sigDel-containing chimeras are soluble trimers and that the sigDel domain functions better than GCN by forming more stable trimeric fiber chimeras.

To test for the ability of the trimers to complex properly with adenoviral penton base protein, recombinant penton base is mixed in solution with the T5S7GCN.TS.PS.LS, and T5S7sigDel.TS.PS.LS trimeric fiber proteins. The resultant penton base/fiber chimera complex is then immunoprecipitated with anti-penton base antibody coupled to protein A-agarose. The precipitated sample is then run on an SDS- PAGE gel and evaluated by Western analysis as described above using the FLAG antibody. Binding of the FLAG antibody indicates that the fiber chimera containing the FLAG epitope complexes with the penton base in solution.

EXAMPLE 2 This example demonstrates the ability of the fiber-sigDel chimeras to incorporate exogenous protein domains larger than peptide tags.

The sequence encoding a modified version of the green fluorescent protein was amplified by PCR using the primers containing restriction sites to allow efficient cloning into the fiber-sigDel chimera plasmids described above in Example 1. Cloning of the GFP sequence in the proper orientation into the Spel site of pAcT5sigDel.TS.PS.LS (Fig.

3B) yields the plasmid, pAcT5sigDel.GFP.TS.PS.LS (Fig. encoding a fiber-sigDel- GFP chimera. The DNA and amino acid sequence of this clone is set forth as SEQ ID NO:3.

This plasmid was then used to produce recombinant protein using the baculovirus expression system, as described above. The solubility of the chimeric fiber proteins (indicative of correct folding) and the ability of the resultant proteins to bind penton base (indicative of trimerization) was confirmed as discussed above. Production of soluble, trimeric protein containing the GFP domains indicates that large, functional protein domains can be incorporated into the fiber-sigDel chimeras as easily as can be the smaller peptide tags. The results predict that such chimeras could also incorporate ligands, such as ScAbs, without significantly interfering with protein function.

EXAMPLE 3 This example describes the construction of recombinant adenovirus vectors containing fiber trimers having non-native trimerization domains.

The NdeI to BamHI fragment is excised from pAcT5S7sigDel.TS.PS.LS (Fig.

3C), to replace the corresponding fragments in pAS pGS HAAV (Fig. 5A), and pAS pGS WO 98/54346 PCT/US98/11024 32 pK7 (Fig. 5B), to produce the final transfer vectors pAS T5S7sigDelpGS.HAAV (Fig.

and pAST5S7sigDelpGS.pK7 (Fig. 5D), respectively. The vectors encode the fibersigDel chimera containing either the RGD or pK7 binding domains at their C-terminus for binding to an a, integrin and heparin sulfate-containing receptors that are expressed by 293 cells.

These vectors are then linearized and then transfected 293 cells had been preincubated with the El, E3, E4-deleted adenovirus AdCMVZ. 11A (GenVec, Inc., Rockville, MD) prior to transfection with the plasmids. Recombination of the E4+ pNS plasmid with the E4-deleted vector results in the rescue of an El-, E3-, E4+ vector capable of replication in 293 cells. The infected/transfected cells are harvested after days and lysed to release virions. The lysate is then used to infect freshly plated cells and to further plaque-purify the recombinant viruses. Plaques cross-contaminated with the original AdCMVZ. 11A stain blue when plaqued in medium containing X-glu substrate.

White plaques (indicating viable vector) are then amplified to produce pure virus stocks of the recombinant adenovirus.

EXAMPLE 4 This example describes the production of targeted adenovirus particles having genomes encoding chimeric fibers. The chimeric fibers represent the Ad5 fiber tail and seven shaft repeats fused to the sigDel trimerization domain from reovirus followed by a high affinity RGD sequence for binding av integrins.

The plasmid, pAS T5S7sigDel.HAAV (Fig. 5C), is cut with the restriction enzyme DrdI, and the large fragment containing all the adenovirus sequences is isolated and purified. This fragment is then electroporated into BJ5183 bacterial cells along with a linearized plasmid, containing the majority of Ad genome prior to the fiber gene with a small overlap of identical sequence with the pAS T5S7sigDel.HAAV plasmid. Upon recombination of the two pieces of DNA, a new plasmid is produced in the bacterial cells through homologous recombination. This plasmid encodes a modified adenovirus genome that is capable of replicating in the appropriate complementing mammalian cell line (El and fiber-complementing). The plasmid DNA from selected colonies is isolated and confirmed to be the correct plasmid by restriction analysis. This plasmid DNA is then used to transform DH5a bacterial cells in order to obtain adequate amounts of DNA for transfection into the fiber-complementing cell line.

One microgram of the plasmid is cut with the appropriate restriction enzyme and transfected into a fiber-complementing cell line, such as the cell line described above. At 0-4 days post-transfection, the cells are induced with zinc, and 1-5 days later the cells are lysed. The lysate is passaged onto fresh fiber-complementing cells. This passage and lysis cycle is repeated until a cytopathic effect develops in the cells. During the cycle, the LacZ WO 98/54346 PCT/US98/11024 33 activity of the cell lysate is also followed, as it should increase as the recombinant vector is amplified. Once an adequate titer of the "mixed" stock is obtained, a final passage onto non-fiber-complementing cells is made to produce a targeted virus lacking a native fiber protein. The resulting virus is then assayed for its ability to bind and enter cells via the interaction of its high affinity RGD sequence with av integrins.

EXAMPLE This example describes four different fiber trimers having non-native trimerization domains. Specifically, the exemplified fiber trimers are chimeras incorporating the knob portion of the NADC-1 fiber, a porcine adenoidal strain. The exemplified trimers, thus, contain known receptor-binding motifs a galectin motif and an RGD motif).

Furthermore, exemplified trimers incorporate mutations known to reduce the affinity of each of the receptor-binding motifs. Finally, this example describes the incorporation of a non-native ligand (FLAG) into an exposed loop of a non-native trimer.

Using PCR, the knob of the NADC-1 fiber gene was amplified from a plasmid containing the full length gene. The PCR product was then cloned into a baculovirus expression plasmid to produce a plasmid which encoded the NADC-1 knob plus an Nterminal polyhistidine tag (the Pig4KN protein) for purification and detection by Western analysis using an anti-polyhistidine antibody. The DNA and amino acid sequences of this clone are set forth at SEQ ID NO:4.

The resultant plasmid, pAcPig4KN (Fig. 6A), was then mutated by site-directed mutagenesis using the two oligonucleotide primer pairs PigD363Es (SEQ ID NO: 10) and PigD363Ea (SEQ ID NO: 11), and PigN437Ds (SEQ ID NO:12) and PigN437Da (SEQ ID NO:13). The former pair of primers was used to produce the plasmid pAcPigKN D363E (Fig. 6B), in which the DNA sequence encoding the RGD integrin binding motif (a.a.

361-363 in the native fiber protein) was mutated to the non-functional sequence RGE.

The second pair of primers was used to produce the plasmid pAcPigKN N437D (FIG.

6C), in which the DNA sequence encoding the native amino acid N 437) was mutated to a D. This mutation has been previously shown to abrogate the binding of another galectin protein to its ligand, galactose (Hirabayashi et al., J. Biol. Chem., 266, 23648-53 (1991)).

A final baculovirus plasmid was constructed to demonstrate the feasibility of incorporating a novel binding motif into an exposed loop on the NADC-1 knob.

Hydrophobicity analysis of the NADC-1 knob protein revealed that the protein sequence immediately prior to the RGD motif was likely to be an exposed loop that would be capable of incorporating additional amino acid sequences polypeptide domains) for the purpose of targeting or purification. Therefore, the plasmid, pAcPig4KN(FLAG) (Fig. 6D), was produced using complementary overlapping oligonucleotides, which WO 98/54346 PCT/US98/11024 34 encoded the FLAG binding domain. The oligonucleotides were annealed and cloned into the plasmid pAcPig4KN (Fig. 11A), which contained a unique, native restriction site, AvrII, just prior to sequence encoding the RGD domain.

The four baculovirus transfer plasmids described above carrying NADC-1 knob genes were used to express recombinant protein in insect cells using the baculovirus expression system. Tn5 insect cells were infected with the recombinant baculovirus clones derived from the plasmids. After three days the cells were pelleted and freezethawed three times in PBS plus protease inhibitors to release the soluble intracellular protein. The debris were pelleted and the cleared lysate was decanted. The remaining pellet was resuspended in PBS.

Lysate and pellet samples were then evaluated by SDS-PAGE and Western analysis to determine whether the recombinant knob proteins were soluble. Western analysis revealed that the majority of all four knob proteins were present in the cell lysate, indicating that they were soluble and correctly folded. These results demonstrate that neither the point mutations introduced into the receptor-binding domains nor the FLAG binding sequence inserted into an exposed loop adversely affected knob folding and solubility.

To investigate whether the chimeric trimers having the NADC-1 knob-FLAG domains can interact with the FLAG antibody, cell lysates are immunoprecipitated using anti-FLAG M2 antibody and then blotted. Western analysis will demonstrate that the NADC-1 knob containing the FLAG epitope is precipitated by the anti-FLAG antibody.

Thus, the NADC-1-fiber trimers are soluble, and each is capable of interacting with the anti-FLAG M2 monoclonal antibody.

EXAMPLE 6 This example describes the synthesis of recombinant AdS-based vector containing an NADC-1 (porcine adenovirus) fiber knob.

Using PCR, the knob of the NADC-1 fiber gene was amplified from a plasmid containing the full length gene. The PCR product was then cloned into the plasmid PNS F5F2K (Fig. 7A) to produce the plasmid, pNS Pig4.SS (Fig. 7B) which encodes the first 7 p-repeats of the Ad5 shaft fused to the NADC-1 knob. The DNA and amino-acid sequences of this clone are set forth at SEQ ID The pNS Pig4.SS plasmid was then used to create a recombinant adenovirus vector. The plasmid was transfected into 293 cells which had been infected with an adenovirus vector lacking the E4 region. Homologous recombination between the plasmid and the vector produced an E4-containing, replication competent vector having the chimeric NADC-1 fiber. The recombinant virus was then plaque purified on 293 cells. Preincubation of Ramos cells (which do not express av integrins but do express WO 98/54346 PCT/US98/11024 receptors for the fiber protein of adenovirus) with recombinant NADC-1 knob blocked the transduction of these cells by the AdZ.PigSS vector, demonstrating that the vector contains a functional NADC knob. The results indicate that chimeric NADC-1 fiber can be correctly synthesized and incorporated into viable virus particles.

EXAMPLE 7 This example describes an Ad5-based adenoviral vector having a chimeric fiber trimer comprising a mutant NADC-1 knob with attenuated receptor-binding ability and containing a functional non-native ligand.

The Apal to BamHI fragment containing the N-D mutation in pAcPig4KN N437D (Fig. 6C) is cloned into the plasmid pAcPig4KN D363E (Fig. 6B) containing the RGD- RGE mutation to create the plasmid pAcPig4KN D363E N437D (Fig. 8A) containing both mutations in the NADC-1 knob gene. Overlapping, complementary oligonucleotide primers encoding the high affinity a, integrin binding domain, are thereafter cloned into the native AvrII site to produce the plasmid pAcPig4KN D363E N437D HAAV (Fig.

8B). The mutated NADC-1 gene fragment EcoRI to BamHI is then cloned into the plasmid pNSPig4.SS (Fig. 7B) to create the plasmid, pNS Pig4 D363E N437D HAAV SS (Fig. 8C). This plasmid is then used to create a recombinant adenovirus vector containing the mutated and a, integrin-targeted NADC-1 knob as described above.

The ability of the double mutation in the NADC-1 knob to block binding to the native cell surface binding sites (galectin and integrin) is confirmed via competition assays. Moreover, the ability of the resultant virus to target cell-surface a, integrin is confirmed using 293 cells, as discussed above.

EXAMPLE 8 This example describes two chimeric blocking proteins able to interfere with native adenoviral receptor binding. In particular, the blocking protein each include a domain having a substrate for the native adenovirus fiber, namely the extracellular domain of the CAR.

The extracellular domain of CAR was amplified from the CAR gene (Bergelson et al., supra; Tomko et al., supra) via PCR. The PCR product was then cloned into a baculovirus expression vector to create the plasmid pACSG2-sCAR (Fig. 9A). The soluble CAR protein (sCAR) also contained a FLAG epitope for purification and for detection by Western analysis. The DNA and amino acid sequences of this sCAR clone are set forth at SEQ ID NO:6.

Western analysis of sCAR produced in insect cells using a baculovirus clone containing sCAR revealed that the protein was secreted from the cell and that some of the protein was retained within the cell.

WO 98/54346 PCT/US98/11024 36 To assess whether the sCAR protein retains the function of the native CAR, radiolabeled adenovirus type 2 were preincubated in a solution containing various concentrations of sCAR and then exposed to 293 cells. The data demonstrated that increasing concentrations of sCAR blocked virus binding to 293 cells. This result demonstrated that the soluble sCAR protein retains the structure and function of the native extracellular domain of CAR. Moreover, these results demonstrate that preincubation with sCAR can ablate native adenoviral receptor binding via the CAR-binding ligand on the adenovirus fiber.

A second sCAR-containing chimera was produced in which DNA sequence encoding an RGD targeting motif was cloned into an Spel site following the C-terminal end of sCAR using complementary, overlapping primers. The chimeric gene retained the FLAG epitope on the C-terminus. The resultant plasmid, SG2-sCAR-HAAV (Fig. 9B), was used to produce recombinant sCAR.RGD protein as was done for sCAR protein described above. The DNA sequence of this clone is set forth at SEQ ID NO:7.

The sCAR.RGD protein was synthesized and secreted from insect cells similarly to the sCAR protein. To assess whether the sCAR.RGD protein retains the function of the native CAR, radiolabeled adenovirus type 2 were preincubated in a solution containing various concentrations of sCAR.RGD and then exposed to Ramos cells, which do not express av integrins but do express receptors for the fiber protein of adenovirus.

Preincubation of radiolabeled adenovirus type 2 with either sCAR or sCAR.RGD blocked virus binding to Ramos cells. This result demonstrates that the sCAR domain present in the sCAR.RGD protein is functional.

To assess whether the sCAR.RGD protein retains the function of the native RGD domain, cell adhesion studies were conducted. Both sCAR.RGD, and sCAR were immobilized onto tissue culture plastic plates, which were subsequently contacted with 293 cells (which express a, integrin). After the cells were incubated on the coated plates, the plates were rinsed, and the number of cells remaining in contact with the plates were assayed. The results showed that cells adhered to plates coated with sCAR.RGD, while they did not adhere to plates coated with sCAR or control plates, demonstrating that the RGD motif present in the sCAR.RGD protein is functional.

EXAMPLE 9 This example demonstrates the inventive method of directing adenoviral targeting using a chimeric blocking protein having a ligand for a cell surface binding site.

An adenovirus vector carrying a lacZ reporter gene is preincubated with either the sCAR.RGD protein or the sCAR protein, described above in Example 8. The resultant complexes are then exposed to either Ramos cells (which express fiber receptor (CAR) but lack a, integrins) or HuVEC cells (which express both CAR and a, integrins) under WO 98/54346 PCT/US98/11024 37 conditions suitable for viral infection. Subsequently, the cells are assayed for lacZ expression, the level of which will correlate to the degree to which the viruses infect the cells. The results will demonstrate that both sCAR and sCAR.RGD effectively block adenovirus transduction of Ramos cells whereas sCAR, but not sCAR.RGD, blocks adenovirus transduction of HuVEC cells, indicating that the Ad/sCAR.RGD complex is targeted to av integrins while avoiding adenoviral-mediated gene delivery to cells via

CAR.

EXAMPLE This example describes two chimeric blocking proteins able to form trimers interfering with native adenoviral receptor binding. In particular, the blocking proteins each include a domain having a substrate for the native adenovirus fiber, namely the extracellular domain of the CAR, and a trimerization domain, namely the sigDel trimerization domain of the Sigma-1 reovirus protein.

The sigDel trimerization domain of the Sigma-1 reovirus protein is amplified by PCR, and the resultant PCR product is cloned into the pAcSG2-sCAR plasmid (Figure 9A). The resultant plasmid, pAcSG2sCAR.sigDel (Fig. 10A) contains a gene chimera encoding the extracellular domain of CAR, a spacer region, the trimerization domain from sigma 1 protein of reovirus, and a FLAG binding domain. An Spel restriction site following the trimerization domain allows for the convenient cloning of targeting domains, such as the high affinity RGD motif which binds a, integrins. The DNA and amino acid sequences of this clone are set forth at SEQ ID NO:8.

PAcsCAR.sigDel was used to make baculovirus. Western analysis of boiled and unboiled cell lysates from baculovirus-infected cells showed that the unboiled chimeric sCAR.sigDel migrated as a trimer.

A second sCAR-containing chimera is produced in which DNA sequence encoding an RGD targeting motif is cloned into an Spel site following the C-terminal end of sCAR.sigDel using complementary, overlapping primers. The resultant plasmid, pAcSG2-sCARsigDel (HAAV) (Fig. 10B), encodes a chimera having the extracellular domain of CAR, a spacer region, the trimerization domain from sigma 1 protein of reovirus, and the high affinity RGD motif which binds a, integrins.

The pAcSG2sCAR.sigDel and pAcSG2-sCARsigDel.RGD (HAAV) plasmids were used to produce recombinant baculovirus which are used to produce the recombinant chimeric protein in insect cells by standard means. Western analysis of boiled and unboiled cell lysates from bacculovirus-infected cells demonstrated that the unboiled sCAR.sigDel protein migrated as a trimer.

To assess the ability of the trimeric sCAR.sigDel and sCARsigDel.RGD proteins to block adenoviral infection, an adenovirus vector carrying a lacZ reporter gene is WO 98/54346 PCT/US98/11024 38 preincubated with either the sCAR.sigDel or the sCARsigDel.RGD trimer or the sCAR monomeric protein. Several concentrations are employed to generate dose-response data.

The resultant complexes are then exposed to 293 cells under conditions suitable for viral infection. Subsequently, the cells are assayed for lacZ expression, the level of which will correlate to the degree to which the viruses infect the cells. The results will demonstrate that the trimeric sCAR.sigDel and sCARsigDel.RGD proteins are more potent in blocking adenovirus binding to via the sCAR protein cells than the sCAR monomers.

EXAMPLE 11 This example demonstrates the inventive method of directing adenoviral targeting using a trimeric blocking protein having a ligand for a cell surface binding site.

An adenovirus vector carrying a lacZ reporter gene is preincubated with either sCAR.sigDel, sCARsigDel.RGD, or sCAR described above. Similarly, the adenovirus can be preincubated with a blocking protein isolated, for example, by phage display. The resultant complexes are then exposed to either Ramos cells or HuVEC cells under conditions suitable for viral infection. Subsequently, the cells are assayed for lacZ expression, the level of which will correlate to the degree to which the viruses infect the cells. The results will demonstrate that, while such proteins will effectively block adenovirus transduction of Ramos cells, the trimers are more potent in blocking adenovirus binding than the sCAR monomers. Moreover, both sCAR and sCAR.sigDel, will block adenovirus transduction of HuVEC cells; however, sCARsigDel.RGD will not effectively block adenovirus transduction of HuVEC cells. Such results strongly suggests that the Ad sCARsigDel.RGD complex is targeted to av integrins while avoiding adenoviral-mediated gene delivery to cells via CAR.

EXAMPLE 12 This example describes the construction and evaluation of mutated fiber knobs each having reduced affinities for native substrates, particularly monoclonal antibodies raised against the native fiber knob.

Using site-directed mutagenesis, separate mutations were introduced into the full length Ad5 fiber gene in a baculoviral vector. The resultant plasmids were then used to generate recombinant baculoviral clones.

Each of the mutants, plus a native Ad5 fiber control, were used to produce protein in infected insect cells. Three days post infection, the cells were harvested and lysed.

Western analysis using polyclonal antisera recognizing the Ad5 fiber revealed the presence of high amounts of fiber protein in lysates from cells infected with each of the vectors. In cells infected with five of the mutant clones (see table 1) (as well as the native fiber gene), the signal was predominantly in the soluble portion of the lysates, indicating WO 98/54346 PCT/US98/11024 that the protein encoded by each mutant was correctly folded. The sequences of the wildtype Ad5 fiber is set forth a SEQ ID NO:9. The amino acids of SEQ ID NO:9 changed by each of these mutations is indicated in Table 1.

Table 1 Mutations Monoclonal Antibodies 2C9 4B8 3D9 CD Loop (449 SGTVQ-GSGSG) IJ Loop (559 GSHN-GSGS) FG Loop (507 SHGKTA-GSGSGS) T533S/T353S (535 TIT-SIS) K506R (506 K-R) C-Term Addition Native Ad5 Fiber Boiled Ad5 Fiber Using Western slot-blot analysis, each of the five soluble mutant fiber proteins, the native Ad5 fiber, and a denatured Ad5 fiber were screened against a panel of four monoclonal antibodies raised against the fiber knob. The signals were detected by chemiluminescence and the strength of signals of each band compared. The results of this assay are set forth in Table 1.

That none of the antibodies recognizes the denatured fiber demonstrates that each binds only correctly folded, trimeric fibers. Furthermore, that none of the mutants exhibited reduced affinity for the 2E5 antibody confirmed that each of the mutant fibers was, indeed, trimeric.

The K506R mutation significantly reduced the affinity of the resultant fiber for the 3D9 antibody without affecting the affinity for any of the other antibodies. The location of this mutation within the fiber knob is indicated in Figs. 15A-15C.

Mutations in the CD, IJ, or FG loops, in which 4-6 amino acids were replaced by altering serines and glycines, significantly reduced the affinity of the resultant mutant trimers for the 2C9 antibody. Moreover, the double mutant T533S/T535S also reduced the affinity of the mutant knob for the 4B8 antibody. The location of each of these mutations within the fiber knob are indicated in Figs. 15D-15F.

These results indicate that the trimeric fiber knobs having reduced affinity for native substrates can be generated. A similar screening protocol can be used to identify mutants having reduced affinity for cellular receptors. For example, a soluble form of sCAR having a FLAG epitope (or other tag), such as described above, can be used as a probe in place of the monoclonal antibodies described above. The blots are then screened WO 98/54346 PCT/US98/11024 with anti-FLAG monoclonal antibodies to detect mutations interfering with fiber-CAR binding.

EXAMPLE 13 This example describes the construction of a recombinant adenovirus containing a short-shafted fiber 8 shaft repeats) and a mutant fiber(5) knob having reduced affinity for its native receptor CAR). Such a fiber permits targeting via a ligand expressed in the penton base.

Using standard recombination techniques, a deletion is introduced into the sequence encoding the fiber shaft. For example, a portion of the mutant fiber knob from the 2 2 d shaft repeat until the end of the coding sequence and containing the K506R mutation (see Example 12) is amplified by PCR from SEQ ID NO:8. The resultant product is used to create the pAS T5S7F5K(R506K) plasmid (Fig. 16). The plasmid, thus, contains a gene encoding a short-shafted fiber with reduced affinity for a native substrate (the 3D9 antibody). An adenovirus having such a short-shafted fiber will be able to bind to cells via the RGD ligand on the penton base. Of course, a similar strategy can be used to create adenoviral vectors having short-shafted fibers with reduced affinity for the CAR.

EXAMPLE 14 This example demonstrates the construction of adenovirus vectors having specific non-native ligands that can be used to purify the vector via affinity chromatography.

The base vector pNSF5F2K (Figure 8A) contains a gene which encodes a chimeric fiber having the shaft of the Ad5 fiber and the knob of the Ad2 fiber protein.

The Ad2 fiber gene contains an Spel restriction site in the region of the knob which encodes the flexible, exposed HI loop of the fiber knob. This Spel restriction site was used to insert sequences which encode the FLAG peptide SEQ ID NO:16 or a DNA/heparin-binding ligand (SEQ ID The base vector pBSSpGS (Figure 11A) encodes a C-terminal 12 amino acid extension (SEQ ID NO: 14). The codons encoding the TS also are a unique Spel site that was used to insert sequences which encode the FLAG peptide (SEQ ID NO:16) or the DNA/heparin-binding polypeptide (SEQ ID NO:15) as described below.

Transfer plasmids (pBSS pGS (RKKK)2 (Figure 11B) and pNSF5F2K(RKKK)2 (Figure 11C)) for introducing the DNA/heparin-binding ligand into the adenoviral genome were created using overlapping oligonucleotides. Sense and antisense oligonucleotides were mixed in equimolar ratios and cloned into the Spel site of pBSS pGS (Fig. 11A) or pNS F5F2K (Fig. 8A) to create the transfer plasmids. Sequencing in WO 98/54346 PCT/US98/11024 41 both directions across the region of the inserts verified that the clones contained the appropriate sequence.

Similarly, transfer plasmids pBSSpGS (FLAG) (Figure 11D) and pNSF5F2K(FLAG) (Figure 11E) for introducing the FLAG ligand (SEQ ID NO: 16) into the adenoviral genome were created. Sequencing in both directions across the region of the inserts verified that the clones contained the appropriate sequence.

The plasmid DNA from the four transfer vectors were linearized with Sall, purified and transfected using calcium phosphate into 293 cells which had been preincubated for 1 h with the El, E3, E4-deleted adenovirus AdCMVZ. 11A (GenVec, Inc., Rockville, MD) a multiplicity of 1 ffu per cell. Recombination of the E4+ pNS plasmid with the E4-deleted vector resulted in the rescue of an El-, E3-, E4+ vector capable of replication in 293 cells. The resultant vectors, AdZ.F2K(RKKK)2, AdZ.F2K(FLAG), AdZ.F(RKKK)2 and AdZ.F(FLAG), were isolated in two successive rounds of plaquing on 293 cells.

Each vector was verified to contain the correct insert by sequencing PCR products derived from virus DNA template using primers spanning the region of the insert DNA.

Restriction analysis of Ad DNA from each of the viruses showed that the viruses were pure and contained the BamHI restriction site unique to the correctly constructed virus.

EXAMPLE This example demonstrates that an adenoviral vector having a non-native ligand can bind a support conjugated to a substrate for that ligand.

The vector AdZ.PK was constructed similarly to the vectors described above; the virus has a fiber protein containing polylysines. AdZ.PK was assayed to determine whether the virus could bind a support having a substrate for polylisine, heparin. 50 ml of heparin-agarose beads (SIGMA) were added to 1.0 ml of phosphate buffers containing 150, 300, 500 and 1000 mM NaCI, respectively. 6600 cpm of either AdZ or AdZ.PK were then added to the saline buffers containing the heparin-agarose beads and rocked for min. The beads were then washed three times with a buffer of equal salinity to the incubation buffer (150, 300, 500, and 1000 mM NaC1, respectively). The bead-associated cpm were then measured and showed the preferential binding of AdZ.PK over AdZ at 150, 300, and 500 mM NaC1. However, at 1000 mM NaCl the binding of AdZ.PK to the beads was much lower and approximately equal to the background binding observed for AdZ.

These results demonstrate that the AdZ.PK vector binds a heparin-linked support material and that binding is ablated by high salt concentration. Therefore, such a support can be used to purify the modified vector by first binding the virus to the support at low salt conditions and then eluting the vector at high salt conditions.

WO 98/54346 PCT/US98/11024 42 EXAMPLE 16 This example demonstrates that an adenoviral vector having a non-native ligand can be purified on a column comprising substrate for that ligand.

20 175 cm2 tissue culture flasks containing 293 packaging cell lines are infected at an m.o.i. of 5 with one of the three vectors: AdZ.PK, AdZ.F2K(RKKK)2 or AdZ.F(RKKK)2 described above. The cells are then incubated for 2 days, after which any remaining adherent cells are then dislodged from the plastic. The removed cells are centrifuged at 3,000 g to form a pellet, the culture medium removed, and the pellet gently washed 2 times with PBS. The cells are then resuspended in a total volume of 5 ml PBS containing 10 mM MgCl 2 The resuspended cells are then freeze-thawed 3 times to release the virus, and the cell debris is then centrifuged at 15,000 g for 15 min. The supernatant is passed over a 3 ml column containing heparin-linked agarose beads. The column is then washed with ml of PBS followed by elution of the virus from the column by a salt step gradient. To elute the virus, 3 ml volumes of buffers containing successively larger concentrations of NaCI (in 100 mM steps) are successively passed over the column, and 1 ml elution volumes are collected (3 ml 200 mM NaCl; 3 ml 300 mM NaCI; 3 ml 400 mM NaCl; up to 2000 mM NaCI).

The fractions, including the runthrough and wash fractions, are then evaluated for adenovirus coat proteins by Western blot, for active virus particles by lacZ transduction levels of A549 cells or by plaque assay, and for overall purity by analytical high performance liquid chromatography (HPLC) as previously described (Shabram, et al, 1997, Hum. Gene Ther. 8, 453-46; Huyghe, et al, 1995, Hum. Gene Ther., 6: 1403-1416; Shabram et al, WO 96/27677). The overall purity of the fractions determined to contain peak adenovirus concentrations is evaluated by running the fractions on HPLC and comparing the profile to a pre-column fraction and a highly purified adenovirus preparation (prepared by 3 successive rounds of purification on CsCl gradients).

EXAMPLE 17 This example describes the production of a pseudo-receptor for constructing a cell line able to replicate adenoviruses lacking native cell-binding function (but targeted for the pseudo-receptor). Specifically, the exemplary pseudo-receptor includes a binding domain from a single-chain antibody (ScFv).

First a vector expressing the ScFv from pHOOK3 (Figure 12A) (Invitrogen), which encodes a ScFv synthesized with a murine Ig signal peptide. The ScFv has an Nterminal HA epitope tag, and its C-terminus is linked to a pair of myc epitopes followed by the PDGF receptor transmembrane anchor. An expression cassette including this WO 98/54346 PCT/US98/11024 43 construct was cloned into plasmid pRC/CMVp-Puro (Fig. 12B) to create the pScHAHK plasmid (Fig. 12C). This plasmid has cloning sites for inserting genes after the CMV promoter and unique Agel and Xbal sites for the addition of cytoplasmic sequences at the C-terminus of the gene.

To demonstrate cell-surface expression of the ScFv pseudo-receptor, either the pNSE4GLP plasmid alone (Figure 12D), which carries a green fluorescent protein gene for detection of tranfectants, or in combination with pScHAHK, were transfected into 293 cells. One day post transfection, the pScHAHK-exhibited surface immunofluorescence using an antibody directed to the HA epitope, demonstrating proper surface expression of the pseudo-receptor.

To demonstrate that the expressed pseudo-receptor is functional, transfected cells were exposed to magnetic CAPTURE-TEC beads conjugated with antigens recognized by the ScFv. Following incubation, the beads were collected in the bottom of a tube using a magnet, washed, and transferred to a culture dish. The culture dishes were then viewed under a fluorescence microscope to identify GFP-expressing cells. No staining was observed from cells transfected only with pNSE4GLP alone, indicating that these cells did not bind the beads. However, cells transfected with pNSE4GLP and pScHAHK were observed in the wells. This result demonstrates that the doubly transfected cells bound to the beads.

EXAMPLE 18 This example describes the production of a pseudo-receptor for constructing a cell line able to replicate adenoviruses lacking native cell-binding function (but targeted for the pseudo-receptor). Specifically, the exemplary pseudo-receptor includes a binding domain from a single-chain antibody recognizing HA.

Anti-HA ScFv was constructed as an N-Term-VL-VH fusion protein. RT-PCR was performed on RNA obtained from hybridomas producing HA antibodies using primers specific for K- or y 2 p- and C-terminus of the VL and VH genes (see Gilliland et al., Tissue Antigens, 47, 1-20 (1996)). After sequencing the resulting PCR products, specific oligonucleotides were designed to amplify the VL-VH fusion in a second round of PCR. The final PCR product was cloned to create the pCANTAB5E(HA) plasmid (Fig. 17A) for production of anti HA ScFv in E. coli. The expressed protein has a Cterminal E peptide for detection of binding to HA-tagged penton base via Western analysis of ELISA assay. Upon transformation of bacterial cells with the pCANTAB5E(HA) plasmid, Western analysis using an antibody recognizing the E peptide revealed a protein of the expected size.

To determine whether the anti-HA ScFv was functional, it was used in protein A immunoprecipitation assays using adenoviral coat proteins (recombinant penton base) WO 98/54346 PCT/US98/11024 44 containing the HA epitope. The anti-HA ScFv was able to precipitate HA-containing penton base proteins. These results indicate the successful construction of the extracellular portion of a pseudo-receptor for binding an adenovirus having a non-native ligand HA).

To create an entire anti-HA pseudo-receptor, the anti-HA ScFv was cloned into the pSCHAHK plasmid in which the HA had been removed to create the pScFGHA plasmid (Fig. 17B). This plasmid will produce an anti-HA pseudo-receptor able to bind recombinant adenoviruses having the HA epitope, similar to adenoviruses described above having FLAG epitopes.

EXAMPLE 19 This example describes the creation of a fiber-expressing cell line for the production of targeted adenovirus particles. The complementing cell line produces a fiber protein with or without additional complementary genes from the adenovirus genome.

The entire adenovirus type 2 fiber gene was amplified from adenovirus type 2 DNA by PCR. The resultant product was cloned into the pCR2.1-TOPO plasmid (Invitrogen) to make the plasmid pCR2.1-TOPO+fiber (Fig. 13A). The fiber2 gene was then excised from the pCR2.1-TOPO+fiber plasmid with the restriction enzymes BamHI and EagI, and it was then subcloned into the plasmid, pKSII (Stratagene), to construct the plasmid pKSII Fiber (Fig. 13B). The fiber2 gene was then excised from the pKSII Fiber plasmid using the restriction enzymes KpnI and EagI, and it was then cloned into the plasmid, pSMTZeo-DBP (Fig. 13C). The resultant plasmid, pSMTZeo-Fiber (Fig. 13D), encoded the entire fiber2 gene under control of the metallothionine promoter. This construct also placed an efficient mRNA splice site before the fiber gene to enhance fiber protein synthesis following induction. The pSMTZeo-Fiber plasmid also contains a Zeo resistance marker to allow selection of cell lines on the antibiotic zeocin.

To produce the cell line, the pSMTZeo-Fiber plasmid is transfected into 293 cells (or some other cell line) with or without additional adenovirus complementing functions.

Individual zeocin-resistant cell colonies are then amplified by standard means and tested for fiber2 production by Western analysis using an anti-fiber2 antibody) before and after induction with zinc, which activates the metallothionine promoter. Selected fiberexpressing clones are then tested for the ability to plaque and/or complement the growth of adenoviruses containing mutated fibers. Clones that adequately complement mutated fibers are suitable for amplifying and growing adenovirus particles having genomes encoding mutant fiber genes.

WO 98/54346 PCT/US98/11024 All references cited herein are hereby incorporated by reference to the same extent as if each reference were individually and specifically indicated to be incorporated by reference and were set forth in its entirety herein.

While this invention has been described with an emphasis on preferred embodiments, it will be obvious to those of ordinary skill in the art that variations of the preferred embodiments can be used and that it is intended that the invention can be practiced otherwise than as specifically described herein. Accordingly, this invention includes all modifications encompassed within the spirit and scope of the invention as defined by the following claims.

WO 98/54346 PCT/US98/11024 46 SEQUENCE LISTING GENERAL INFORMATION:

APPLICANT:

NAME: GENVEC, INC.

STREET: 12111 PARKLAWN DRIVE CITY: ROCKVILLE STATE: MD COUNTRY: US POSTAL CODE (ZIP): 20852 TELEPHONE: (301)816-0396 TELEFAX: (301)816-0085 NAME: WICKHAM, THOMAS J.

STREET: 2106 HUTCHISON GROVE COURT CITY: FALLS CHURCH STATE: VA COUNTRY: US POSTAL CODE (ZIP): 22043 NAME: KOVESDI, IMRE STREET: 7713 WARBLER LANE CITY: ROCKVILLE STATE: MD COUNTRY: US POSTAL CODE (ZIP): 20855 NAME: ROELVINK, PETRUS W.

STREET: 17502 GALLAGHER WAY CITY: OLNEY STATE: MD COUNTRY: US POSTAL CODE (ZIP): 20832 NAME: EINFELD, DAVID STREET: 17502 GALLAGHER WAY CITY: OLNEY STATE: MD COUNTRY: US POSTAL CODE (ZIP): 20832 NAME: BROUGH, DOUGLAS E.

STREET: 3900 SHALLOWBROOK LANE CITY: OLNEY STATE: MD COUNTRY: US POSTAL CODE (ZIP): 20832 NAME: LIZONOVA, ALENA STREET: 5329 RANDOLPH ROAD CITY: ROCKVILLE STATE: MD COUNTRY: US POSTAL CODE (ZIP): 20852 NAME: YONEHIRO, GRANT STREET: 4395 KENTBURY DRIVE CITY: BETHESDA STATE: MD COUNTRY: US POSTAL CODE (ZIP): 20814 (ii) TITLE OF INVENTION: ALTERNATIVELY TARGETED ADENOVIRUS (iii) NUMBER OF SEQUENCES: 18 WO 98/54346 PCT/US98/11024 (iv) COMPUTER READABLE FORM: MEDIUM TYPE: Floppy disk COMPUTER: IBM PC compatible OPERATING SYSTEM: PC-DOS/MS-DOS SOFTWARE: PatentIn Release #1.0, (vi) PRIOR APPLICATION DATA: APPLICATION NUMBER: US 60-047849 FILING DATE: 28-MAY-1997 (vi) PRIOR APPLICATION DATA: APPLICATION NUMBER: US 60-071668 FILING DATE: 16-JAN-1998 INFORMATION FOR SEQ ID NO:1: SEQUENCE CHARACTERISTICS: LENGTH: 960 base pairs TYPE: nucleic acid STRANDEDNESS: unknown TOPOLOGY: unknown (ii) MOLECULE TYPE: DNA (genomic) (ix) FEATURE: NAME/KEY: CDS LOCATION:1..957 (xi) SEQUENCE DESCRIPTION: SEQ ID NO:1: Version #1.30 (EPO) AAG CGC GCA AGA Lys Arg Ala Arg CCG TCT GAA GAT ACC TTC AAC CCC GTG Pro Ser Glu Asp Thr Phe Asn Pro Val 10 TAT CCA Tyr Pro TAT GAC ACG Tyr Asp Thr TTT GTA TCC Phe Val Ser GAA ACC Glu Thr GGT CCT CCA Gly Pro Pro ACT GTG CCT TTT CTT ACT CCT CCC 96 Thr Val Pro Phe Leu Thr Pro Pro 25 CCC AAT GGG TTT Pro Asn Gly Phe

CAA

Gin 40 GAG AGT CCC CCC Glu Ser Pro Pro

GGG

Gly GTA CTC TCT Val Leu Ser TTG CGC Leu Arg CTA TCC GAA CCT Leu Ser Glu Pro

CTA

Leu 55 GTT ACC TCC AAT Val Thr Ser Asn

GGC

Gly ATG CTT GCG CTC Met Leu Ala Leu ATG GGC AAC GGC Met Gly Asn Gly

CTC

Leu TCT CTG GAC GAG Ser Leu Asp Glu GGC AAC CTT ACC Gly Asn Leu Thr CAA AAT GTA ACC Gin Asn Val Thr GTG AGC CCA CCT Val Ser Pro.Pro AAA AAA ACC AAG Lys Lys Thr Lys TCA AAC Ser Asn ATA AAC CTG Ile Asn Leu

GAA

Glu 100 ATA TCT GCA CCC Ile Ser Ala Pro

CTC

Leu 105 ACA GTT ACC TCA GAA GCC CTA Thr Val Thr Ser Glu Ala Leu 110 ACT GTG Thr Val ATG CAA Met Gin 130 GCT GCC GCC Ala Ala Ala 115 GCA CCT CTA Ala Pro Leu 120 ATG GTC GCG GGC Met Val Ala Gly

AAC

Asn 125 ACA CTC ACC Thr Leu Thr TCA CAG GCC CCG Ser Gin Ala Pro CTA ACC GTG CAC GAC Leu Thr Val His Asp 135 TCC AAA CTT AGC ATT Ser Lys Leu Ser Ile 140 WO 98/54346 PCT/US98/11024 GCC ACC CAA GGA CCC Ala Thr Gin Gly Pro 145 GTC TCG GCG CTC GAG Val Ser Ala Leu Glu ACA GTG TCA GAA Thr Val Ser Glu

GGA:

Gly 155 AAG CTA GCA TCA Lys Leu Ala Ser

AGG

Arg 160 AAG ACG TCT CAA Lys Thr Ser Gin CAC TCT GAT ACT His Ser Asp Thr ATC CTC Ile Leu 175 CGG ATC ACC Arg Ile Thr GAG CAA AGT Glu Gin Ser 195

CAG

Gin 180 GGA CTC GAT GAT Gly Leu Asp Asp

GCA

Ala 185 AAC AAA CGA ATC Asn Lys Arg Ile ATC GCT CTT Ile Ala Leu 190 GCT CAA CTT Ala Gin Leu 576 624 CGG GAT GAC TTG Arg Asp Asp Leu GCA TCA GTC AGT Ala Ser Val Ser GCA ATC TCC AGA TTG GAA Ala Ile Ser Arg Leu Glu

AGC

Ser 215 TCT ATC GGA GCC Ser Ile Gly Ala CAA ACA GTT GTC Gin Thr Val Val

AAT

Asn 225 GGA CTT GAT TCG Gly Leu Asp Ser GTT ACC CAG TTG Val Thr Gin Leu

GGT

Gly 235 GCT CGA GTG GGA Ala Arg Val Gly

CAA

Gin 240 CTT GAG ACA GGA Leu Glu Thr Gly GCA GAC GTA CGC Ala Asp Val Arg GAT CAC GAC AAT Asp His Asp Asn CTC GTT Leu Val 255 GCG AGA GTG Ala Arg Val CTA TCA ACT Leu Ser Thr 275

GAT

Asp 260 ACT GCA GAA CGT Thr Ala Glu Arg

AAC

Asn 265 ATT GGA TCA TTG Ile Gly Ser Leu ACC ACT GAG Thr Thr Glu 270 GAT TTC GAA Asp Phe Glu 816 864 CTG ACG TTA CGA Leu Thr Leu Arg ACA TCC ATA CAA Thr Ser Ile Gin TCT AGG Ser Arg 290 GGA TCC GGC GGC Gly Ser Gly Gly

ACT

Thr 295 AGT GGC GGC GAC Ser Gly Gly Asp TAC AAG GAC GAC GAC Tyr Lys Asp Asp Asp 300 CTT GGC TCT AGA TAA Leu Gly Ser Arg

GAC

Asp 305 AAG GGC CCT AGG Lys Gly Pro Arg GGC GCC Gly Ala 310 CGC CGC GCC Arg Arg Ala

TCC

Ser 315 INFORMATION FOR SEQ ID NO:2: SEQUENCE CHARACTERISTICS: LENGTH: 633 base pairs TYPE: nucleic acid STRANDEDNESS: unknown TOPOLOGY: unknown (ii) MOLECULE TYPE: DNA (genomic) (ix) FEATURE: NAME/KEY: CDS LOCATION:1..630 (xi) SEQUENCE DESCRIPTION: SEQ ID NO:2: ATG AAG CGC GCA AGA CCG TCT GAA GAT ACC TTC Met Lys Arg Ala Arg Pro Ser Glu Asp Thr Phe 325 330 AAC CCC GTG Asn Pro Val TAT CCA Tyr Pro 335 TAT GAC ACG GAA ACC GGT CCT CCA ACT GTG CCT TTT CTT ACT CCT CCC WO 98/54346 WO 9854346PCTUS98/1 1024 T yr

TTT

Phe

GCA

Ala

ACT

Thr 385

ATC

Ile

GCT

Ala

ACA

Thr

GTG

Val

AAT

Asn 465

ACC

Thr

GAT

Asp Asp

GTA

Val

TCA

Ser 370

ATC

Ile

GCT

Al a

CAA

Gin

GTT

Val1

GGA

Gly 450

CTC

Leu

ACT

Thr

TTC

Phe Thr

TCC

Ser 355

AGG

Arq

CTC

Leu

CTT

Leu

CTT

Leu

GTC

Val 435

CAA

Gin

GTT

Val1

GAG

Glu

GAA

Glu G1u 340

CCC

Pro

GTC

Val1

CGG

Arg

GAG

Giu

GCA

Al a 420

AAT

Asn

CTT

Leu

GCG

Al a

CTA

Leu

TCT

Ser 500 Thr

AAT

Asn

TCG

Ser

ATC

Ile

CAA

Gin 405

ATC

Ile

GGA

Gly

GAG

Glu

AGA

Arg

TCA

Ser 485

AGG

Arg Gly

GGG

Gly

GCG

Ala

ACC

Thr 390

AGT

Ser

TCC

Ser

CTT

Leu

ACA

Thr

GTG

Val 470

ACT

Thr Pro

TTT

Phe

CTC

Leu 375

CAG

Gin

CGG

Arg

AGA

Arg

GAT

Asp

GGA

Gly 455

GAT

Asp

CTG

Leu Pro

CAA

Gin 360

GAG

Glu

GGA

Gi y

GAT

Asp

TTG

Leu

TCG

Ser 440

CTT

Leu.

ACT

Thr

ACG

Thr Thr 345

GAG

Glu

AAG

Lys

CTC

Leu

GAC

Asp

GAA

Giu 425

AGT

Ser

GCA

Al a

GCA

Al a

TTA

Leu Val

AGT

Ser

ACG

Thr

GAT

Asp

TTG

Leu 410

AGC

Ser

GTT

Val

GAC

Asp

GMA

Glu

CGA

Arg 490

ACT

Thr Pro

CCC

Pro

TCT

Ser

GAT

Asp 395

GTT

Val

TCT

Ser

ACC

Thr

GTA

Val

CGT

Arg 475

GTA

Val

AGT

Ser Phe

CCC

Pro

CAA

Gin 380

GCA

Ala

GCA

Ala

ATC

Ile

CAG

Gin

CC

Arg 460

MAC

As n

ACA

Thr

GGC

Gi y Lieu

~GGG

Gly 365

ATA

Ile

P.AC

Asn

TCA

Ser

GGA

Gi y

TTG

Leu 445

GTT

Val

ATT

Ile

TCC

Ser

GCC

Cl y Thr 350

GGA

Gly

CAC

His

AMA

Lys

GTC

Val

GCC

Ala 430

GGT

Gly

GAT

Asp

GGA

Gly

ATA

Ile

GAC

Asp 510 Pro

GGG

Gi y

TCT

Ser

CGA

Arg

AGT

Ser 415

CTC

Leu

GCT

Al a

CAC

His

TCA

Ser

CMA

Gin 495

TAC

Tyr Pro

CTA

Leu

GAT

Asp

ATC

Ile 400

GAT

Asp

CMA

Gin

CGA

Arg

GAC

Asp

TTG

Leu 480

GCG

Ala

MAG

Lys 144 192 240 288 336 384 432 480 528 576 624 GCA TCC CCC GGC Gly Ser Cly-Gly 505 GAC GAC GAC GAC MAG GGC CCT AGG GGC GCC CGC CGC GCC TCC CTT GGC Asp

TCT

Ser (2) Asp Asp Asp Lys Gly Pro Arg Gly Ala Arg 515 520 AGA TMA Arg 530 INFORMATION FOR SEQ ID NO:3: SEQUENCE CHARACTERISTICS: LENGTH: 1704 base pairs TYPE: nucleic acid STRANDEDNESS: unknown TOPOLOGY: unknown (ii) MOLECULE TYPE: DNA (qenomic) (ix) FEATURE: NAME/KEY: CDS LOCATION:l. .1701 (xi) SEQUENCE DESCRIPTION: SEQ ID NO:3: Arq Ala Ser Leu Gly 525 WO 98/54346 PCT/US98/1 1024 ATG AAG CGC GCA AGA CCG TCT GAA GAT ACC TTC AAC CCC GTG TAT CCA Met Lys Arg Ala.Arq Pro Ser Glu 215 Asp 220 Thr Phe Asn Pro Val Tyr Pro 225

TTG

Leu 260

AAA

Lys

CAA

Gin

ATA

Ile

ACT

Thr

ATG

Met 340

GCC

Ala

GTC

Val

CGG

Arg

GAG

Clu

GCA

Ala 420

AAI

Asn

CTT

Let GAC I Asp

GTA

Val 245

CGC

Arg

ATG

Met

AAT

Asn

AAC

Asn

GTG

Val 325

CAA

Gin

ACC

Thr

TCG

Ser

ATC

Ile

CAA

Gin 405

ATC

Ile

GGA

Cly

GAG

a Glu kCG hr 230

TCC

Ser

CTA

Leu

GGC

ly

GTA

Vai

CTG

Leu 310

GCT

Ala

TCA

Ser

CAA

Gin

GCG

Ala

ACC

Thr 390

AGT

Ser

TCC

Ser

CTT

Leu

ACA

Thr GAA A Glu T CCC P Pro P TCC C Ser C AAC C Asn C ACC I Thr 295

GAA

Glu

GCC

Ala CAG 4 Gin

GGA

Cly

CTC

Leu 375

CAG

Gin

CGG

Arg

AGA

Arg

GAT

Asp

GGA

Cly 455

CC

'hr

AT

~sn

;AA

flu

;GC

fly kCT 'hr

%TA

Ile

GCC

Ala

GCC

Ala

CCC

Pro 360

GAG

Glu

GGA

Gly

CAT

Asp

TTG

Leu

TCG

Ser 440

CTT

Let CGT C Gly I GGG I Gly I

CCT

Pro 265

CTC

Leu

GTG

Val

TCT

Ser

CCA

Ala

CCC

Pro 345

CTC

Leu

AAG

Lys

CTC

Leu

GAC

Asp

CAA

Glu 425

AGT

Ser

CCA

a Ala

:CT

?ro

[TT

?he 250

:TA

Leu

TCT

Ser

AGC

Ser

SCA

Ala

CCT

Pro 330

CTA

Leu

ACA

Thr

ACG

Thr

GAT

Asp

TTG

Leu 410

AGC

Ser

GTI

Val

CAC

Asr

CCA

Pro 235

CAA

Gin

GTT

Val

CTG

Leu

CCA

Pro

CCC

Pro 315

CTA

Leu

ACC

Thr

GTC

Val

TCT

Ser

GAT

Asp 395

GTT

Val

TC

Sei

ACC

Thi

GTI

Va ACT G Thr V GAG P Clu S ACC 1 Thr 5

GAC

Asp

CCT

Pro 300

CTC

Leu

ATC

Met

GTG

Va1

TCA

Ser

CAA

Gin 380

CCA

Ala

CCA

Ala

ATC

Ile

CAC

Gin k CGC i Arq 460

;TG

al

LGT

jer

:CC

er

;AG

flu 285

CTC

Leu

ACA

rhr

GTC

1al

CAC

His

CAA

Clu 365

ATA

Ile

AAC

Asn

TCA

Ser

GGP

Gl

TTC

Let 44E

GT

Va CCT T Pro E CCC C Pro I AAT C Asn C 270 GCC C Ala

AAA~

Lys

GTT

Val

GCG

Ala

CAC

Asp 350

GGA

Cly

CAC

His

AAA

Lys

CTC

Val

GCC

Ala 430

GGT

G Cly P GAT L Asp

'TT

'he

CC

?ro

;GC

fly 3GC ;iy

~AA

Lys

ACC

rhr

GGC

Gly 335

TCC

Ser

AAG

Lys

TCT

Ser

CGA

Arg

AGT

Ser 415

CTC

Leu

GCT

Ala

CAC

His CTT I Leu 1 240 GGG C Cly ATC C Met I

AAC

Asn

ACC

Thr TCA 4 Ser 320

AAC

Asn

AAA

Lys

CTA

Leu

CAT

Asp

ATC

Ile 400

CAT

Asp

CAA

Gin

CGA

Arg G CAC Asp

CT

hr 3TA lal

'TT

Leu

CTT

Leu

AAG

Lys 305

GAA

Glu

ACA

Thr

CTT

Leu

CCA

Ala

ACT

Thr 385

ATC

Ile

CCT

Ala

ACA

Thr

GTG

Val

AAT

Asn

CCT

Pro

CTC

Leu

GCG

Ala

ACC

Thr 290

TCA

Ser

GCC

Ala

CTC

Leu

AGC

Ser

TCA

Ser 370

ATC

Ile

CCT

Ala

CAA

Gin

CTT

Val

GGA

Gly 450

CTC

Leu

CCC

Pro

TCT

Ser

CTC

Leu 275

TCC

Ser

AAC

Asn

CTA

Leu

ACC

Thr

ATT

Ile 355

AGG

Arg

CTC

Leu

CTT

Leu

CTT

Leu

CTC

Val 435

CAA

Gin

CTT

Val 96 144 192 240 288 336 384 432 480 528 576 624 672 720 768 816 465 GCG AGA GTG GAT ACT CCA CAA CCT AAC ATT GGA TCA TTC ACC ACT GAG WO 98/54346 PCTIUS98/1 1024 Ala Arg Val Asp Thr Ala Glu 470

I

CTA

Leu

TCT

Ser 500

GAA

Glu

GTG

Val

ACA

Thr

CCT

Pro

TGC

Cys 580

AGC

Ser

GAT

Asp

ACC

Thr

GGA

Gly

GTG

Val 660

AAG

Lys

TAT

Tyr

AAC

Asn

TCA

Ser 485

AGG

Arg

CTG

Leu

AAT

Asn

TAC

Tyr

GTG

Va1 565

TTT

Phe

GCC

Ala

GAC

Asp

CTG

Leu

AAG

Asn 645

TAG

Tyr

ATC

Ile

CAA

Gli

CAT

His ACT CTG Thr Leu GGA TCC Gly Ser TTC ACT Phe Thr GGG CAC Gly His 535 GGA AAG Gly Lys 550 CCA TGG Pro Trp TCC AGA Ser Arg ATG CCC Met Pro GGG AAC Gly Asn 615 GTG AAT Val Asn 630 ATT GTC Ile Leu ATG ATG Ile Met AGA CAC Arq His GAG AAG Gin Asn 695 TAC CTG Tyr Leu 710

ACG

Thr

GGC

ly

GGC

Gly 520

AAA

Lys

GTC

Leu

CCA

Pro

TAC

Tyr

GAG

Glu 600

TAG

Tyr

AGA

Arg

GGC

Gly

GCC

Ala

AAG

Asn 680

ACT

Thr

TCC

Ser TTA CGA Leu Arg 490 GGC ACT Gly Thr 505 GTG GTG Val Val TTT TCT Phe Ser ACC CTG Thr Leu ACA GTG Thr Leu 570 CCA GAG Pro Asp 585 GGG TAT Gly Tyr AAG ACC Lys Thr ATC GAG Ile Glu CAC AAG His Lys 650 GAC AAG Asp Lys 665 ATT GAG Ile Glu CCA ATG Pro Ile ACC GAG Thr Gin krg 175 3TA Jal

%GA

Erg

CA

Pro

GTC

Vai

AAA

Lys 555

GTG

Val

CAT

His

GTG

Vai

CGC

Arg

TTG

Leu 635

GTG

Leu

CAA

Gin

GAT

Asp

GGC

Gly

TCI

Ser 715 ACA TCG 2 Thr Ser GGA GGT Gly Gly ATT CTC Ile Leu 525 AGC GGA Ser Gly 540 TTC ATG Phe Ile ACT ACC Thr Thr ATG AAG Met Lys GAG GAG Gin Glu 605 GCT GAA Ala Glu 620 AAG GGG Lys Gly GAA TAG Glu Tyr AAG AAT Lys Asn GGA TCC Gly Ser 685 GAG GGG Asp Gly 700 GCC CTG Ala Leu

%TA

lie

GGA

ly 510

GTG

Val

GAG

Glu

TGG

Cys

TTG

Phe

GAG

Gin 590

AGA

Arg

GTG

Va1

ATT

Ile

AAG

Asn

GGC

Gly 670

GTG

Val

CCT

Prc

TCI

Ser 1 CAA C Gin I 495

ATG

Met

GAA

Glu

GGT

Gly

ACC

Thr

ACC

Thr 575

CAT

His

ACC

Thr

AAG

Lys

GAG

Asp

TAT

Tyr 655

ATG

Ile

GAG

Gin

GTG

Vai

AAA

Lys 180 3CG kIa I

%GC

Ser I

CTG

Leu

GAA

Glu

ACT

Thr 560

TAT

Tyr

GAG

Asp

ATG

Ile

TTG

Phe

TTT

Phe 640

AAC

Asn

AAG

Lys

CTG

Leu

GTC

Leu

GAT

Asp 720 3AT ksp

AG

Lys

GAT

Dsp

GGT

ly 545

GGA

Gly

GGG

Gly

TTT

Phe

TTT

Phe

GAA

Glu 625

AAG

Lys

TCC

Sex

GTC

Val

GCC

Aia

CT(

Let

CC

Prc %sn Ile Giy Ser Leu Thr Thr

TTC

Phe

GGC

Gly

GGC

Gly 530

GAT

Asp

AAG

Lys

GTG

Val

TTG

Phe

TTC

Phe 610

GGT

Gly

GAA

Glu

CAC

His

AAG

Asn

GAG

Asp 690

GCA

i Pro

AAG

3 Asn lu

GAA

lu

GAG

lu 515

GAT

Asp

GCC

Ala

CTC

Leu

GAG

Gin

AAG

Lys 595

AAA

Lys

GAG

Asp

GAT

Asp

AAT

Asn

TTG

Phe 675

CAT

His

GAG

Asp

GAA

Glu 864 912 960 1008 1056 1104 1152 1200 1248 1296 1344 1392 1440 1488 1536 AAG AGA GAC CAC ATG GTC CTG GTG GAG TTT GTG ACC GGT GGT GGG ATC Lys Arg Asp His Met Val Leu Leu Giu Phe Val Thr Ala Ala Gly Ile 1584 725 730 735 WO 98/54346 PCT/US98/11024

ACA

Thr 740 CAT GGC ATG GAC His Gly Met Asp CTG TAC AAG GGT Leu Tyr Lys Gly

GGA

Gly 750 GGT AGA TCT ACT Gly Arg Ser Thr GGC GGC GAC TAC Gly Gly Asp Tyr GAC GAC GAC GAC Asp Asp Asp Asp

AAG

Lys 765 GGC CCT AGG GGC Gly Pro Arg Gly GCC CGC Ala Arg 770 1632 1680 1704 CGC GCC TCC Arg Ala Ser GGC TCT AGA TAA Gly Ser Arg INFORMATION FOR SEQ ID NO:4: SEQUENCE CHARACTERISTICS: LENGTH: 1830 base pairs TYPE: nucleic acid STRANDEDNESS: unknown TOPOLOGY: unknown (ii) MOLECULE TYPE: DNA (genomic) (ix) FEATURE: NAME/KEY: CDS LOCATION:1..1827 (xi) SEQUENCE DESCRIPTION: SEQ ID NO:4: ATG AGA Met Arg 570 GGA TCT CAC CAT Gly Ser His His CAT CAC CAT GGC His His His Gly

GAA

Glu 580 GAT GGA GCT TTG Asp Gly Ala Leu

TCC

Ser 585

CCC

Pro CTG ACA AAA ACC Leu Thr Lys Thr GAG GCC AAC GTC Glu Ala Asn Val 605 TTA GTC Leu Val 590 TAT CCC ACC Tyr Pro Thr TGG ACG GGG CCT Trp Thr Gly Pro ACC TTC TCG GGG Thr Phe Ser Gly AAT TCC CCA TCT Asn Ser Pro Ser GGC ATT Gly Ile 615 CTC AGA CTG Leu Arg Leu TCT GTA CAA Ser Val Gin 635 CTC AGC AGA ACC Leu Ser Arg Thr

GGG

Gly 625 GGC ACG GTC ATT Gly Thr Val Ile GGC ACC CTG Gly Thr Leu 630 ACC CTG GGC Thr Leu Gly 192 240 GGT AGC CTC ACG Gly Ser Leu Thr CCC AGT ACC GGT Pro Ser Thr Gly

CAG

Gin 645 ATG AAC Met Asn 650 CTT TAC TTT GAC Leu Tyr Phe Asp GAC GGC AAT GTG Asp Gly Asn Val

CTG

Leu 660 TCT GAG AGC AAC Ser Glu Ser Asn

CTC

Leu 665 GTC CGA GGG TCC Val Arg Gly Ser GGA ATG AAA GAC Gly Met Lys Asp

CAA

Gin 675 GAT ACC CTG GTG Asp Thr Leu Val 336 CCC ATT GCC AAT Pro Ile Ala Asn CAG TAC CTG ATG Gin Tyr Leu Met AAC CTC ACT GCA Asn Leu Thr Ala TAC CCT Tyr Pro 695 CGC CTC ATA Arg Leu Ile

CAG

Gin 700 ACC CTA ACT TCC Thr Leu Thr Ser

AGC

Ser 705 TAC ATT TAC ACA Tyr Ile Tyr Thr CAA GCG CAC Gin Ala His 710 CTT GAC CAC AAT AAC AGT GTG GTG GAC ATC AAG ATA GGG CTC AAC ACA WO 98/54346 PCT/US98/11024 53 Leu Asp His Asn Asn Ser Val Val Asp Ile Lys Ile Gly Leu Asn Thr 715 720 725 GAC CTG AGG CCC ACT GCG GCC TAC GGC CTA AGC TTT ACC ATG ACC TTC 528 Asp Leu Arg Pro Thr Ala Ala Tyr Gly Leu Ser Phe Thr Met Thr Phe 730 735 740 ACT AAC TCT CCC CCC ACC TCA TTT GGT ACC GAC CTG GTG CAA TTT GGC 576 Thr Asn Ser Pro Pro Thr Ser Phe Gly Thr Asp Leu Val Gin Phe Gly 745 750 755 760 TAC CTG GGT CAG GAT AGC TCC CCC TCC TTC CTG AGA GAA CTT CCC CTT 624 Tyr Leu Gly Gin Asp Ser Ser Pro Ser Phe Leu Arg Glu Leu Pro Leu 765 770 775 GCA TCC GAG GCG GGC TAC TTT GGC AAA CTG GCA GCT GCC TCT GAG GAA 672 Ala Ser Glu Ala Gly Tyr Phe Gly Lys Leu Ala Ala Ala Ser Glu Glu 780 785 790 ATG CCA GCC CCT CCT GAG GCC CAG ACG CAG GAC CAA GCA GCT GAG GAG 720 Met Pro Ala Pro Pro Glu Ala Gin Thr Gin Asp Gin Ala Ala Glu Glu 795 800 805 CCC CCG GCT CCT GCT GAG GCT GAG GCC CCC GCT CCT GCT GAG GCT GAG 768 Pro Pro Ala Pro Ala Glu Ala Glu Ala Pro Ala Pro Ala Glu Ala Glu 810 815 820 GCT GAG GCT GAA CCG CCC CGA AAA CCC CCT AGG GGT GAC CTG GCC GCC 816 Ala Glu Ala Glu Pro Pro Arg Lys Pro Pro Arg Gly Asp Leu Ala Ala 825 830 835 840 CTA TAC AAT AGG GTC CAC AGC GAC ACC CGC GCA GAG GAC ACA CCA ACC 864 Leu Tyr Asn Arg Val His Ser Asp Thr Arg Ala Glu Asp Thr Pro Thr 845 850 855 AGC CCC GAG TTG GTC ACA ACC TTG CCA GAC CCC TTT GTC CTC CCC CTA 912 Ser Pro Glu Leu Val Thr Thr Leu Pro Asp Pro Phe Val Leu Pro Leu 860 865 870 CCC GAC GGA GTC CCA ACC GGT GCG AGC ATT GTG TTG GAA GGT ACC CTC 960 Pro Asp Gly Val Pro Thr Gly Ala Ser Ile Val Leu Glu Gly Thr Leu 875 880 885 ACA CCC TCC GCT GTG TTT TTT ACC CTG GAT CTG GTG ACC GGG CCC GCC 1008 Thr Pro Ser Ala Val Phe Phe Thr Leu Asp Leu Val Thr Gly Pro Ala 890 895 900 AGT CTG GCG CTG CAC TTT AAC GTG CGC CTC CCA CTG GAA GGC GAA AAG 1056 Ser Leu Ala Leu His Phe Asn Val Arg Leu Pro Leu Glu Gly Glu Lys 905 910 915 920 CAC ATT GTG TGC AAC TCC AGA GAG GGT AGC AGC AAC TGG GGC GAA GAA 1104 His Ile Val Cys Asn Ser Arg Glu Gly Ser Ser Asn Trp Gly Glu Glu 925 930 935 GTA AGA CCG CAG GAG TTC CCC TTT GAA AGG GAA AAG CCA TTC GTC CTG 1152 Val Arg Pro Gin Glu Phe Pro Phe Glu Arg Glu Lys Pro .Phe Val Leu 940 945 950 GTC ATT GTC ATC CAA AGT GAC ACA TAC CAG ATC ACT GTG AAC GGG AAG 1200 Val Ile Val Ile Gin Ser Asp Thr Tyr Gin Ile Thr Val Asn Gly Lys 955 960 965 CCT CTG GTG GAT TTT CCA CAG AGA CTA CAG GGC ATT ACC CGT GCC TCC 1248 Pro Leu Val Asp Phe Pro Gin Arg Leu Gin Gly Ile Thr Arg Ala Ser 970 975 980 WO 98/54346 PCT/US98/11024

CTA

Leu 985 TCC GGA GAC Ser Gly Asp CTT GTG Leu Val 990 ACA ACC Thr Thr 1005 TTT ACC CGG Phe Thr Arg TTG TTA CCA Leu Leu Pro TTG ACA- Leu Thr 995 CCC CCC Pro Pro 1010 ATG TAC CCA Met Tyr Pro GCA GCT CCC Ala Ala Pro CCC GGA Pro Gly 1000 CTG GAC Leu Asp 1015 GAC CCC CGT CCC Asp Pro Arg Pro GTA ATC Val Ile CCA GAT GCC Pro Asp Ala .1020 TAT GTG CTC Tyr Val Leu AAT CTG CCC ACC GGA Asn Leu Pro Thr Gly 1025 CTG ACG CCT Leu Thr Pro 1030 GCC GAA TTT Ala Glu Phe AGA ACA CTC CTC Arg Thr Leu Leu 1035 ACC GTC ACG Thr Val Thr GGA ACC Gly Thr 1040 CCC ACG CCC Pro Thr Pro

CTC

Leu 1045 TTT ATT GTG Phe Ile Val 1050 AAT CTG GTC Asn Leu Val TAC GAT Tyr Asp 1055 TTA CAC TAT Leu His Tyr GAT TCC AAA AAT GTG Asp Ser Lys Asn Val 1060 GCC CTC CAC TTT AAT GTC GGC TTC ACC TCT Ala Leu His Phe Asn Val Gly Phe Thr Ser GAC AGC Asp Ser 1075 AAA GGC CAC ATC Lys Gly His Ile 1080 1065 1070 GCC TGC AAT GCC Ala Cys Asn Ala AGA ATG Arg Met 1085 AAT GGC ACA Asn Gly Thr TGG GGA AGT GAA ATC Trp Gly Ser Glu Ile 1090 ACA GTG Thr Val 1095 1296 1344 1392 1440 1488 1536 1584 1632 1680 1728 1776 1824 1830 TCT GAT TTC Ser Asp Phe CCC TTT Pro Phe 1100 CAA AGG GGA Gin Arg Gly AAA CCC Lys Pro 1105 TTC ACT CTG Phe Thr Leu CAG ATT CTC Gin Ile Leu 1110 ACC AGA GAG GCA GAC TTC CAA Thr Arg Glu Ala Asp Phe Gin 1115 GTC CTC Val Leu 1120 GTA GAT AAA CAA CCT TTA ACC Val Asp Lys Gin Pro Leu Thr 1125 GAC CAA ATC AAA TAT GTA CAC Asp Gin Ile Lys Tyr Val His 1140 CAG TTT CAA TAC Gin Phe Gin Tyr 1130 AGG CTG AAG GAA CTG Arg Leu Lys Glu Leu 1135 ATG TTT GGC CAT GTT GTG CAA ACC CAC CTG Met Phe Gly His Val Val Gin Thr His Leu GAA CAC CAA Glu His Gin 1155 GTG CCA GAT Val Pro Asp 1160 1145 1150 ACT CCA GTT TTT Thr Pro Val Phe TCT ACT GCG GGA Ser Thr Ala Gly 1165 GTT TCG AAA GTT TAC CCT Val Ser Lys Val Tyr Pro 1170 CAG ATA Gin Ile 1175 CTG TAG Leu INFORMATION FOR SEQ ID SEQUENCE CHARACTERISTICS: LENGTH: 2253 base pairs TYPE: nucleic acid STRANDEDNESS: unknown TOPOLOGY: unknown (ii) MOLECULE TYPE: DNA (genomic) (ix) FEATURE: NAME/KEY: CDS LOCATION:1..2250 WO 98/54346 WO 9854346PCT/US98/1 1024 (xi) SEQUENCE DESCRIPTION: SEQ ID NO: ATG AAG CGC GCA AGA CCG TCT GAA GAT ACC TTC AAC CCC GTG TAT CCA Met Lys Arg Ala Arg Pro Ser Glu Asp Thr Phe Asn Pro Val

TAT

Tyr

TTT

Phe

TTG

Leu

AAA.

Lys 675

CAA

Gin

ATA

Ile

ACT

Thr

ATG

Met

GCC

Ala 755

AAA

Lys

AAC

As n

TGT

Cys

GGT

Gly

TAC

T yr 835

GAC

GTA

Val1

CGC

A.rg 660

ATG

Met

AAT

As n

AAC

As n

GTG

Val

CAA

Gin 740

ACC

Thr

ACC

Thr

GTC

Val

CTC

Leu

AGC

Ser 820

TTI

PhE

ACG

Thr

TCC

Ser 645

CTA

Leu

GGC

Gly

GTA

Val

CTG

Leu

GCT

Ala 725

TCA

Ser

CAA

Gin

TTA

Le u

ACC

Thr

AGC

Ser 805

CTC

Leu

GAG

Asp

GAA

Glu 630

~CCC

Pro

TCC

Ser

AAC

Asn

ACC

Thr

GAA

Glu 710

GCC

Ala

CAG

Gin

GGA

Gly

GTC

Val

TTC

Phe 790

AGA

Arg

ACG

Thr GC7 Ala 615

ACC

Thr

AAT

As n

GAA

Glu

GGC

Gly

ACT

Thr 695

ATA

Ile

GCC

Ala

GCC

Ala

CCC

Pro

TAT

T yr 775

TCG

Ser

ACC

Thr

AAC

Asn

GAC

Asp GGT C Gly P GGG T1 Gly I CCT C Pro I

CTC

Leu 680

GTG

Val

TCT

Ser

GCA

Ala

CCG

Pro

CTC

Leu 760

CCC

Pro

GGG

Gly

GGG

Gly

CCC

Pro

GGC

Gly 840

CT

>ro

'TT

~he

~TA

~eu 665 rCT Ser

%GC

Ser

GCA

Al a

CCT

Pro

CTA

Leu 745

ACA

Thr

ACC

Thr

GAG

Glu

GGC

Gly

AG]

Sei 825 AA1 Asr CCA A Pro TI 6 CAA G Gin C 650 GTT I Val I)

CTG

Leu

CCA

Pro

CCC

Pro

CTA

Leu 730

ACC

Thr

GTG

Val

CTG

Leu

AAT

Asn

ACG

Thr 810

ACC

Thr

GTG

IVal

LCT

'hr )35

;AG

;lu .cc ~hr 3AC k.sp

CCT

Pro

CTC

Le u 715

ATG

M4et

GTG

Val

TCA

Ser

TGG

Trp

TCC

Ser 795

GTC

Val

GGI

Giy

CTC

Let 620 GTG C Val

AGT

Ser

TCC

Ser

GAG

Glu

CTC

Leu 700

ACA

Thr

GTC

Val

CAC

His

GAA

Giu

ACG

Thr 780

CCA

Pro

ATT

Ile

CAG

Gin

TCT

Ser

~CT

?ro Pro

I.AT

GCC

Al a 685

~AA

Lys

GTT

Val

GCG

Ala

GAC

Asp

GGA

Gly 765

GGG

Gly

TCT

Ser

GGC

Gly

ACC

Thr

GAG

Glu 845

~TTT

Phe

CCT

Pro

GGC

Gi y 670

GGC

Gly

AAA

Lys

ACC

Thr

GGC

Gi y

TCC

Ser 750

AAG

Lys

CCT

Pro

GGC

Gly

ACC

Thr

CTG

Leu 830

AGC

Ser

CTT

Leu

GGG

Gly 655

ATG

Met

AAC

Asn

ACC

Thr

TCA

Ser

AAC

Asn 735

AAA

Lys

CTA

Leu

GCT

Ala

ATT

Ile

CTG

Leu 815

GGC

Gly

AAC

Asn

%.CT

rhr 640

GTA

Val1

CTT

Le u

CTT

Leu

AAG

Lys

GAA

Glu 720

ACA

Thr

CTT

Leu

GCC

Ala

CCC

Pro

CTC

Leu 800

TCT

Ser

ATG

Met

CTC

Let Tyr Pro 625 CCT CCC Pro Pro CTC TCT Leu Ser GCG CTC Ala Leu ACC TCC Thr Ser 690 TCA AAC Ser Asn 705 GCC CTA Ala Leu CTC ACC Leu Thr AGC ATT Ser Ile CTG ACA Leu Thr 770 GAG GCC Glu Ala 785 AGA CTG *Arg Leu GTA CAA *Vai Gin AAC CTT *Asn Leu *GTC CGA 1Val Arg 850 96 144 192 240 288 336 384 432 480 528 576 624 672 720 768 GGG TCC TGG GGA ATG AAA GAG CAA GAT ACC CTG GTG ACT CCC ATT GCC Gly Ser Trp Gly Met Lys Asp Gin Asp Thr Leu Val Thr Pro Ile Ala 855 86086 WO 98/54346 PCT/US98/11024 56 AAT GGG CAG TAC CTG ATG CCC AAC CTC ACT GCA TAC CCT CGC CTC ATA 816 Asn Gly Gin Tyr Leu Met Pro Asn Leu Thr Ala-Tyr Pro Arg Leu Ile 870 875 880 CAG ACC CTA ACT TCC AGC TAC ATT TAC ACA CAA GCG CAC CTT GAC CAC 864 Gin Thr Leu Thr Ser Ser Tyr Ile Tyr Thr Gin Ala His Leu Asp His 885 890 895 AAT AAC AGT GTG GTG GAC ATC AAG ATA GGG CTC AAC ACA GAC CTG AGG 912 Asn Asn Ser Val Val Asp Ile Lys Ile Gly Leu Asn Thr Asp Leu Arg 900 905 910 CCC ACT GCG GCC TAC GGC CTA AGC TTT ACC ATG ACC TTC ACT AAC TCT 960 Pro Thr Ala Ala Tyr Gly Leu Ser Phe Thr Met Thr Phe Thr Asn Ser 915 920 925 930 CCC CCC ACC TCA TTT GGT ACC GAC CTG GTG CAA TTT GGC TAC CTG GGT 1008 Pro Pro Thr Ser Phe Gly Thr Asp Leu Val Gin Phe Gly Tyr Leu Gly 935 940 945 CAG GAT AGC TCC CCC TCC TTC CTG AGA GAA CTT CCC CTT GCA TCC GAG 1056 Gin Asp Ser Ser Pro Ser Phe Leu Arg Glu Leu Pro Leu Ala Ser Glu 950 955 960 GCG GGC TAC TTT GGC AAA CTG GCA GCT GCC TCT GAG GAA ATG CCA GCC 1104 Ala Gly Tyr Phe Gly Lys Leu Ala Ala Ala Ser Glu Glu Met Pro Ala 965 970 975 CCT CCT GAG GCC CAG ACG CAG GAC CAA GCA GCT GAG GAG CCC CCG GCT 1152 Pro Pro Glu Ala Gin Thr Gin Asp Gin Ala Ala Glu Glu Pro Pro Ala 980 985 990 CCT GCT GAG GCT GAG GCC CCC GCT CCT GCT GAG GCT GAG GCT GAG GCT 1200 Pro Ala Glu Ala Glu Ala Pro Ala Pro Ala Glu Ala Glu Ala Glu Ala 995 1000 1005 1010 GAA CCG CCC CGA AAA CCC CCT AGG GGT GAC CTG GCC GCC CTA TAC AAT 1248 Glu Pro Pro Arg Lys Pro Pro Arg Gly Asp Leu Ala Ala Leu Tyr Asn 1015 1020 1025 AGG GTC CAC AGC GAC ACC CGC GCA GAG GAC ACA CCA ACC AGC CCC GAG 1296 Arg Val His Ser Asp Thr Arg Ala Glu Asp Thr Pro Thr Ser Pro Glu 1030 1035 1040 TTG GTC ACA ACC TTG CCA GAC CCC TTT GTC CTC CCC CTA CCC GAC GGA 1344 Leu Val Thr Thr Leu Pro Asp Pro Phe Val Leu Pro Leu Pro Asp Gly 1045 1050 1055 GTC CCA ACC GGT GCG AGC ATT GTG TTG GAA GGT ACC CTC ACA CCC TCC 1392 Val Pro Thr Gly Ala Ser Ile Val Leu Glu Gly Thr Leu Thr Pro Ser 1060 1065 1070 GCT GTG TTT TTT ACC CTG GAT CTG GTG ACC GGG CCC GCC AGT CTG GCG 1440 Ala Val Phe Phe Thr Leu Asp Leu Val Thr Gly Pro Ala Ser Leu Ala 1075 1080 1085 1090 CTG CAC TTT AAC GTG CGC CTC CCA CTG GAA GGC GAA AAG CAC ATT GTG 1488 Leu His Phe Asn Val Arg Leu Pro Leu Glu Gly Glu Lys His Ile Val 1095 1100 1105 TGC AAC TCC AGA GAG GGT AGC AGC AAC TGG GGC GAA GAA GTA AGA CCG 1536 Cys Asn Ser Arg Glu Gly Ser Ser Asn Trp Gly Glu Glu Val Arg Pro 1110 1115 1120 CAG GAG TTC CCC TTT GAA AGG GAA AAG CCA TTC GTC CTG GTC ATT GTC 1584 Gin Glu Phe Pro Phe Glu Arg Glu Lys Pro Phe Val Leu Val Ile Val WO 98/54346 PCT/US98/11024 57 1125 1130 1135 ATC CAA AGT GAC ACA TAC CAG ATC ACT GTG AAC GGG AAG CCT CTG GTG 1632 Ile Gin Ser Asp Thr Tyr Gin Ile Thr Val Asn Gly Lys Pro Leu Val 1140 1145 1150 GAT TTT CCA CAG AGA CTA CAG GGC ATT ACC CGT GCC TCC CTA TCC GGA 1680 Asp Phe Pro Gin Arg Leu Gin Gly Ile Thr Arg Ala Ser Leu Ser Gly 1155 1160 1165 1170 GAC CTT GTG TTT ACC CGG TTG ACA ATG TAC CCA CCC GGA GAC CCC CGT 1728 Asp Leu Val Phe Thr Arg Leu Thr Met Tyr Pro Pro Gly Asp Pro Arg 1175 1180 1185 CCC ACA ACC TTG TTA CCA CCC CCC GCA GCT CCC CTG GAC GTA ATC CCA 1776 Pro Thr Thr Leu Leu Pro Pro Pro Ala Ala Pro Leu Asp Val Ile Pro 1190 1195 1200 GAT GCC TAT GTG CTC AAT CTG CCC ACC GGA CTG ACG CCT AGA ACA CTC 1824 Asp Ala Tyr Val Leu Asn Leu Pro Thr Gly Leu Thr Pro Arg Thr Leu 1205 1210 1215 CTC ACC GTC ACG GGA ACC CCC ACG CCC CTC GCC GAA TTT TTT ATT GTG 1872 Leu Thr Val Thr Gly Thr Pro Thr Pro Leu Ala Glu Phe Phe Ile Val 1220 1225 1230 AAT CTG GTC TAC GAT TTA CAC TAT GAT TCC AAA AAT GTG GCC CTC CAC 1920 Asn Leu Val Tyr Asp Leu His Tyr Asp Ser Lys Asn Val Ala Leu His 1235 1240 1245 1250 TTT AAT GTC GGC TTC ACC TCT GAC AGC AAA GGC CAC ATC GCC TGC AAT 1968 Phe Asn Val Gly Phe Thr Ser Asp Ser Lys Gly His Ile Ala Cys Asn 1255 1260 1265 GCC AGA ATG AAT GGC ACA TGG GGA AGT GAA ATC ACA GTG TCT GAT TTC 2016 Ala Arg Met Asn Gly Thr Trp Gly Ser Glu Ile Thr Val Ser Asp Phe 1270 1275 1280 CCC TTT CAA AGG GGA AAA CCC TTC ACT CTG CAG ATT CTC ACC AGA GAG 2064 Pro Phe Gin Arg Gly Lys Pro Phe Thr Leu Gin Ile Leu Thr Arg Glu 1285 1290 1295 GCA GAC TTC CAA GTC CTC GTA GAT AAA CAA CCT TTA ACC CAG TTT CAA 2112 Ala Asp Phe Gin Val Leu Val Asp Lys Gin Pro Leu Thr Gin Phe Gin 1300 1305 1310 TAC AGG CTG AAG GAA CTG GAC CAA ATC AAA TAT GTA CAC ATG TTT GGC 2160 Tyr Arg Leu Lys Glu Leu Asp Gin Ile Lys Tyr Val His Met Phe Gly 1315 1320 1325 1330 CAT GTT GTG CAA ACC CAC CTG GAA CAC CAA GTG CCA GAT ACT CCA GTT 2208 His Val Val Gin Thr His Leu Glu His Gin Val Pro Asp Thr Pro Val 1335 1340 1345 TTT TCT ACT GCG GGA GTT TCG AAA GTT TAC CCT CAG ATA CTG TAG 2253 Phe Ser Thr Ala Gly Val Ser Lys Val Tyr Pro Gin Ile Leu 1350 1355 1360 INFORMATION FOR SEQ ID NO:6: SEQUENCE CHARACTERISTICS: LENGTH: 795 base pairs TYPE: nucleic acid STRANDEDNESS: unknown TOPOLOGY: unknown WO 98/54346 WO 9854346PCTIUS98/1 1024 58 (ii) MOLECULE TYPE: DNA (genomic) (ix) FEATURE: NAME/KEY: CDS LOCATION:l. .792 (xi) SEQUENCE DESCRIPTION: SEQ ID NO:6: ATG GCG CTC CTG CTG TGC TTC GTC CTC CTC TGC GGA GTA GTC GAT TTC Met Ala Leu Leu Leu Cys Phe Val Leu Leu Cys Sly Val Val Asp Phe

GCC

Ala

AAA(

Lys

GAC

Asp 800

CAC

Gin

GAT

Asp

GAT

Asp

TCA

Ser

GCA

Ala 880

AGA

Arg

AAA

Lys

AAA

Lys

ACT

Thr

ACA

Thr 960

%GA

krq

~GGG

Gly 785

CAG

Gln

PAAG

Lys

SAC

Asp

CTC

Le u

GAT

Asp 865

AAT

Asn

TGT

Cys

TGT

Cys

TTG

Leu

TCA

Ser 945

TAC

Tyr

%GT

Ser 770

GAA

Giu

GGA

Gly

GTG

Val

TAC

Tyr

AAA

Lys 850

ATT

Ile

AAG

Lys

TAC

Tyr

GAA

Giu

TCT

Ser 930

TCT

Ser

AGC

Ser 755

TTG

Leu

ACT

Thr

CCG

Pro

GAT

Asp

TAT

Tyr 835

TCT

Ser

GCC

Gly

AAG

Lys

GTT

Val

CCA

Pro 915

GAC

Asp

STT

*Val

TGT

*Cys

ACT

Ser

CC

Al a

CTC

Leu

CAA

Gin 820

CCA

Pro

CCT

Sly

ACA

Thr

ATT

Ile

CAT

Asp 900

AAA

Lys

TCA

Ser

ATA

Ile

ACA

Thr

ATC

Ile

TAT

Tyr

SAC

Asp 805

GTG

Val1

CAT

Asp

SAT

Asp

TAT

Tyr

CAT

His 885

SCA

Sly

SAA

Giu

CAS

Cmn

TCT

Ser

GTC

Val1 965

ACT

rhr

CTG

Leu 790

ATC

Ile

ATT

Ile

CTG

Leu

SCA

Al a

CAS

Gin 870

CTG

Leu

TCT

Ser

CCT

Sly

AAA

Lys

GTA

Val1 950

AGA

Arg

ACTC

Thr 775

CCC

Pro

GAC

Giu

ATT

Ile

AAA.

Lys

TCA.

Ser 855

TSC

Cys

STA

Val1

CAA

Glu

TCA

Ser

ATG

Met 935

AAA

Lys

AAC

Asn 760

'CT

Pro

TGC

:ys rCC rrp

TTA

Leu

GGC

Sly 840

ATA

Ile

AAA

Lys

CTT

Val1

CAA

Giu

CTT

Leu 920

CCC

Pro

AAT

Asn

ASP

Arg

GAA

Giu

AAA

Lys

CTC

Leu

TAT

T yr 825

CCA

Arg

AAT

As n

GTS

Val

CTT

Leu

ATT

Ile 905

CCA

Pro

ACT

Thr

CC

Ala

GTC

Val

GAG

Giu

~TTT

Phe

PITA

Ile 810

TCT

Ser

STA

Val1

CTA

Val1

AAA

Lys

CTT

Val1 890

GSA

Sly

TTA

Leu

TCA

Ser

TCT

Ser

GCC

Sly 970

ATS

Met

ACS

Thr 795

TCA

Ser

GSA

Sly

CAT

His

ACS

Thr

AAA

Lys 875

AAS

Lys

ACT

Ser

CAC

Gin

TG

Trp

TCT

Ser 955

TCT

Ser

ATT

Ile 780

CTT

Leu

CCA

Pro

SAC

Asp

TTT

Phe

AAT

As n 860

CCT

Aila

CCT

Pro

SAC

Asp

TAT

Tyr

TTA

Leu 940

GAG

Giu

SAT

Asp 765

GAA

Giu

ACT

Ser

GCT

Al a

AAA

Lys

ACS

Thr 845

TTA

Leu

CCT

Pro

TCA

Ser

TTT

Phe

SAG

Slu 925 G CA Ala

TAC

*Tyr

CAC

Gin

AAA

Lys

CCC

Pro

CAT

Asp

ATT

Ile 830

ACT

Ser

CAA

Gin

SST

Sly

GST

Siy

AAS

Lys 910

TCS

Trp

GAA

Giu

TCT

Ser

TC

Cys

CC

Ala

SAA

Clu

AAT

Asn 815

TAT

Tyr

AAT

Asn

CTG

Leu

STT

Val

C

Ala 895

ATA

Ile

CAA

Sin

ATS

Met

GCC

Sly

CTC

Leu 975 96 144 192 240 288 336 384 432 480 528 576 624 672 720 TTG CGT CTA AAC GTT GTC CCT CCT TCA AAT AAA GCT GSA TCT GSA TCC Leu Arg Leu Asn Val Val Pro Pro Ser Asn Lys Ala Sly Ser Sly Ser WO 98/54346 PCT/US98/11024 GGC TCA GGG TCT Gly Ser Gly Ser 995 GAC TAT AAA GAT Asp Tyr Lys Asp 1010 ACT AGT GGG GCC CAG CCG GCC CTG CAG GCG GCC GCA Thr Ser Gly Ala Gin Pro Ala Leu Gin Ala Ala Ala 1000 1005 GAC GAC GAT AAG TGA Asp Asp Asp Lys 1015 INFORMATION FOR SEQ ID NO:7: SEQUENCE CHARACTERISTICS: LENGTH: 834 base pairs TYPE: nucleic acid STRANDEDNESS: unknown TOPOLOGY: unknown (ii) MOLECULE TYPE: DNA (genomic) (ix) FEATURE: NAME/KEY: CDS LOCATION:1..831 (xi) SEQUENCE DESCRIPTION: SEQ ID NO:7: ATG GCG CTC CTG Met Ala Leu Leu

CTG

Leu 270 TGC TTC GTG CTC Cys Phe Val Leu

CTG

Leu 275 TGC GGA GTA GTG Cys Gly Val Val GAT TTC Asp Phe 280 GCC AGA AGT Ala Arg Ser AAA GGG GAA Lys Gly Glu 300 AGT ATC ACT ACT Ser Ile Thr Thr

CCT

Pro 290 GAA GAG ATG ATT Glu Glu Met Ile GAA AAA GCC Glu Lys Ala 295 AGT CCC GAA Ser Pro Glu ACT GCC TAT CTG Thr Ala Tyr Leu

CCG

Pro 305 TGC AAA TTT ACG Cys Lys Phe Thr GAC CAG GGA CCG CTG GAC Asp Gin Gly Pro Leu Asp 315 GAG TGG CTG ATA Glu Trp Leu Ile CCA GCT GAT AAT Pro Ala Asp Asn CAG AAG GTG GAT CAA GTG ATT ATT TTA TAT TCT Gin Lys Val Asp Gin Val Ile Ile Leu Tyr Ser 330 335 340 GGA GAC AAA ATT Gly Asp Lys Ile TAT *240 Tyr 345 GAT GAC TAC TAT Asp Asp Tyr Tyr

CCA

Pro 350 GAT CTG AAA GGC Asp Leu Lys Gly

CGA

Arg 355 GTA CAT TTT ACG Val His Phe Thr AGT AAT Ser Asn 360 GAT CTC AAA Asp Leu Lys TCA GAT ATT Ser Asp Ile 380 GGT GAT GCA TCA Gly Asp Ala Ser AAT GTA ACG AAT Asn Val Thr Asn TTA CAA CTG Leu Gin Leu 375 CCT GGT GTT Pro Gly Val GGC ACA TAT CAG Gly Thr Tyr Gin

TGC

Cys 385 AAA GTG AAA AAA Lys Val Lys Lys GCA AAT Ala Asn 395 AAG AAG ATT CAT Lys Lys Ile His

CTG

Leu 400 GTA GTT CTT GTT Val Val Leu Val CCT TCA GGT GCG Pro Ser Gly Ala

AGA

Arg 410 TGT TAC GTT GAT Cys Tyr Val Asp

GGA

Gly 415 TCT GAA GAA ATT Ser Glu Glu Ile

GGA

Gly 420 AGT GAC TTT AAG ATA Ser Asp Phe Lys Ile WO 98/54346 PCT/US98/11024 AAA TGT GAA CCA Lys Cys Glu Pro GAA GGT TCA CTT Glu Gly Ser Leu

CCA

Pro 435 TTA CAG TAT GAG Leu Gin Tyr Glu TGG CAA Trp Gin 440 528 576 AAA TTG TCT Lys Leu Ser ACT TCA TCT Thr Ser Ser 460 TCA CAG AAA ATG Ser Gin Lys Met

CCC

Pro 450 ACT TCA TGG TTA GCA GAA ATG Thr Ser Trp Leu Ala Glu Met 455 GTT ATA TCT GTA Val Ile Ser Val

AAA

Lys 465 AAT GCC TCT TCT Asn Ala Ser Ser TAC TCT GGG Tyr Ser Gly ACA TAC Thr Tyr 475 AGC TGT ACA GTC Ser Cys Thr Val

AGA

Arg 480 AAC AGA GTG GGC Asn Arg Val Gly GAT CAG TGC CTG Asp Gin Cys Leu

TTG

Leu 490 CGT CTA AAC GTT Arg Leu Asn Val

GTC

Val 495 CCT CCT TCA AAT Pro Pro Ser Asn

AAA

Lys 500 GCT GGA TCT GGA Ala Gly Ser Gly GGC TCA GGG TCT Gly Ser Gly Ser AGA GCC TGC GAC Arg Ala Cys Asp CGC GGC GAT TGT Arg Gly Asp Cys TTT TGC Phe Cys 520 GGT ACT AGT Gly Thr Ser GCC CAG CCG GCC Ala Gin Pro Ala CAG GCG GCC GCA Gin Ala Ala Ala GAC TAT AAA 816 Asp Tyr Lys 535 GAT GAC GAC GAT AAG TGA Asp Asp Asp Asp Lys 540 INFORMATION FOR SEQ ID NO:8: SEQUENCE CHARACTERISTICS: LENGTH: 1194 base pairs TYPE: nucleic acid STRANDEDNESS: unknown TOPOLOGY: unknown (ii) MOLECULE TYPE: DNA (genomic) (ix) FEATURE: NAME/KEY: CDS LOCATION:1..1191 (xi) SEQUENCE DESCRIPTION: SEQ ID NO:8: ATG GCG Met Ala 280 CTC CTG CTG TGC Leu Leu Leu Cys

TTC

Phe 285 GTG CTC CTG TGC Val Leu Leu Cys

GGA

Gly 290 GTA GTG GAT TTC Val Val Asp Phe

GCC

Ala 295 AGA AGT TTG AGT Arg Ser Leu Ser ACT ACT CCT GAA Thr Thr Pro Glu

GAG

Glu 305 ATG ATT GAA AAA Met Ile Glu Lys

GCC

Ala 310 AAA GGG GAA ACT Lys Gly Glu Thr TAT CTG CCG TGC Tyr Leu Pro Cys

AAA

Lys 320 TTT ACG CTT AGT Phe Thr Leu Ser CCC GAA Pro Glu 325 GAC CAG GGA Asp Gin Gly CTG GAC ATC GAG Leu Asp Ile Glu CTG ATA TCA CCA GCT GAT AAT Leu Ile Ser Pro Ala Asp Asn 340 CAG AAG GTG GAT CAA GTG ATT ATT TTA TAT TCT GGA GAC AAA ATT TAT Gin Lys Val Asp Gin Val Ile Ile Leu Tyr Ser Gly Asp Lys Ile Tyr WO 98/54346 PCT/US98/11024 345 GAT GAC TAC 350

AAA

355 TTT TAT CCA GAT CTG GGC CGA GTA CAT ACG AGT AAT Asp 2

GAT

Asp 375

TCA

Ser

GCA

Ala

AGA

Arg

AAA

Lys

AAA

Lys 455

ACT

Thr

ACA

Thr

TTG

Leu

GGC

Gly

GAG

Glu 535

GGA

Gly

GAT

Asp

TTG

Leu %sp 360

CTC

Leu

GAT

Asp

AAT

Asn

TGT

Cys

TGT

Cys 440

TTG

Leu

TCA

Ser

TAC

Tyr

CGT

Arg

TCA

Ser 520

AAG

Lys

CTC

Leu

GAC

Asp

GA;

Glx Tyr 'I

AAA')

Lys

ATT

Ile

AAG

Lys

TAC

Tyr 425

GAA

Glu

TCT

Ser

TCT

Ser

AGC

Ser

CTA

Leu 505

GGG

Gly

ACG

Thr

GAT

Asp

TTG

Leu L AGC Ser 585 [yr

[CT

3er

GGC

ly

AAG

Lys 410

GTT

Va1

CCA

Pro

GAC

Asp

GTT

Val

TGT

Cys 490

AAC

Asn

TCI

Ser

TCT

Ser

GA']

AsF

GT

Val 57(

TC

Se: Pro A GGT G Cly P ACA 'I Thr '9 395 ATT C Ile I

GAT

Asp

AAA(

Lys

TCA

Ser

ATA

Ile 475

ACA

Thr

CTT

Val

ACT

Thr

CAA

Gin

GCA

Ala 555 r' GCA L Ala T ATC r Ile sp

;AT

~sp )80

'AT

~yr

:AT

Us 3GA 3ly

GAA

lu

CAG

Gin 460

TCT

Ser

GTC

Va1

GTC

Val

AGA

Arg

ATA

Ile 540

AAC

Asr

TC.

Se3

GG

Ci' Leu L 365 GCA I Ala S CAC 'I Gin C CTG C Leu

TCT

Ser

GGT

Gly 445

AAA

Lys

GTA

Vai

AGA

Arg

CCT

Pro

GGA

Cly 525

CAC

His

AAA

Lys I GTC Val k GCC y' Ala 'ys

'CA

;er

'GC

;ys

TA

lal

;AA

lu 130

TCA

Ser

ATG

Met

AAA

Lys

AAC

Asn

CCT

Pro 510

CGT

Gly

TC'I

Ser

CGI

Arc

AG

Se

CT(

Le 59( c iy Arg Vai His Phe Thr Ser Asn ATA A Ile A AAA G Lys V 4 GTT C Vai I 415 GAA I Glu I CTT C Leu I CCC 2 Pro

AAT

Asn

AGA

Arg 495

TCA

Ser

GGT

Gly

GAT

Asp

ATC

Ile C GAT Asp 575

CAA

u Gin 0

AT

sn

;TG

'ai 00

[TT

,eu

~TT

le

CA

?ro

%CT

rhr

GCC

Ala 480

GTG

Vai

AAT

Asn

GCA

Ala

ACT

Thr

ATC

Ile 560

GCT

Ala

ACI

Thi GTA 2 Vai 385

AAA

Lys

GTT

Val

GGA

Gly

TTA

Leu

TCA

Ser 465

TCT

Ser

GGC

Cly

AAA

Lys

TCA

Ser

ATC

Ile 545

GCT

Ala

CAA

Gin G CTT Val kCG rhr

AA

Lys

AAG

Lys

AGT

Ser

CAG

Gin 450

TGG

Trp

TCT

Ser

TCT

Ser

CCT

Ala

AGG

Arg 530

CTC

Leu

CTT

Leu

CTI

Leu

GTC

Val

AAT

Asn

GCT

Ala

CCT

Pro

GAC

Asp 435

TAT

Tyr

TTA

Leu

GAG

Glu

GAT

Asp

GGA

Cly 515

GTC

Vai

CGG

Arq

GAG

Glu

GCA

Ala

AA]

Asr 595 TTA C Leu G CCT C Pro C 4 TCA C Ser C 420

TTT

Phe

GAG

Glu

GCA

Ala

TAC

Tyr

CAG

Gin 500

TCT

Ser

TCC

Ser

ATC

Ile

CAA

Gin

ATC

Ile 580

CGA

Gly

AA

;ln

GT

fly 105

GT

;iy

.AG

Lys

TGG

Trp

GAA

Glu

TCT

Ser 485

TGC

Cys

GGA

Gly

GCG

Ala

ACC

Thr

AG'I

Se.

561

TCC

Se

CT'

Le

CTG

Leu 390

CTT

Val

GCG

Ala

ATA

Ile

CAA

Gin

ATG

Met 470

GGG

Cly

CTC

Leu

TCC

Ser

CTC

Leu

CAG

Gin 550

CGG

Arq

AGA

Arq

GAT

u Asp 288 336 384 432 480 528 576 624 672 720 768 816 864 912 960 1008 TCC ACT CTT ACC CAC TTG GGT CCT CGA GTG GCA CAA CTT GAG ACA CGA Ser Ser Val Thr Gin Leu Gly 605 Ala Arg Val Gly Gin Leu Glu Thr Gly WO 98/54346 PCT/US98/11024

CTT

Leu 615 GCA GAC GTA CGC Ala Asp Val Arg

GTT

Val 620 GAT CAC GAC AAT Asp His Asp Asn CTC GTT GCG AGA GTG GAT Leu- Val Ala Arg Val Asp 625 630 ACT GCA GAA CGT Thr Ala Glu Arg ATT GGA TCA TTG Ile Gly Ser Leu ACT GAG CTA TCA Thr Glu Leu Ser ACT CTG Thr Leu 645 1056 1104 1152 ACG TTA CGA Thr Leu Arg ATG CAG GCG Met Gin Ala 665 ACA TCC ATA CAA Thr Ser Ile Gin GAT TTC GAA TCT Asp Phe Glu Ser AGG ACT AGT Arg Thr Ser 660 GCC GCA GAC TAT Ala Ala Asp Tyr GAT GAC GAC GAT Asp Asp Asp Asp AAG TGA 1194 Lys 675 INFORMATION FOR SEQ ID NO:9: SEQUENCE CHARACTERISTICS: LENGTH: 1743 base pairs TYPE: nucleic acid STRANDEDNESS: unknown TOPOLOGY: unknown (ii) MOLECULE TYPE: DNA (genomic) (ix) FEATURE: NAME/KEY: CDS LOCATION:1..1743 (xi) SEQUENCE DESCRIPTION: SEQ ID NO:9: ATG AAG CGC GCA AGA CCG TCT GAA GAT ACC TTC Lys Arg Ala Arg Pro Ser Glu Asp Thr 10 Phe AAC CCC GTG Asn Pro Val TAT CCA 48 Tyr Pro TAT GAC ACG Tyr Asp Thr ACC GGT CCT CCA Thr Gly Pro Pro GTG CCT TTT CTT Val Pro Phe Leu ACT CCT CCC Thr Pro Pro GTA CTC TCT Val Leu Ser TTT GTA TCC CCC AAT GGG TTT Phe Val Ser Pro Asn Gly Phe CAA GAG AGT CCC CCT Gin Glu Ser Pro Pro GTT ACC TCC AAT GGC Val Thr Ser Asn Gly TTG CGC Leu Arg CTA TCC GAA CCT Leu Ser Glu Pro ATG CTT GCG CTC Met Leu Ala Leu

AAA

Lys ATG GGC AAC GGC Met Gly Asn Gly TCT CTG GAC GAG Ser Leu Asp Glu GGC AAC CTT ACC Gly Asn Leu Thr

TCC

Ser CAA AAT GTA ACC Gin Asn Val Thr GTG AGC CCA CCT Val Ser Pro Pro AAA AAA ACC AAG Lys Lys Thr Lys TCA AAC Ser Asn ATA AAC CTG Ile Asn Leu ACT GTG GCT Thr Val Ala 115 ATA TCT GCA CCC Ile Ser Ala Pro

CTC

Leu 105 ACA GTT ACC TCA Thr Val Thr Ser GAA GCC CTA Glu Ala Leu 110 ACA CTC ACC Thr Leu Thr GCC GCC GCA CCT Ala Ala Ala Pro

CTA

Leu 120 ATG GTC GCG GGC Met Val Ala Gly ATG CAA TCA CAG GCC CCG CTA ACC GTG CAC GAC TCC AAA CTT AGC ATT Met Gin Ser Gin Ala Pro Leu Thr Val His Asp Ser Lys Leu Ser lie WO 98/54346 WO 9854346PCT/US98/1 1024 GCC ACC CAA GGA CCC CTC ACA GTG TCA Ala Thr Gin Gly Pro

ACA

Thr

GCC

Ala

AAA

Lys

GCT

Al a

GGT

Gly 225

GGA

Gly

GGA

Gly

AGT

Ser

GGC

Gly

AAA

Lys 305

GTT

Val

GCC

Ala

AAC

Asn

TCA

Ser

TCA

Ser

TCA

Ser

GAG

Glu

CCT

Pro 210

CCA

Pro

GCC

Al a

GGA

Gi y

TAT

T yr

CCT

Pro 290

GGG

Gly

AAG

Asn

ATT

Ile

ACA

Thr

AAC

Asn 370

GGC

Gly

CCC

Pro

CCC

Pro 195

TTG

Leu

GGT

Gly

TTG

Leu

CTA

Leu

CGG

Pro 275

CTT

Leu

GTT

Leu

GTA

Leu

AAT

Asn

AAT

Asn 355

AAG

Lys

CCC

Pro

CCT

Pro 180

ATT

Ile

CAT

His

GTG

Val

GGT

Gly

AGG

Arg 260

TTT

Phe

TTT

Phe

TAG

Tyr

AGC

Ser

GGA

Ala 340

CCC

Pro

GCT

Ala

CTC

Leu 165

CTA

Leu

TAT

Tyr

GTA

Val

ACT

Thr

TTT

Phe 245

ATT

Ile

GAT

Asp

ATA

Ile

TTG

Leu

ACT

Thr 325

GGA

Gly

CTC

Leu

ATG

Met Leu Thr 150 ACC ACC Thr Thr ACT ACT Thr Thr ACA CAA Thr Gin ACA GAG Thr Asp 215 ATT AAT Ile Asn 230 GAT TGA Asp Ser GAT TCT Asp Ser GGT CAA Ala Gin AAG TGA Asn Ser 295 TTT AGA Phe Thr 310 GGG AAG Ala Lys GAT GGG Asp Gly AAA AGA Lys Thr GTT GGT Val Pro 375 Val

ACCG

Thr

GGG

Ala

AAT

Asn 200

GAG

Asp

AAT

Asn

CAA

Gin

CAA

Gin

AAG

Asn 280

GCC

Al a

GGT

Al a

GGG

Gly

GTT

Leu

AAA

Lys 360

AAA

Lys Ser

GAT

Asp

ACT

Thr 185

GGA

Gly

GTA

Leu

ACT

Thr

GGG

Gly

AAG

Asn 265

GAA

Gin

GAC

His

TGA

Ser

TTG

Leu

GAA

Giu 345

ATI

Ile

GTT

Leu

GAA

Giu

AGC

Ser 170

GGT

Gly

AAA

Lys

AAC

Asn

TGG

Ser

AAT

Asn 250

AGA

Arg

CTA

Leu

AAG

As n

AAC

Asn

ATG

Met 330

*TTT

*Phe

GGG

Gly

GGA

Gly

GGA

G

1 y 155

PAGT

Ser

AGC

Ser

CTA

Leu

ACT

Thr

TTG

Leu 235

ATG

Met

CGC

Arg

AAT

Asn

TTG

Leu

AAT

Asn 315

TTT

Phe

GGT

Gly

CAT

His

ACT

Thr 140

AAG

Lys

ACC

Thr

TTG

Leu

GGA

Giy

TTG

Leu 220

CAA

Gin

CAA

Gin

CTT

Leu

CTA

Leu

GAT

Asp 300

TGG

Ser

GAG

Asp

TGA

*Ser

GGC

Gly

*GGC

Gly 380

CTA

CTT

Leu

GGC

G

1 y

CTA

Leu 205

ACC

Thr

ACT

Thr

CTT

Leu

ATA

Ile

AGA

Arq 285

ATT

Ile

AAA

Lys

GGT

Ala

GCT

Pro

CTA

Leu 365

CTT

Leu

GCC

Aia

ACT

Thr

ATT

Ile 190

AAG

Lys

GTA

Val

AAA

Lys

AAT

As n

GTT

Leu 270

CTA

Leu

AAC

As n

AAG

Lys

ACA

Thr

AAT

Asn 350

*GAA

*Glu

AGT

Ser

CTG

Leu

A.TC

Ile 175

GAC

Asp

TAG

Tyr

GGA

Al a

GTT

Val1

GTA

Val 255

GAT

Asp

GGA

Gi y

TAG

T yr

CTT

Leu

GCC

Al a 335

GCA

Al a

TTT

Phe

TTT

Phe

GAA

Gin 160

ACT

Thr

TTG

Leu

GGG

Giy

ACT

Thr

ACT

Thr 240

GCA

Ala

GTT

Val1

GAG

Gin

AAG

As n

GAG

Giu 320

ATA

Ile

CCA

Pro

GAT

Asp

GAG

Asp 480 528 576 624 672 720 768 816 864 912 960 1008 1056 1104 1152 1200 AGC ACA GGT GCC ATT AGA GTA GGA AAC AAA AAT AAT GAT AAG CTA ACT Ser Thr Gly Ala Ile Thr Val Gly Asn Lys Asn Asn Asp Lys Leu Thr WO 98/54346 PCT/US98/11024 TTG TGG ACC ACA Leu Trp Thr Thr

CCA

Pro 405 GCT CCA TCT CCT AAC TGT AGA CTA AAT Ala Pro Ser Pro Asn Cys Arg Leu Asn 410 GCA GAG Ala Glu 415 AAA GAT GCT Lys Asp Ala CTT GCT ACA Leu Ala Thr 435 CTC ACT TTG GTC Leu Thr Leu Val

TTA

Leu 425 ACA AAA TGT GGC Thr Lys Cys Gly AGT CAA ATA Ser Gin Ile 430 GCT CCA ATA Ala Pro Ile GTT TCA GTT TTG Val Ser Val Leu

GCT

Ala 440 GTT AAA GGC AGT Val Lys Gly Ser TCT GGA Ser Gly 450 ACA GTT CAA AGT Thr Val Gin Ser CAT CTT ATT ATA His Leu Ile Ile

AGA

Arg 460 TTT GAC GAA AAT Phe Asp Glu Asn

GGA

Gly 465

AGA

Arg GTG CTA CTA AAC Val Leu Leu Asn AAT GGA GAT CTT Asn Gly Asp Leu 485 TCC TTC CTG GAC Ser Phe Leu Asp GAA TAT TGG AAC Glu Tyr Trp Asn ACT GAA GGC ACA Thr Glu Gly Thr

GCC

Ala 490 TAT ACA AAC GCT Tyr Thr Asn Ala GTT GGA Val Gly 495 1248 1296 1344 1392 1440 1488 1536 1584 1632 1680 1728 1743 TTT ATG CCT Phe Met Pro AAA AGT AAC Lys Ser Asn 515 CTA TCA GCT TAT Leu Ser Ala Tyr

CCA

Pro 505 AAA TCT CAC GGT Lys Ser His Gly AAA ACT GCC Lys Thr Ala 510 AAA ACT AAA Lys Thr Lys ATT GTC AGT CAA Ile Val Ser Gin

GTT

Val 520 TAC TTA AAC GGA Tyr Leu Asn Gly CCT GTA Pro Val 530 ACA CTA ACC ATT Thr Leu Thr Ile

ACA

Thr 535 CTA AAC GGT ACA Leu Asn Gly Thr

CAG

Gln 540 GAA ACA GGA GAC Glu Thr Gly Asp

ACA

Thr 545 ACT CCA AGT GCA Thr Pro Ser Ala TCT ATG TCA TTT Ser Met Ser Phe TGG GAC TGG TCT Trp Asp Trp Ser

GGC

Gly 560 CAC AAC TAC ATT His Asn Tyr Ile AAT GAA Asn Glu 565 ATA TTT GCC Ile Phe Ala

ACA

Thr 570 TCC TCT TAC ACT Ser Ser Tyr Thr TTT TCA Phe Ser 575 TAC ATT GCC Tyr Ile Ala CAA GAA Gin Glu 580 INFORMATION FOR SEQ ID SEQUENCE CHARACTERISTICS: LENGTH: 36 base pairs TYPE: nucleic acid STRANDEDNESS: single TOPOLOGY: unknown (ii) MOLECULE TYPE: DNA (genomic) (xi) SEQUENCE DESCRIPTION: SEQ ID TGCATGCATA CTAGTCCTAG ATTCGAAATC CGCTTG INFORMATION FOR SEQ ID NO:11: SEQUENCE CHARACTERISTICS: LENGTH: 38 base pairs WO 98/54346 PCT/US98/11024 TYPE: nucleic acid STRANDEDNESS: single TOPOLOGY: unknown (ii) MOLECULE TYPE: DNA (genomic) (xi) SEQUENCE DESCRIPTION: SEQ ID NO:11: GCTCTAGAGG AGGTGGTGCA TCAAGGGTCT CGGCGCTC 38 INFORMATION FOR SEQ ID NO:12: SEQUENCE CHARACTERISTICS: LENGTH: 29 base pairs TYPE: nucleic acid STRANDEDNESS: single TOPOLOGY: unknown (ii) MOLECULE TYPE: DNA (genomic) (xi) SEQUENCE DESCRIPTION: SEQ ID NO:12: CCGGATCCCT ACAGTATCTG AGGGTAAAC 29 INFORMATION FOR SEQ ID NO:13: SEQUENCE CHARACTERISTICS: LENGTH: 31 base pairs TYPE: nucleic acid STRANDEDNESS: single TOPOLOGY: unknown (ii) MOLECULE TYPE: DNA (genomic) (xi) SEQUENCE DESCRIPTION: SEQ ID NO:13: GGGCACCATG GCGAAGATGG AGCTTTGTCC C 31 INFORMATION FOR SEQ ID NO:14: SEQUENCE CHARACTERISTICS: LENGTH: 12 amino acids TYPE: amino acid STRANDEDNESS: not relevant TOPOLOGY: not relevant (ii) MOLECULE TYPE: peptide (xi) SEQUENCE DESCRIPTION: SEQ ID NO:14: Gly Ser Gly Ser Gly Ser Gly Ser Gly Ser Thr Ser 1 5 INFORMATION FOR SEQ ID SEQUENCE CHARACTERISTICS: LENGTH: 8 amino acids TYPE: amino acid STRANDEDNESS: not relevant TOPOLOGY: not relevant (ii) MOLECULE TYPE: peptide (xi) SEQUENCE DESCRIPTION: SEQ ID Arg Lys Lys Lys Arg Lys Lys Lys WO 98/54346 PCT/US98/11024 66 1 INFORMATION FOR SEQ ID NO:16: SEQUENCE CHARACTERISTICS: LENGTH: 8 amino acids TYPE: amino acid STRANDEDNESS: not relevant TOPOLOGY: not relevant (ii) MOLECULE TYPE: peptide (xi) SEQUENCE DESCRIPTION: SEQ ID NO:16: Asp Tyr Lys Asp Asp Asp Asp Lys 1 INFORMATION FOR SEQ ID NO:17: SEQUENCE CHARACTERISTICS: LENGTH: 14 amino acids TYPE: amino acid STRANDEDNESS: not relevant TOPOLOGY: not relevant (ii) MOLECULE TYPE: peptide (xi) SEQUENCE DESCRIPTION: SEQ ID NO:17: Pro Lys Ala Arg Arg Pro Ala Gly Arg Thr Trp Ala Gin Pro 1 5 INFORMATION FOR SEQ ID NO:18: SEQUENCE CHARACTERISTICS: LENGTH: 15 amino acids TYPE: amino acid STRANDEDNESS: not relevant TOPOLOGY: not relevant (ii) MOLECULE TYPE: peptide (xi) SEQUENCE DESCRIPTION: SEQ ID NO:18: Arg Pro Ile Asp Asp Phe Asp Gin Gly Trp Gly Pro Ile Thr Tyr 1 5 10

Claims (38)

1. A trimer comprising three monomers, each of said monomers having an amino terminus of an adenoviral fibre protein and each of said monomers having a trimerization domain comprising an adenoviral fibre knob lacking a native substrate- binding amino acid and having a non-native amino acid differing in charge or molecular weight from said native amino acid.

2. The trimer of claim 1, wherein said native amino acid is substituted with said non-native amino acid.

3. The trimer of claim 1 or 2, wherein said native substrate-binding amino acid is within a P-sheet.

4. The trimer claim 1 or 2, wherein said native substrate-binding amino acid is within a loop connecting two P-sheets.

5. A trimer comprising three monomers, each of said monomers having an amino terminus of an adenoviral fibre protein and each of said monomers having a trimerization domain from trout axonal dynein, parainfluenza virus hemagglutanin, or the sigma-1 protein of reovirus. S*

6. A trimer comprising three monomers, each of said monomers having an amino terminus of an adenoviral fibre protein and each of said monomers having a trimerization domain comprising a modified leucine-zipper motif.

7. The trimer of claim 6, wherein said leucine-zipper motif is modified by substituting one or more leucine residues with isoleucine.

8. The trimer of claim 6 or 7, wherein said trimerization domain is derived from the yeast GCN4p-II trimer.

9. The trimer of any of claims 1-8, which is not a ligand for a native mammalian cell-surface binding site.

The trimer of any of claims 1-9, wherein at least one of said three monomers has a non-native polypeptide interfering with the binding of said trimer to T its native cell-surface binding site. 68

11. A composition of matter comprising a trimer of any of claims 1-10 and an adenoviral penton base.

12. The composition of claim 11, wherein said penton base has a non- native ligand.

13. An adenovirus having the trimer of any of claims 1-10.

14. The adenovirus of claim 13, which does not productively infect 293 cells.

The adenovirus of claim 13 or 14, having a non-adenoviral ligand.

16. The adenovirus of claim 15, wherein said ligand binds a substrate other than a native mammalian adenoviral receptor.

17. The adenovirus of claim 15 or 16, wherein said ligand binds a substrate other than a native cell-surface protein.

18. The adenovirus of claim 16 or 17, wherein said substrate is present on the surface of a cell.

19. The adenovirus of claim 16 or 17 wherein said substrate is present within an affinity column.

20. The adenovirus of claim 16 or 17, wherein said substrate is present on a blood-borne molecule.

21. A cell line expressing a non-native cell-surface receptor to which an adenovirus having a ligand for said receptor binds, wherein said cell-surface receptor is derived from an antibody molecule.

22. The cell line of claim 21, wherein said antibody molecule comprises a single chain antibody.

23. The cell line of claim 21 or 22, wherein said antibody molecule recognizes hemagglutinin. 69

24. The cell line of any of claims 21-23, which can support viral growth for at least 10 passages.

A method of propagating an adenovirus comprising infecting a cell line of any of claims 21-24 with an adenovirus, maintaining said cell line, and recovering the adenoviruses produced within said cell line.

26. A method of purifying an adenovirus having a ligand for a substrate from a composition comprising said adenovirus, wherein said method comprises exposing said composition to said substrate such that said adenovirus selectively binds to said substrate, separating said substrate from said composition without removing said adenovirus from said substrate, and eluting said adenovirus from said substrate.

27. A method of inactivating in a fluid an adenovirus having a ligand recognizing a fluid-borne substrate by exposing said virus to said substrate such that said ligand binds said substrate, thereby adsorbing said virus from said fluid.

28. The method of claim 27, wherein said fluid is blood or lymph.

29. A chimeric blocking protein having a substrate for an adenovirus fibre, wherein said substrate is the extracellular domain of the CAR cell-surface protein.

30. The chimeric blocking protein of claim 29, further having a ligand.

31. The chimeric blocking protein of claim 30, wherein said ligand recognizes a substrate present on a cell surface binding site.

32. A method of interfering with adenoviral targeting comprising incubating an adenovirus with the chimeric blocking protein of any of claims 29-31 in a solution such that said chimeric blocking protein binds the fibre of the adenovirus.

33. A trimer substantially as hereinbefore defined with reference to the accompanying examples.

34. An adenovirus substantially as hereinbefore defined with reference to the accompanying examples.

A cell line substantially as hereinbefore defined with reference to the accompanying examples.

36. A method of purifying an adenovirus having a ligand for a substrate from a composition substantially as hereinbefore defined with reference to the accompanying examples.

37. A chimeric blocking protein substantially as hereinbefore defined with reference to the accompanying examples.

38. A method of inactivating in a fluid an adenovirus having a ligand recognizing a fluid-borne substrate substantially as hereinbefore defined with reference to the accompanying examples. o •o *oo ooo** o• oooo* *oo