Enhancement of Calcium Sensitivity of Lipocortin I in Phospholipid Binding Induced by Limited Proteolysis and Phosphorylation at the Amino Terminus as Analyzed by Phospholipid Affinity Column Chromatography*

A phospholipid column was prepared by coating sil- iconized porous glass beads with phospholipids. The analysis of the Ca" requirement of lipocortin I and its derivatives in the binding to phospholipids was carried out with this column. The Ca2+ concentration required for 50% binding to the phospholipid column at room temperature was about 30 1tM for lipocortin I, while that was reduced to 15 PM when lipocortin I was phosphorylated by the epidermal growth factor receptor/ kinase, and a further reduction in the Cas+ requirement was observed with proteolytic cleavage at the N-ter-minal region. Cathepsin D and calpain I (low calcium- requiring form of calcium-activated neutral protease) rapidly cleaved human placental lipocortin I at Trp-12 and Lys-26, respectively. These N-terminal-truncated proteins required only 5 PM Ca2+ for 50% binding to the phospholipid column. This enhancement of Ca2+ sensitivity by limited proteolysis was also observed for porcine lung

A phospholipid column was prepared by coating siliconized porous glass beads with phospholipids. The analysis of the Ca" requirement of lipocortin I and its derivatives in the binding to phospholipids was carried out with this column. The Ca2+ concentration required for 50% binding to the phospholipid column at room temperature was about 30 1 t M for lipocortin I, while that was reduced to 15 PM when lipocortin I was phosphorylated by the epidermal growth factor receptor/ kinase, and a further reduction in the Cas+ requirement was observed with proteolytic cleavage at the N-terminal region. Cathepsin D and calpain I (low calciumrequiring form of calcium-activated neutral protease) rapidly cleaved human placental lipocortin I at Trp-12 and Lys-26, respectively. These N-terminal-truncated proteins required only 5 PM Ca2+ for 50% binding to the phospholipid column. This enhancement of Ca2+ sensitivity by limited proteolysis was also observed for porcine lung lipocortin I. Essentially the same results were obtained when the ea2+ sensitivities of the modified lipocortins I were analyzed using dispersed phospholipid vesicles instead of the phospholipid affinity column. Equilibrium dialysis indicated that the release of the N-terminal region markedly increased the affinity of lipocortin I for Ca" in the presence of phosphatidylserine, without any appreciable change of the number of Ca2'-binding sites. Limited proteolysis by endogenous proteases such as calpain may be an important regulatory mechanism for the Ca2+ sensitivity of lipocortin I in phospholipid binding.
Subsequent studies have revealed that lipocortin I can associate with phospholipid vesicles in a Ca2+-dependent manner and that the phospholipase Az inhibitory activity is due to the interaction of lipocortin I with phospholipid substrates rather than to a direct inhibition of the enzyme (3)(4)(5).
* The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked "advertisement" in accordance with 18 U.S.C. Section 1734 solely to indicate this fact. In addition to lipocortin I, several cytosolic Ca2+/phospholipid-binding proteins such as lipocortin 11 (calpactin I heavy chain), endonexins, proteins 1-111, calelectrins, chromobindins, and placental anticoagulant proteins have been identified in many organs (6)(7)(8)(9)(10)(11). Recent studies on their primary structures have shown that each Ca2+/phospholipid-binding protein has a specific small N-terminal domain connected to a C-terminal core domain. The C-terminal core domain is highly homologous among these proteins and composed of four copies of a 70-amino acid repeating unit. These repeating units are thought to contain Ca2+/phospholipid-binding sites (12)(13)(14)(15)(16)(17)(18).
No similarity has been found in the amino acid sequence of the N-terminal domain of these proteins. The N-terminal domain may confer a biological activity specific to each protein. In lipocortins I and 11, this domain consists of 30 amino acids and has interesting biochemical properties. The Nterminal tail of lipocortin I1 contains a binding site for a 10-kDa protein (p10, calpactin I light chain) and phosphorylation sites by pp60"'" and protein kinase C (Ca'+/phospholipiddependent enzyme) (7, 19,20). Proteolysis and phosphorylation at the tail are reported to cause a loss in the ability to associate with p10 and a decrease in the Ca2+ sensitivity of lipocortin I1 for phospholipid binding (21)(22)(23).
The N-terminal region of lipocortin I contains phosphorylation sites for the EGF' receptor/kinase and protein kinase C, and is extremely susceptible to proteolysis similar to lipocortin I1 (5,(24)(25)(26). In contrast with lipocortin 11, phosphorylation of lipocortin I by the EGF receptor/kinase is reported to enhance the Ca2+ sensitivity for phospholipid binding (27).
In this study, we tested whether the proteolytic modification Y . Ando, unpublished data. in cytosolic fractions of various cells (for reviews, Refs. 28-30). Since calpain is known to associate with plasma membrane in a Ca2+-dependent manner (31)(32)(33), the interaction between lipocortin and calpain may occur on the cytoplasmic surface of the plasma membrane in vivo. In order to analyze the Ca2+ requirement for the phospholipid binding of lipocortins, we performed affinity chromatography using a newly developed phospholipid-CPG column employing a decreasing Ca2+ gradient system. EXPERIMENTAL PROCEDURES3  Determination of Partial Amino Acid Sequence of Proteolyzed Lipocortin I-Cathepsin D-treated (35.5 kDa) and calpain I-treated (34 kDa) human placental lipocortins I were subjected to 17 and 10 cycles of automated Edman degradation, respectively. The sequential release of phenylthiohydantoin derivatives indicated that the N terminus was FIENE-EQEYVQTVKSSK for cathepsin D-treated lipocortin I and SSKGGPGSAV for calpain I-treated lipocortin I. These sequences were consistent with the predicted sequences of human lipocortin I commencing with

Degradation of Lipocortin I by Cathepsin D and Calpain I-
Phe-13 and Ser-27, respectively (12). Cathepsin D-treated 35.5-kDa lipocortin I and calpain I-treated 34-kDa lipocortin I are hereafter called des 1-12 lipocortin I and des 1-26 lipocortin I, respectively. For the untreated lipocortin I (37 kDa), the N terminus was blocked and no sequential release of phenylthiohydantoin derivatives was observed. This finding indicated that the affinity-purified lipocortin I was intact and not truncated at the N terminus (5). to the affinity column in the presence of 100 p M Ca2+, and eluted with linear gradients of decreasing CaC12 and correspondingly increasing EGTA at room temperature (Fig. 7). Intact lipocortin I was eluted at 32 p~ Ca", while the elution  of des 1-26 lipocortin I was markedly retarded and was observed a t 4 p~ Ca2+. When the mixture of des 1-12 lipocortin I and des 1-26 lipocortin I was applied to the column, they were eluted in a single protein peak, but SDS-PAGE analysis showed that des 1-12 lipocortin I was eluted slightly earlier than des 1-26 lipocortin I (data not shown).

Comparison of ea2+ Requirements for Phospholipid Binding
T o see whether limited proteolysis reduces the Ca2+ requirement for the binding to phospholipid vesicles as well as that for the binding to the affinity column, we studied the Ca2+ sensitivity in the association of lipocortin I with dispersed PS vesicles (Fig. 9). Human placental lipocortin I or des 1-26 lipocortin I was incubated with PS vesicles a t varying free calcium concentrations ranging from 0 to 100 pM. As expected from the experiments using the phospholipid affinity column (Fig. 7), about 80% of des 1-26 lipocortin I was found to be associated with the liposomes a t 10 ~L M Ca2+, and the halfmaximal association was observed a t about 5 p~ (Fig. 9). This was appreciably less than the Ca2+ concentration required for intact lipocortin I, which was 28 pM (Fig. 9).
Equilibrium Dialysis-In order to examine whether the enhancement of Ca2+ sensitivity of des 1-26 lipocortin I in phospholipid binding had an effect on Ca2+ affinity or stoichiometry, equilibrium dialysis of human placental lipocortin I and des 1-26 lipocortin I against 45Ca2+ was performed in the presence or absence of PS vesicles (Fig. 10) but not when PS was absent (Fig. 1OA) values of 90 and 18 ~L M , respectively (Fig. 1OB).

Phospholipid-CPG Affinity Chromatography of Intact, Protease-treated, and Phosphorylated Lipocortin I Prepared from
Porcine Lung-In order to test whether the enhancement of Ca2+ sensitivity is observed only with human lipocortin I or also with lipocortin I of other animal species, we purified porcine lung lipocortin I and investigated the effect of modifications of the N terminus on the Ca2+ sensitivity.
Phosphorylated porcine lung lipocortin I was prepared using the EGF receptor/kinase isolated from A431 cells. About 1% of native 37-kDa lipocortin I was phosphorylated, but none of des 1-26 lipocortin I was phosphorylated under our experimental conditions. This observation is consistent with previous reports that Tyr-21 is phosphorylated by the EGF receptor/kinase (5, 44). Porcine lung lipocortin I (37 kDa), 32P-labeled lipocortin I (37 kDa), and des 1-26 lipocortin I (34 kDa) were applied to the phospholipid-CPG column and eluted with a decreasing Ca2+ gradient a t 4 "C or a t room temperature. Fig. 11 shows one of these experiments. Phosphorylated lipocortin I was eluted between intact and des 1-26 lipocortins I as shown by Cerenkov counting (Fig. 1I.A) and autoradiography (Fig. 1lC). While phosphorylation of lipocortin I decreased to half the amount of Ca2+ required for 50% association of the protein with the phospholipid column, much greater reduction of the Ca2+ requirement was observed for des 1-26 lipocortin I. The Ca2+ concentrations required for 50% binding of lipocortin I and its derivatives to the phospholipid-CPG column are summarized in Table I. The order of Ca2+ sensitivity was des 1-26 lipocortin I 2 des 1-12 lipocortin I > phosphorylated lipocortin I > intact (native) lipocortin I , and this order was the same in experiments performed a t both 4 "C and room temperature.

DISCUSSION
Lipocortins have been purified from various cells and organs such as A431 cells, intestinal mucosa, lung, and placenta (5,20,24,44). Since the N-terminal regions of lipocortins are extremely sensitive to proteolytic cleavage, a rapid purification with fewer steps is highly recommended. In this paper, we have developed a phospholipid-CPG column and utilized it for affinity chromatography of lipocortins, instead of using repeated precipitation and extraction of lipocortins from the cellular particulate fraction, which is the standard method for the purification of lipocortins (5, 20, 24, 44). The intact (native) form of lipocortin I was rapidly purified from human placenta and porcine lung with phospholipid-CPG column chromatography with a good yield.
For the analysis of Caz+ dependence for lipocortins, phospholipid-CPG column chromatography with CaC12/EGTA gradients was found to be very useful. Human placental lipocortin I and des 1-26 lipocortin I required 31 p~ and 4.2 pM Ca2', respectively, for 50% binding to the affinity column (Table I), and these values were comparable to those required for half-maximal association with PS vesicles: about 28 p~ Ca2+ for intact lipocortin I and 5 pM Ca2+ for des 1-26 lipocortin I (Fig. 9). Lipocortin I1 required lower concentration of Ca2+ for phospholipid binding than lipocortin I in our experiments using the affinity column (Fig. 2), and such results have been reported using PS vesicles (7, 27). An affinity column containing only PC did not adsorb lipocortins as PC liposomes did not (7). These findings collectively suggest that phospholipid-CPG columns can be substituted for phospholipid vesicles and enable more precise and quantitative characterization of the Ca2+-dependent phospholipid association of lipocortins.
Occurrence of des 1-12 lipocortin I and des 1-26 lipocortin I was often reported from several organs (5,44,47). Intracellular proteases which produce such lipocortin I derivatives have not yet been identified. Our results show cathepsin D and calpain I are possible candidates for the production of des 1-12 lipocortin I and des 1-26 lipocortin I, respectively.
In this paper, we first demonstrated using affinity column chromatography (Figs. 7 and 8), vesicle binding analysis (Fig.  9), and equilibrium dialysis (Fig. 10) that the proteolytic modification of both human and porcine lipocortins I at the N-terminal regions leads to a remarkable enhancement of the Ca2+ sensitivity in the binding to phospholipids. Scatchard analysis revealed that this enhancement was due to an effect on Ca2+ affinity (Fig. 1OB). Intact lipocortin I has four Ca2+binding sites with an apparent K d of 90 pM, and these results are in agreement with the studies of Schlaepfer and Haigler (27). Release of the N-terminal tail substantially increased the affinity of lipocortin 1 for ca2+ ( K d , 18 pM), and the four Ca2+-binding sites were preserved.
Since cathepsin D is a lysosomal enzyme and works under acidic conditions, the degradation of lipocortin I by cathepsin D could be artifactual. On the other hand, calpain I is an intracellular, nonlysosomal, neutral proteinase, and it has been reported to associate with plasma membrane in the presence of micromolar concentration of Ca2+ (31)(32)(33). This property of calpain I led several investigators to postulate that calpain may play a significant role in signal transduction in the cell membrane (29)(30)(31)(32). We speculate that a stimulation-coupled transient rise in intracellular Ca2+ causes the translocation of lipocortin I and calpain I from the cytosol to the plasma membrane, where lipocortin I is cleaved by calpain I, and that the cleaved lipocortin I with higher affinity for Ca2+/phospholipid remains bound to the membrane even after the cytosolic Ca" concentration declines to the basal level. This implies that the intracellular location of lipocortin I can be modulated by the limited proteolysis.
Phosphorylation of lipocortin I by the EGF receptor/kinase reduced by half the amount of Ca2+ required for 50% association of the protein with the phospholipid column, as reported by Schlaepfer and Haigler (27). They also reported that phosphorylated lipocortin I at Tyr-21 was 10-fold more sensitive to proteolytic cleavage at Lys-26 than the native protein was, and suggested that there may be a successive modification of the N terminus of lipocortin I in uiuo; i.e. phosphorylation, followed by proteolysis. Our results suggest that the Ca2+ sensitivity of lipocortin I increases during the course of the sequential reaction, and the proteolyzed lipocortin I may have a profound and long term effect on cellular functions when compared to the phosphorylated lipocortin I.
Interestingly, with lipocortin 11, the opposite effect of the modification of the N-terminal domain is reported; i.e. phosphorylation and proteolysis at the N terminus of lipocortin I1 cause a loss in the ability to associate with p10 and then a decrease in Ca2+ sensitivity (21)(22)(23). Overall, these findings, including our results, suggest that the N-terminal domain of lipocortins regulates the Ca"/phospholipid binding of the Cterminal core domain in a specific manner for each protein. , " " -.