Clathrin light chains are calcium-binding proteins.

Clathrin light chains have been purified to near homogeneity. When analyzed by sodium dodecyl sulfate gel electrophoresis followed by silver stain for proteins, no bands corresponding to light chains were detected. As calmodulin and troponin C are known to behave in the same manner on silver staining, the possibility that clathrin light chains were Ca2+-binding proteins was investigated. Light chains fixed to nitrocellulose filters were found to bind 45Ca2+ in the presence of 5 mM Mg2+. The Ca2+-binding capacity of the light chains was further investigated, using gel filtration and equilibrium dialysis. The light chains were shown to bind, in the presence of 3 mM Mg2+, 1 mol of Ca2+ per mol of light chain with a Kd of 25-55 microM. Nitrocellulose binding and gel filtration studies showed that light chains present in triskelions are still capable of binding Ca2+, in this case with a calculated Kd of 45 microM.

Coated vesicles, which are cellular organelles containing an outer protein lattice, are involved in intracellular protein transport. The vesicle coat can be dissociated into basic structural units termed triskelions which are composed of three clathrin heavy chains and three clathrin light chains. There are two types of light chains, a and b, which in bovine brain have molecular weights of 36,000 and 33,000, respectively.
Cellular Ca2+ is a general second messenger involved in multitudes of cellular functions (1). Evidence suggests that receptor-mediated endocytosis may be regulated by calcium (2,3). Because of this, a number of investigators have studied the possible involvement of calmodulin in this system. Clathrin light chains, along with a number of other coated vesicle associated proteins, have been shown to bind calmodulin ( 4 , and it has been suggested that calmodulin may facilitate the recruitment of clathrin components to the plasma membrane for the assembly of coated pits (5). In this paper we report that clathrin light chains are Ca2+-binding proteins. This binding has a dissociation constant in the micromolar range and can be demonstrated in the presence of high concentrations of M$+.

MATERIALS AND METHODS
Purification of Clathrin Light Chains-Clathrin light chains were purified from fresh bovine brain according to a modified procedure of Lisanti et al. (6). Fresh tissue was homogenized in 0.1 M MES' (pH 6.5), 1 mM EGTA, 0.5 mM MgC12, 0.02% NaN3, 100 mg/liter crude soybean trypsin inhibitor, 100 mg/liter phenylmethylsulfonyl fluoride, and 200 mg/liter benzamidine and centrifuged at 16,000 X g for 1 h. The supernatant was then centrifuged at 48,000 X g for 2 h. The resultant pellet was resuspended in 50 mM Tris (pH 8.0) containing 1 mM EDTA and fresh protease inhibitor and placed in a boiling water bath for 5 min. The boiled sample was centrifuged at 100,000 X g for 1 h. CaC12 was added to the supernatant to a final concentration of 1.25 mM, and the sample was applied to a calmodulin-Sepharose affinity column equilibrated in 20 mM Tris (pH 7.5), 0.5 M NaCI, 20 mM 2-mercaptoethanol, and 0.25 mM CaCl2. The column was washed with the same buffer until all unbound protein was eluted, and a 0.25 mM CaC12 to 0.1 mM EGTA gradient was applied. The clathrin light chains were eluted in 0.1 mM EGTA. The fractions containing the light chains were pooled and applied to a Sephadex G-100 column equilibrated in 20 mM Tris (pH 7.5), 0.5 M NaCl, 20 mM 2-mercaptoethanol, and 0.1 mM EGTA. The peak fractions containing the light chains were pooled, and the light chains were stored at -20 "C.
Purification of Triskelions-Triskelions were purified from fresh bovine brain essentially by the method of Pearse and Robinson (7).
Nitrocellulose Binding-Calcium-binding proteins were detected on nitrocellulose blots by %! a2+ autoradiography. Purified light chains (10-20 pg) or triskelions (20-40 pg) were electrophoresed on SDS-polyacrylamide gels according to the method of Laemmli (9). Electrophoretic transfer of the proteins from the polyacrylamide gels to nitrocellulose membranes was performed according to the method Sephadex G-25 Column Chromatography-Calcium binding was measured by the technique of Hummel and Dreyer (11). Clathrin light chains (1 mg) or triskelions (1.5 mg) were dialyzed extensively against 50 mM Tris (pH 7.0) and 3 mM MgCl, and then applied to a Sephadex G-50 column equilibrated in the same buffer to remove trace amounts of EGTA. The peak fractions were pooled and applied to a 0.5 X 20-cm Sephadex G-25 column equilibrated in 50 mM Tris (pH 7.01, 3 mM MgCl,, 1 p M CaCl,, and a trace amount of %a2+ (60 cpmllr!): Equzllbrium Dialysis-Clathrin light chains (1 mg/ml) were dialyzed with three changes of buffer against 10 mM MOPS (pH 7.0), 1 mM dithiothreitol, and 3 mM MgC12. A 0.3-ml sample of protein was then dialyzed for 24 h at 4 "C in the same buffer containing &Ca2+ and varying amounts of CaC12. Aliquots of the solution inside and outside the dialysis tubing were counted, and the protein concentration was determined.
Protein determinations were by the method of Bradford (12), and SDS-PAGE was performed according to the method of Laemmli (9) using 10 or 5-15% gradient slab gels as indicated.

RESULTS
Clathrin light chains were purified from fresh bovine brain using calmodulin affinity and gel filtration chromatographic  techniques. Routinely, about 8-10% of the protein applied to the calmodulin affinity column was eluted in the EGTAcontaining buffer. This peak was pooled and applied to a Sephadex G-100 column. The light chains eluted as a sharp peak consisting of 80-9096 of the applied protein. This peak contained clathrin light chains which were homogeneous as judged by SDS gel electrophoresis and Coomassie Blue protein staining (Fig. 1, lune A). Interestingly, when the light chains were analyzed by SDS-PAGE followed by silver staining for protein (13,14) no bands corresponding to the light chains could be detected (Fig. 1, lane B). As reported by Schleicher and Watterson (15), in many of the procedures for silver staining proteins on SDS gels, calcium-binding proteins such a calmodulin and troponin C are not detected. However, if the gel is pretreated with glutaraldehyde, the calcium-binding proteins can be detected with silver stain. It appeared that our clathrin light chains had silver staining characteristics similar to those of these calcium-binding proteins (Fig. 1). This figure also shows the near homogeneity of the light chain preparation.
The behavior of the light chains on silver staining raised the possibility that the light chains were calcium-binding proteins. Calcium binding was initially investigated using the nitrocellulose binding method of Maruyama et al. (8). By this method both light chains were found to bind Ca2+ (Fig. 2) in a buffer containing 150 mM KCl, 3 mM M e , and submicromolar Ca2+ concentrations. Analysis of triskelions by the nitrocellulose binding method (Fig. 2, lane B ) showed that

FIG. 2. Detection of calcium-binding proteins in purified light chain preparations and triskelion preparations by .'%a2+ autoradiography of nitrocellulose membranes.
Purified light chains and triskelions were electrophoresed on 5-15% SDS-polyacrylamide gels and then transferred to nitrocellulose membranes as described under "Materials and Methods." The membranes were labeled with 45Ca2+, autoradiographed, and then stained with Amido Black stain. Lanes A and B, amido black stain and autoradiograph of purified light chains (15 pg) on nitrocellulose membranes, respectively; lanes C and D, amido black stain and autoradiograph of triskelion preparation (50 pg), respectively. light chains derived from triskelions in the absence of a boiling step also bound Ca2+. This result suggested that the binding of Ca2+ to the purified light chains was not an artifact of the purification procedure. Neither the clathrin heavy chain nor any of the other proteins in the triskelion preparation were found to bind calcium.
Although bovine brain light chains have been reported to contain tightly bound nucleotide (16), the ultraviolet spectrum of our purified light chains showed an absorbance maximum at 280 nm (data not shown). This information demonstrated that our purified light chains did not contain bound nucleotide, thus favoring the proposed direct binding of Ca2+ to the light chains and not to associated nucleotide.
To confirm further our finding that clathrin light chains were Ca2+-binding proteins, the Hummel-Dreyer gel filtration method (11) was used. The typical gel filtration elution pattern obtained (Fig. 3A) was characteristic of Ca2+-binding proteins. The peak of 45Ca2+ coincided with the protein peak and was resolved from the resultant trough. A similar pattern was obtained when 10-fold higher concentrations of calcium were used. Assuming that 1 mol of Ca2+ binds per mol of light chain, a Kd of 26 p~ was calculated at both Ca2+ concentrations. Intact triskelions also showed the characteristic gel filtration pattern (Fig. 3B). The calculated Kd was 45 p~, assuming that 3 mol of calcium bound per mol of triskelion. As the triskelion preparation was relatively impure (Fig. 2,  lune C ) , this value is most probably an underestimate of the actual affinity. Since the gel filtration binding was performed in buffer containing 3 mM M$+, it followed that clathrin light chains contained specific Ca2+-binding sites. These sites appeared to be available for binding Ca2+ when the light chains were either in the free state or when in triskelions, where they were bound to the clathrin heavy chain.

Clathrin Light Chains Are Calcium-binding Proteins
Equilibrium dialysis experiments showed that the light chains bound, in the presence of 3 mM MgC12, 1 mol of Ca2+ per mol of light chain with a Kd of 54 p M (Fig. 4). The measured Kd tended to vary between 25 and 55 ptM for different preparations. We believe this to be due to slightly different ratios of the two light chains in the various preparations. However, we are uncertain whether the two light chains differ in their Ca2+-binding properties. In the absence of M$+, the light chains bound 1 mol of ca2+ with a Kd of 17.0 p~.

DISCUSSION
We have purified the clathrin light chains from bovine brain to near homogeneity and have shown that both light We have also shown that the light chains can bind Ca2+ either in the "free state" or when bound to the clathrin heavy chain in the form of triskelions. This finding suggests that Ca2+ binding by the light chains may occur when the triskelions are present in coat structures.
The binding affinity of the light chains for Ca2+ is less than that reported for other Ca2+-binding proteins such as calmodulin or troponin C but is in the same range as those reported for other Ca2+-binding, membrane-associated proteins. Calmodulin and myosin light chains belong to the Ca2+-binding protein family characterized by an " E F hand structure (17).
This family of proteins are in the 10-20 kDa range and most of them are soluble cytosolic proteins. A new Ca2+-binding family has been speculated to exist; this includes the calelectrins, endonexin, and the sarcoma virus tyrosine kinase substrate, p36 (18). These proteins share some sequence homology, have molecular masses predominantly in the 30-40 kDa range, and their affinity for Ca2+ is relatively low; a Kd = 100 p M for p36 (19) and a Kd = 70-80 p M for endonexin (20).
These proteins have been suggested to possess similar cellular functions; all are thought to be involved in various aspects of intracellular transport. The molecular masses of bovine brain clathrin light chains (36 and 33 kDa) and their affinity for Ca2+ (45 pM) are well within the range of these proteins. Also, the light chains are involved in intracellular transport, predominantly in endocytosis. Thus, we suggest that clathrin light chains are also members of this family of novel Ca2+binding proteins.
The low affinities for Ca2+ characterized by this family may be related to the observation that most of these proteins are membrane-associated. The concentration of Ca2+ at the endoplasmic face of the plasma membrane may be significantly higher than the measured mean value for cytosolic Ca2+ (21). The lower affinities may then be a reflection of the fluctuations of Ca2+ concentration at the membrane and not in the cytosol. Tupper and Bodine (3) showed a decrease in the endocytosis of the epidermal growth factor receptor by normal W138 human fibroblasts when the extracellular Ca2+ concen-

Clathrin Light Chains
Are Calcium-binding Proteins trations were lowered. They suggested that this decrease was due to a change in surface membrane localized Ca*+. Thus, it may be the local membrane Ca2+ concentration which is the important regulator of clathrin light chain function.
The physiological significance of our observation that clathrin light chains can bind Ca2+ as well as calmodulin awaits further elucidation. The light chains have been proposed to play a regulatory role in coated vesicle function. They have a stimulatory effect on a coated vesicle-associated protein kinase activity (22) and are required for the activity of an "uncoating" ATPase (23). They have also been conjectured to modulate clathrin-membrane interaction (24). The effect of Ca2+ on light chain function in these reported studies is unclear; however, if the light chains also share functional similarities with the calelectrins and other members of this calcium-binding family, the light chains may be important in coated pit formation at the plasma membrane. In view of our current findings, the modulatory role of Caz+ in clathrin light chain function deserves further investigation.