AU750395B2 - A method for selectively controlling membrane protein display and protein secretion in eukaryotic cells - Google Patents

Description

WO 99/16875 PCT/US98/20364

-I-

A METHOD FOR SELECTIVELY CONTROLLING MEMBRANE PROTEIN DISPLAY AND PROTEIN SECRETION IN EUKARYOTIC CELLS Related Applications This application is related to provisional application Serial No. 60/047,352, filed May 21, 1997, and provisional application Serial No. 60/053,218, filed July 18, 1997, both of which are herein incorporated by reference in their entirety. This application also claims priority to U.S. Provisional Application 60/060,559, which is herein incorporated by reference in its entirety. All other articles, publications, patents and documents referred to throughout this application are hereby incorporated by reference in their entirety.

Technical Field The present invention relates to methods for selectively modulating the sequestration of integral membrane or secretory proteins into transport vesicles and the trafficking of such vesicles between the endoplasmic reticulum, Golgi bodies, plasma membrane and other membrane compartments. The present invention, based on the discovery of a new biological phenomena, provides methods and compositions for use in identifying relevant integral membrane and secretory proteins, as well as agents that inhibit the sequestration of such proteins in transport vesicles. Related methods and compositions can be used to modulate the secretion and cell membrane display of various disease-related proteins.

WO 99/16875 PCT/US98/20364 -2- Acknowledgment of Federal Support The research and discoveries described herein were supported by grants from the National Institutes of Health: NIH P01-DK38979, NIH R01-HL28560, NIH R01-DK43812, and NIH R29-DK47072.

Background of the Invention In all eukaryotic cells, a central and common process in integral membrane protein and secretory protein delivery is vesicular transport from their site of synthesis, the endoplasmic reticulum, to and through the Golgi apparatus, and ultimately to specific plasma membrane or internal membrane domains. See, Darnell et al., Molecular Cell Biology (1990, Scientific American Books) at Chapter 17, Plasma-Membrane, Secretory, and Lysosome Proteins: Biosynthesis and Sorting. General mechanisms by which this process is achieved are now understood in reasonable detail, and involve complex pathways of coated vesicle budding, transport, and fusion (reviewed in Kreis and Pepperkok,1994; Schmid and Damke,1995; Rothman and Wieland,1996; Schekman and Orci,1996; Bannykh and Balch,1997). Collectively, these processes link the synthesis and folding of membrane proteins in the endoplasmic reticulum to a process of sequential vesicular transport through the various Golgi compartments and ultimately their targeted delivery to specific plasma membrane or organelle domains. The molecular mechanisms that mediate these processes are less well understood. Typically, budding vesicles are encased in a closely adherent protein shell that favors their extrusion from a planar membrane compartment. Three general types of coats are now recognized: Clathrin/AP, COPI, and COPII. These coats are similar in many ways. They all form easily discernible electron dense layers about their respective vesicles; they are of relatively simple and uniform composition; and they assemble onto vesicles under the control of ARF like GTP-binding proteins.

Homologues of the spectrin-ankyrin membrane cytoskeleton of erythrocytes have WO 99/16875 PCT/US98/20364 -3recently emerged as candidates for a fourth class of vesicle or Golgi coat protein.

Immunologic, gel electrophoretic, and functional data strongly imply the presence of a novel isoforms of spectrin (termed PIE* and PIII) associated with Golgi membranes (Beck et al.,1994; for review see Devarajan and Morrow,1996; Devarajan et al.,1996; Morrow et al.,1997). Such isoforms of spectrin have been identified as components of the dynactin complex required for microtubule based vesicular transport (Holleran et al., 1996). Unique homologues of ankyrin, an adapter protein that links Na,K-ATPase and other transport proteins to the spectrin skeleton at the plasma membrane (for reviews see Bennett,1992; Devarajan and Morrow,1996; Morrow, et al.,1997), also exist in association with the Golgi. The Golgi associated ankyrins include a novel small ankyrin (Ank,,, 9 that has been cloned and characterized (Devarajan, et al.,1996), and larger isoforms that so far are only identified immunologically (Beck et al.,1997). Additional ankyrins also associate with other internal membrane compartments such as lysosomes (Hoock et al.,1997). While spectrin differs from the coatomer proteins in that it does not form geometrically precise coats and is more difficult to visualize by electron microscopy, as with other coat proteins spectrin's association with Golgi membranes is stimulated by ARF in a PtdInsP,-dependent manner (Godi et al.,1997). Evidence disclosed below indicates a direct and specific role for spectrin-ankyrin and other adapter proteins in mediating the transport of integral and secretory proteins. This spectrin-ankyrin-adapter protein trafficking/tethering system is collectively termed

SAATS.

Summary of the Invention The present invention is based on our discovery of a covert intracellular processing and trafficking system that mediates the sequestration of integral membrane and secretory proteins into transport vesicles for transport from the endoplasmic reticulum to the cis-Golgi apparatus, from the cis-Golgi to the medial-Golgi apparatus, from the medial-Golgi to the trans-Golgi apparatus and from the trans-Golgi apparatus to 4 the cell membrane or other cellular compartments. We identify specific adapter proteins and functionally active sites within components of the SAATS system that in isolation demonstrate pharmacologic activity, with the ability to control specifically the display or trafficking of selected membrane proteins. General methods are presented that will allow specific transport inhibitors for a selected membrane or secreted protein to be identified, and several examples of the application of these methods are presented.

Throughout this specification the word "comprise", or variations such as "comprises" or "comprising", will be understood to imply the inclusion of a stated element, integer or step, or group of elements, integers or steps, but not the exclusion of any other element, integer or step, or group of elements, integers or steps.

Any discussion of documents, acts, materials, devices, articles or the like which has been included in the present specification is solely for the purpose of providing a context for the present invention. It is not to be taken as an admission that any or all of these matters form part of the prior art base or were common general knowledge in the field relevant to the present invention as it 20 existed in Australia before the priority date of each claim of this application.

o.

.go• g.o.

4a Brief Description of the Drawings Figure 1: Diagram of the spectrinlankyrinladapter protein trafficking/tethering system (SAATS) and one way protein transport may be blocked. General schematic by which dynactin moves a vesicle from the endoplasmic reticulum (ER) to the Golgi. The transported tubular vesicular structure/vesicle emanating from the exit portals of the endoplasmic reticulum Golgi intermediate compartment (ERGIC) are linked to dynactin by SAATS, which is able to simultaneously bind the vesicle and dyNactin through an interaction with centractin, which is part of the dynactin complex. Centractin is an actin related protein, or ARP that binds to spectrin. In wild-type cells, various proteins bind to vesicular associated spectrin directly or indirectly via adapter proteins such as ankyrin or by binding to another membrane carrier protein that is subsequently bound by SAATS (such as for secreted proteins).

L, lumen of the ER. Selected protein transport is achieved by blockage of its attachment to SAATS, such as in a cell transfected with a cDNA encoding a spectrin plN-5 construct. In such cells, certain integral membrane or secretory proteins are not sequestered into the vesicle for transport from the endoplasmic S"reticulum to the Golgi.

*o*o o0* a *o o *oO WO 99/16875 PCT/US98/20364 Figure 2: A series of appropriate and representative PIZ* spectrin constructs useful for blocking membrane protein display and for determining suitable SAATS targets for pharmacologic agent identification. Some of the important functional domains are also depicted. Ankyrin binding occurs in repeat unit 15. The amino-terminal domain (green) bestows actin and presumably ARP binding on spectrin complexes. Membrane association domains (MAD1 and MAD2) are also shown. A third membrane association domain (MAD3) also resides somewhere between repeat units 3 and 9. Similar constructs prepared from other spectrins involved in protein trafficking PIII spectrin) can also be prepared.

Figure 3: Illustration of the use of cultured Madin Darby Canine Kidney (MDCK) cells to monitor the surface expression of Na,K-ATPase by indirect immunofluorescence and immunoprecipitation. In this example, the expression of amino terminal PI spectrin constructs in MDCK cells disrupts the native Golgi spectrin-ankyrin skeleton and Na,K-ATPase assembly. The intracellular distribution of wild type Golgi spectrin was monitored using Mab VIIIC7, which reacts with full length Golgi spectrin but not the shorter FLAG tagged PIN- construct (Harris et al.,1985; Devarajan, et al.,1996; Godi, et al.,1997). The distribution of PIZ* spectrin in cells transfected with PI, -c was monitored by Mab VD4, which does not react with the PI, 1-c peptide. AnkG119 was monitored by the antibody "Jasmin" (Devarajan, et al.,1996); a-Na,K-ATPase by a Mab from Upstate Biotechnology. Note the dispersal of the endogenous Golgi spectrin, Ankl 9 and a-Na,K-ATPase by PIN. (but not by PI All shortened spectrin constructs containing region I and MAD 1 displayed a similar effect, whereas constructs lacking these regions of PI spectrin did not significantly alter the native distributions of pIE* spectrin, AnkG1 19, or Na,K-ATPase. Bar 10pam. The interaction of Na,K-ATPase with the 14-15 repeat unit of spectrin occurs via an ankyrin G1 19 adapter protein intermediate, as detected by co-immunoprecipitation. In wt cells, WO 99/16875 PCT/US98/20364 -6- PIE* spectrin, AnkG119, centractin, and Na,K-ATPase are co-precipitated from detergent lysates by Mab VIIIC7 (lane IP). Soluble allplI spectrin (as detected here by immunoblotting for all spectrin), the form found predominately at the plasma membrane, is also present in these detergent lysates (lane lys) but is not part of the precipitable Golgi spectrin complex. Control experiments with preimmune or irrelevant antibody did not precipitate any of these components (data not shown). Similarly, immunoprecipitates with Mab VIIIC7 of the PIN- transfected cells also demonstrated co-precipitation of AnkGil 9 and Na,K-ATPase with PI* spectrin, indicating that despite the dispersal of these elements induced by the PIN. construct, these elements remained associated in the detergent lysates (data not shown). Western blots after SDS-PAGE, visualized by ECL.

Figure 4: Illustration of alternative assays demonstrating the impairment of Na,K-ATPase trafficking from the ER induced by SAATS inhibition. Dispersal of Golgi spectrin by PIN- blocks Na,K-ATPase transport to the medial Golgi and its incorporation into a detergent stable skeletal complex. The intracellular distribution of the endogenous PIS* spectrin (determined by Mab VIIIC7, which does not react with the expressed PIN. peptide), and Na,K-ATPase in MDCK cells transfected with plINspectrin. Golgi dependent glycosylation of P-Na,K-ATPase is blocked by the PINspectrin peptide. Control MDCK cells (wt) or cells transfected with the pIN-5 construct (PIN5) were analyzed by Western blot using a P-Na,K-ATPase specific antibody (Mab B 1-13). The apparent MW of the various products are depicted. Note the broad band above 50 kDa (gly) representing mature glycosylated P-Na,K-ATPase, and the intense band at 44 kDa in the pIN, cells (core), representing the ER dependent core P-Na,K-ATPase glycosylation product. It thus appears that the pINs spectrin peptide inhibits mature glycosylation of P-Na,K-ATPase, a process characteristic of medial Golgi processing, but does not interfere with the formation of the core glycosylation product, which occurs in the ER. In this experiment, approximately half of the PIN, cells are expressing the PIN. construct, as judged by immunofluorescence (data not shown), which WO 99/16875 PCT/US98/20364 -7appears to account for most of the mature glycosylated product observed in the IN.5 cell lines. The band 106 kDa in the wild type cells is inconstant in its appearance, and may represent a cross-linked adduct of P-Na,K-ATPase with a-Na,K-ATPase or other protein, as is commonly observed after reduced SDS-PAGE in MDCK cells (Morrow et al.,1989). Fraction of Na,K-ATPase present in the soluble and detergent insoluble pools of MDCK cells transfected with different PI spectrin constructs. Fxl (soluble fraction) is that material solubilized by 0.5% Triton X-100 in 100 mM NaC1; Fx2 (cytoskeletal fraction) is the material solubilized by 0.5% Triton X-100 and 250 mM

NH

4

SO

4 (Devarajan et al.,1994). Shown are Western blots for a-Na,K-ATPase after SDS-PAGE. The results of three separate determinations of a-Na,K-ATPase extractability of wt and two transfected MDCK cell lines were quantified in-duplicate by densitometry, and all analyses averaged. Error bars represent 1SD. Note the substantial loss of fully assembled and detergent insoluble Na,K-ATPase in the PIN5 expressing cells.

Figure 5: Illustration of selective blockage of Na,K-ATPase and VSV-G transport, but not E-cadherin, by pIN-5. Comparison of the effect of PIN5 spectrin on Golgi and the assembly of different proteins in MDCK cells. Wild-type and transfected cells are shown. The localization ofVSV- G protein (E,F) was measured after transient infection. E-cadherin was monitored with a Mab from Transduction labs, and also by the extent to which the precursor peptide was proteolyzed (inset, Western blot) from 135 kDa to 120 kDa (EC) (a process that occurs in the trans-Golgi) (Shore and Nelson, 1991). There was no significant difference in the extent of E-cadherin processing or its level of assembly at the plasma membrane in the PIN 5 s line vs. wt cells. Despite the disruption of Na,K-ATPase and VSV-G transport (and the wt Golgi spectrin skeleton, Fig. the Golgi appears to remain largely intact as measured by the distribution of P-COP and by the presence of normal appearing juxtanuclear Golgi structures in uranyl acetate and lead stained electron microscopy (arrows).

WO 99/16875 PCT/US98/20364 -8- Bar 10p in 0.5i in Original magnification 63,000x.

Figure 6: Illustrates the rescue of Na,K-ATPase trafficking by inclusion of a specific SAATS binding domain that in this case, binds to ankyrin. MDCK cells were transfected with either PIN-5 as above, or a with 3IN-5.15. This latter construct incorporates the ankyrin binding domain of spectrin into the PIN_ peptide. Note the complete restoration of Na,K-ATPase transport, as measured in this case by its surface display and secondarily by a reduction in cell size (which is a consequence of cell swelling due to a deficiency of plasma membrane Na,K-ATPase in the pIN-s transfected cells.

Figure 7: Illustrates how to make SAATS block VSV-G transport (but not Na,K-ATPase). MDCK cells infected with vesicular stomatitis virus were used to monitor the trafficking of VSV-G protein. The transport of this protein is normally to the basolateral membrane, similar to that for Na,K-ATPase (as shown in the wt cells and in Figure This transport is blocked by SAATS inhibition with the PIN. construct, as is Na,K-ATPase (cf. Figure However, unlike Na,K-ATPase, VSV-G trafficking is not restored by the inclusion of the ankyrin binding domain of spectrin (repeats 14-15) in the construct. Thus, it is possible to selectively block VSV-G transport without affecting Na,K-ATPase transport.

Figure 8: Illustrates the identification of a small peptide sequence responsible for association of Na,K-ATPase with ankyrin. Schematic representation of the five cytoplasmic domains of a-Na,K-ATPase and their relationship to the ankyrin binding peptide sequences identified here. Codon positions defining each peptide are shown. Previous studies have established broad reactivity of cytoplasmic domains II and III with ankyrin, with domain II contributing most of the binding activity (Devaraj an, et al.,1994; Jordan et al.,1995). Each depicted peptide (II-IIC) was prepared as a fusion WO 99/16875 PCT/US98/20364 -9construct with SjGST, and examined for its ability to bind at various concentrations I either purified ANK1 (from human red cells) or kidney ankyrin (ANK3) derived from whole MDCK cell lysates. Results from a single experiment are shown. The top panel shows Coomassie blue stained SDS-PAGE analysis of each peptide, as well as the entire MDCK cells extract applied to the affinity column to detect ANK3 binding. To detect ANK1 binding, purified erythrocyte ankyrin was used. Peptide IIA, sequence -SYYQEAKSSKIMESFKNMVPQQALV-, represents the minimal active sequence detected. We term this the minimal ankyrin binding domain (MAB).

Figure 9: Illustrates that the specific deletion of a SAATS attachment sequence in an integral membrane protein selectively blocks its transport alone. In this case, MDCK cells were transiently transfected with either wild-type Na,K-ATPase to which a FLAG epitope tag had been added to its NH2-terminus, or by a similarly FLAG-tagged mutant Na,K-ATPase in which codons 142 to 166 had been deleted. These residues correspond to the minimal ankyrin binding domain (MAB) identified in Figure 8. In both transfections, the cells were also co-transfected with wt P-Na,K-ATPase, so as to assure sufficient P-Na,K-ATPase to pair with the oc-Na,K-ATPase. (top) Flag tagged wt-Na,K-ATPase is delivered normally to the plasma membrane. (bottom) Flag tagged Na,K-ATPase lacking MAB does not assemble at the plasma membrane, and eventually is targeted for degradation in lysosomes. In these cells, the mutant Na,K-ATPase is the only protein whose transport is disrupted.

Figure 10: Illustrates selective blockage of Na,K-ATPase transport in normal cells by the expression of a small peptide inhibitor. Wild-type MDCK cells were transiently transfected with green fluorescent protein linked to the 25 residue MAB peptide. This is the same sequence whose deletion caused Na,K-ATPase to be retained in the ER and ultimately lysosomes in Figure 9. After approximately 1-2 days in culture, cells were fixed and stained for GFP (using anti-GFP antibodies, left panel) or for WO 99/16875 PCT/US98/20364 Na,K-ATPase (right panel, red). Note that the expression of GFP alone in MDCK cells (wt) did not affect Na,K-ATPase distribution, or cell size (which is dependent on Na,K-ATPase function. However, when the GFP carried the 25 residue MAB peptide, the binding of the endogenous Na,K-ATPase to SAATS was inhibited, resulting in its selective accumulation in the ER and reduced levels on the plasma membrane (two examples are shown, labeled Cells containing GFP-MAB are also markedly swollen, as expected. No other protein distributions were affected in these cells.

Figure 11: Illustrating the three-dimensional structure of MAB, which is well suited for the rational design of small molecule agents with similar pharmacologic action as MAB. This structure was determined using carrier mediated crystallization based on the active GST-MAB peptide shown in Figure 8 (Zhang et al.,1997a).

Three-axis views of the structure of the minimal ankyrin binding domain of a-Na,K-ATPase. The basic structural motif is that of a seven residue "loop" on a "stalk" composed of antiparallel P-strands. Surface accessibility depictions showing the amphipathic surfaces formed by the "loop". Ribbon diagrams demonstrating the back-bone contour. It is envisioned that the seven residue loop interacts with the P-hairpin tips of one or more ankyrin repeat units. The arrow in marks the depth of the structural plane shown in (E and It is anticipated that small molecule congeners mimicking the 7 residue loop structure will prove to be effective and selective inhibitors of Na,K-ATPase trafficking.

Figure 12: Illustrating another integral membrane protein of significant medical interest that is linked to SAATS by a different ankyrin interaction. (A) Both CFTR (cystic fibrosis transmembrane conductance regulator) and its close homologue TNR (Transmembrane-Nucleotide binding -Regulatory domain) protein are present in the soluble (Fxl) and cytoskeletal pools (Fx2) of confluent MDCK cells.

When these cells are surface labeled with biotin from either the apical surface (lane A) or WO 99/16875 PCT/US98/20364 -11the basolateral surface (lane it is also apparent that both CFTR and TNR are highly polarized, and unlike Na,K-ATPase, are expressed predominately in the apical domain of these polarized epithelial cells. Despite the fact that CFTR and TNR are apical proteins, they nevertheless binds components of the SAATS system, which is responsible for their trafficking. Either whole cell lysates of cells extracted with RIPA buffer, or Fx 1 and Fx 2, were immunoprecipitated with irrelevant antibody control or with anti-CFTR antibody The washed immunoprecipitates were then analyzed by SDS-PAGE and examined by western blotting for either ankyrin. Note that both ankyrins (AnkGl 19 and Ank R) co-immunoprecipitate with CFTR, in a fashion exactly analogous to the way Na,K-ATPase binds SAATS. However, CFTR and Na,K-ATPase bind at distinct sites on SAATS, enabling their selective modulation.

Figure 13: Illustrating both down and up-regulation of integrin display in cultured endothelial cells by SAATS. SV40 transformed murine endothelial cells were transfected with PIN, or PIN-s,5, and the surface display of three different integrins was measured by flow cytometry. Note the blockage of alpha V, beta 3, and beta 1 integrins by the pINs peptide. Conversely, the pIN-5,5 peptide markedly enhances the surface display of alpha V and beta 1 integrin (but not beta 3 integrin). This enhanced display (as opposed to blockage) is presumably achieved by enhancing the efficiency of SAATS directed cargo loading of vesicular transport by choosing peptides or agents that enhance the binding of a given membrane protein to SAATS Figure 14: Illustrating additional examples of SAATS regulation of integrin display in cultured endothelial cells. Experiment was carried out as in Figure 13. Note that alpha-V integrin is also modulated by SAATS. Also note that in this example, PIN-2,15 was used to rescue beta-1 expression. PIN.2 is an even more broadly blocking inhibitor of SAATS trafficking (see Figure but in this instance, supranormal levels of beta-1 can be achieved by inclusion of repeats 14,15 into the pIN2 peptide.

WO 99/16875 PCT/US98/20364 -12- Figure 15: Illustration of PECAM (CD31) surface modulation by SAATS inhibitory peptides employing two different epitope tags, FLAG and GFP. Cells were transfected as above, and surface display of PECAM, a cell-cell adhesion molecule of the IgG superfamily, was monitored by flow cytometry. Note that for either construct, inclusion of the 1IN-5 sequences indicated strong blockage of transport, indicating their involvement with SAATS. In both cases, surface display was rescued or accentuated by the plIN5,15 peptide. This effect was a bit more pronounced with the N-terminal FLAG tagged construct vs. the COOH-terminal GFP tagged construct, presumably due to interference of the GFP with the COOH-terminal 15th repeat unit needed to restore SAATS binding activity to PECAM.

Figure 16: Illustrates the modulation of CD45 and TNFR-1 in T-lymphocytes by SAATS inhibitors. In this experiment, Jurkat T-lymphocytes were transfected with the constructs indicated, including p 1 4 15 alone. Note the marked down regulation of a documented ankyrin binding protein, by constructs lacking repeat 15, as well as by the pji4.15 peptide itself (which lacks the constitutive Golgi targeting signal, see Fig.

These experiments also illustrate that TNFR-1 display is upregulated (rather than blocked) by PIN.2 or 1IN_5, suggesting that like E-cadherin, its attachment to SAATS is mediated either directly or indirectly by sequences contained within PIN.2.

Figure 17: Illustrates the modulation of Fas and Fas-L in T-lymphocytes by SAATS. Experiments were as before. Note the changes in both Fas and Fas-L due to and P'N-5,15 peptides.

WO 99/16875 PCT/US98/20364 -13- Figure 18: Illustrates the use of antibody or injected small peptides to Smodulate Na,K-ATPase in wild-type MDCK cells. Clusters ofwild-typeMDCK cells were micro injected with either GST alone; or (B).with the PIN4 peptide generated as a recombinant fusion peptide with GST, and then stained for Na,K-ATPase by indirect immunofluorescence. While micro injected GST is without effect, note the loss of Na,K-ATPase staining intensity in the injected cells on the left side of The consequences of impaired Na,K-ATPase delivery to the membrane are also apparent in the larger size of cells with insufficient Na,K-ATPase, similar to the changes seen when is transfected. These results indicate that selective blockage of Na,K-ATPase can be achieved by exogenous small peptides. Similar experiments were also carried out using various monoclonal antibodies. In these experiments, the cells were subconfluent, and all cells in an isolated cluster were injected. Control injections of buffer alone are without effect on the distribution of Na,K-ATPase. Microinjection of Mab IID2, which reacts exclusively with al spectrin (not considered to be a component of SAATS), is also without effect on Na,K-ATPase. Microinjection ofMab VIIIC7, which reacts with PI spectrin (Harris et al.,1986), leads to severe disruption of Na,K-ATPase delivery, with extensive accumulation within the cells and cell swelling. In this micrograph, the loss of membrane Na,K-ATPase is so severe, that the membrane borders cannot be discerned. This experiment illustrates the effectiveness of SAATS specific monoclonal antibodies in blocking SAATS activity.

Figure 19: Presents the full-length cDNA sequence of a novel isoform of PIII spectrin that may be an additional component of SAATS. A partial cDNA clone identified as an expressed sequence (EST) was identified as a spectrin family member, extended to complete the sequence by 5'RACE PCR amplification, and sequenced. (B) The resulting full-length sequence is predicted to encode a novel isoform of spectrin, termed PII (Morrow,1997), that is similar to both PI and PII, but shows significant differences in selected regions. Sequence similarity between the various WO 99/16875 PCT/US98/20364 -14beta-spectrins. Also shown is a phylogenetic dendritogram showing the approximate relationship of PIII spectrin to PI and PII spectrins.

Figure 20: Illustrative examples of SAATS inhibitory peptide constructs, in two different vectors: pcDNA3 (FLAG tagged) and enhanced green fluorescent protein (eGFP).

Figure 21: Figure 4B represents a model of how Na,K-ATPases may interact with one or more ankyrin repeat units.

Figure 22: Modulation of surface display of the Cystic Fibrosis Transmembrane Conductance Regulator (CFTR) by SAATS. MDCK cells were transfected with either the PIN5 or PINs,s constructs and then surface labeled with biotin.

After labeling, cells were solubilized, and the total CFTR in each culture immunoprecipitated with anti-CFTR antibodies. The precipitates were then analyzed by SDS-PAGE, transferred to nitrocellulose, and surface CFTR detected by avidin labeling.

Total CFTR was also determined in these cellular extracts by western blotting (data not shown). In all cell lines, there was minimal change in the total CFTR in the cells.

However, there was marked diminution of surface display of CFTR induced by the SAATS blocker PIN-5; whereas the PIN.515 construct markedly enhanced the surface display of CFTR. Collectively these results document the involvement of SAATS in CFTR trafficking, and suggest a possible therapeutic approach to the control of disorders such as cystic fibrosis, or the control of related proteins such as the multi-drug resistance gene product (MDR).

Figure 23: Schematic representation of the Spectrin Heterodimer.

WO 99/16875 PCT/US98/20364 Modes of Carrying Out the Invention The following discussion presents a general description of the SAATS system and discusses ways in which this system can be modulated for general classes of integral membrane and secretory proteins as well as for specific proteins in order to affect cell membrane display and secretion.

The generality in terms of which classes of membrane proteins utilize this pathway remains to be fully determined. However, given the guidance presented herein, a skilled artisan will know how to determine which proteins participate in the SAATS and analogous intracellular transport systems, how to identify the vesicular skeletal proteins and associated adapter proteins that mediate protein transport from the ER to and through the Golgi apparatus, how to identify small molecule and other agents that mediate such transport, and how to utilize such agents for therapeutic purposes.

Table IA: Summary of mammalian spectrins gene Isoforms chromosome GenBank Comment al Sl lq21 M61877 No splice forms identified. Found in red cells brain all S1 to S8 9q34.1 U83867 (SI) Found in most tissues; three regions of (see ref. (Cianci et J05243 (S2) alternative splicing, some tissue al.,1997)) specific. S6-S8 believed to exist but not proven (Cianci, et al.,1997).

PI S S2 Possibly 14q23-q24 J05500 (SI) Predominant in red cells, muscle. S2 S3 M37884 (S2) isoform adds MAD2/PH domain PIP, binding. Putative PI,3 may be Golgi form, but this is unproven.

PII SI S2 2p21 M96803 (SI) Found in most tissues. 1PIS2 partially characterized(Lombardo et al.,1994b).

Other splice forms may exist, but not well documented.

III S1 AB002300 Partially identified by EST match. Full Ssequence shown in Figure 19 WO 99/16875 PCTIUS98/20364 -16- Table III: Interactions involving spectrin Membrane Interactiong with spertrin band 3 (anion exchangers) linker /cofactor ankyrin tissue RBC, kidney, brain Na/K-ATPase ankyrin kidney amiloride-sensitive Na' channel voltage-gated Na' channel ABGP 205 inositol triphosphate receptor CD44 (gp85) ankyrin ankyrin ankyrin ankyrin ankyrin ankyrin ankyrin kidney brain brain brain lymphoma endothelium brain CD44-like I l6kD protein CAMS related to LI/neurofascin selcted referencs (Bennett and Brantonjl 977; Bennett and Stenbuckj1979b; Bennett and Stenbuck, I 979a; Morgans and Kopito,1993) (Nelson and Veshnock, 1987; Koob el al., 1988; Morrow, et al., 1989; Nelson and Hamnmerton,1989) (Smith el at.,1991) (Srinivasan et aI.,1988; Srinivasan et al.,1992) (Luna and Hitt,1992) (Joseph and Samanta, 1993) (Lokeshwar and Bourguignon,1I992a; Lokeshwar et al., 1994) (Bourguignon et al., 1992) (Davis and Bennett, 1993; Davis ei a.,I 1993) (Smith et al., 1993) (Hemming et al.,l 1995) (Bourguignon et al.,1986) (Lombardo et al., I 994a) (Sinard et aI.,1994) (Pitcher el aLl.,992; Touhara etal., 1994) (Bourguignon ei a.,l 1985; Lokeshwar and Bourguignon,l 992b) (Molday et al., 1990) (Cianci et ,1995) (Rotin el al., 1994) (Pollerberg et al., 1987) (Mombers et al., 1977; Mombers et a.,l 1979; Pradhan el al.,l 199 1; Diakowski and Sikorski,l 994) (Andrews and Fox, 1992; Lopez el al., 1992) (Rimm el al., 1995) H/K-ATPasc glycophorin C Th1y-i cadherin stomnatin bg subunits of trimeric G-proteins C D45 (gp 180) (tyrosine phosphatase) cGMP-gated cation channel dynamin (GTPase) epithelia] Na+ Channel N-CAM 180 Phospholipid ankyrin protein 4.1 protein 4.1 ce(E)-catenin adducin direct direct gastric cells erythrocytes lymphocytes kidney RBC and others lymphocytes retina (rods) MEL cells,brain MDCK cells brain RBC et al direct direct direct direct direct GP Ib-IX to actin cadherin to actin ABP-280 platelets MDCK cells a(E)-catenin tCOKPC2 *InerTIlnS WIMf Spea~rin SUBSTTUTE SHEET (RULE 26) WO 99/16875 WO 9916875PCTIUS98/20364 adducin protein 4.1 intermediate filaments direct tubulin direct D plectin synapsin (kher Interactns 1 0 synapsin to actin tropomyosin to actin dematin (protein 4.9) to actin protein 4.1 and glycophorin to p55 ankyrin to vimentin spectrin to IP3 direct direct direct direct direct actin protein 4.1 ankyrin direct

RBC.

RIBC

brain, RI3C (avjan) brain gliomna cells brain brain (axons) brain RB3C, muscle

RBC

RBC

RBC (avian) brain, etc (Gardner and Bennett, 1987; Mische ei aL,1987) (Cohen and Foley, 1980; Cohen and Korsgren, 1980; Cohen and Foley, 1982; Wolfe el al., 1982) (Mangeat and Burridge, 1984; Langley and Cohen,1986; Frappicr el al.,1987; Herrmann and Wiche,1987; Frappier el al., 1992) (Fach et al., 1985; Riederer and Goodman,1 990) (Henrmann and Wiche,1987) (Sikorski et 99 1) (Hayes ei al., 1995) (Bahier and Greengard, 1987; Petrucci and Morrow, 1987) (Fowler and Bennett, 1984; Fowler el al.,1993) (Siegel and Branton,1985; Husain-Chishti el al., 1988) (M arfatia et al., 1994; Hemming, et td.,1995; M arfati a et al., 1995) (Georgatos and Marchesi,1985; Georgatos et nI.,1985) (Hyvonen et al., 1995; Wang and Shaw, 1995) The practice of the present invention thus generally involves the modulation of specific interactions between SAATS and the recognized cargo or carrying proteins of the ER and Golgi. Also the elucidation of particular binding sites (which are often complex and involve homologues of spectrin adapter proteins) is described in order to identify targets for blocking specific interactions. In some instances, the SAATS system can discriminate between H,K-ATPase, Na,K-ATPase, CFTR and other transport molecules, or can be modulated at the level of binding domains common to various classes of 25 proteins.

SUBSTITUTE SHEET (RULE 26) WO 99/16875 PCT/US98/20364 -18- For example, identifying the binding domain for CFTR, the gene product responsible for cystic fibrosis, in SAATS by techniques disclosed herein, would allow early analysis of possible transport enhancing agents that might ameliorate the clinical severity of this disorder by enhancing the delivery of mutant (but functional) CFTR to the plasma membrane. In Figures 12 and 22 the binding of CFTR to ankyrin components of SAATS is illustrated as well as its enhanced delivery to the membrane by the psINs5 construct.

An important discovery that remains under investigation is the actual gene structure of the spectrin(s) that participate in SAATS. Without intending to be bound by the identification of any particular spectrin species, we believe that PI spectrin or homologues of PI spectrin such as P3II spectrin are participants in the SAATS process because they share extensive immunologic and functional similarity to the spectrin of SAATS, and because peptides based on PI spectrin (Winkelmann et al.,1990) are functionally active as discussed below. A discussion of a particular Golgi-associated spectrum called plI*, so designated because it is still a somewhat uncharacterized putative splice form of PI spectrin, is found in our earlier publication, Devarajan et al., JCB 133:819-830 (1996) and in Beck et al. (1994).

The participation of various adapter proteins, other than ankyrin, such as protein 4.1 homologues, adducin homologues, catenin homologues, and others are contemplated.

The skilled artisan can explore the role of such adapter proteins by techniques disclosed herein in order to determine their participation in the modulation of the interaction of SAATS with proteins trafficking through the ER and Golgi. The action of similar pathways acting on the delivery of proteins from the medial-Golgi and trans-Golgi to the plasma membrane or other internal membrane-bound compartments is also likely, but remains to be more fully explored with techniques disclosed herein.

The methods outlined here offer a novel approach to controlling the activity of specific membrane proteins, be they receptors, adhesion receptors, ion channels, or transporters. While most conventional therapeutic approaches target the activity of a given protein, the approach outlined herein targets the very appearance of a protein at WO 99/16875 PCT/US98/20364 -19the cell surface or in a secretory vesicle. This strategy is somewhat similar in concept to approaches that attempt to specifically suppress the synthesis of a given protein, such as by anti-sense RNA or by specific transcriptional regulators. However, the methods outlined herein are fundamentally different in that they: 1) do not attempt to suppress the synthesis of a given protein, only its delivery to the correct cellular or tissue compartment; 2) do not require intracellular expression of RNA, obviating many hurdles inherent in such a task; and 3) lend themselves to high-throughput in vitro screening assays, and should be amenable to regulation by small molecule effectors. As such, the methods outlined here offer a novel therapeutic approach that is complementary to and synergistic with other leading drug development strategies.

As used herein, the following terms have the meanings ascribed: a. As used herein, Transport vesicle means any small membrane bounded structure containing protein and involved in the movement of such protein from one membrane compartment to another.

b. As used herein, Cell membrane means a lipid and protein containing bilayer structure bounding any cellular compartment including the surface of the cell, the so called plasma membrane.

c. As used herein, Adapterprotein means any protein that provides for a direct or indirect linkage to the spectrin backbone of SAATS or to any other multifunctional proteins that directly or indirectly determine the specificity of cargo or cargo protein carrier capture by SAATS.

d. As used herein, an agent is said to be randomly selected when the agent is chosen randomly without considering the specific sequences of the vesicular skeletal protein and associated adapter protein. An example of randomly selected agents is the use a chemical library or a peptide combinatorial library, or a growth broth of an organism or plant extract.

e. As used herein, an agent is said to be rationally selected or rationally designed WO 99/16875 PCT/US98/20364 when the agent is chosen on a nonrandom basis that takes into account the sequence of the target site and/or its conformation in connection with the agent's action. Agents can be rationally selected or rationally designed by utilizing the peptide sequences that make up the relevant vesicular skeletal protein or associated adapter protein. For example, a rationally selected peptide agent can be a peptide whose amino acid sequence is identical to a fragment or relevant domain of a spectrin protein.

f. As used herein, "cell" absent any further designation, means any eukaryotic cell population, including primary and transformed immortalized cell lines, that express spectrin, or an equivalent protein. Madin Darby canine kidney (MDCK) cells can be employed in the claimed methods. Transfection of some cell lines with broadly active SAATS inhibitors, such as the PIN- or PIN2 constructs, may lead to poor cell survival and the activation of apoptosis or lethal swelling. This problem by be often circumvented by selecting less broadly active SAATS agents (eg. larger constructs, or constructs in which only small portions of spectrin, ankyrin, or one of the adapter proteins has been deleted or modified). This strategy of scanning mutagenesis will minimize the global disruption cellular transport that accompanies the most severe disruptors of SAATS, and improve cell viability so that determinations can be made on less hardy cell lines. This technique also has the advantage of more precisely pinpointing the locus of SAATS activity for a given protein. Transfection levels may also be modified or controlled to manage poor cell survival.

If the induction of apoptosis becomes a problem with some broadly active SAATS constructs, improvement in the survivability of the cell based assays may be accomplished by the inclusion of available apoptosis inhibitors in the cell medium, such as caspase inhibitors or Fas blocking antibodies (see U.S. Patents 5,632,994, 5,656,725, 5650,491 and 5,635, 187). Cells may also be engineered to express increased levels of bcl-2 (see U.S. Patent 5,650,491).

g. As used herein, "stringent conditions" refers to conditions those commonly defined and available, such as those defined by Sambrook et al (Molecular Cloning: A WO 99/16875 PCT/US98/20364 -21- Laboratory Approach, Cold Spring Harbor Press, NY, 1989) or Ausubel et al. (Current Protocols in Molecular Biology, Greene Publishing Co., NY, 1995).

Hybridization is a function of sequence identity (homology), G+C content of the sequence, buffer salt content, sequence length and duplex melt temperature among other variables. See, Maniatis et al. Molecular Cloning, Cold Spring Harbor Laboratory, Cold Spring Harbor, N.Y. 1982). With similar sequence lengths, the buffer salt concentration and temperature provide useful variables for assessing sequence identity (homology) by hybridization techniques. For example, where there is at least percent homology, hybridization is commonly carried out at 680 C. in a buffer salt such as 6XSCC diluted from 20XSSC. See Maniatis et al. The buffer salt utilized for final Southern blot washes can be used at a low concentration, 0.1XSSC and at a relatively high temperature, e.g. 680 and two sequences will form a hybrid duplex (hybridize). Use of the above hybridization and washing conditions together are defined as conditions of high stringency or highly stringent conditions. Moderately high stringency conditions can be utilized for hybridization where two sequences share at least about 80 percent homology. Here, hybridization is carried out using 6XSSC at a temperature of about 50-551 C. A final wash salt concentration of about 1-3XSSC and at a temperature of about 60-68' C are used. These hybridization and washing conditions define moderately high stringency conditions.

In particular, specific hybridization refers to conditions in which a high degree of complementarity exists between a nucleic acid comprising the sequence of at least one of the Figures and another nucleic acid. With selective hybridization, complementarity will generally be at least about 75%, 80%, 85%, preferably about 90-100%, or most preferably about 95-100%.

I. General Description The present invention relates to a method for selectively controlling the transport of integral membrane and secretory proteins by controlling their sequestration into transport WO 99/16875 PCT/US98/20364 -22vesicles that mediate protein trafficking between the ER, Golgi, plasma membrane, and other membrane compartments. The present patent application expands on the discovery that specific isoforms of spectrin and various adapter proteins, such as ankyrin, associate with vesicle and Golgi membranes by multiple interactions (Devarajan et al., 1994; Devarajan and Morrow, 1995; Devarajan and Morrow, 1996; Devarajan et al.,1996a; Devarajan et al.,1996b; Devarajan et al.,1997; Godi et al.,1997). These interactions are necessary for the transport of broad classes (and possibly all) membrane and secretory proteins from the ER through the intermediate compartment to the medial-Golgi (Devarajan, et al.,1996a; Devarajan, et al.,1997; Godi, et al.,1997). Moreover, by modifying a specific interaction site within spectrin or in one of its associated adapter proteins, the transport and ultimately the rate of delivery of a specific protein to its final membrane target can be significantly altered or even blocked.

We call this newly discovered system mediating transport the spectrin-ankyrin-adapter protein trafficking system (SAATS). Since all membrane and secretory proteins transit this pathway, and must be selectively and sequentially collected into different transport vesicles during their journey from the ER to the plasma membrane (or to other membrane compartments), this system has wide potential applicability and generality. We contemplate that the great diversity of proteins originating in the ER is accommodated by SAATS by a combinatorial process involving many weak but specific interactions created by the polyvalent binding capacity of spectrin and ankyrin and other adapter proteins, as well as by the pairing of potentially non-binding proteins in oligomeric complexes with binding proteins. The control of secretory protein trafficking, via recognized cargo loading proteins resident in the ER, we envision also to .be mediated by this pathway, in which the cargo loading membrane proteins bind to the spectrin/adapter protein system.

Thus, by selectively controlling the binding of a specific membrane protein, or the binding of a specific set of cargo proteins to SAATS, the delivery of that specific protein into transport vesicles will be modified (blocked or enhanced), as will the WO 99/16875 PCT/US98/20364 -23delivery of that protein (but not other proteins) to the plasma membrane (or other targeted membrane compartment). This SAATS system and its functions are illustrated in Fig. 1.

We also believe that similar but distinct SAATS systems operate to mediate protein transport between the ER and Golgi, as well from the Golgi to the plasma membrane or other target post-Golgi). Therefore, the methods described here apply as well to the process of targeted membrane delivery post Golgi. An example of such a process would be spectrin linked microtubule mediated axonal transport of specific neuronal receptors destined for the pre or post-synaptic membranes.

In general, several methods are available for controlling the binding of selective proteins (or lipids) to SAATS using methods and technologies known in the biomedical field but heretofore applied for other purposes.

a) Using recombinant cDNA methodology, for example, the expression of PI spectrin related peptides and homologues from other spectrin related genes that preserve the MAD1 and actin binding motifs (Lombardo et al.,1994) of PI spectrin, but delete its ankyrin binding domain (which binds Na,K-ATPase) (Kennedy et al.,1991), selectively blocks the delivery of Na,K-ATPase to the plasma membrane (Devarajan, et al.,1996a; Devarajan, et al.,1997). Because other proteins participate in the SAATS transport processed, this opens the possibility of gene therapeutic approaches to modulating SAATS function in order to introduce genes that encode modified spectrin molecules that either inhibit or enhance the sequestration of selected membrane and secretory proteins into transport vesicles and, ultimately, their display on or secretion through the plasma membrane.

b) Certain polyclonal and monoclonal antibodies to spectrin block the transport of Na,K-ATPase (Figure 18) or VSV-G protein (Harris et al.,1986; Devarajan, et al.,1996a; Godi, et al.,1997). While of limited therapeutic interest since such agents generally act extracellularly, we envision such agents as important tools for use in in vitro transport assays and in immunomicroscopy.

c) Inhibitors of ADP-ribosylation factor (ARF) block the binding of the WO 99/16875 PCTIUS98/20364 -24- PH domain of PI spectrin via a PtdInsP 2 -dependent process, implying that SAATS may also contribute to the organization and trafficking of specific phospholipids. Agents modulating PtdInsP2 phospholipid display, or ARF activity directly, may also thus form the basis for a class of therapeutics. An examples of such an agent would be a permeant non-hydrolyzable homologue of GTP, such as a permeant GTP-y-S.

d) We have prepared recombinant peptides representing specific functional domains of pI spectrin (Kennedy, et al.,1991; Lombardo, et al.,1994; Weed et al., 1996) that also block transport (see figures, also Godi, et al.,1997) of selected integral membrane and secretory proteins. These demonstrate the potential activity of small peptide or peptide mimetic drugs.

e) Many of the interactions involving SAATS, based on our understanding of the spectrin skeleton that underlies the plasma membrane (for review, see Morrow et al.,1997), are likely to be subject to post-translational regulation by pathways involving protein kinase C, Ca", calmodulin, casein kinase, etc. Agents that therefore modulate these regulatory pathways, while less specific, may also form the basis of potential SAATS oriented therapeutics.

f) We envision that small molecule modulators of SAATS specific interactions will form the basis for a broad and effective category of therapeutics. Such small molecule agents would either block or enhance the delivery of specific proteins to the plasma membrane or other sites. We have developed several in vitro assays that detect specific interactions between SAATS and other proteins, based on in vitro binding, surface plasmon resonance, and genetic screening (two-hybrid systems). Many of these in vitro assays are amenable to high-throughput screening, facilitating the identification W of lead compounds. We have also developed sophisticated algorithms that allow us to model the three-dimensional structure of certain spectrin repeat units, together with the variant sequences that mediate specific binding interactions (Cianci and Morrow, 1997; Gallagher et al.,1997; Stabach et al.,1997). We envision that the availability of these sophisticated 3-D protein models will facilitate rational drug design initiatives.

WO 99/16875 PCT/US98/20364 The specific examples presented below are illustrative only and are not intended to limit the scope of the invention.

II. Specific Embodiments The following discussion describes specific embodiments of our invention and provides procedures for: identifying integral membrane and secretory proteins whose vesicular transport to or through the Golgi is mediated by the SAATS or analogous systems; identifying which vesicular skeletal protein, spectrin, or associated adapter proteins, ankyrin, are specifically involved in mediating sequestration of particular integral membrane and secretory proteins into a transport vesicle; identifying peptides and small molecules that modulate sequestration and/or vesicle transport of such proteins; and utilizing agents that modulate the sequestration and/or vesicular transport of such proteins for therapeutic and research purposes.

EXAMPLES

Example 1: Identification of integral membrane and secretory proteins whose vesicular transport is mediated by SAATS The identification of broad inhibitors of SAATS may generally be approached by comparing the ability of a series of overlapping and complimentary spectrin peptides to disrupt the transport in cultured cell systems of any particular integral membrane or secretory protein. This information allows identification of a minimal region within spectrin that is responsible for the cryptic sorting and docking of the targeted protein to the SAATS system. Other potential blocking agents also follow from knowledge of adapter proteins that bind to this region of spectrin, as discussed below. The following steps are preferred.

WO 99/16875 PCT/US98/20364 -26- Step Al: Select a model cultured cell line that expresses the membrane or secretory protein that is being evaluated, and for which an assay exists that will allow the detection of surface or secreted target proteins. Examples would be immunofluorescent detection of surface protein; flow cytometry; surface labeling assays, or transport assays using pulsed 3 S methionine labeling. Assays that monitor the transport of a specific protein may also be performed by the use of GFP labeled proteins and time lapse vital flourescent microscopy. Also envisioned is the detection of a specific transport or electrical activity associated with the surface display or secretion of the targeted protein.

As a specific example, cultured Madin Darby Canine Kidney cells (MDCK) may be used to monitor the surface expression of Na,K-ATPase by indirect immunofluorescence (See Figs 3 through 10, 18).

Step A2: Prepare using standard transfection methods of the model cell line a series of individual clonal lines or lines transiently transfected or lines infected using retroviral vectors or other virus based infection strategies, each expressing one of a complimentary and overlapping series of PIZ* spectrin cDNA constructs of varying lengths under the control of a strong eukaryotic expression vector. A series of suitable PI* spectrin constructs is listed in Figure 2. For example, stable MDCK cell lines expressing the P'N- spectrin peptide (See Figs. 3 and others), which competitively displaces or substitutes for the endogenous SAATS system, as one appropriate line.

The transformation of appropriate cell hosts with an rDNA (recombinant DNA) molecule of the present invention is accomplished by well known methods that typically depend on the type of vector used and host system employed. With regard to transformation of prokaryotic host cells, electroporation and salt treatment methods are typically employed, see, for example, Cohen et al., Proc Acad Sci USA (1972) 69:2110; and Maniatis et al., Molecular Cloning. A Laboratory Manual, Cold Spring Harbor Laboratory, Cold Spring Harbor, NY (1982). With regard to transformation of vertebrate cells with vectors containing rDNAs, electroporation, cationic lipid or salt treatment WO 99/16875 PCT/US98/20364 -27methods are typically employed. See, for example, Graham et al., Virol (1973) 52:456; Wigler et al., Proc Natl Acad Sci USA (1979) 76:1373-76.

Successfully transformed cells, cells that contain an rDNA molecule of the present invention, can be identified by well known techniques. For example, cells resulting from the introduction of an rDNA of the present invention can be cloned to produce single colonies. Cells from those colonies can be harvested, lysed and their DNA content examined for the presence of the rDNA using a method such as that described by Southern, JMol Biol (1975) 98:503, or Berent et al., Biotech (1985) 3:208 or the proteins produced from the cell assayed via an immunological method. If tags such as green fluorescent protein are employed in the construction of the recombinant DNA, the transfected cells may also be detected in vivo by the fluorescence of such molecules by cell sorting.

Step A3: Evaluate each clonal line or cells expressing the desired construct (in assays in which transient expression is used, e.g. Figure 9) for disruption of the surface display or secretion of the selected protein. Expression of full length PIE* spectrin (or other full length spectrins) will not disrupt SAATS. Expression of at least one of the truncated constructs shown in Figure 2 or analogous constructs for other spectrins and other vesicular skeletal proteins will act in a dominant negative way to disrupt SAATS. A comparison of the inhibitory ability of different constructs will reveal the boundaries of an inhibitory peptide or cDNA construct encoding such inhibitory peptide. For example, inhibition of surface display ofNa,K-ATPase in MDCK cells by expression of the PI-5 spectrin construct (Fig. 3).

Step A4: Knowledge of the minimal inhibitory peptide can be further refined by repeating steps A2 and A3, above, using PIE* spectrin constructs or similar constructs prepared from other spectrins in which specific regions deleted in the inhibitory construct are returned to the inhibitory construct, thereby identifying the WO 99/16875 PCT/US98/20364 -28regions required to rescue surface display or secretion of the selected (targeted) protein.

By this approach, minimal sequences required for docking the targeted protein to SAATS can be identified.

For example, the surface display of Na,K-ATPase in MDCK cells, which is inhibited by PIN, is rescued by attaching pi spectrin repeats 14 and 15 to the PIN, construct (See Fig. 6 and others). These results indicate that PI spectrins, 14-15 repeat region is the site of docking (either directly or indirectly) of Na,K-ATPase to SAATS.

Step A5: A further refinement of the SAATS docking site for the targeted protein may be obtained by determining whether it attaches directly to spectrin at the site identified in Steps A1-A4, or whether an intermediary adapter protein is utilized. In most instances, this can be readily determined by in vitro binding or co-immunoprecipitation assays (eg Figure in which the interaction of the targeted protein with spectrin and/or adapter proteins (such as ankyrin, protein 4.1, adducin, a-catenin, see TABLE IB) is evaluated. Such methods are well known in the art as exemplified by Harlow et al., 1988.

For example, the interaction of Na,K-ATPase with the 14-15 repeat unit of spectrin is via an ankyrin G1 19 adapter protein intermediate, as detected by coimmunoprecipitation (Figure and by direct binding assays indicating that ankyrin binds to this region of spectrin (Kennedy et al.,1991) and that Na,K-ATPase binds to ankyrin (Morrow et al.,1989) and at a specific locus within Na,K-ATPase (Devarajan, et al.,1994).

Preparation of Spectrin and associated cDNA constructs Recombinant peptides are prepared as fusion proteins with glutathione-S-transferase, using the prokaryotic expression vector pGEX (Smith and Johnson,1988; Kennedy et al.,1991). Spectrin clones used for these are identical to sequences published for PII spectrin (Hu et al.,1992; Chang et al.,1993; Chang and WO 99/16875 PCT/US98/20364 -29- Forget, unpublished), PI spectrin (Winkelmann et al.,1988; Winkelmann et al.,1990b; Winardi et al., 1993), and PIII spectrin as shown in figure 19. In general, the desired segment of cDNA is excised by EcoR1 or other suitable restriction enzymes from pGEM purified from agarose gels by electroelution and subcloned into the appropriate pGEX vector, usually pGEX-2T. In the examples shown (Figure 2 and (Kennedy, et al., 1991; Kennedy et al., 1994; Lombardo, et al.,1994b)) the polymerase chain reaction was typically used in conjunction with Pfu polymerase and selected primers to amplify regions of the spectrin clones to facilitate plasmid construction and in some cases linker-insertion mutagenesis (Sambrook et al.,1989; Kennedy, et al.,1991). Occasionally Taq polymerase was used to facilitate cloning of PCR products into the pCRII vector of the TA cloning kit (Invitrogen). The validity of all PCR generated constructs is confirmed by DNA sequencing. In general, it is useful but not required if PCR primers are engineered to have BamHI sites 5' and EcoRI sites 3' to ease directional subcloning into pGEX-2T. Two to five additional bases were usually added at the ends of each primer to enhance their susceptibility to restriction endonucleases. The amino acid boundaries for the amplification of the spectrin structural repeats are selected based upon the phasing of Drosophila a-spectrin (Winograd et al.,1991). Peptides representing the unique alternatively spliced COOH-terminal region of the PI spectrin found in brain and muscle are prepared from plasmid pMSP6AC (Winkelmann, 1990) by digestion with Xma I. The resulting 767 bp fragment containing the alternatively spliced unique C-terminus of spectrin 3IIII is gel purified and subcloned into pGEX-2T at the Sma I site. Other typical details of primer selection and plasmid construction are summarized as in Kennedy, et al., 1991; Lombardo, et al.,1994b. The various PI spectrin peptides are expressed as above from existing human PI spectrin constructs (Winkelmann et al.,1990a; Winkelmann, et al.,1990b; Weed et al.,1996). Constructs for eukaryotic cell expression are typically prepared in the pcDNA3- vector (Invitrogen) or other comparable vector incorporating a suitable eukaryotic promoter. With the pcDNA3 vector, stable MDCK cell lines can be established by selection with G418 after WO 99/16875 PCT/US98/20364 lipofectamine transfection (Weed, et al.,1996). Transient experiments can also be carried out in this system or in several other systems well suited for transient assays, such as COS cells. For work with VSV-G, newly confluent cells are infected for 60 min. with VSV-G in Transwell® filters (100 ml, 108 titer virus Pimplikar et al.,1994). Infected cells are incubated for 3 hours and fixed prior to analysis.

Protein expression and purification Recombinant proteins are expressed in E. Coli strains HB101, DH5a or the protease deficient strain CAG-456 and purified as before (Smith and Johnson,1988; Kennedy, et al., 1991). When required, peptides were further purified by gel filtration on a 1 x 100 cm column of Sephacryl S-200 HR in the presence of 1 M NaBr, 10 mM NaHPO,, 0.5 mM EGTA, 0.5 mM EDTA, 15 mM sodium pyrophosphate, 1 mM NaN,, mM DTT, pH 8.0. Purified proteins were dialyzed into 20 mM Tris, 150 mM NaC1, 1 mM EDTA, 1 mM NaN,, 1 mM DTT, 0.2 mM PMSF, 25% glycerol, pH 7.5 and stored at -80 Bovine brain spectrin is prepared from fresh brain membranes by low ionic strength extraction of demyelinated brain membranes followed by gel filtration on Sephacryl-S-500 HR (Harris, et al.,1985; Bennett,1986) Antibodies and immunodetection procedures.

Constructs epitope labeled by the 8 aa FLAG marker (IBI Scientific) were detected with Mab M5 (IBI Scientific); PI spectrin antibodies included: Mab VIIIC7, Mab VD4, Mab IVC9, and Mab IVF8 (Harris, et al.,1986); Pab C19 against 19 COOH-terminal residues of PIl1 spectrin (gift, V. Marchesi); and Pab MUS1 against region III of PI2 spectrin (Malchiodi-Albedi et al.,1993). Pab 10D detected PII spectrin 25 (Devarajan, et al.,1996). Other antibodies were: Mab P-COP, I. Melman; Pab E-cadherin, Transduction labs; Mab c-Na,K-ATPase, UBI; Mab P-Na,K-ATPase, M. Caplan; Mab VSV-G, J. Rose; and Pab centractin, E. Holzbaur. Cells were grown on glass cover slips or culture dishes, washed x3 with PBS, fixed 15 min. in acetone, blocked 30 min. in goat WO 99/16875 PCT/US98/20364 -31serum, and incubated with 1° antibody in 2% BSA in PBS 10% goat serum at RT for min. All spectrin Mab's were used at 1:100; E-cadherin at 1:5000; b-COP, 10D, C19 at 1:200. Detection used 2° antibodies conjugated to CY3 or CY2 (Vector).

Immunoprecipitations.

Confluent MDCK cells were PBS washed, lysed 20 min. at 4°C by gentle rocking in IP buffer (10 mM Tris-HC1, pH 7; 150 mM NaCl; 5mM EDTA; ImM EGTA; 2 mg/ml BSA; 0.5% deoxycholate; 1% NP-40; 0.5 mM pefablock; ImM PMSF; and ImM leupeptin. The lysate (2 ml) was decanted and centrifuged 1 min. at 10,000xg, precleared by 60 min. incubation with 25 ml of non-immune rabbit serum with 200 ml of a 50% Protein A Sepharose (Pharmacia). The remaining supernatant was treated 4 hours at 4° with 200 ml of 1° antibody, pelleted by Protein A Sepharose, and analyzed.

Example 2: Identification of Na,K-ATPase as an integral membrane protein participating in SAATS In wild type MDCK cells, Golgi spectrin and Ank, 9 are localized with the Golgi, while a-Na,K-ATPase is predominately distributed at the plasma membrane where it is tethered via the plasma membrane form of ankyrin to aIlpII spectrin (Morrow, et al.,1989; Nelson and Hammerton,1989) (Figure 3).

To determine the intracellular distribution of Golgi spectrin, AnkG and c-Na,K-ATPase, cells were grown on glass cover slips or culture dishes, washed x3 with PBS, fixed 15 min. in acetone, blocked 30 min. in goat serum, and incubated with 1° antibody in 2% BSA in PBS 10% goat serum at RT for 60 min. The intracellular distribution of wt Golgi spectrin was monitored using Mab VIIIC7, which reacts with full length Golgi spectrin but not the shorter FLAG tagged PIN 5 construct (Harris et al.,1985; Devarajan, et al.,1996; Godi, et al.,1997). The distribution of pI* spectrin in cells transfected with pli-c was monitored by Mab VD4, which does not react with the PI, -c peptide. AnkG 19 was monitored by the antibody "Jasmin" (Devarajan, et al.,1996); WO 99/16875 PCT/US98/20364 -32a-Na,K-ATPase by a Mab from Upstate Biotechnology. Detection was by indirect immunofluorescence.

Electron microscopy was performed by fixing cell monolayers in situ for 1 hr (Kamovsky's fixative 18). Fixed monolayers were washed in 0.1 M Na-cacodylate, pH 7.4, and postfixed in 1% Os0 4 in 0.1M s-collidine. After dehydration in graded ethanol and washing with propylene oxide, samples were embedded in Epox-812 (Ernest Fullman, Inc. Ultrathin sections stained with aqueous uranyl acetate and lead citrate were viewed on a Zeiss EM-910 at 80 kV.

In Figure 3A, the results of the indirect-immunofluorescence labeling are presented. Note the dispersal of the endogenous Golgi spectrin, Ank;Gl and a-Na,K-ATPase by PIN5 (but not by All shortened spectrin constructs containing region I and MAD1 displayed a similar effect, whereas constructs lacking these regions of PI spectrin did not significantly alter the native distributions of PIZ* spectrin, AnkGl 19, or Na,K-ATPase. Bar The distribution of the endogenous Golgi spectrin-ankyrin skeleton, as well as membrane assembly of Na,K-ATPase, was disrupted by PI spectrin constructs that contained the region I/MAD 1 Golgi targeting signal. (Fig. 3A). In the presence of such peptides, endogenous Golgi PIS* spectrin and AnkG,, 9 were dispersed into coarse punctate cytosolic patches. (Figs. 3A 4A). Even more dramatic is the effect of these peptides on the distribution of Na,K-ATPase, which becomes widely dispersed throughout the cytoplasm in a pattern resembling the endoplasmic reticulum, with very little protein identifiable at the cell surface (Figs 3A, 4A). The p13, expressing cells are also swollen compared to control cells (mean diameter 29±6.7m vs. 17±3.7m), and grow more slowly presumably due to insufficient plasma membrane Na,K-ATPase (or other membrane proteins, see below). This effect on Na,K-ATPase distribution is not evident with constructs lacking region I/MAD1 Fig. 3A), nor is it apparent in cells expressing full length PI1 spectrin (that contains the ankyrin-Na,K-ATPase binding domain but lacks MAD2 and PtdInsP 2 -dependent Golgi binding).

WO 99/16875 PCT/US98/20364 -33- To determine the means of interaction between Na,K-ATPase and spectrin, coimmunoprecipitation studies are performed. Briefly, confluent MDCK cells were PBS washed and lysed 20 min. at 40C by gentle rocking in IP buffer (10 mM Tris-HCl, pH 7; 150 mM NaCl; 5mM EDTA; 1mM EGTA; 2 mg/ml BSA; 0.5% deoxycholate; 1% NP-40; 0.5 mM pefablock; ImM PMSF; and ImM leupeptin. The lysate (2 ml) was decanted and centrifuged 1 min. at 10,000xg, precleared by 60 min. incubation with ml of non-immune rabbit serum with 200 ml of a 50% Protein A Sepharose (Pharmacia).

The remaining supernatant is treated 4 hours at 4o with 200 ml of 1o antibody, pelleted by Protein A Sepharose, analyzed by Western blots after SDS-PAGE and visualized by ECL. The co-immunoprecipitation studies demonstrate that the interaction of Na,K- ATPase with the 14-15 repeat unit of spectrin occurs via an ankyrin Gl 19 adapter protein intermediate (see Figure 3B). In wt cells, PIE* spectrin, AnkGl 19, centractin, and Na,K-ATPase are co-precipitated from detergent lysates by Mab VIIIC7 (lane IP).

Soluble aIIII spectrin (as detected here by immunoblotting for all spectrin), the form found predominately at the plasma membrane, is also present in these detergent lysates (lane lys) but is not part of the precipitable Golgi spectrin complex. Control experiments with preimmune or irrelevant antibody did not precipitate any of these components.

Similarly, immunoprecipitates with Mab VIIIC7 of the pIN. transfected cells also demonstrated co-precipitation ofAnkG 19 and Na,K-ATPase with PIE* spectrin, indicating that despite the dispersal of these elements induced by the PIN5 construct, these elements remained associated in the detergent lysate.

As exemplified by Figure 3B, immunoprecipitation with Mab VIIIC7 reveals that a fraction of a-Na,K-ATPase exists in a complex with PIZ* spectrin, Ankl,, 9 and centractin (Fig. 3B). Centractin (ARP-1) is an actin related protein that is part of the dynactin complex. This pool of a-Na,K-ATPase associated with PIZ* spectrin, Ank,,,, and centractin appears to be distinct from the pool of plasma membrane a-Na,K-ATPase, since all spectrin was not present in these complexes (Fig. 3B), and presumably represents Na,K-ATPase not yet assembled at the plasma membrane.

WO 99/16875 PCT/US98/20364 -34- Na,K-ATPase is blocked at the ER to medial-Golgi transition by fllN., spectrin Normally, after synthesis and core glycosylation in the ER, p-Na,K-ATPase undergoes further glycosylation in the medial Golgi and then is vectorally incorporated as an a,P-Na,K-ATPase heterodimer into a detergent insoluble IIpII spectrin and ankyrin cortical skeletal lattice (Nelson and Veshnock,1986; Morrow, et al.,1989; Nelson and Hammerton,1989; Mays et al.,1995). The distribution of Na,K-ATPase in the pI-, expressing cells suggested that the exit of Na,K-ATPase from the ER might be blocked (Fig. 4A).

Western blotting confirmed that the spectrin PIN, polypeptide blocked Golgi mediated glycosylation of P-Na,K-ATPase (Fig. 4B). Control MDCK cells (wt) or cells transfected with the PIN construct (PIN.) were analyzed by Western blot using a P-Na,K-ATPase specific antibody (Mab B 1-13). The apparent MW of the various products are depicted. Note the broad band above 50 kDa (gly) representing mature glycosylated p-Na,K-ATPase, and the intense band at 44 kDa in the PIN- cells (core), representing the ER dependent core P-Na,K-ATPase glycosylation product. It thus appears that the PIN5 spectrin peptide inhibits mature glycosylation of P-Na,K-ATPase, a process characteristic of medial Golgi processing, but does not interfere with the formation of the core glycosylation product, which occurs in the ER. In this experiment, approximately half of the PIN, cells are expressing the PIN, construct, as judged by immunofluorescence which appears to account for most of the mature glycosylated product observed in the pINs cell lines. The band =106 kDa in the wild type cells is inconstant in its appearance, and may represents a cross-linked adduct of p-Na,K-ATPase with a-Na,K-ATPase or other protein, as is commonly observed after reduced SDS-PAGE in MDCK cells (Morrow et al.,1989).

While these experiments did not seek to quantitatively measure the degree of transport inhibition, the results of several separate determinations suggest that the efficiency of P-Na,K-ATPase transport inhibition by the pIN., peptide is quite high, since the level of processed P-Na,K-ATPase that can be detected correlates inversely with the WO 99/16875 PCT/US98/20364 fraction of cells in these cultures that actually express the recombinant peptide. Also blocked in these cells was the transition of a-Na,K-ATPase to detergent insolubility (Fig.

4C), indicating a failure of Na,K-ATPase to join a stable plasma membrane associated spectrin lattice. Shown are Western blots for a-Na,K-ATPase after SDS-PAGE. Fxl (soluble fraction) is that material solubilized by 0.5% Triton X-100 in 100 mM NaCI; Fx2 (cytoskeletal fraction) is the material solubilized by 0.5% Triton X-100 and 250 mM

NH

4

SO

4 (Devarajan et al.,1994). The presence of the pl,, polypeptide, with its competent ankyrin binding domain but lacking region IMAD 1 sequences (cf Fig 2C), only marginally impacted the assembly of Na,K-ATPase into a detergent insoluble skeleton. Collectively, these data indicate that truncated PI spectrin peptides containing the Golgi targeting signal, but lacking downstream sequences including the ankyrin/Na,K-ATPase binding domain, block the ER to medial Golgi transport of both subunits of Na,K-ATPase.

Na,K-ATPase interaction with SAATS occurs via an adapter protein As set forth above, co-immunoprecipitation of wild-type cells with antibodies to PI spectrin precipitate complexes containing spectrin, ANKG119, centractin, and Na,K-ATPase (Figure Independent studies using in vitro assays have determined that a-Na,K-ATPase binds directly to ankyrin (Morrow, et al.,1989; Devarajan, et al.,1994), and that a small 25 residue region of Na,K-ATPase, codons 142 to 166, accounts for most of this binding activity (this region is termed MAB, for minimal ankyrin binding domain) (Figure 8) (Zhang, et al.,1997a). Other studies have also established that ankyrin binds to the 15th repeat unit of spectrin (Kennedy, et al.,1991).

The requirement for ankyrin binding to SAATS in Na,K-ATPase transport can then be demonstrated by i) showing that the ankyrin binding domain of spectrin is required for transport, as evidenced by the rescue of Na,K-ATPase delivery to the plasma membrane in cells carrying P1-5,15 spectrin construct (Figure 2) in which the ankyrin binding domain has been returned to the otherwise plN5 peptide which blocks SAATS WO 99/16875 PCT/US98/20364 -36mediated transport of Na,K-ATPase (Figure and, ii) by showing that the removal of the specific ankyrin binding sequence from Na,K-ATPase renders the protein incapable of being transported from the ER (Figure Further proof of the importance of this linkage to an adapter protein may also be obtained by inhibiting its transport selectively by the expression (or micro-injection) of specific inhibitory peptides that block the interaction of (in this case) a-Na,K-ATPase with ankyrin (Figures 10, 18).

In Figure 6, MDCK cells were transfected with either PIN5 as above, or a with PIN-5,15. This latter construct incorporates the ankyrin binding domain of spectrin into the peptide. Note the complete restoration of Na,K-ATPase transport, as measured in this case by its surface display and secondarily by a reduction in cell size (which is a consequence of cell swelling due to a deficiency of plasma membrane Na,K-ATPase in the pINS transfected cell.

In Figure 9, MDCK cells were transiently transfected with either wild-type Na,K-ATPase to which a FLAG epitope tag had been added to its NH2-terminus, or by a similarly FLAG-tagged mutant Na,K-ATPase in which codons 142 to 166 had been deleted. These residues correspond to the minimal ankyrin binding domain (MAB) identified in Figure 8. In both transfections, the cells were also co-transfected with wt P-Na,K-ATPase, so as to assure sufficient P-Na,K-ATPase to pair with the a-Na,K-ATPase. (top) Flag tagged wt-Na,K-ATPase is delivered normally to the plasma membrane. (bottom) Flag tagged Na,K-ATPase lacking MAB does not assemble at the plasma membrane, and eventually is targeted for degradation in lysosomes. In these cells, this is the only protein whose transport apparently is disrupted.

Example 3: Identification of VSV-G as an integral membrane protein participating in SAATS WO 99/16875 PCT/US98/20364 -37- Infection with vesicular stomatitis virus demonstrated that VSV-G transport to the plasma membrane, which is packaged exclusively by the COPII coat see review Schekman and Orci, 1996), like Na,K-ATPase, was blocked by the PI., spectrin construct (Fig. 5E,F). The pattern of VSV-G staining observed suggests a block prior to the intermediate compartment. Wild-type and pINs transfected cells (A,C,E,G) are shown. The localization ofVSV- G protein was measured after transient infection. E-cadherin was monitored with a Mab from Transduction labs, and also by the extent to which the precursor peptide was proteolyzed (inset, Western blot) from 135 kDa to 120 kDa (EC) (a process that occurs in the trans-Golgi/Shore and Nelson,1991). There was no significant difference in the extent of E-cadherin processing or its level of assembly at the plasma membrane in the PIN.5 line vs. wt cells. Despite the disruption of Na,K-ATPase and VSV-G transport (and the wt Golgi spectrin skeleton, Fig. the Golgi appears to remain largely intact as measured by the distribution of P-COP and by the presence of normal appearing juxtanuclear Golgi structures in uranyl acetate and lead stained electron microscopy (arrows). Bar 10. in in Original magnification 63,000x.

Unlike a-Na,K-ATPase, VSV-G protein is transported by SAATS without using ankyrin as its adapter protein, as illustrated by its failure to be rescued by return of the spectrin ankyrin binding domain (repeat 15) to the inhibitory PIN, construct (Figure 7).

MDCK cells infected with vesicular stomatitis virus were used to monitor the trafficking of VSV-G protein. Newly confluent cells are infected for 60 min. with VSV-G in Transwell* filters (100 ml, 108 titer virus Pimplikar et al.,1994). Infected cells are incubated for 3 hours and fixed prior to analysis. Mab VSV-G Rose) was used as the primary antibody in 2% BSA in PBS 10% goat serum at room temperature for minutes after blocking with goat antiserum for 30 minutes. The transport of this protein is normally to the basolateral membrane, similar to that for Na,K-ATPase (as shown in the wt cells and in Figure This transport is blocked by SAATS inhibition with the construct, as is Na,K-ATPase (cf. Figure However, unlike Na,K-ATPase, WO 99/16875 PCT/US98/20364 -38- VSV-G trafficking is not restored by the inclusion of the ankyrin binding domain of spectrin (repeats 14-15) in the construct. Thus, it is possible to selectively block VSV-G transport without affecting Na,K-ATPase transport.

Example 4: Determination that E-cadherin sequestration is not blocked by spectrin E-cadherin, another type I basolaterally targeted membrane protein was assayed for its interactions with SAATS. Wild-type (Figure 5D) and PIN. transfected cells (Figure 5C) were assayed for the distribution of E-cadherin. E-cadherin was monitored with a Mab from Transduction labs as the primary antibody in 2% BSA in PBS 10% goat serum at room temperature for 60 minutes after blocking with goat antiserum for 30 minutes., and also by the extent to which the precursor peptide was proteolyzed (inset, Western blot) from 135 kDa to 120 kDa (EC) (a process that occurs in the trans-Golgi Shore and Nelson,1991). There was no significant difference in the extent of E-cadherin processing or its level of assembly at the plasma membrane in the pIN. line vs. wt cells. Despite the disruption ofNa,K-ATPase and VSV-G transport (and the wt Golgi spectrin skeleton, Fig. the Golgi appears to remain largely intact as measured by the distribution of P-COP and by the presence of normal appearing juxtanuclear Golgi structures in uranyl acetate and lead stained electron microscopy (arrows). Bar 10lO in 0.5pi in Original magnification (G,H) 63,000x., was fully expressed at the plasma membrane (Fig. Western blot analysis of E-cadherin processing, which involves a protease cleavage in the trans-Golgi compartment that reduces the precursor protein from 135 to 120 kDa (Shore and Nelson,1991), could detect no increase in uncleaved E-cadherin pools (Fig. 5D inset). Thus, unlike Na,K-ATPase and VSV-G, the transport of E-cadherin was not blocked at the ER to Golgi transition, indicating that its attachment to SAATS occurs within the PIN-5 peptide, or else is assembled by a separate pathway.

WO 99/16875 PCT/US98/20364 -39- Example 5: Identification of a5, aV, p3, and pi Integrins as integral membrane proteins participating in SAATS in endothelial cells.

A rapid way to evaluate the participation of a variety of integral membrane proteins in SAATS trafficking is to use flow cytometry to monitor their surface display in cells transfected with the various constructs shown in Figure 2, or related constructs.

Examples of this approach are shown in Figures 13 and 14. Note, that in Figure 14, another somewhat more severe blocker of SAATS function, the plN.2 peptide, was used with comparable results In Figure 13, SV40 transformed murine endothelial cells were transfected with or P 1 and the surface display of three different integrins was measured by flow cytometry. Note the blockage of alpha V, beta 3, and beta 1 integrins by the PIN,_ peptide. Conversely, the pIN.5.5 peptide markedly enhances the surface display of alpha V and beta 1 integrin (but not beta 3 integrin). This enhanced display (as opposed to blockage) is presumably achieved by enhancing the efficiency of SAATS directed cargo loading of vesicular transport by choosing peptides or agents that enhance the binding of a given membrane protein to SAATS.

In Figure 14, experiments were carried out as in Figure 13. SV40 transformed murine endothelial cells were transfected with PIN,, PIN.5,5s, or plN.2,15 and the surface display of three different integrins was measured by flow cytometry. Note that alpha-V integrin is also modulated by SAATS. Also note that in this example, plN-2.,5 was used to rescue beta-1 expression. PIN2 is an even more broadly blocking inhibitor of SAATS trafficking (see Figure but in this instance, supranormal levels of beta-1 can be achieved by inclusion of repeats 14,15 into the pIN2 peptide.

Example 6: Identification of other proteins participating in SAATS transport in lymphocytes.

Using flow cytometry in conjunction with the methods outlined here, other WO 99/16875 PCT/US98/20364 proteins can be identified as modulated by SAATS, and those modulated by the adapter protein ankyrin are also apparent, since only those modulated by ankyrin will be rescued by the inclusion of repeats 14,15 in the inhibitory spectrin construct. In Figure PECAM (CD31) is shown to be ankyrin/SAATS transported in lymphocytes. Cells were transfected as above, and surface display of PECAM, a cell-cell adhesion molecule of the IgG superfamily, was monitored by flow cytometry. Note that for either construct, inclusion of the PIN.5 sequences resulted in strong blockage of transport, indicating their involvement with SAATS. In both cases, surface display was rescued or accentuated by the pIN5,15 peptide. This effect was a bit more pronounced with the N-terminal FLAG tagged construct vs. the COOH-terminal GFP tagged construct, presumably due to interference of the GFP with the COOH-terminal 15th repeat unit needed to restore SAATS binding activity to PECAM.

In Figure 16, CD45 is shown to be SAATS regulated with ankyrin its putative adapter, while TNFR-1 is modulated by SAATS, but does not involve ankyrin as its adapter molecule. In this experiment, Jurkat T-lymphocytes were transfected with the constructs indicated, including PI4.15 alone. Note the marked down regulation of a documented ankyrin binding protein, by constructs lacking repeat 15, as well as by the p 1 i 4 5 peptide itself (which lacks the constitutive Golgi targeting signal, see Fig. 2).

These experiments also illustrate that TNFR-1 display is upregulated (rather than blocked) by PIN2 or PIN.5, suggesting that like E-cadherin, its attachment to SAATS is mediated either directly or indirectly by sequences contained within PIN-2.

In Figure 17, Fas and Fas-L are shown to be SAATS dependent, but only Fas appears to utilize ankyrin. Jurkat T-lymphocytes were transfected with the constructs indicated. Experiments were as before. Note the changes in both Fas and Fas-L due to 1I,.5 and PIN-5,15 peptides.

Example 7: Selective inhibition of transport of a single protein.

By utilizing a specific inhibitor of the docking of a single integral membrane WO 99/16875 PCT/US98/20364 -41protein to SAATS, rather than a more general blocker of SAATS function, great specificity may be achieved. For example, the ankyrin binding domain sequence of a-Na,K-ATPase is highly conserved across all species of Na,K-ATPase, but is not found in any other protein (Zhang et al., 1997b). Expression of this peptide sequence alone in cells (Figure 10) selectively blocks the transport of Na,K-ATPase, without blocking the transport of any other known protein. Thus, by the identification of selective inhibitors of protein docking to SAATS, great specificity can be achieved.

In Figure 10, wild-type MDCK cells were transiently transfected with green fluorescent protein linked to the 25 residue MAB peptide. This is the same sequence whose deletion caused Na,K-ATPase to be retained in the ER and ultimately lysosomes in Figure 9. After approximately 1-2 days in culture, cells were fixed and stained for GFP (using anti-GFP antibodies, left panel) or for Na,K-ATPase (right panel, red). Note that the expression of GFP alone in MDCK cells (wt) did not affect Na,K-ATPase distribution, or cell size (which is dependent on Na,K-ATPase function. However, when the GFP carried the 25 residue MAB peptide, the binding of the endogenous Na,K-ATPase to SAATS was inhibited, resulting in its selective accumulation in the ER and reduced levels on the plasma membrane (two examples are shown, labeled "MAB").

Cells containing GFP-MAB are also markedly swollen, as expected. No other protein distributions were affected in these cells.

In Figure 18, clusters of wild-type MDCK cells were micro injected with either GST alone; or (B).with the 3I-.4 peptide generated as a recombinant fusion peptide with GST, and then stained for Na,K-ATPase by indirect immunofluorescence. While micro injected GST is without effect, note the loss ofNa,K-ATPase staining intensity in the injected cells on the left side of The consequences of impaired Na,K-ATPase delivery to the membrane are also apparent in the larger size of cells with insufficient Na,K-ATPase, similar to the changes seen when p1N, is transfected. These results indicate that selective blockage of Na,K-ATPase can be achieved by exogenous small peptides. Similar experiments were also carried out using various monoclonal antibodies.

WO 99/16875 PCT/US98/20364 -42- In these experiments, the cells were subconfluent, and all cells in an isolated cluster were injected. Control injections of buffer alone are without effect on the distribution of Na,K-ATPase. Microinjection of Mab IID2, which reacts exclusively with al spectrin (not considered to be a component of SAATS), is also without effect on Na,K-ATPase. Microinjection ofMab VIIIC7, which reacts with PI spectrin (Haris et al.,1986), leads to severe disruption of Na,K-ATPase delivery, with extensive accumulation within the cells and cell swelling. In this micrograph, the loss of membrane Na,K-ATPase is so severe, that the membrane borders cannot be discerned.

This experiment illustrates the effectiveness of SAATS specific monoclonal antibodies in blocking SAATS activity.

Example 8: The cystic fibrosis transmembrane conductance regulator participates in SAATS transport.

Figure 12 illustrates another integral membrane protein of significant medical interest that is linked to SAATS by different ankyrin interaction. Both CFTR (cystic fibrosis transmembrane conductance regulator) and its close homologue TNR (Transmembrane-Nucleotide binding -Regulatory domain) protein are present in the soluble (Fxl) and cytoskeletal pools (Fx2) of confluent MDCK cells. When these cells are surface labeled with biotin from either the apical surface (lane A) or the basolateral surface (lane it is also apparent that both CFTR and TNR are highly polarized, and unlike Na,K-ATPase, are expressed predominately in the apical domain of these polarized epithelial cells. Despite the fact that CFTR and TNR are apical proteins, they nevertheless binds components of the SAATS system, which is responsible for their trafficking. Either whole cell lysates of cells extracted with RIPA buffer, or Fx 1 and Fx 2, were immunoprecipitated with irrelevant antibody control or with anti-CFTR antibody The washed immunoprecipitates were then analyzed by SDS-PAGE and examined by western blotting for either ANKI,,, or ANKR. Note that both ankyrins co-immunoprecipitate with CFTR, in a fashion exactly analogous to the way WO 99/16875 PCT/US98/20364 -43- Na,K-ATPase binds SAATS. However, CFTR and Na,K-ATPase bind at distinct sites on SAATS, enabling their selective modulation.

Figure 22 illustrates the modulation of surface display of the Cystic Fibrosis Transmembrane Conductance Regulator (CFTR) by SAATS. MDCK cells were transfected with either the PIN or PIN-5.s constructs and then surface labeled with biotin.

After labeling, cells were solubilized, and the total CFTR in each culture immunoprecipitated with anti-CFTR antibodies. The precipitates were then analyzed by SDS-PAGE, transferred to nitrocellulose, and surface CFTR detected by avidin labeling.

Total CFTR was also determined in these cellular extracts by western blotting (data not shown). In all cell lines, there was minimal change in the total CFTR in the cells.

However, there was marked dimunition of surface display of CFTR induced by the SAATS blocker PIN,_; whereas the PIN-5,15 construct markedly enhanced the surface display of CFTR. Collectively these results document the involvement of SAATS in CFTR trafficking, and suggest a possible therapeutic approach to the control of disorders such as cystic fibrosis, or the control of related proteins such as the multi-drug resistance gene product (MDR).

Example 9: Selective enhancement of integral membrane protein display by SAATS The rate of exit of integral membrane and secretory proteins from the ER and Golgi appears to be at least partially kinetically regulated by their rate and efficiency of docking to SAATS. Therefore, upregulation of the surface display or secretion of a given protein, as well as its blocking, can be achieved by modulation of SAATS. This is illustrated in Figures 13 to 16, where it can be noted that for those molecules utilizing ankyrin as an adapter complex, supranormal levels are achieved by augmenting their transport with the PIN-5,15 peptide, which increased ankyrin dependent transport at the expense of ankyrin independent transport the fall in TNFR-1 with PIN5,15 expression, Figure 16).

WO 99/16875 PCT/US98/20364 -44- Example 10: Determination of which vesicular skeletal proteins and associate adapter proteins are specifically involved in mediating sequestration of a given cell membrane or secretory proteins into transport vesicles.

The general method described herein is based on the discovery that the delivery of a given membrane protein or class of membrane proteins to the plasma membrane of eukaryotic cells may be inhibited or augmented by altering the binding of such proteins to the spectrin/ankyrin/adapter protein trafficking system (SAATS). The secretion of proteins that bind to so-called cargo receptors in the endoplasmic reticulum may also be controlled by controlling the binding of such cargo receptor proteins to SAATS.

As disclosed here, SAATS is a cryptic protein sorting apparatus operating to concentrate specific proteins in the endoplasmic reticulum and to facilitate their inclusion into vesicles undergoing transport or diffusion between the ER and Golgi compartments. A similar pathway is also active between the Golgi and the plasma membrane of many cells, but the most general use of this method of control is achieved by blocking SAATS at the ER to cis Golgi transition. The method is implemented by 1) identifying to which component of the SAATS system a given integral membrane protein binds, discussed as follows; and 2) identifying a peptide, small molecule analog, antibody, or antisense oligonucleotide that will disrupt in vivo the binding interaction between the target protein and SAATS. This latter aspect of the present invention is discussed in the next section of this specification.

Previous efforts have only identified binding sites for a small number of what we now know to be SAATS ligands (See Table 1B). We believe this is due to the fact that many interactions will occur in solution with only low affinity, affinities that will become significant in the restricted environment of SAATS-membrane interaction. Specifically, once entropic contributions of to the binding energy are largely accommodated by way of a high affinity SAATS-membrane binding (such as via MAD1 or MAD2) (Lombardo et al.,1994), then other interactions between membrane proteins and SAATS will be WO 99/16875 PCT/US98/20364 facilitated. Assays to detect such interactions must therefore be sensitive to detecting low affinity but specific interactions.

Standard techniques that measure protein-protein interactions at equilibrium are thus well suited to detect the potentially weak interactions (Ka 10- 4 to 10 5 that may typically characterize binary interactions between membrane proteins and many ligand sites in SAATS. Typical of such assays would be equilibrium dialysis; fluorescent based assays in which the change in intrinsic fluorescence of a protein is monitored as a function of binding; assays based on osmolarity or light scattering; and others commonly known to practitioners of the art. In addition, surface plasmon resonance, such as implemented in Pharmacia's BiaCore instrument, is also well suited to detecting specific macromolecular interactions withing SAATS, often even in unpurified solutions prepared from cell or bacterial lysates. Other ways to enhance detection sensitivity is by employing radioactive labeled compounds, or other sensitive detection techniques such as enhanced chemiluminescence in conjunction with standard microtiter or blot overlay techniques. Genetic selection assays are also useful, such as two-hybrid assays in yeast, or phage epitope display library screening.

Example 11: Identification of peptides and small molecules that modulate sequestration and/or vesicle transport of integral membrane and secretory proteins In general, several approaches can be used to identify an agent able to block (or augment) the interaction between a target protein and SAATS. In general, this method may be approached by either: A) identifying broad inhibitors of SAATS function that disrupt a class or multiple classes of membrane or secretary proteins, such as peptide agents or expressed recombinant genes that encode specific functional domains in SAATS, and that block the docking of classes of membrane proteins by competitive inhibition with the endogenous SAATS system; or B) following a drug design strategy that can be either rational, with inhibitory agents determined based on detailed WO 99/16875 PCT/US98/20364 -46knowledge of binding sites mediating the attachment of the target protein to SAATS, or random, based on in vitro surrogate assays of SAATS function. This latter approach is based on identifying compounds that block interactions with SAATS as measured by in vitro binding assays.

Examples of approach A) were presented in Section A of this detailed description, above. Examples utilizing approach B are outlined below, together with a general protocol for identifying such agents.

B) Strategies based on rational drug design. To block or modulate the display of membrane protein X using a rational drug design approach, detailed knowledge of its SAATS binding site, up to and including its three dimensional structure, will be required. Completion of Steps A4 and A5 above provide the starting point for a systematic approach.

Step BI: Determine whether protein X when mature is monomeric or homopolymeric, or whether it exists in its mature state as a heterocomplex with another membrane protein. This step may not be necessary if protein X can be shown to bind SAATS directly, but many membrane and all secretary proteins will only attach to SAATS via their interaction with another membrane protein. If this is the case, either agents must be identified that will prevent the formation of the heterocomplex, or the complex will be controlled as a unit the display of all components of the functional heterocomplex will be controlled together by controlling their SAATS assembly.

Standard protein biochemical and biophysical techniques can be used to make these determinations, including co-immunoprecipitation from non-ionic detergent extracts of cultured cells expressing protein X.

For example, the demonstration of direct binding of a-Na,K-ATPase (as opposed to P-Na,K-ATPase) to ankyrin (Morrow, et al.,1989, Devarajan, 1994) can be demonstrated by a variety of in vitro and in vivo techniques, including co-sedimentation, gel overlay binding assays, and surface plasmon resonance. In these assays see WO 99/16875 PCT/US98/20364 -47- Morrow et al. 1989), direct binding of ankyrin to the a-Na,K-ATPase, but not P-Na,K-ATPase can be demonstrated. However, since P-Na,K-ATPase binds stoichiometrically to the alpha subunit, both are efficiently transported by SAATS, but this transport is dependent on the binding of a-Na,K-ATPase to SAATS see Figure 4 demonstrating failure of P-Na,K-ATPase transport).

Step B2: Determination of the specific binding site within protein X that links it to SAATS. Many conventional methods may be used to accomplish this. A common one is the use of deletional mutagenesis using a series of recombinant peptides, often generated as fusion peptides linked to glutathione-S-transferase (GST). For example, a deletional analysis of GST-fusion peptides representing the cytoplasmic domains of Na,K-ATPase (Devarajan et al.,1994a; Devarajan et al.,1994b).

(Go to Step B3 or B3 alternate).

Step B3: Use of conventional structure determination methods to resolve the 3-D structure of the binding domain. If a complete structure determination of protein X is not possible, a useful method for determining the relevant structure of the binding site is carrier mediated crystallization, in which the binding domain of interest is crystallized as a fusion protein with GST (for example, see Lim et al.,1994). Subsequent determination of the structure of this complex yields the structure of the key binding site.

Other structural determination methods, such as multidimensional NMR, are of course applicable as well.

For example, the determination of the 3-D structure of the minimal ankyrin binding domain in a-Na,K-ATPase by crystallization of the active GST-fusion peptide would be useful, as we have recently demonstrated (Zhang, et al.,1997b). (See Figs. 4B, 4C, and 11). To prove the efficacy of agents based on the structure of such as binding domain, it is also useful to demonstrate the specific inhibition of the SAATS WO 99/16875 PCT/US98/20364 -48transport of the native protein using the binding domain alone (as illustrated in Figures 18), as well as the ability of the binding domain to bind in vitro to its ligand receptor site with an affinity and stoichiometry comparable to that of the native protein.

Step B4: Use of the structure determined above to design candidate small molecules that will inhibit the binding interaction and assay of these compounds by their ability to block the in vitro interaction between the active site of protein X and its SAATS binding site. Alternatively, if the goal is to identify small molecules that augment a given SAATS interaction, so as to identify a therapeutic that enhances the display or transport or secretion of a given protein (such as would be desirable in the treatment children with cystic fibrosis, Figure 12B), molecules that augment in vitro the interaction between protein X and SAATS would be sought.

Step B3 alternate: Alternatively, an empirical approach identifying lead compounds based on a surrogate measures of SAATS function is possible. Since all or most proteins exit the ER compartment via a direct or indirect binding to SAATS, in vitro assays based on specific binding sites for each membrane or secretory protein of interest provide an ideal basis for high-throughput in vitro screens of SAATS function. For example, high-throughput screens in microtiter plates, seeking to identify all agents that block or diminish the binding of spectrin to AnkG1 19, or AnkG1 19 to protein X, would be one such suitable assay. Alternatively, one could prepare similar assays in which protein X or ankyrin or other SAATS component was immobilized onto an inert substrate (eg. nitrocellulose, immobilon, 96 well titer plate, microchip, etc), and various agents compared with respect to their ability to block or enhance a the binding of a soluble protein to the immobilized protein. The details of the implementation of such assays is well known in the art. The key feature of these assays, as applied to SAATS, is that the direct binary interaction between an integral membrane protein, or even an oligomeric WO 99/16875 PCT/US98/20364 -49complex of such proteins even partially purified, can be used as a suitable target for drug discovery since i)such proteins can be immobilized even in impure form; ii) purified and well characterized components of SAATS can be monitored for binding to such immobilized or impure proteins; and iii) the presence of the binary interaction is itself a well proven and good surrogate for in vivo SAATS activity.

Example 12: Strategies based on regulation of spectrin binding to the Golgi complex We envision that naturally the rate of transport of specific secretory and integral membrane proteins from the ER to the plasma membrane will be regulated by control of the binding to SAATS. Several levels of post-translational control of the spectrin cortical cytoskeleton have been identified, and we anticipate that several of these will also be active in regulating SAATS (Morrow, et aL.,1997). In addition, regulation of SAATS also appears to share at least one regulatory pathway that also regulates the assembly of other coatamer components in the secretory pathway, and that is modulation of SAATS affinity for the Golgi by ADP-ribosylation factor (ARF) (Godi et al., 1997).

In one embodiment of the present invention, to evaluate the activity of a candidate compound to modulate interaction of SAATS with selected integral membrane or secretory proteins, such a compound is mixed with the selected protein and whatever vesicular skeletal protein or adapter protein is relevant. After mixing under conditions that allow association of these components, the mixture is analyzed to determine if the agent inhibited or enhanced binding of the selected protein to its SAATS binding partner.

Inhibitors and enhancers are thus confirmed as being able to modulate this interaction. A skilled artisan can readily employ numerous art-known techniques for determining whether a particular agent modulates the binding of such selected proteins to their SAATS binding partner. Agents can be further tested for the ability to modulate binding using a cell-free assay system or a cellular assay system. Example 14 provides one such method that can be used to assay for relevant activity.

WO 99/16875 PCT/US98/20364 As used herein, an agent is said to inhibit SAATS binding activity when the agent reduces the binding of a selected integral membrane or secretory protein to its SAATS binding partner. The preferred inhibitor will selectively inhibit such binding specifically, not affecting the interaction of any other proteins to SAATS. Further, the preferred inhibitor will reduce such binding by more than about 50%, more preferably by more than about 90%, most preferably eliminating substantially all binding of the selected protein to SAATS.

As used herein, an agent is said to enhance SAATS binding activity when the agent increases the binding of a selected integral membrane protein to SAATS. The preferred binding enhancer will increase SAATS binding activity by more than about more preferably by more than about 90%, most preferably more than doubling the level of binding or transport of such proteins to SAATS.

The preferred inhibitors and enhancers will be selective for a specific species of integral membrane or secretory protein. The agents of the present invention can be, as examples, peptides, small molecules, and vitamin derivatives, as well as carbohydrates.

A skilled artisan can readily recognize that there is no limit as to the structural nature of the agents of the present invention. One class of agents of the present invention are peptide agents whose amino acid sequences are chosen based on the amino acid sequence of various spectrin homologues. Small peptide agents can serve as competitive inhibitors of SAATS binding, sequestration and transport.

The peptide agents of the invention can be prepared using standard solid phase (or solution phase) peptide synthesis methods, as is known in the art. In addition, the DNA encoding these peptides may be synthesized using commercially available oligonucleotide synthesis instrumentation and produced recombinantly using standard recombinant production systems. The production using solid phase peptide synthesis is necessitated if non-gene-encoded amino acids are to be included.

Another class of agents of the present invention are antibodies immunoreactive with critical epitopes of vesicular skeletal proteins such as spectrin and associated adapter WO 99/16875 PCT/US98/20364 -51proteins such as ankyrin. Antibodies are obtained by immunization of suitable mammalian subjects with peptides, containing as antigenic regions, those portions of, spectrin, which binds the selected integral membrane or secretory. Critical regions for the transport of Na,K-ATPase include the domains identified in Figure 3 or the cytoplasmic 29 residues of VSV-G. Such agents can be used in competitive binding studies to identify second generation inhibitory agents.

Example 13: Use of agents that modulate the sequestration and/or vesicular transport of integral membrane and secretory proteins for therapeutic purposes We contemplate that the methods described above will be useful in a wide variety of therapeutic contexts. For example, the following Table 2 illustrates several possibilities.

WO 99/16875 PCT/US98/20364 -52- Table 2: Therapeutics envisioned for the following disorders General Area Cancer Cardiovascular Degenerative Disorders Dermatologic/ cosmetic Genetic Disease Hematological Disease Immune-modulation Infectious Disease Inflammation Malabsorption Nutritional disorders Neurological Disorders Transplantation Diabetes Specific Conditions Growth factor and Growth factor receptor modulation.

Multi-drug resistance transporter down-regulation.

Hormone receptor down-regulation in breast, endocrine, and prostatic tumors. Adhesion molecule modulation; Integrin modulation.

Hypertension; Cholesterol and hyperlipidemia conditions, anti-arrhythmic agents Ca" channel antagonists), Thrombosis control; anti-platelet agents; endothelial surface modifying agents; restenosis control Osteoporosis, Psoriasis, scleroderma, cutaneous hypersensitivity, poison ivy, Cystic fibrosis (eg D508 mutation reduces transport of CFTR to membrane.) Transfusion incompatibilities Type I Diabetes; Arthritis; Asthma; Allergy; Scleroderma; lupus; Viral infections of all types, including influenza and HIV. Enteric and other bacterial infections. Malaria, protozoan infections, clamydia.

Chronic ulcers; diabetes; wound healing; osteomyelitis; inflammatory cell/ endothelial cell adhesion; fibrosis control.

Hemochromatosis; Celiac disease/non-tropical sprue; gluten enteropathy; ulcer disease Depression; Manic-depressive illness; Schizophrenia; Parkinson's disease; Neural regeneration; APP protein display and secretion in Alzheimer's Disease.

Epitope modulation; HL-A antigen display in grafts.

Moderate immunosuppression. Bone marrow transplantation autoimmiune control. (Type Insulin receptor modulation; glucagon regulation.

Comment Blockage of MDR to sensitize tumors favorable early therapeutic target.

Modification of membrane channels, thrombin receptors, endothelial adhesion receptors.

BMP regulation, calcitonin/PTH balance regulation; calcium homeostasis.

Topical application may offer early therapeutic advantage Enhanced binding to SAATS may ameliorate disease without correcting the mutation.

surface epitope display modulation; pro inflammatory cytokines; moderate immunoregulators Inhibition of surface display of viral proteins. Infections mediated by binding to tissue receptors.

Modification of selectin display, endothelial adhesion receptors, cytokine secretion and cytokine receptors. Interleukins and receptors, Tumor necrosis factor and receptors. Collagen secretion.

transferrin receptor modulation, immunogens and immunogen receptors; mucosal IgA display.

Neuropeptide receptor modulation; Neurotransmitter uptake and secretion; axonal guidance molecule display and receptors. Axonal transport modulators also.

antigen presentation control; epitope display, cross-reactivity suppression Glucose transporter and Insultin receptor control.

SUBSTITUTE SHEET (RULE 26) WO 99/16875 PCT/US98/20364 -53- Reproductive biology Sperm/ovum receptors, contraception, infertility, maternal/fetal immune disorders Endocrine Disorders Goiter, hyper/hypo thyroidism, growth hormone and receptors, pituitary disorders Traumatic and Ischemic to accelerate, e.g. recovery of mislocated Injury Na,K-ATPase following ischemia/heart, kidney, brain, salvage of myocardium, recovery of infarction, limitation of infarct.

The administration of agents identified by the foregoing techniques will be dependent on their intended purpose, to enhance or inhibit the transport of a selected integral membrane protein or secretory protein, and on their chemical nature, peptide or small molecule. For example, to treat cystic fibrosis, as noted above, by identifying the SAATS binding domain for CFTR would allow the identification of agents that selectively increase the delivery of mutant (but functional) CFTR to the plasma membrane. Appropriate agents might optimally be delivered by inhalation therapy (see U.S. Patents 5,669,376 and 5,655,516).

For vascular regulators, an appropriate agent can be administered via parenteral, subcutaneous, intravenous, intramuscular, intraperitoneal, transdermal, or buccal routes.

Alternatively, or concurrently, administration may be by the oral route. The dosage administered will be dependent upon the age, health, and weight of the recipient, kind of concurrent treatment, if any, frequency of treatment, and the nature of the effect desired.

As described below, there are many methods that can readily be adapted to administer such agents once identified by the methods herein described. The determination of optimal ranges of effective amounts of each component is within the ordinary level of skill in the therapeutic art and is based on the intended use. In addition to the particular agent of choice, the compositions of the present invention may contain other ingredients, such as suitable pharmaceutically acceptable carriers comprising SUBSTITUTE SHEET (RULE 26) WO 99/16875 PCT/US98/20364 -54excipients and auxiliaries which facilitate processing of the active compounds into preparations which can be used pharmaceutically for delivery to the site of action.

Suitable formulations for parenteral administration include aqueous solutions of the active compounds in water-soluble variants, for example, water-soluble salts. In addition, suspensions of the active compounds, and as appropriate, oily injection suspensions may be administered. Suitable lipophilic solvents or vehicles include fatty oils, for example, sesame oil, or synthetic fatty acid esters, for example, ethyl oleate or triglycerides. Aqueous injection suspensions may contain substances which increase the viscosity of the suspension and include, for example, sodium carboxymethyl cellulose, sorbitol, and/or dintran.

Optionally, the suspension may also contain stabilizers. Liposomes can also be used to encapsulate the agent for delivery into the cell. The pharmaceutical formulation for systemic administration according to the invention may be formulated for enteral, parenteral or topical administration. Indeed, all three types of formulations may be used simultaneously to achieve systemic administration of the active ingredient.

Suitable formulations for oral administration include hard or soft gelatin capsules, pills, tablets, including coated tablets, elixirs, suspensions, syrups or inhalations and controlled release variants thereof.

The agents of the present invention that modulate sequestration and trafficking of selected integral membrane and secretory proteins can be provided alone, or in combination with another appropriate pharmacoactive. For example, an agent of the present invention that reduces the presence of various cell surface receptors involved in the attachment of viral envelopes might be administered in combination with other anti-viral agents. As used herein, two agents are said to be administered in combination when the two agents are administered simultaneously or are administered independently in a fashion such that the agents will act at the same time.

Example 14: Identification of the spectrin or adaptor protein domain(s) WO 99/16875 PCT/US98/20364 responsible to binding and transport of an integral membrane or secretory protein.

As set forth in Figure 2, PIS spectrin contains a series of functional domains that have been shown to confer or may confer the ability to bind specific adaptor proteins and/or proteins to be transported. Similarly, ankyrin contains a number of repeats (13 to 24 repeat units characteristic of most ankyrins), with each ankyrin-repeat structure composed of two alpha helices and a P-hairpin loop. The complexity of both the individual repeats and the number of repeats in the structure of both structural proteins such as spectrins and adaptor proteins such as ankyrin may allow for the varied ability of both molecules to bind to multiple distinct integral membrane or secretory proteins.

As an example, in ankyrins, multiple repeat units create a structure in which interactions between the helices form a central core structure, while the tips of the exposed p-hairpin turns provide putative protein-protein interaction surfaces (See Figure 21). In one scenario, the seven residue loop within MAB of a-Na,K-ATPase interacts with a specific site in ankyrin created by the tips of one or more of these P-hairpin turns.

Since a multiplicity of potential binding pockets would be created by the 13 to 24 repeat units characteristic of most ankyrins, specific and unique binding sites presumably also exist for the other short peptide sequences responsible for ankyrin binding activity in other proteins.

The ability of adaptor proteins to bind to spectrins or integral membrane or secretory proteins to bind directly to spectrins can be assayed to identify the spectrin domain, or combinations of domains responsible for specific binding of a given adaptor protein, adaptor protein and integral membrane or secretory protein or integral membrane or secretory protein directly. Likewise, the ability of an given integral membrane or secretory protein to bind to an adaptor protein such as ankyrin can be assayed to identify the ankyrin domain or combination of domains responsible for specific integral membrane or secretory protein binding. Identification of the specific domains responsible for binding then allows for the development of peptide based molecules, or mimetopes with similar structures, to modulate or disrupt specific binding of the many WO 99/16875 PCT/US98/20364 -56protein-protein interactions involved in SAATS.

As an example, to determine the identity of ankyrin domains responsible for specific binding to a given integral membrane, deletional analysis of the ankyrin repeat region is used to produce a series of fragments of the protein. The fragments are then tested for their ability to bind to the integral'membrane protein of interest. Any means commonly available of detecting the ability of the ankyrin fragment to bind to the specific integral membrane protein can be used. For instance, cell free systems can be used wherein the binding is detected by the ability of the ankyrin fragment, detected by radiolabelling, fluorescent tag, or specific antibody, to be co-immunoprecipitated by antibodies to the protein to which binding is suspected. Alternatively, the ability of the ankyrin fragment to co-sediment with permeabilized cells or membrane preparations enriched in the proteins of interest the co-sedimentation of ankyrin with Na,K-ATPase enriched membranes, Morrow et al., 1989) is another method of detecting binding. Other methods include binding to immobilized ligands; co-sedimentation, co-elution on gel filtration chromatography; surface plasmon resonance, fluorescent energy transfer; osmolarity; light-scattering; resonance raman spectroscopy, NMR spectroscopy, yeast two-hybrid systems; expression cloning, and phage display. Other methods will also be known to the practitioner of the art. Since individual ankyrin repeat units, if that is the target of the assay, are often insoluble or unstable, it will often be of value to examine smaller internal sequences, such as those from the beta-hairpin loops of the ankyrin repeat structure, for binding activity.. High through-put assays can also be developed to detect binding the a library of ankyrin fragments comprising all individual repeat units plus all possible combinations of the repeat units. Examples of such assays would include incorporating single, two-mer, and trimer combinations of the beta-hairpin loop sequences of all known ankyrin repeat units into pseudo-libraries suitable for screening in the yeast two-hybrid or CHO two-hybrid KISS assays (Fearon et al.,1992) against the ligand of interest. Similar strategies could also be employed utilizing the same peptide sequences, generated either recombinantly or by synthetic methods, WO 99/16875 PCT/US98/20364 -57immobilized on nitrocellulose or immobilon or similar membranes, or on plastic multiwell plates, or with the sequences incorporated into bacteriophage so as to create a phage display library suitable for binding to the protein of interest. (Scott, 1992) It should be understood that the foregoing discussion and examples present merely present a detailed description of certain preferred embodiments. It therefore should be apparent to those of ordinary skill in the art that various modifications and equivalents can be made without departing from the spirit and scope of the invention. All articles, patents and patent applications that are identified above are incorporated by reference in their entirety.

Additional References The following articles are hereby incorporated by reference in their entirety: Andrews, R. K. and J. E. Fox. 1992. Identification of a region in the cytoplasmic domain of the platelet membrane glycoprotein Ib-IX complex that binds to purified actin-binding protein. J Biol Chem 267:18605-11.

Bihler, M. F. and P. Greengard. 1987. Synapsin I bundles F-actin in a phosphorylation-dependent manner. Nature 326:704-707.

Bannykh, S. I. and W. E. Balch. 1997. Membrane dynamics at the endoplasmic reticulum-Golgi interface. J. Cell Biol. 138:1-4.

Beck, J. A. Buchanan and W. J. Nelson. 1997. Golgi membrane skeleton: identification, localization and oligomerization of a 195 kDa ankyrin isoform associated with the Golgi complex. J. Cell Sci 110:1239-1249.

Beck, K. J. A. Buchanan, V. Malhotra and W. J. Nelson. 1994. Golgi spectrin: identification of an erythroid beta-spectrin homolog associated with the Golgi complex. J. Cell Biol. 127:707-23.

WO 99/16875 PCT/US98/20364 -58- Bennett, V. 1986. Purification of brain analogs of red blood cell membrane skeletal proteins: Ankyrin, protein 4.1 (synapsin), spectrin, and spectrin subunits. Meth. Enzymol. 134:55-69.

Bennett, V. 1992. Ankyrins. Adaptors between diverse plasma membrane proteins and the cytoplasm.

J Biol Chem 267:8703-6.

Bennett, V. and D. Branton. 1977. Selective association of spectrin with the cytoplasmic surface of human erythrocyte plasma membranes. Quantitative determination with purified (32P)spectrin. J Biol Chem 252:2753-63.

Bennett, V. and P. J. Stenbuck. 1979a. Identification and partial purification of ankyrin, the high affinity membrane attachment site for human erythrocyte spectrin. JBiol Chem 254:2533-41.

Bennett, V. and P. J. Stenbuck. 1979b. The membrane attachment protein for spectrin is associated with band 3 in human erythrocyte membranes. Nature 280:468-73.

Bourguignon, L. V. B. Lokeshwar, J. He, X. Chen and G. J. Bourguignon. 1992. A CD44-like endothelial cell transmembrane glycoprotein (GP116) interacts with extracellular matrix and ankyrin. Mol Cell Biol 12:4464-71.

Bourguignon, L. S. J. Suchard and E. L. Kalomiris. 1986. Lymphoma Thy-1 glycoprotein is linked to the cytoskeleton via a 4.1-like protein. J Cell Biol 103:2529-2540.

Bourguignon, L. S. J. Suchard, M. L. Nagpal and J. J. Glenney. 1985. A T-lymphoma transmembrane glycoprotein (gpl80) is linked to the cytoskeletal protein, fodrin. J Cell Biol 101:477-87.

Chang, J. C. and B. G. Forget. unpublished. Molecular cloning reveals novel isoforms of PII spectrin.

Chang, J. A. Scarpa, R. L. Eddy, M. G. Byers, A. S. Harris, J. S. Morrow, P. Watkins, T. B. Shows and B. G. Forget. 1993. Cloning of a portion of the chromosomal gene for human b fodrin, the non-erythroid form ofb spectrin. Genom 17:287-293.

Cianci, C. P. G. Gallagher, B. G. Forget and J. S. Morrow. 1995. Unique subsets of proteins bind the SH3 domain of al spectrin and Grb2 in differentiating mouse erythroleukemia (MEL) cells. Molec.Biol.Cell WO 99/16875 PCT/US98/20364 -59- 5:271 a (abstract).

Cianci, C. Z. Zhang and J. S. Morrow. 1997. Brain and muscle express a unique alternative transcript of all spectrin. submitted Cohen, C. M. and S. F. Foley. 1980. Spectrin-dependent and -independent association of F-actin with the erythrocyte membrane. Journal of Cell Biology 86:694-8.

Cohen, C. M. and S. F. Foley. 1982. The role of band 4.1 in the association of actin with erythrocyte membranes. Biochim Biophys Acta 688:691-701.

Cohen, C. M. and C. Korsgren. 1980. Band 4.1 causes spectrin-actin gels to become thixiotropic.

Biochemical Biophysical Research Communications 97:1429-35.

Davis, J. Q. and V. Bennett. 1993. Ankyrin-binding activity of nervous system cell adhesion molecules expressed in adult brain. J Cell Sci Suppl 17:109-17.

Davis, J. T. McLaughlin and V. Bennett. 1993. Ankyrin-binding proteins related to nervous system cell adhesion molecules: candidates to provide transmembrane and intercellular connections in adult brain. J Cell Biol 121:121-33.

Devarajan, P. and J. S. Morrow (1996). The Spectrin Cytoskeleton and Organization of Polarized Epithelial Cell Membranes. Membrane Protein-Cvtoskeleton Complexes: Protein Interactions. Distributions and Functions. W. J. Nelson, Ed. New York, Academic Press. 97-128.

Devarajan, D. A. Scaramuzzino and J. S. Morrow. 1994. Ankyrin binds to two distinct cytoplasmic domains ofNa,K-ATPase a subunit. Proc. Natl. Acad. Sci (USA) 91:2965-2969.

Devarajan, P. R. Stabach, A. S. Mann, T. Ardito, M. Kashgarian and J. S. Morrow. 1996.

Identification of a small cytoplasmic ankyrin (AnkG 119) in kidney and muscle that binds PIE* spectrin and associates with the Golgi apparatus. J. Cell Biol. 133:819-830.

Devarajan, P. R. Stabach. M. A. De Matteis and J. S. Morrow. 1997. Na,K-ATPase Transport from ER to Golgi Spectrin-Ankyrin G119 Skeleton in MDCK Cells. Proc. Natl. Acad. Sci., USA. 94: WO 99/16875 PCT/US98/20364 Diakowski, W. and A. F. Sikorski. 1994. Brain spectrin interacts with membrane phospholipids. Acta Biochimica Polonica 41:153-4.

Fach, B. S. F. Graham and R. A. Keates. 1985. Association of fodrin with brain microtubules. Can J. Biochem. CellBiol. 63:372-381.

Fearon, E. T. Finkel, M. L. Gillison, S. P. Kennedy, J. F. Casella, G. F. Tomaselli, J. S. Morrow and C. V. Dang. 1992. Karyoplasmic interaction selection strategy (KISS): A general strategy to detect protein-protein interactions in mammalian cells. Proc. Natl. Acad. Sci (USA) 89:7958-7962.

Fowler, V. M. and V. Bennett. 1984. Erythrocyte membrane tropomyosin. Purification and properties.

JBiol Chem 259:5978-89.

Fowler, V. M. A. Sussmann, P. G. Miller, B. E. Flucher and M. P. Daniels. 1993. Tropomodulin is associated with the free (pointed) ends of the thin filaments in rat skeletal muscle. J Cell Biol 120:411-20.

Frappier, J. Derancourt and L. A. Pradel. 1992. Actin and neurofilament binding domain of brain spectrin beta subunit. EurJBiochem 205:85-91.

Frappier, F. Regnouf and L. A. Pradel. 1987. Binding of brain spectrin to the neurofilament subunit protein. EurJ Biochem 169:651-7.

Gardner, K. and V. Bennett. 1987. Modulation of spectrin-actin assembly by erythrocyte adducin.

Nature 328:359-362.

Georgatos, S. D. and V. T. Marchesi. 1985. The binding of vimentin to human erythrocyte membranes: a model system for the study of intermediate filament-membrane interactions. Journal of Cell Biology 100:1955-61.

Georgatos, S. D. C. Weaver and V. T. Marchesi. 1985. Site specificity in vimentin-membrane interactions: intermediate filament subunits associate with the plasma membrane via their head domains.

Journal of Cell Biology 100:1962-7.

Godi, I. Santone, P. Pertile, P. Devarajan, P. R. Stabach, J. S. Morrow, G. Di Tullio, R. Polishuck, WO 99/16875 PCT/US98/20364 -61- T. C. Petrucci, A. Luini and M. A. De Matteis. 1997. Role in membrane tiansport and ARF-dependence of the Golgi-associated spectrin. submitted Harlow, E. and D. Lane. 1988. Antibodies: A Laboratory Manual. Cold Spring Harbor Laboratory.

Harris, A. J. P. Anderson, P. D. Yurchenco, L. A. D. Green, K. J. Ainger and J. S. Morrow. 1986.

Mechanisms ofcytoskeletal regulation: Functional and antigenic diversity in human erythrocyte and brain beta spectrin. J. Cellular Biochem. 30:51-70.

Harris, A. L. A. D. Green, K. J. Ainger and J. S. Morrow. 1985. Mechanisms of cytoskeletal regulation Functional differences correlate with antigenic dissimilarity in human brain and erythrocyte spectrin. Biochim. Biophys. Acta 830:147-158.

Hayes, N. F. E. Holmes, J. Grantham and A. J. Baines. 1995. A60, an axonal membrane-skeletal spectrin-binding protein. Biochem Soc Trans 23:54-8.

Hemming, N. D. J. Anstee, M. A. Staricoff, M. J. A. Tanner and N. Mohandas. 1995. Identification of the membrane attachment sites for protein 4.1 in the human erythrocyte. J. Biol. Chem 270:5360-5366.

Herrmann, H. and G. Wiche. 1987. Plectin and IFAP-300K are homologous proteins binding to microtubule-associated proteins 1 and 2 and to the 240-kilodalton subunit of spectrin. J. Biol. Chem 262:1320-1325.

Holleran, E. M. K. Tokito, S. Karki and E. L. F. Holzbaur. 1996. Centractin (ARP1) associates with spectrin revealing a potential mechanism to link dynactin to intracellular organelles. J. Cell Biol.

135:1815-1829.

Hoock, T. L. L. Peters and S. E. Lux. 1997. Isoforms of ankyrin-3 that lack the NH2-terminal repeats associate with mouse macrophage lysosomes. J Cell Biol 136:1059-70.

Hu, M. Watanabe and V. Bennett. 1992. Characterization of human brain cDNA encoding the general isoform of b-spectrin. J. Biol. Chem 267:18715-18722.

Husain-Chishti, A. Levin and D. Branton. 1988. Abolition of actin-bundling by phosphorylation WO 99/16875 PCT/US98/20364 -62of human erythrocyte protein 4.9. Nature 334:718-721.

Hyvonen, M. J. Macias, M. Nilges, H. Oschkinat, M. Saraste and M. Wilmanns. 1995. Structure of the binding site for inositol phosphates in a PH domain. Embo J 14:4676-85.

Jordan, B. Puschel, R. Koob and D. Drenckhahn. 1995. J. Biol. Chem 270:29971-29975.

Joseph, S. K. and S. Samanta. 1993. Detergent solubility of the inositol trisphosphate receptor in rat brain membranes. Evidence for association of the receptor with ankyrin. J Biol Chem 268:6477-86.

Kennedy, S. S. L. Warren, B. G. Forget and J. S. Morrow. 1991. Ankyrin binds to the repetitive unit of erythroid and non-erythroid b-spectrin. J. Cell Biol. 115:267-277.

Kennedy, S. S. A. Weed, B. G. Forget and J. S. Morrow. 1994. A Partial Structural Repeat Forms the Heterodimer Self-Association Site of all b-Spectrins. J. Biol. Chem 269:11400-11408.

Koob, M. Zimmermann, W. Schoner and D. Drenckhahn. 1988. Colocalization and coprecipitation of ankyrin and Na+,K+-ATPase in kidney epithelial cells. Eur J Cell Biol 45:230-7.

Kreis, T. E. and R. Pepperkok. 1994. Coat proteins in intracellular membrane transport. Curr Opin Cell Biol 6:533-7.

Langley, R. Jr. and C. M. Cohen. 1986. Association of spectrin with desmin intermediate filaments. J Cell Biochem 30:101-9.

Lokeshwar, V. B. and L. Y. Bourguignon. 1992a. The lymphoma transmembrane glycoprotein (CD44) is a novel guanine nucleotide-binding protein which regulates GP85 (CD44)-ankyrin interaction. J Biol Chem 267:22073-8.

Lokeshwar, V. B. and L. Y. Bourguignon. 1992b. Tyrosine phosphatase activity of lymphoma (GP180) is regulated by a direct interaction with the cytoskeleton. J Biol Chem 267:21551-7.

Lokeshwar, V. N. Fregien and L. Y. Bourguignon. 1994. Ankyrin-binding domain of CD44(GP85) is required for the expression of hyaluronic acid-mediated adhesion function. J Cell Biol WO 99/16875 PCT/US98/20364 -63- 126:1099-109.

Lombardo, C. D. L. Rimm, E. Koslov and J. S. Morrow. 1994a. Human recombinant alpha-catenin binds to spectrin. Molec.Biol.Cell 5 (supp):47a.

Lombardo, C. S. A. Weed, S. P. Kennedy, B. G. Forget and J. S. Morrow. 1994b. PII-Spectrin (fodrin) and PI2-Spectrin (muscle) Contain NH2- COOH-terminal Membrane Association Domains (MADI MAD2). J. Biol. Chem 269:29212-29219.

L6pez, J. B. Leung, C. C. Reynolds, C. Q. Li and J. E. B. Fox. 1992. Efficient plasma membrane expression of a functional platelet glycoprotein Ib-IX complex requires the presence of its three subunits. J.

Biol. Chem 267:12851-12859.

Luna, E. J. and A. L. Hitt. 1992. Cytoskeleton--plasma membrane interactions. Science. 258:955-964.

Malchiodi-Albedi, M. Ceccarini, J. C. Winkelmann, J. S. Morrow and T. C. Petrucci. 1993. The 270 kDa splice variant oferythrocyte b-spectrin (PI22) segregates in vivo and in vitro to specific domains of cerebellar neurons. J Cell Sci 106:67-78.

Mangeat, P. H. and K. Burridge. 1984. Immunoprecipitation of nonerythrocyte spectrin within live cells following microinjection of specific antibodies: relation to cytoskeletal structures. J. Cell Biol.

98:1363-1377.

Marfatia, S. R. A. Leu, D. Branton and A. H. Chishti. 1995. Identification of the protein 4.1 binding interface on glycophorin C and p55, a homologue of the Drosophila discs-large tumor suppressor protein. JBiol Chem 270:715-9.

Marfatia, S. R. A. Lue, D. Branton and A. H. Chishti. 1994. In vitro binding studies suggest a membrane-associated complex between erythroid p55, protein 4.1, and glycophorin C. J Biol Chem 30 269:8631-4.

i" Mays, R. K. A. Siemers, B. A. Fritz, A. W. Lowe, G. van Meer and W. J. Nelson. 1995. Hierarchy of mechanisms involved in generating Na/K-ATPase polarity in MDCK epithelial cells. J Cell Biol 130:1105-15.

WO 99/16875 PCT/US98/20364 -64- Mische, S. M. S. Mooseker and J. S. Morrow. 1987. Erythrocyte adducin: A calmodulin regulated actin bundling protein that stimulates spectrin-actin binding. J. Cell Biol. 105:2837-2845.

Molday, L. N. J. Cook, U. B. Kaupp and R. S. Molday. 1990. The cGMP-gated cation channel of bovine rod photoreceptor cells is associated with a 240-kDa protein exhibiting immunochemical cross-reactivity with spectrin. J Biol. Chem 265:18690-18695.

Mombers, P. W. van Dijck, L. L. van Deenen, J. de Gier and A. J. Verkleij. 1977. The interaction of spectrin actin and synthetic phospholipids. Biochim Biophys Acta 470:152-60.

Mombers, A. J. Verkleij, J. de Gier and L. L. van Deenen. 1979. The interaction of spectrin-actin and synthetic phospholipids. II. The interaction with phosphatidylserine. Biochim Biophys Acta 551:271-81.

Morgans, C. W. and R. R. Kopito. 1993. Association of the brain anion exchanger, AE3, with the repeat domain of ankyrin. Cell Sci 105:1137-1142.

Morrow, J. S. (1997). Spectrins. Guidebook to the Cvtoskeletal and Motor Proteins. T. Kreis and R.

Vale, Ed. London, Oxford University Press. second, ed. in press.

Morrow, J. C. Cianci, T. Ardito, A. Mann and M. T. Kashgarian. 1989. Ankyrin links fodrin to alpha Na/K ATPase in Madin-Darby canine kidney cells and in renal tubule cells. J. Cell Biol. 108:455-465.

Morrow, J. D. L. Rimm, S. P. Kennedy, C. D. Cianci, J. H. Sinard and S. A. Weed (1997). Of Membrane Stability and Mosaics: The Spectrin Cytoskeleton. Handbook of Physiology. J. Hoffman and J.

Jamieson, Ed. London, Oxford. 485-540.

Nelson, W. J. and R. W. Hammerton. 1989. A membrane-cytoskeletal complex containing Na+,K+-ATPase, ankyrin, and fodrin in Madin-Darby canine kidney (MDCK) cells: implications for the biogenesis of epithelial cell polarity. J. Cell Biol. 108:893-902.

Nelson, W. J. and P. J. Veshnock. 1986. Dynamics of membrane-skeleton (fodrin) organization during development of polarity in Madin-Darby canine kidney epithelial cells. J. Cell Biol. 103:1751-1765.

WO 99/16875 PCT/US98/20364 Nelson, W. J. and P. J. Veshnock. 1987. Ankyrin binding to (Na+ K+)ATPase and implications for the organization of membrane domains in polarized cells. Nature 328:533-536.

Petrucci, T. C. and J. S. Morrow. 1987. Synapsin-I: An actin bundling protein under phosphorylation control. J. Cell Biol. 105:1355-1363.

Pimplikar, S. E. Ikonen and K. Simons. 1994. Basolateral protein transport in streptolysin O-permeabilized MDCK cells. J Cell Biol 125:1025-35.

Pitcher, J. J. Inglese, J. B. Higgins, J. L. Arriza, P. J. Casey, C. Kim, J. L. Benovic, M. M. Kwatra, M. G. Caron and R. J. Lefkowitz. 1992. Role of beta gamma subunits of G proteins in targeting the beta-adrenergic receptor kinase to membrane-bound receptors. Science. 257:1264-7.

Pollerberg, G. K. Burridge, K. E. Krebs, S. R. Goodman and M. Schachner. 1987. The 180-kD component of the neural cell adhesion molecule N-CAM is involved in a cell-cell contacts and cytoskeleton-membrane interactions. Cell Tiss. Res. 250:227-236.

Pradhan, P. Williamson and R. A. Schlegel. 1991. Bilayer/cytoskeleton interactions in lipid-symmetric erythrocytes assessed by a photoactivatable phospholipid analogue. Biochem 30:7754-8.

Riederer, B. M. and S. R. Goodman. 1990. Association of brain spectrin isoforms with microtubules.

FEBS Lett. 277:49-52.

Rimm, D. E. R. Koslov, P. Kebriaei, C. D. Cianci and J. S. Morrow. 1995. a,(E)-catenin is a novel actin binding and bundling protein mediating the attachment of F-actin to the membrane adhesion complex.

Proc. Natl. Acad. Sci (USA) 92:8813-8817.

Rothman, J. E. and F. T. Wieland. 1996. Protein sorting by transport vesicles. Science. 272:227-34.

Rotin, D. Bar-Sagi, H. O'Brodovich, J. Merilainen, V. P. Lehto, C. M. Canessa, B. C. Rossier and G. P. Downey. 1994. An SH3 binding region in the epithelial Na+ channel (alpha rENaC) mediates its localization at the apical membrane. Embo J 13:4440-50.

Sambrook, E. F. Fritsch and T. Maniatis (1989). Molecular Cloning: A laboratory manual. New r- WO 99/16875 PCT/US98/20364 -66- York, Cold Spring Harbor Laboratory Press.

Schekman, R. and L. Orci. 1996. Coat proteins and vesicle budding. Science. 271:1526-33.

Schmid, S. L. and H. Damke. 1995. Coated vesicles: a diversity of form and function. Faseb J.

9:1445-53.

Scott, J. K. 1992. Discovering peptide ligands using epitope libraries. Trends Biochem Sci 17:241-5.

Shore, E. M. and W. J. Nelson. 1991. Biosynthesis of the cell adhesion molecule uvomorulin (E-cadherin) in Madin-Darby canine kidney epithelial cells. J. Biol. Chem 266:19672-19680.

Siegel, D. L. and D. Branton. 1985. Partial purification and characterization of an actin-bundling protein, band 4.9, from human erythrocytes. J Cell Biol 100:775-85.

Sikorski, A. G. Terlecki, I. S. Zagon and S. R. Goodman. 1991. Synapsin I-mediated interaction of brain spectrin with synaptic vesicles. J Cell Biol 114:313-8.

Sinard, J. G. W. Stewart, A. C. Argent and J. S. Morrow. 1994. Stomatin binding to adducin: A novel link between transmembrane ion transport and the cytoskeleton. Molec.Biol.Cell 5 (supp):421a.

Smith, D. B. and K. S. Johnson. 1988. Single-step purification of polypeptides expressed in Escherichia coli as fusions with glutathione S-transferase. Gene 67:31-40.

Smith, P. A. L. Bradford, E. H. Joe, K. J. Angelides, D. J. Benos and G. Saccomani. 1993. Gastric parietal cell H(+)-K(+)-ATPase microsomes are associated with isoforms of ankyrin and spectrin. Am JPhysiol 264:C63-70.

Smith, P. G. Saccomani, E. H. Joe, K. J. Angelides and D. J. Benos. 1991. Amiloride-sensitive sodium channel is linked to the cytoskeleton in renal epithelial cells. Proc Natl Acad Sci US A 88:6971-5.

Srinivasan, M. Lewallen and K. J. Angelides. 1992. Mapping the binding site on ankyrin for the voltage-dependent sodium channel from brain. J Biol Chem 267:7483-9.

WO 99/16875 PCT/US98/20364 -67- Srinivasan, Y. E. J. Davis, V. Bennett and K. Angelides. 1988. Ankyrin and spectrin associate with voltage-dependent Na channels in brain. Nature 333:177-180.

Touhara, J. Inglese, J. A. Pitcher, G. Shaw and R. J. Lefkowitz. 1994. Binding of G protein beta gamma-subunits to pleckstrin homology domains. J. Biol. Chem 269:10217-10220.

Wang, D. S. and G. Shaw. 1995. The association of the C-terminal region of beta I sigma II spectrin to brain membranes is mediated by a PH domain, does not require membrane proteins, and coincides with a inositol-1,4,5 triphosphate binding site. Biochem Biophys Res Commun 217:608-15.

Weed, S. P. R. Stabach, C. E. Oyer, P. G. Gallagher and J. S. Morrow. 1996. The Lethal Hemolytic Mutation in PI2 Spectrin Providence Yields a Null Phenotype in Neonatal Skeletal Muscle. Lab.

Investig. 74:1117-1129.

Winardi, M. Reid, J. Conboy and N. Mohandas. 1993. Molecular analysis of glycophorin C deficiency in human erythrocytes. Blood 81:2799-803.

Winkelmann, J. J. G. Chang, W. T. Tse, A. L. Scarpa, V. T. Marchesi and B. G. Forget. 1990a.

Full-length sequence of the cDNA for human erythroid beta-spectrin. J. Biol. Chem 265:11827-11832.

Winkelmann, J. F. F. Costa, B. L. Linzie and B. G. Forget. 1990b. Beta spectrin in human skeletal muscle. Tissue-specific differential processing of 3' beta spectrin pre-mRNA generates a beta spectrin isoform with a unique carboxyl terminus. J Biol Chem 265:20449-54.

Winkelmann, J. T. L. Leto, P. C. Watkins, R. Eddy, T. B. Shows, A. J. Linnenbach, K. E. Sahr, N. Kathuria, V. T. Marchesi and B. G. Forget. 1988. Molecular cloning of the cDNA for human erythrocyte beta-spectrin. Blood 72:328-334.

Winograd, D. Hume and D. Branton. 1991. Phasing the conformational unit of spectrin. Proc.

Natl. Acad. Sci (USA) 88:10788-10791.

Wolfe, L. K. M. John, J. C. Falcone, A. M. Byme and S. E. Lux. 1982. A genetic defect in the binding of protein 4.1 to spectrin in a kindred with hereditary spherocytosis. N.Eng.J. Med. 307:1367-74.

WO 99/16875 PCT/US98/20364 -68- Zhang, P. Devarajan, A. L. Dorfinan and J. S. Morrow. 1997a. Structure of the Ankyrin Binding Domain of a-Na,K-ATPase. submitted Zhang, P. Devarajan and J. S. Morrow. 1997b. Crystal Structure of Ankyrin Binding Domain of Alpha Subunit. Am. Crystal Assoc. Newslet. winter:July 19-25, St. Louis, MO (abstract).

Claims (33)

1. A method to modulate the intracellular sequestration of a selected integral membrane or secretory protein into a transport vesicle for transport from the endoplasmic reticulum to the cis-Golgi apparatus, or from the cis-Golgi to the medial-Golgi apparatus, or from the medial-Golgi to the trans-Golgi apparatus or from the trans-Golgi apparatus to the plasma cell membrane, comprising the steps of: selecting a compound effective to enhance or inhibit the binding of the selected protein to the transport vesicle; and contacting the cell with said compound, whereby the sequestration of the protein is enhanced if binding is enhanced or inhibited if binding is inhibited.

2. The method of claim 1, wherein the binding and sequestration of the selected protein is inhibited and wherein said compound is selected from the group consisting of: i) a modified or truncated skeletal component of the transport vesicle, ii) an associated adapter protein, or iii) a small molecule mimetic of either i) or ii).

3. The method of claim 1, wherein the sequestration of the selected protein is inhibited and wherein said compound is a modified or truncated adapter protein, or a small molecule mimetic thereof, which protein or mimetic mediates binding or association of the selected protein to a skeletal component of the transport vesicle.

4. The method of claim 3, wherein the adapter protein is ankyrin and the skeletal component of the transport vesicle is spectrin. The method of claim 4, wherein the ankyrin is AnkG 19. WO 99/16875 PCTIUS98/20364

6. The method of claim 2 or 3, wherein the skeletal component or associated adapter protein is selected from the group consisting of ADP-ribosylation factor (ARF), ankyrin, AnkGl19 spectrin, spectrin PI, spectrin .PIE*, centractin and tubulin and small molecule mimetics of any of the foregoing.

7. The method of any of claims 1-5, wherein the transport of the selected protein from the endoplasmic reticulum to the cis-Golgi apparatus is inhibited.

8. A method to determine whether a candidate compound inhibits or enhances the transport of a selected integral membrane or secretory protein via a transport vesicle from the endoplasmic reticulum to the cis-Golgi apparatus, or from the cis-Golgi to the medial-Golgi apparatus, or from the medial-Golgi to the trans-Golgi apparatus or from the trans-Golgi apparatus to the cell membrane, comprising the steps of: contacting a cell in which the selected protein is expressed with the candidate compound; and determining whether intracellular vesicular transport of the selected protein is enhanced or inhibited.

9. The method of claim 8 wherein the determination that inhibition of intracellular transport has occurred is made by: i) detecting a substantial decrease in the presence of the selected protein in the cell membrane or in the extracellular medium; or ii) detecting the dispersed intracellular presence of the selected protein outside of its normal pattern of transport to and through the Golgi apparatus. A method to determine whether the vesicular transport of a selected integral membrane protein or secretory protein from the endoplasmic reticulum to and through the Golgi apparatus is mediated by the binding of the selected protein to a WO 99/16875 PCT/US98/20364 -71- component of the skeleton of the vesicle, comprising the steps of: treating cells that express the selected protein with an agent that interferes with the sequestration of integral membrane proteins or secretory protein into transport vesicles; and determining whether the selected protein is found substantially in its normal distribution pattern in said cells, whereby an abnormal distribution pattern indicates mediated transport of the vesicle.

11. The method of claim 10 in which said agent is one of a series of progressively more truncated portions of a protein that is a structural component of the vesicular skeleton or is an associated adapter protein, and said cells separately are transfected with cDNAs encoding said series of vesicular skeleton or associated adapter proteins.

12. The method of claim 11, wherein said vesicular skeleton or associated adapter proteins are selected from the group consisting of ADP-ribosylation factor (ARF), ankyrin, AnkG119 spectrin, spectrin PI, spectrin PIE*, centractin and tubulin.

13. The method of claim 10, wherein the step of determining whether the distribution pattern of the selected protein is substantially normal includes an assessment of whether there is: i) a substantial decrease in the presence of the selected protein in its normal intracellular distribution pattern; ii) or a substantial decrease in the presence of the selected protein on the outer cell membrane or in the extracellular media of said cells.

14. The method of claim 12, wherein said vesicular skeleton protein is spectrin. The method of claim 14, wherein said spectrin is pI Z* spectrin. WO 99/16875 PCT/US98/20364 -72-

16. The method of claim 12, wherein said adapter protein is ankyrin.

17. The method of claim 16, wherein said ankyrin is AnkGl 19.

18. A group of PIE* spectrin cDNA constructs of varying lengths that each encode one of a complimentary and overlapping series of truncated PIT spectrin peptides.

19. A set of individual clonal cell lines each expressing one of the PIE spectrin cDNA constructs of claim 18. An isolated and purified nucleic acid fragment encoding a PIII spectrin.

21. The nucleic acid of claims 20, wherein the Pill spectrin comprises the amino acid sequence of Figure 19B.

22. The nucleic acid of claim 20 comprising the sequence of Figure 19A.

23. An isolated nucleic acid that specifically hybridizes to the nucleic acid of claim 22 under stringent conditions.

24. A vector comprising a nucleic acid of any one of claims 20 to 23. A host cell comprising the vector of claim 24.

26. A method to ameliorate the effects of a disease caused by an relative excess 73 of a particular integral membrane protein, comprising the administration of a compound in a unit dose that is effective to inhibit the intracellular sequestration of a selected integral membrane or secretory protein into a transport vesicle for transport from the endoplasmic reticulum to the cis-Golgi apparatus, or from the cis-Golgi to the medial-Golgi apparatus, or from the medial-Golgi to the trans- Golgi apparatus or from the trans-Golgi apparatus to the cell membrane.

27. A method to ameliorate the effects of a disease caused by an relative deficit in the amount of a particular integral membrane protein, comprising the administration of a compound in a unit dose that is effective to enhance the intracellular sequestration of a selected integral membrane or secretory protein into a transport vesicle or the transport thereof from the endoplasmic reticulum to the cis-Golgi apparatus, or from the cis-Golgi to the medial-Golgi apparatus, or from the medial-Golgi to the trans-Golgi apparatus or from the trans-Golgi apparatus to the cell membrane.

28. A method of determining whether vesicular transport of a selected protein is mediated by the binding of the protein to a component of spectrin- ankyrin-adapter protein trafficking system (SAATS) comprising: 20 treating cells that express the selected protein with an agent that interferes with sequestration of proteins into transport vesicles; obtaining a distribution pattern of the selected protein in the cells; obtaining a distribution pattern of the selected protein in untreated cells that express the protein; and, comparing the distribution pattern of the protein in the treated cells and the untreated cells, wherein a difference in the distribution pattern indicates that transport of the protein is mediated by the binding of the protein to SAATS.

29. The method of claim 28, wherein the component of SAATS is spectrin or 30 an adapter protein.

30. The method of claim 29, wherein the adapter protein is selected from the group consisting of ankyrin, protein 4.1, adducin, and catenin.

31. The method of claim 30, wherein the ankyrin is ankyrin G119 (Ank

32. The method of claim 29, wherein the spectrin is selected from the group consisting of spectrin p1 spectrin p111, and spectrin 13I*.

33. The method of claim 28, wherein treating the cells with an agent comprises transfecting the cells with nucleic acid encoding the agent.

34. The method of claim 28 or 33, wherein the agent is a truncated component of SAATS.

35. The method of claim 34, wherein the truncated component of SAATS is a truncated spectrin or a truncated adapter protein.

36. The method of claim 35, wherein the truncated adapter protein is selected from the group consisting of truncated ankyrin, truncated protein 4.1, truncate adducin, and truncated catenin.

37. The method of claim 36, wherein the truncated ankyrin is truncated Ank G119. 20 38. The method of claim 35, wherein the truncated spectrin is selected from the group consisting of truncated spectrin 1p, truncated spectrin 11pill, and truncated spectrin p3I*.

39. The method of claim 28, wherein vesicular transport is selected from the group consisting of transport from endoplasmic reticulum to the cis-Golgi apparatus, transport from cis-Golgi to the medial-Golgi apparatus, transport from the medial Golgi to the trans-Golgi apparatus, transport from the trans- Golgi apparatus to the cell membrane, and transport from the trans-Golgi apparatus to other cellular compartments. Dated this thirtieth day of May 2002 Yale University Patent Attorneys for the Applicant: F B RICE CO