The Involvement of Ankyrin in the Regulation of Inositol 1,4,5-Trisphosphate Receptor-mediated Internal Ca2+ Release from Ca2+ Storage Vesicles in Mouse T-lymphoma Cells*

Mouse T-lymphoma cells contain a unique type of internal vesicle which bands at the relatively light density of 1.07 g/cc. These vesicles do not contain any detectable Golgi, endoplasmic reticulum, plasma mem- brane, or lysosomal marker protein activities. Binding of [3H]inositol 1,4,5-trisphosphate (1P3) to these internal vesicles reveals the presence of a single, high affin- ity class of IP3 receptor with a dissociation constant (&) of 1.6 2 0.3 IIM. Using a panel of monoclonal and polyclonal antibodies against IPS receptor, we have established that the IPS receptor (-260 kDa) displays immunological cross-reactivity with the rat brain IP3 receptor. Polymerase chain reaction analysis of first- strand cDNAs from both mouse T-lymphoma cells and rat brain tissues reveals that the IPS receptor tran- script in mouse T-lymphoma cells belongs to the short form (non-neuronal form) and not the long form (neu- ronal form) detected in rat brain tissue. Scatchard plot analysis shows that high affinity binding occurs between ankyrin and the IPS receptor with a Kd of 0.2 nM. Most importantly, the binding of ankyrin to the light density vesicles

to these internal vesicles reveals the presence of a single, high affinity class of IP3 receptor with a dissociation constant (&) of 1.6 2 0.3 IIM. Using a panel of monoclonal and polyclonal antibodies against IPS receptor, we have established that the IPS receptor (-260 kDa) displays immunological cross-reactivity with the rat brain IP3 receptor. Polymerase chain reaction analysis of firststrand cDNAs from both mouse T-lymphoma cells and rat brain tissues reveals that the IPS receptor transcript in mouse T-lymphoma cells belongs to the short form (non-neuronal form) and not the long form (neuronal form) detected in rat brain tissue.
Scatchard plot analysis shows that high affinity binding occurs between ankyrin and the IPS receptor with a Kd of 0.2 nM. Most importantly, the binding of ankyrin to the light density vesicles significantly inhibits IP3 binding and IP3-induced internal Ca2+ release. These findings suggest that the cytoskeleton plays a pivotal role in the regulation of IPS receptormediated internal Ca" release during lymphocyte activation.
Lymphocyte activation is initiated as a consequence of ligand-receptor binding which often generates the onset of phospholipase C activity (1)(2)(3). Hydrolysis of phosphatidylinositol 4,5-bisphosphate by phospholipase C results in the formation of two intracellular second messengers, diacylglycerol, and inositol 1,4,5-trisphosphate (IP3)' (4). Diacylglycerol is an essential cofactor in activating members of the protein kinase C family of serine/threonine kinases (5). IPS is a physiological ligand known to mediate internal Ca2+ release from intracellular Ca2+ storage sites by binding to a specific receptor on certain intracellular membrane vesicles (4).
The IP3 receptor has been identified by a number of investigators based on the specific binding of [3H]IPs to internal vesicles in smooth muscle cells (6) and in several nonmuscle cell types such as liver (7), adrenal cortical cells (8,9), brain cerebellum (lo), and cerebellum Purkinje fibers (11). The primary intracellular storage sites for internal Ca2+ were originally thought to be part of the endoplasmic reticulum (12)(13)(14). However, subcellular fractionation studies revealed that the distribution of Ca2+-pumping, Ips-responsive organelles does not correlate with markers for plasma membrane, endoplasmic reticulum, mitochondria, Golgi apparatus, nor any other known organelles (15). Consequently, the IP3-responsive vesicles appear to be a unique type which have been designated as "calciosomes" (16). Recently, an IPS receptor has been reported to exist on the plasma membrane of human T-lymphocytes (17). Since the structural and functional properties of this plasma membrane-associated IP3 receptor have not been fully established, one cannot preclude the possibility that this surface IPS receptor may be structurally or functionally related to several different plasma membrane receptors for inositol 1,3,4,5-tetrakisphosphate (IP,) or inositol hexakisphosphate (IPS) (18,19). Consequently, the subcellular location of the IP3 receptor still remains to be determined. In mouse T-lymphoma cells, activation of phospholipase C by either a Gi,-like protein (3) or tyrosine kinase(s) (2) has also been shown to generate IP3. Although an Ips appears to be required for inducing internal Ca2+ release, the formation of receptor patching/capping (3,20) and the general activation of lymphocytes (1,2), very little is known at the present time concerning the nature of the IP3 receptor and IP3-inducible internal storage sites in lymphocytes. In this study, we have isolated and partially characterized a mouse T-lymphoma IP3 receptor from a unique type of internal vesicle which bands at the relatively light density of 1.07 g/ml. This IPS receptor, which shares immunological cross-reactivity with the brain IP, receptor, displays both high affinity IP, binding and Ca2' ion channel properties. Most importantly, the interaction between mouse T-lymphoma IP3 receptor and ankyrin, an important membrane-associated cytoskeletal protein, significantly inhibits IP3 binding and IPS-mediated internal Ca2+ release.

Cell Culture
The mouse T-lymphoma BW 5147 cell line (an AKR/J lymphoma line) were grown at 37 "C in 5% C02/95% air using Dulbecco's

Cellular Fractionation
The cells (suspended in 50 ml of ice-cold buffer consisting of 15 mM KCI, 1.5 mM Mg(OAc)z, 1 mM dithiothreitol (DTT) and 10 mM HEPES (pH 7.0)) were disrupted by nitrogen cavitation in an Artisan homogenizer (Artisan Industries, Inc., Waltham, MA) held at 0 "C using a pressure of 60 p s i . for 15 min. After disruption, 0.10 volume of 700 mM KC1,40 m M Mg(OAc)z, 1 mM DTT, and 400 mM HEPES (pH 7.0) was added, and nuclei were removed by centrifugation at 500 X g,, for 4 min. The resulting supernatant was layered on a discontinuous sucrose gradient consisting of 0, 15, 25, 35, 40, and 50% sucrose (w/w) in a buffer containing 10 mM HEPES (pH 7.0), 50 mM KCl, 1 mM DTT, and 2 mM MgCl,. The gradient was centrifuged in a Beckman SW28 rotor at 25,000 rpm for 16 h. The membranous materials located in various sucrose layers were collected for further biochemical analyses including enzyme marker assays, IPB binding assays, Ca2+ flux measurement, and immunoblotting techniques as described below.
Enzyme Marker Assays Na+/K+-ATPase activity was used as a specific enzyme marker for plasma membrane (21,22). NADPH-dependent cytochrome c reductase and sulfatase C activities were used as independent markers of endoplasmic reticulum (9,23). Galactosyltransferase activity, which was used as a Golgi marker was assayed according to the procedures described previously (24). P-N-Acetylglucosaminidase was used as a lysosomal marker (25,26). Protein concentrations were determined using the Bio-Rad protein assay kit.

Transmission Electron Microscopy
Lymphoma light density vesicles collected from the 15-25% sucrose interface (according to the procedures described above), were fixed with 2% glutaraldehyde in phosphate-buffered saline (pH 7.3), postfixed with 1% OsO,, dehydrated through a graded ethanol series, and embedded in Spurrs embedding medium. Ultrathin sections were cut on a Sorvall MT2-B ultramicrotome, stained with uranyl acetate and lead citrate, and examined using a JEOL electron microscope at 80 kV.

Immunoblotting Techniques
Isolated lymphoma light density vesicles (collected from 15-25% sucrose interface) and brain IP, receptor (obtained from rat cerebellar membranes) (10) were either directly spotted on a sheet of nitrocellulose paper or analyzed by a 7.5% polyacrylamide gel electrophoresis followed by transferring to nitrocellulose sheets. Subsequently, the nitrocellulose sheets were incubated with the following various immunoreagents such as monoclonal mouse anti-IP3 receptor (IPR.1; it recognizes the C terminus cytoplasmic domain of IP3 receptor; full characterization of this monoclonal antibody will be published elsewhere'), polyclonal rabbit anti-brain IP3 receptor (a gift from Dr. Thomas Sudhof, University of Texas Southestern Medical School, Dallas, TX), monoclonal mouse anti-calsequestrin or sheep anticalreticulin (kindly provided by Dr. D. H. MacLennan, university of Toronto, Canada). Nonimmune sheep serum, mouse serum, and rabbit serum were used as controls. Immunolabeled nitrocellulose sheets were subsequently incubated with biotinylated secondary antibodies and avidin H plus biotinylated horseradish peroxidase to form avidin biotin complex and to develop color reaction in the presence of diaminobenzidine with hydrogen peroxide.

Immunofluorescence Staining
Mouse T-lymphoma cells fixed in 2% paraformaldehyde were either surface-labeled with monoclonal mouse anti-IPa receptor antibody or rendered permeable by freezing (at -20 "C) and thawing (at room temperature) in the presence of 90% ethanol followed by staining with monoclonal mouse anti-IPS receptor antibody. These samples were then labeled with fluorescein-conjugated goat anti-mouse IgG. To detect nonspecific antibody binding, cells were incubated with nonimmune mouse serum followed by fluorescein-conjugated goat anti-mouse IgG. No staining was observed in such control samples.
H. Jin and L. Y. W. Bourguignon, submitted for publication.

Purification of IPS Receptor
Lymphoma light density vesicles, collected from the 15-25% sucrose interface (according to the procedures described above), were solubilized by adding Triton X-100 to a final concentration of 1% (v/ v). Subsequently, the solubilized material was passed through a heparin-agarose column which was washed with 20 ml of a buffer (50 mM Tris-HC1 (pH 8.3), 1 mM EDTA, 1 mM P-mercaptoethanol) plus 0.1% Triton X-100 and 0.25 M NaC1. The materials bound to the heparin-agarose were eluted with 3 ml of 50 mM Tris-HC1 (pH 7.71, 1 mM 0-mercaptoethanol, 0.1% Triton X-100,0.5 M NaC1; and then applied to an anti-IP3 receptor (IPR.l)-conjugated affinity column. The IP, receptor was eluted from the column with a solution containing 0.05 M diethylamine (pH ll.O), 10 mM EDTA, and 0.05% Triton X-100. Purity of the IPS receptor preparations was confirmed by SDS-polyacrylamide gel electrophoresis and silver staining. Purified IPS receptor was subsequently used for the incorporation into phospholipids (liposomes) and Ca2+ flux measurements as described below.
[3H]IP3 Binding Assay Specific [3H]IP3 binding was determined by the method described by Guillemette (8). Aliquots of each fraction (e.g. materials from the interface of 0-15, 15-25, 25-35, 35-40, and 40-50% sucrose layers) were incubated for 10 min at 4 "C in a 0.5 ml of a medium containing 25 mM NazHP04, 100 mM KCI, 20 mM NaCI, 1 mM sodium EDTA, 1 mg/ml bovine serum albumin, and 0.05 pCi of [,H]IP3 (34.2 Ci/mmol, Amersham) at pH 7.4. In some cases, low density vesicles were pretreated with either ankyrin (10 pg/ml) or monoclonal anti-IP, receptor antibody (IPR.l; 10 pg/ml) followed by [3H]IP3 binding. Binding was estimated in the presence of various concentrations of unlabeled 1P3 ranging from lo-" M to 1O"j M. The binding reaction was terminated by adding 2.5 ml of cold phosphate-buffered saline (pH 7.4) and filtrating through GF/B glass fiber filters that had been presoaked in phosphate-buffered saline containing 1% bovine serum albumin. The filter-associated radioactivity was analyzed by liquid scintillation counting.

Ankyrin Binding Assay
Human erythrocyte ankyrin was purified by the procedure of Bennett and Stenbuck (27) and labeled with Nalz5I using IODO-GEN beads. '"1-Ankyrin (1-10 ng) was incubated with a nitrocellulose sheet coated with purified IP, receptor (obtained from anti-IP3 receptor affinity column chromatography according to the procedures described above) in a binding solution containing 20 mM Tris-HC1 (pH 7.4), 150 mM NaCl, 0.05% Triton X-100 and 0.1% bovine serum albumin for 30 min at room temperature. Following incubation, the nitrocellulose sheets were washed five times with the same binding solution and analyzed by dot assays followed by autoradiographic analysis or counted in a y counter (for Scatchard plot analysis). Background or nonspecific binding was determined by including a large excess of unlabeled ankyrin (at least 10-100-fold excess) in both dot assays and Schatchard plot analysis. The results were expressed as "specific binding" in which the background level of binding was subtracted. 10 mM phosphocreatine/ creatine kinase (10 units/ml) (Boehringer Mannheim), 3.75 FM ruthenium red, 1 mM Mg-ATP, 0.5 mM EGTA. CaClz was added to this solution to generate a range of free Ca2+ concentration between 100 and 300 nM. Subsequently, "Ca" (5-10 pCi/ml; 50 Ci/pg; Amersham) and light density vesicles (0.5 mg/ml) were added to the reaction mixture at 30 "C for 25 min. In Ca2+ release experiments, IP, (10-100 nM) was added to these 45Ca2+-containing vesicles. The maximal amount of Ca2+ release occurred 10 s after the addition of IP,. In some cases, low density vesicles were pretreated with either ankyrin (10 pglml) or monoclonal anti-IP, receptor antibody (IPR.l; 10 pg/ ml) followed by the addition of IPS (10-100 nM) to the reaction mixture for ca2' flux measurements. The amount of Ca2+ released from the light density vesicles was determined hy a filtration method using Millipore filters (HAWP, 0.45 pm) and washing with a buffer containing 120 mM KC1 and 20 mM Tris-HEPES (pH 7.2).

Ca" Flux Measurement in IP3 Receptor-containing Phospholipid
Vesicles (Liposomes)-Purified IPS receptor (obtained from the pro-IP3 Receptor and Ankyrin Interaction in T-lymphoma Cells cedures described above) (100 pg/ml) was incorporated into phosphatidylcholine/phosphatidyserine vesicles (a ratio of 1:1 IPS receptor:phospholipids) in 1% CHAPS followed by dialysis against a buffer containing 20 mM Tris-HC1 (pH 7.4) at 25 "C, 100 mM NaCl, 100 mM KCI, 2 mM 8-mercaptoethanol) for 72 h. These IPS receptor-containing phospholipid vesicles (liposomes) were used to measure either IP, binding or IPS-induced Ca2+ release. The Ca2+ flux measurement was initiated by adding 2 pCi of T a 2 + to the receptor-containing liposomes in the presence or absence of IP3 (10 nM) at 30 "C. The Ca2+ flux measurement was terminated by adding a solution containing 0.5 mM CaCI2, 5 mM MgSO,, and 100 pg/ml heparin and external 45Ca2+ was removed by Dowex 50WX (Sigma) (28). The intravesicular 46 Ca2+ was counted by liquid scintillation counting.

SDS-Polyacrylamide Gel Electrophoresis and Autoradiographic
Analyses Electrophoresis was conducted using a 7.5 or 10% SDS-polyacrylamide gel electrophoresis slab gel and the discontinuous buffer system described by Laemmli (30). For autoradiographic analysis, all gels were vacuum-dried and exposed to Kodak x-ray film (X-Omat XAR-5).

Subcellular Localization of IP3 Receptor in Mouse Tlymphoma Celk
In striated muscle intracellular Ca2+ is released from a specialized intracellular organelle, the sarcoplasmic reticulum (31). This structure contains an ATP-dependent Ca2+ pump and several Ca2+-binding proteins such as calsequestrin (32) and calreticulin (33). Ca2+ is released from the sarcoplasmic reticulum through a ryanodine-sensitive channel (homotetramer, molecular mass of each subunit is -565 kDa) which may be modulated by small molecules and calmodulin as well as by muscle contraction (34-37). The observations of a membrane organelle with a Ca" pump, Ca2+-binding proteins and a Ca2+ release channel led to the hypothesis that nonmuscle cells may possess a special sarcoplasmic reticulumlike organelle generally referred to as the "calciosome" (16).
In order to determine (i) whether mouse T-lymphoma cells actually contain "calciosome-like" structures, and (ii) how internal Ca" release is regulated in lymphocytes, we decided to establish a convenient and effective cellular fractionation procedure to isolate and identify those vesicles responsible for internal CaZ+ release. Mouse T-lymphoma cells were homogenized and fractionated by differential centrifugation followed by density gradient centrifugation on a discontinuous sucrose gradient as described under "Materials and Methods." Our results show the initial separation of various lymphoma membranes including soluble proteins (fraction A: 0-15% sucrose interface) (Fig. la), "light density" membrane vesicles (fraction B: 15-25% sucrose interface) (Fig. la); Golgi membranes (fraction C: 25-35% sucrose interface) (Fig. 1, a and e); endoplasmic reticulum ( Fig. 1, a and f ) and plasma membranes (fraction D: 35-45% sucrose interface) (Fig. 1, a and   g); and lysosomal membranes and other large particles (membrane-bound ribosomes) (fraction E: 40-50% sucrose layer) ( Fig. 1, a and h ) as determined by enzyme marker analyses (Fig. 1, e-h).
Further analysis of the various sucrose gradient fractions shows that the IP3 binding site is preferentially located in the light density vesicle fraction (fraction B, 15-25% sucrose interface) (Fig. Id) which represents approximately 10% of total membrane protein. Since there is a large amount of protein in both fractions D and E, the specific activity of IP3 binding detected in these fractions is very low compared with that in the B band (Fig. Id). Recently, a plasma membraneassociated IP3 receptor has been reported in Jurkat lymphocytes by Snyder and co-workers (17). It is possible that the IP3 binding sites detected in C and D may represent a plasma membrane-associated species of IP3 receptor. The biochemical similarities and/or differences between the "light density vesicle" IP3 receptor and the plasma membrane-associated IPS receptor awaits further analysis. A number of previous studies suggest that the intracellular IP3 binding sites are primarily located at the endoplasmic reticulum (12-14). The fact that the majority of IP3 binding is detected in the light density vesicle fraction and not in the ER may indicate that there is a unique cellular location (possibly in a cell-specific manner) for IPB receptors in lymphocytes.
In this paper, we have primarily focused on the high specific activity class of IPS binding sites detected in the light density vesicles (fraction B with a density of approximately 1.07 g/ ml) (Fig. 1, a and d). Morphological analysis using electron microscopic techniques reveals that these light density vesicles are smooth and heterogeneous in size; and occasionally they are present in tubular shapes (Fig. lb). These vesicles have negligible amounts of marker enzyme activities derived from Golgi, endoplasmic reticulum, plasma membrane and lysosomes (Fig. 1, e-h). We have also determined that these low density vesicles are different from endosome structures using mannose 6-phosphate receptor internalization as a marker assay (25) (data not shown). Furthermore, using specific antibodies raised against calsequestrin (Fig. IC (i)) and calreticulin (Fig. IC (ii)), we have found that the light density vesicles (fraction B, 1 5 2 5 % sucrose interface) contain protein molecules immunologically cross-reactive with these two Ca*+-binding proteins (molecular masses of 63 and 55 kDa, respectively). We believe that this Western blot data are specific based on the facts that (a) the molecular masses of both calsequestrin (63 kDa) and calreticulin (55 kDa) in lymphocyte light density vesicles are similar to those found in sarcoplasmic reticulum of striated muscle (32, 33); and ( b ) nonimmune serum gives a negligible nonspecific reaction (data not shown). In addition, we have tested whether IPB can induce the release of Caz+ from the light density vesicles.

TABLE I
Effects of monoclonal anti-IPS receptor on "Ca2+ release from lymphoma light density vesicles The amount of Ca2+ released for A23187-treated vesicle was designated as the maximal level of Ca2+ release signal and experimental figures were compared to this maximal release. The incubation time for '%a2+ release was 10 s. The data shown are the averages of triplicate determinants. which varied bv less than 5% (n = 5).  (Fig. 2). The binding affinity of the light density vesicles for [3H]IP3 is comparable to those reported for internal vesicles isolated from other cell types (6).

Immunological Analyses of Mouse T-lymphoma IPS Recep-
tor-The IP3 receptor in nonmuscle cells has been shown to be a homotetramer (2500 amino acid subunits, molecular mass -260 kDa) with limited homology to the ryanodine receptor in striated muscle (38, 39). Using a polyclonal rabbit anti-IPS receptor antibody, we have determined that the T-lymphoma light density vesicles (Fig. 3, A and B ) contain an IPS receptor analogous to brain IP3 receptor (Fig. 3, D and E ) with a molecular mass of -260 kDa. In addition, a monoclonal mouse anti-IP3 receptor antibody (IPR.l) which does not interfere IP3 binding (Table 11) and recognizes a short sequence at the C terminus cytoplasmic side of IP3 receptor was also used.
Similar immunocross-reactivity was observed using this newly developed IP3 receptor antibody (Fig. 3, C and F).
Using immunofluorescence staining, we have found that the IP3 receptor is preferentially associated with numerous vesicular structures located in the cytoplasm (Fig. 4B). Al lymphoma light density vesicle membrane proteins and rat brain cerebellum membrane proteins. A , total lymphoma light density vesicle membrane proteins. B, immunoblot of lymphoma light density vesicle membrane proteins with polyclonal rabbit anti-rat brain IP3 receptor antibody. (As a control, nonimmune rabbit serum was used. No staining was detected on these samples (data not shown).) C, silver staining of purified lymphoma IPS receptor obtained from monoclonal mouse anti-IPS receptor-conjugated affinity column.
(As a control, nonimmune mouse serum-conjugated column was used. No protein was detected on these columns (data not shown).) D, total rat brain cerebellum membrane proteins. E, immunoblot of rat brain cerebellum membrane proteins with polyclonal rabbit anti-rat brain IPS receptor antibody. F, silver staining of purified rat brain IPS receptor obtained from monoclonal mouse anti-IPR receptor-conjugated affinity column.
obvious endoplasmic reticulum-like network structure is observed. In order to verify that our immunofluorescence staining is specific for the intracellular IP3 receptor and not surface-bound IP3 receptor, we also carried out this immunolabeling procedure using intact cells (i.e. without any permeabilization procedures) as a control. Our data clearly indi-

-L o S ( I P 3 ) ( M )
TABLE I1 Effects of ankyrin and monoclonal anti-IPS receptor on [3HlIP~ binding in low density vesicles Low density vesicles (either pretreated with ankyrin (10 pg/ml) or monoclonal anti-IPS receptor antibody (10 pg/ml) or without any treatment) were incubated with [3H]IP3 as described under "Materials and Methods." cate that no cell surface label is detected using our newly developed monoclonal anti-IP3 antibody (IPR.l) (Fig. 4A).
Most importantly, binding of this monoclonal anti-IPS receptor antibody to 260-kDa protein-containing low density vesicles induces a significant amount of internal ca'+ release analogous to IP3-mediated Ca2+ stimulation (Table I). We believe this anti-IP3 receptor antibody-induced Ca2+ release is specific, since nonimmune normal mouse IgG shows no stimulation on Ca2+ release activity (Table I). There is no additive effect on the stimulation of internal Ca2+ release if both a monoclonal anti-IP3 receptor antibody (IPR.l) and IPS are added together to the low density vesicles (Table I). These results suggest that the epitope which this monoclonal anti-IP3 antibody (IPR.l) recognizes must be very close to the proposed Ca2+ channel region in the C terminus of the IP3 receptor protein (40). Also, the binding of this antibody to IPS receptor possibly induces a conformational change of the receptor which mimics the effect of IP3 binding to the receptor. Together, these findings clearly indicate that the lymphoma 260-kDa protein is an IP3 receptor-like molecule (Fig.  3, C and F).
Demonstration of IP3-induced Ca" Flux in Lipid Vesicles Reconstituted with the Putative 260-kDa ZP3 Receptor-The first direct demonstration that IP3 induces the openings of Ca'+ channels was obtained in planar lipid bilayers into which vesicles made from aortic smooth muscle sarcoplasmic retic-

FIG. 4. Immunofluorescence staining of intracellular IPS receptors in mouse T-lymphoma cells using fluorescein-conjugated monoclonal anti-IP3 receptor.
A, staining of surface exposed IPS receptor (note that only a background level of label was detected); B, staining of intracellular IPS receptor (note that a significant amount of label was detected in vesicular structures). ulum were incorporated (28,40). Subsequently, purified brain IP3 receptor has been shown to mediate Ca2+ flux in reconstituted lipid vesicles (28). In this study we have used anti-IP3 receptor-affinity chromatography to obtain a highly purified 260-kDa protein (IP, receptor-like molecule) fraction (Fig.  3C). Subsequently, this purified 260-kDa protein was incorporated into synthetic phospholipid vesicles (liposomes). Our data indicate that Ca2+ flux activity in these 260-kDa proteincontaining lipid vesicles responds to IPS at the normal in uiuo concentrations (-10 nM) (Fig. 5). This functional evidence further supports the contention that the lymphoma 260-kDa protein contains both the IP3 binding site and the Ca2+ ion channel property.
Detection of the IP3 Receptor Transcripts by MAPPing Techniques-Since two different IP3 receptor transcripts derived by alternative splicing have been identified and shown to be expressed in a tissue-specific manner (neuronal uersus non-neuronal tissues) (29), total RNA materials isolated from mouse T-lymphoma cells and rat brain cerebellum were analyzed for the presence of such specific transcripts by MAPPing techniques. MAPPing was developed to analyze RNAs in small numbers of cells. This technique utilizes reverse transcription of total cellular RNA to synthesize complementary DNA, followed by the polymerase chain reaction to specifically amplify DNA fragments of interest.
Previously, MAPPing techniques from various rat tissues have revealed two distinct IP3 receptor transcripts (29). A long form (also considered to be neuronal-specific) contains a 120-nucleotide insert between the two CAMP-dependent protein kinase phosphorylation consensus sequences, and is predominantly detected in adult brain tissues. A short form (also considered to be a non-neuronal type) lacks the insert, and is primarily found in fetal brain and peripheral tissues (29). Specific oligonucleotide primers were utilized in the PCR reactions to differentiate the long neuronal type from the short non-neuronal type of the IP3 receptor transcripts. Our results with mouse T-lymphoma cells indicate the existence OO -/ , 10 of the short form of the transcript (Fig. 6, lane 1 ); whereas the IP3 receptor transcripts in the rat brain cerebellum appear to display the long form (Fig. 6, lane 2). These PCR products were further cloned using a TA cloning kit (Invitrogen Corp.) and sequenced. The nucleotide sequence data confirms that the PCR-amplified fragments represent the segment of the IPS receptor cDNA reported previously (29). As a control, the house keeping gene GAPDH primers were used to verify the specificity and sensitivity of MAPPing techniques (Fig. 6,  lane 3 ) . This finding suggests that only the non-neuronal form of the IP3 receptor is synthesized in mouse T-lymphoma cells.

Interaction of IP3 Receptor and the Cytoskeleton
Cytoskeleton proteins such as ankyrin have been shown to be involved in regulating a number of cellular activities including receptor patching and capping (41)(42)(43)(44)(45)(46), cell adhesion (47)(48)(49), organelle movement, cell motility, protein secretion, and cell division (50,51). Putney et al. (7) have reported that IP3 receptor-containing vesicles may be attached to the plasma membrane through cytoskeletal elements such as actin. Van Bennett and co-workers (52) have also reported that a complex consisting of IP3 receptor, ankyrin, and GP180 can be isolated from brain tissue. In this study, we have used a newly developed in uitro assay which involves the use of nitrocellulose papers coated with purified lymphoma IP3 receptor (obtained from anti-IP3 antibody (IPR.l) affinity column chromatography as shown in Fig. 3C) to determine IP3 receptor's binding to '251-ankyrin. Our data clearly indicate that the lymphoma IP3 receptor binds to ankyrin (Fig.  7). T o further establish the specificity and affinity of IP3 receptor's binding to ankyrin, we have incubated lZ5I-ankyrin with IP3 receptor-coated nitrocellulose sheets in the absence ( Fig. 7B (i)) and presence of various concentrations of unlabeled ankyrin (Fig, 7B (ii) and (iii)). Scatchard plot analysis reveals the presence of a single high affinity class of ankyrin binding sites on IPS receptor with a dissociation constant ( K d ) of 0.2 nM (Fig. 7A).

IPS Receptor and Ankyrin Interaction in T-lymphoma Cells
Finally and most importantly, we have found that the binding of ankyrin to the IPS receptor in light density vesicles significantly inhibits IP3 binding (Table 11) and IP3-stimulated internal Ca2+ release (Fig. 8, A and B). This ankyrinmediated inhibitory effect on IPS-inducible Ca2+ release appears to be very selective, since ankyrin fails to block monoclonal anti-IP3 receptor-stimulated internal Ca2+ release (Fig.   8, C and D). I t is possible that ankyrin binding competes or overlaps with IPS binding sites (Table 11). This may explain 1251 -Ankyrin Bound ( f m o l l g ) the inhibitory effect on IP, function in these vesicles (Fig. 8,   A and B ) . Alternatively, ankyrin binding may cause some conformational changes at the regulatory domain(s) of IPa receptor resulting in a decrease in IP3 binding as well as a reduction in the potency of IP3 in releasing Ca2+ as shown previously by protein kinase A-mediated phosphorylation in cerebellar membranes (53). Currently, we are using in uitro mutagenesis and deletion mutation techniques to define further the ankyrin-binding domain(s) on mouse T-lymphoma IP3 receptor.