c-Src Binds αII Spectrin's Src Homology 3 (SH3) Domain and Blocks Calpain Susceptibility by Phosphorylating Tyr1176 *

Spectrin is a ubiquitous heterodimeric scaffolding protein that stabilizes membranes and organizes protein and lipid microdomains on both the plasma membrane and intracellular organelles. Phosphorylation of β-spectrin on Ser/Thr is well recognized. Less clear is whether α-spectrin is phosphorylatedin vivo and whether spectrin is phosphorylated on tyrosine (pTyr). We affirmatively answer both questions. In cultured Madin-Darby canine kidney cells, αII spectrin undergoesin vivo tyrosine phosphorylation. Enhancement of the steady state level of pTyr-modified αII spectrin by vanadate, a phosphatase inhibitor, implies a dynamic balance between αII spectrin phosphorylation and dephosphorylation. Recombinant peptides containing the Src homology 3 domain of αII spectrin (but not the Src homology 3 domain of αI spectrin) bind specifically to phosphorylated c-Src in Madin-Darby canine kidney cell lysates, suggesting that this kinase is responsible for its in vivo phosphorylation. pTyr-modified αII spectrin is resistant to maitotoxin-induced cleavage by μ-calpain in vivo. In vitro studies of recombinant αII spectrin peptides representing repeats 9–12 identify two sites of pTyr modification. The first site is at Tyr1073, a residue immediately adjacent to a region encoded by alternative exon usage (insert 1). The second site is at Tyr1176. This residue flanks the major site of cleavage by the calcium-dependent protease calpain, and phosphorylation of Tyr1176 by c-Src reduces the susceptibility of αII spectrin to cleavage by μ-calpain. Calpain cleavage of spectrin, activated by Ca2+ and calmodulin, contributes to diverse cellular processes including synaptic remodeling, receptor-mediated endocytosis, apoptosis, and the response of the renal epithelial cell to ischemic injury. Tyrosine phosphorylation of αII spectrin now would appear to also mediate these events. The spectrin skeleton thus forms a point of convergence between kinase/phosphatase and Ca2+-mediated signaling cascades.

Spectrin is a ubiquitous heterodimeric scaffolding protein that stabilizes membranes and organizes protein and lipid microdomains on both the plasma membrane and intracellular organelles. Phosphorylation of ␤-spectrin on Ser/Thr is well recognized. Less clear is whether ␣-spectrin is phosphorylated in vivo and whether spectrin is phosphorylated on tyrosine (pTyr). We affirmatively answer both questions. In cultured Madin-Darby canine kidney cells, ␣II spectrin undergoes in vivo tyrosine phosphorylation. Enhancement of the steady state level of pTyr-modified ␣II spectrin by vanadate, a phosphatase inhibitor, implies a dynamic balance between ␣II spectrin phosphorylation and dephosphorylation. Recombinant peptides containing the Src homology 3 domain of ␣II spectrin (but not the Src homology 3 domain of ␣I spectrin) bind specifically to phosphorylated c-Src in Madin-Darby canine kidney cell lysates, suggesting that this kinase is responsible for its in vivo phosphorylation. pTyr-modified ␣II spectrin is resistant to maitotoxin-induced cleavage by -calpain in vivo. In vitro studies of recombinant ␣II spectrin peptides representing repeats 9 -12 identify two sites of pTyr modification. The first site is at Tyr 1073 , a residue immediately adjacent to a region encoded by alternative exon usage (insert 1). The second site is at Tyr 1176 . This residue flanks the major site of cleavage by the calciumdependent protease calpain, and phosphorylation of Tyr 1176 by c-Src reduces the susceptibility of ␣II spectrin to cleavage by -calpain. Calpain cleavage of spectrin, activated by Ca 2؉ and calmodulin, contributes to diverse cellular processes including synaptic remodeling, receptor-mediated endocytosis, apoptosis, and the response of the renal epithelial cell to ischemic injury. Tyrosine phosphorylation of ␣II spectrin now would appear to also mediate these events. The spectrin skeleton thus forms a point of convergence between kinase/phosphatase and Ca 2؉ -mediated signaling cascades.
Post-translational protein modification regulates cellular function. Although in many instances the role of such modification is well understood, less clear is the impact of protein modification on cytoskeletal dynamics. This is particularly true for spectrin, a large multifunctional scaffolding molecule positioned at the interface between membrane and cytosol. Pro-duced by seven distinct genes, the spectrin family segregates as two subunits, ␣ and ␤. Heterodimeric ␣II␤II spectrin is the most common, expressed in most if not all vertebrate cells. By binding to an array of integral membrane and cytosolic proteins and to acidic phospholipids, spectrin links membrane protein and lipid microdomains to the actin and microtubule filamentous skeleton (for a review, see Ref. 1).
Several pathways of post-translational regulation impact spectrin. The ␤I, ␤II, and ␤III isoforms are phosphorylated on Ser and Thr (2)(3)(4)(5); spectrins ␤IV (6) and ␤V (7) may also be similarly phosphorylated, although no data yet exists for these proteins. The functions ascribed to ␤-spectrin phosphorylation include destabilization of the erythrocyte membrane skeleton (8 -10), disassembly of the skeleton during mitosis (4), and the control of Golgi stability (11). However, the mechanism of these effects remains elusive. Less clear is whether spectrin can be tyrosine-phosphorylated and whether the ␣-subunit of spectrin is ever covalently phosphorylated. One report indicates that ␤-spectrin can be tyrosine-phosphorylated when incubated in vitro with purified insulin receptor kinase (12). Another report has appeared indicating that both spectrin subunits are tyrosine-phosphorylated when incubated in vitro with a spleen protein tyrosine kinase (13). A recent report, appearing after the present study was first submitted for publication, indicates that ␣II spectrin is tyrosine-phosphorylated in cultured COS-7 cells in vivo (14).
Other regulatory pathways impacting spectrin include the action of calcium, calmodulin, calcium-activated proteolysis, and the regulation of its Golgi binding by ARF1, a small GTPase. ARF1 acts by modifying phosphatidylinositol 4,5bisphosphate levels in Golgi membranes, a substrate for the pleckstrin homology domain of ␤I⌺2 and ␤III spectrin (15). Calcium binds directly to two EF-hand domains near the COOH terminus of ␣-spectrin (16) with unknown functional consequences. Calmodulin binds to a non-homologous sequence inserted into the 11th repeat unit of vertebrate ␣II spectrin (17). Binding at this site modifies the susceptibility of the nearby Tyr 1176 -Gly 1177 bond to cleavage by -calpain (18) and renders the adjacent ␤II spectrin subunit susceptible to -calpain cleavage at Gln 1441 -Ser 1442 (18). 1 After calpain cleavage, spectrin's self-association, actin binding, and membrane binding properties are modified (20,21). Spectrin proteolysis by calpain has been correlated with processes involved with secretion and endocytosis in epithelial cells (22)(23)(24), opacification of the vertebrate lens (25), synaptic plasticity and long-term potentiation in the central nervous system (26 -28), and various central nervous system pathologies including neurotoxic or ischemic injury (e.g. see Refs. 29 -33).
In the present study we demonstrate that ␣II spectrin is subject to tyrosine phosphorylation in vivo in cultured Madin-Darby canine kidney (MDCK) 2 cells and that such phosphorylation bestows on ␣II spectrin resistance to the calpain-mediated proteolysis that follows maitotoxin exposure. In vitro studies document at least two sites of tyrosine phosphorylation in ␣II spectrin. One is at the site of calpain cleavage, the other flanks a short sequence encoded by alternative exon utilization. Both sites are adjacent to the SH3 domain of spectrin, a locus that we show binds specifically to c-Src in MDCK cell extracts. Collectively, these data together with a recent independent report that appeared after this study was first submitted (14) establish a role for tyrosine phosphorylation of ␣II spectrin as a modifier of the calpain sensitivity of spectrin and reveal the spectrin cortical skeleton as a point of convergence between tyrosine kinase/phosphatase and Ca 2ϩ -mediated signal transduction pathways.

MATERIALS AND METHODS
In Vivo Determination of ␣II Spectrin Tyrosine Phosphorylation-MDCK cells were cultured to confluence with Dulbecco's modified Eagle's medium (DMEM) as before (34). In some experiments when blockage of endogenous phosphatase was desired, cells were washed three times with serum-free DMEM and exposed to 0.1 mM pervanadate (freshly prepared from aliquots of 100 mM activated sodium orthovanadate (Sigma) dissolved in water and 100 mM H 2 O 2 ) in Opti-MEM I at 37°C for 30 min. Control cells were incubated in Opti-MEM I only. For immunoprecipitation, pervanadate-treated and -untreated MDCK cells were washed three times in ice-cold PBS, lysed gently at 4°C for 15 min with 1 ml of modified RIPA lysis buffer (Tris-HCl 50 mM, pH 7.4, 1% Nonidet P-40, 0.25% sodium deoxycholate, 150 mM NaCl, 1 mM PMSF, 1 mg/ml each of aprotinin and leupeptin in a 1:100 dilution of Calbiochem phosphatase inhibitor mixture set II (catalog number 524625). After centrifugation at 14,000 rpm for 15 min at 4°C, supernatants were analyzed by SDS-PAGE or incubated (0.5 ml) with 3 l of antibody overnight at 4°C. Immune complexes were captured with 25 l of packed ImmunoPure Plus immobilized protein G beads (Pierce) (2 h, 4°C), washed four times with modified RIPA lysis buffer, and analyzed by SDS-PAGE. For Western blotting, proteins transferred to nitrocellulose membranes were blocked for 1 h with either 5% (w/v) dry skim milk in TBST buffer (20 mM Tris-HCl, pH 7.5, 150 mM NaCl, and 0.1% Tween 20) or in blocking buffer (1 M glycine, 5% (w/v) dry skim milk, 1% (w/v) bovine serum albumin, and 5% (v/v) bovine calf serum). Proteins were detected with primary antibodies in TBST at room temperature for 1 h at a typical dilution of 1:1000 or as specified. The anti-glutathione S-transferase (GST) antibody (Amersham Biosciences) was used at 1:2000. Detection employed goat anti-rabbit/anti-mouse conjugated horseradish peroxidase at 1:10,000 and enhanced chemiluminescence (ECL, Amersham Biosciences).
Maitotoxin Induced Activation of Calpain Resulting in ␣II Spectrin Proteolysis from Vanadate-pretreated MDCK Cells-MDCK cells were grown to confluence as above and treated with 5 mM activated sodium orthovanadate in Opti-MEM I without H 2 O 2 . Cells were then washed with Opti-MEM I one time, and 100 nM maitotoxin (Calbiochem) diluted in Opti-MEM I was applied to cells for various periods at 37°C. Cells were harvested by washing one time in ice-cold PBS then scraped immediately thereafter into 250 l of one time sample (4% SDS, 60 mM Tris, pH 6.8, 10% v/v glycerol, 1 mM PMSF, 10 g/ml aprotinin, 10 g/ml leupeptin, 0.2 mM DTT, 10 g/ml protease arrest (Calbiochem), 1 g/ml pepstatin, 1 mM EDTA, 125 mM NaCl). Samples were heated to 100 C°f or 20 min and analyzed by SDS-PAGE and Western blotting. The ␣II spectrin bdp-1 antibody (35) was used at 1:10,000 in TBST (.01% v/v Tween 20) for blotting.
c-Src Binding Assays-Confluent MDCK cells were treated with pervanadate as above at 37°C and solubilized in binding buffer (20 mM HEPES (pH 7.5), 25 mM KCl, 120 mM NaCl, 2 mM EGTA, 2 mM EDTA, 0.2 mM DTT, 0.5% Triton X 100, and 0.1 mM PMSF). This lysate was incubated with GST fusion peptides representing the SH3 domains of ␣I or ␣II spectrin or control peptides bound to glutathione-agarose for 3 h at 4 C°, washed four times in above buffer, and placed in sample buffer for immunoblotting with Ant-c-Src (SRC2 Sc-18 Santa Cruz) at 1:1000 dilution in TBST.

Tyrosine Phosphorylation of ␣II Spectrin in MDCK Cells
Retards Its Susceptibility to Calpain Digestion-The phosphotyrosine content of ␣II spectrin in confluent MDCK cells was examined by immunoprecipitation and Western blotting with anti-phosphotyrosine antibodies (Fig. 1A). Immunoprecipita-FIG. 2. Vanadate protects ␣II spectrin from calpain-mediated breakdown in vivo. Cultured MDCK cells were incubated up to 30 min with either buffer alone or with 5 mM vanadate and then lysed directly into an SDS-containing buffer. Precautions were taken to avoid post-lysis proteolysis (see "Materials and Methods"). The extent of the spectrin breakdown product (␣II-bdp1) was then evaluated with an antibody specific for this calpain cleavage product of spectrin. Note that in the absence of maitotoxin, quiescent cells contain nearly undetectable levels of ␣II-bdp1. Cleavage is stimulated by the activation of calpain by maitotoxin and delayed by the tyrosine phosphatase inhibitor vanadate, a treatment that increases the level of tyrosine-phosphorylated ␣II spectrin. By 30 min of incubation however, the protection afforded by vanadate is lost (for reasons that remain uncertain). tion was conducted under conditions that allow both subunits of the strongly associated spectrin heterodimer (␣II␤II) to be precipitated. Spectrin was abundant in the precipitates (Fig.  1A, left panel), and the immunoprecipitated spectrin was reactive with antibodies to phosphotyrosine (P-Y on Fig. 1A, center  panel). The level of anti-pTyr reactivity was increased by pretreatment of the cells with vanadate, a tyrosine phosphatase inhibitor, but a phosphotyrosine-modified spectrin was detectable even in untreated cells. While we cannot exclude the presence of trace pTyr in ␤II spectrin that is also present in these precipitates, the preponderance of pTyr-modified spectrin, identified in multiple experiments (n ϭ 6), appears to be the ␣II subunit based on the size of the pTyr immunoreactive band at Ϸ284 kDa. Tyrosine phosphorylation of ␣II spectrin was also verified by immunoprecipitation with anti-pTyr followed by Western blotting (Fig. 1A, right panel). Anti-pTyr precipitates revealed the clear presence of ␣II spectrin, as well as a second immunoreactive band at Ϸ100 kDa. The identity of the latter band is unknown and was not further characterized; it likely represents a tyrosine-phosphorylated proteolytic product of ␣II spectrin. Although in most experiments the level of ␣II spectrin proteolysis was low relative to intact spectrin, it was noted that with vanadate treatment, the abundance of a spectrin breakdown fragment at Ϸ150 kDa was consistently reduced (e.g. the heavily loaded lysate lanes shown in Fig. 1, A and B). This 150-kDa fragment is typically generated by calpain proteolysis of ␣II spectrin at Tyr 1176 (17). To verify relationship between the level of this spontaneous spectrin break-down product and the degree of tyrosine phosphorylation, the extent of ␣II spectrin proteolysis was evaluated using an antibody specific for the ␣II spectrin proteolytic fragment generated by calpain cleavage at Tyr 1176 (35). This antibody detected an Ϸ32% reduction in the level of calpain-cleaved ␣II spectrin derived from cultured MDCK cells after they had been exposed to vanadate for 30 min (Fig. 1B). To more thoroughly evaluate this putative in vivo protection of ␣II spectrin from calpain cleavage, additional experiments were conducted using maitotoxin to stimulate calpain in quiescent MDCK cells (Fig. 2). Maitotoxin is a marine toxin that opens L-type calcium channels and specifically activates calpain but not caspase (39). It thus stimulates the generation of the characteristic 150-kDa ␣II spectrin breakdown product (␣II-bdp1) (40). As with the extraction assays in Fig. 1, incubation of MDCK cells with vanadate before maitotoxin exposure protected ␣II spectrin from calpain cleavage (Fig. 2). This effect directly correlated with the degree of tyrosine phosphorylation as measured by pTyr antibody. The in vivo pTyr state of ␣II spectrin is thus inversely proportional to its susceptibility to calpain cleavage. Surprisingly, protection was not absolute since after incubations exceeding 30 min, even in vanadate-treated cells, the level of calpain-cleaved ␣II spectrin rose to control levels (Fig. 2). We do not know the genesis of this biphasic effect but note that prolonged exposure to maitotoxin is toxic to cells and, thus presumably at longer times, creates a state no longer representative of the in vivo environment.
Finally, it is interesting to note that when there is immediate harvest of untreated confluent MDCK cells into a SDS buffer, with care taken to block postextraction proteolysis by the early and liberal use of effective protease inhibitors, one achieves cell lysates that are nearly devoid of spontaneous ␣II spectrin breakdown product (Fig. 2, time zero). These results are in contrast to observations made following alternative lysis procedures, such as preincubation in RIPA buffer (Fig. 1A), and suggests that the presence of ␣II spectrin bdp-1 in quiescent cultured cells as detected by ourselves (Fig. 1A) and by others (14) probably represents an in vitro (rather than in vivo) proteolytic event that must occur rapidly following cell lysis.
c-Src Binds to the SH3 Domain of ␣II Spectrin in Vivo-The SH3 domain of ␣II spectrin flanks the site of calpain cleavage and thus is an attractive candidate for the docking of kinases or phosphatases involved in the control of ␣II spectrin cleavage by calpain. An earlier report has documented by yeast two-hybrid assay that isoform A of low-molecular-weight phosphotyrosine phosphatase binds to this site (14). We have searched for direct binding partners in MDCK lysates by GST pull-down assays with recombinant GST-spectrin peptides containing only the SH3 domain of either ␣I or ␣II spectrin (Fig. 3). Although the SH3 domains of ␣I and ␣II spectrin share 75% homology, only the SH3 domain of ␣II spectrin bound c-Src in the MDCK cell lysates. Moreover, c-Src only bound when it was derived from vanadate-treated cells, a condition that leads to its autophosphorylation. Although these studies do not exclude the possibility of an unrecognized intervening adapter protein, it seems likely that the tyrosine phosphorylation of c-Src itself also determines its affinity for the SH3 domain of ␣II spectrin. Similar requirements for the binding of c-Src have also been noted in other settings (for review, see Ref. 41). The kinase c-Src thus makes an interesting complement to the phosphotyrosine phosphatase that also binds to this SH3 domain in ␣II spectrin (14).
Recombinant ␣II Spectrin Peptides Are Phosphorylated in Vitro on Tyr 1073 and Tyr 1176 -Observations in MDCK cells indicated that ␣II spectrin could be tyrosine-phosphorylated and that this event correlated with reduced ␣II spectrin proteolysis by endogenous -calpain. While other mechanisms are certainly possible, the simplest explanation of this linkage would be if Tyr 1176 is the target of phosphorylation, a modification that would render it a less attractive substrate for -calpain (42). In earlier work we have demonstrated that the critical Tyr 1176 -Gly 1177 bond targeted by -calpain probably occurs in a highly exposed loop juxtaposed between helix C and the calmodulin-binding domain within the 11th repeat unit of ␣II spectrin, making it likely that subtle conformational features of this region exert significant effects on its suitability as a calpain target (43). This region of ␣II spectrin targeted by calpain is also interesting in that adjacent to the calpain cleavage site and the calmodulin-binding domain is a 20-residue sequence (insert 1) encoded by alternative exon usage (44,45). Upstream of this insert is the spectrin SH3 domain, a motif common to many kinase and phosphatase signal transduction cascades (46), and as shown above, a site that binds c-Src. We therefore focused on this region of ␣II spectrin. A recombinant GST fusion peptide (SACC) encompassing ␣II spectrin repeats 9 -12 was prepared (Fig. 4A). In silico analysis by NetPhos 2.0 (47) predicted that of the eight tyrosine residues in this peptide, only two were high-probability substrates for tyrosine phosphorylation. One was indeed Tyr 1176 , the other was Tyr 1073 (Fig.  2B). Interestingly, Tyr 1073 is the first residue after insert 1, and its phosphorylation potential is enhanced by sequences in insert 1 (Fig. 4B). When the purified recombinant GST-SACC peptide was incubated in vitro with the tyrosine kinase c-Src and ␥-[ 32 P]ATP, it became strongly labeled in a dose-dependent manner (Fig. 4C). Parallel studies using c-Src alone or a GST control peptide incorporated minimal phosphate, confirming that the SACC peptide was a favorable substrate for tyrosine phosphorylation.
To identify the specific sites of tyrosine phosphorylation in SACC, a set of smaller recombinant peptides were examined that individually represented each of the structural and func-tional motifs in SACC (Fig. 5A). These peptides were incubated in vitro with c-Src, and pTyr was measured by immunoblotting with anti-phosphotyrosine antibody (Fig. 5B). Both the SA and Sϩ peptides (representing ␣II spectrin repeats 9 and 10, Ϯ insert 1) were readily phosphorylated. Both of these peptides contained Tyr 1073 . Conversely, peptide S was not phosphorylated; it encompassed most of spectrin repeats 9 and 10 and included five tyrosine residues and the SH3 domain but lacked Tyr 1073 . Since Tyr 1073 is the only tyrosine present in peptides SA and Sϩ that is not present in peptide S, we conclude that Tyr 1073 is one target of spectrin tyrosine phosphorylation in vitro. Surprisingly, despite the impact of insert 1 on the calculated phosphorylation potential of Tyr 1073 , we found that its presence or absence had little effect on phosphorylation of Tyr 1073 in these assays (cf. Fig. 5B, lanes 2 and 3). Alternative phosphorylation conditions or kinases that might reveal the predicted impact of insert 1 were not explored in the present study.
Examination of the recombinant peptides CC and C (encom- FIG. 6. Tyrosine phosphorylation of Tyr 1176 inhibits cleavage by -calpain. A, SDS-PAGE analysis of phosphorylated and non-phosphorylated ␣II spectrin peptides subjected to proteolysis with increasing amounts of -calpain. Peptides are designated as in Fig. 5. PY marks the phosphorylated peptide, and bdp represents its major -calpain-generated cleavage product. Peptide C, which lacks the calpain sensitivity region, is neither phosphorylated nor sensitive to calpain. B, the relative extent of digestion was quantified by densitometry. In every instance, phosphorylation of Tyr 1076 retards susceptibility to calpain. passing the calpain cleavage site and calmodulin-binding domain in repeat 11 of ␣II spectrin) revealed a second site of tyrosine phosphorylation. These peptides differ only in the presence of the calpain cleavage domain (Fig. 5A) with its single tyrosine, Tyr 1176 , at the calpain cleavage site. Peptide CC, but not C, was readily tyrosine-phosphorylated (Fig. 3C), revealing this critical tyrosine as a second favored site of phosphorylation.
Phosphorylation of Tyr 1176 Retards ␣II Spectrin Peptide Susceptibility to Digestion by -Calpain-After tyrosine phosphorylation of the recombinant peptides SACC and CC, both containing residue Tyr 1176 , their susceptibility to in vitro proteolysis by increasing amounts of -calpain was assayed by SDS-PAGE (Fig. 6A). These results were quantified by densitometry of the stained gels (Fig. 6B). For both peptides, at every level of calpain, phosphorylation reduced the fraction of peptide proteolysed. At maximum concentrations of -calpain, tyrosine phosphorylation of SACC or CC afforded a 30 -40% reduction in the extent of cleavage by -calpain. Conversely, a spectrin peptide representing the calmodulin-binding domain, but lacking the calpain-sensitive Tyr 1076 site, was not cleaved by calpain (Fig. 6A, panel C). These results cannot be due to the presence of c-Src itself in the reaction mixture since this kinase (without ATP) was included in all controls. It is also worth noting that the significant reductions in calpain sensitivity observed here probably underestimate the degree of resistance conferred by tyrosine phosphorylation since it is unlikely that stoichiometric phosphorylation of Tyr 1176 was achieved.

DISCUSSION
The studies presented here establish that ␣II spectrin i) is tyrosine-phosphorylated in cultured MDCK cells and that the level of this phosphorylation is increased by treatment with the tyrosine phosphatase inhibitor vanadate; ii) that the in vivo susceptibility of endogenous ␣II spectrin to cleavage at Tyr 1176 by -calpain is inversely proportional to its level of tyrosine phosphorylation; iii) that phosphorylated c-Src binds to the SH3 domain of ␣II spectrin; iv) that the targets of in vitro phosphorylation with c-Src kinase are residues Tyr 1073 and Tyr 1176 in ␣II spectrin; and v) that phosphorylation of Tyr 1176 retards the susceptibility of this site to cleavage by -calpain in vitro. Collectively, these results establish the presence of at least two sites of tyrosine phosphorylation in ␣II spectrin and identify a novel regulatory mechanism acting through c-Src on the -calpain-mediated processing pathway of spectrin. Given the growing recognition of the importance of the cytoskeleton in determining cellular function, both in organelles and at the plasma membrane, these findings have implications for our understanding of membrane skeletal control in several settings and indicate that spectrin may be a key point of signal convergence between tyrosine kinase/phosphatase and Ca 2ϩ -mediated signal cascades.
Calpain cleavage of proteins is important in many cellular and pathologic processes such as long-term potentiation and synaptic remodeling, glutamate-induced neurotoxicity, ischemic cellular injury, apoptosis, platelet activation, exocrine secretion, neutrophil activation, mitosis, progesterone and estrogen receptor modulation, and the regulation of a variety of kinases such as protein kinase C, phosphorylase kinase, myosin light chain kinase, calmodulin-dependent kinase and phosphatase, and other signal transduction pathways (for reviews, see Refs. 33, 42, 48 -51). As cytoskeletal proteins are common substrates for calcium-dependent proteases, regulating the degree of calpain cleavage through tyrosine phosphorylation of the substrate represents an important mechanism for modulating these cellular and pathologic events. While it is likely that the novel pathway for the control of spectrin breakdown reported here will be utilized in a variety of physiologic and pathologic settings, its role in neuronal function may be particularly important. Spectrin and other cytoskeletal-associated proteins interact with the glutamate-gated N-methyl-D-aspartate receptor (NMDA receptor) (for review, see Ref. 52). NMDA receptors, a class of glutamate-gated cation channels with high Ca 2ϩ conductance, mediate fast transmission and plasticity of central nervous system excitatory synapses. Spectrin associates with the NR2 cytosolic subunit of the NMDA receptor (19), an activity regulated by phosphorylation of the NR2 subunit. Calpain proteolysis of NR2 disrupts its association with spectrin, conversely its phosphorylation by c-Src protects it from calpain cleavage (36). Our present results suggest a related pathway by which tyrosine phosphorylation of ␣II spectrin might further protect the integrity of the NMDA receptor complex by retarding proteolysis of the spectrin membrane skeleton. While only a single study has explored the impact of phosphatase inhibitors on glutamate receptor and spectrin stability (38) and found no protection of spectrin breakdown, our present results suggest that this issue deserves a more complete exploration. Regardless, the results reported here establish a novel post-translational modification in ␣II spectrin (tyrosine phosphorylation) and demonstrate a clear coupling between this modification and is susceptibility to calpain-mediated proteolysis, both in vivo and in vitro.