Identification of the Phosphorylation Sites of Clathrin Light Chain LCb*

Clathrin light chains, LC, and Lcb, are products of two closely related genes whose mRNAs undergo dif- ferential splicing to result in at least four different light chain isoforms. The physiological significance of clathrin light chain diversity remains unclear. To date, the only evidence for a functional distinction of LC, and LC), is the preferential phosphorylation of LCb, which takes place at serine residues and is mediated by coated vesicle-associated casein kinase 11. As a first step toward determining the function of light chain diversity, we have mapped the in vitro phosphoryla- tion sites on Lcb. We use r3'P]ATP to phosphorylate LCb within coated vesicles, followed by sequencing of 32P-labeled chymotryptic peptides thereof, to identify serine residues at positions 11 and 13 as the phospho- rylation sites. We find that phosphorylation of Lcb within coated vesicles can be inhibited by four mono- clonal antibodies specific for different epitopes of the clathrin light chains.

Clathrin is the principal protein component of the coat of coated pits and vesicles, structures which regulate receptormediated endocytosis and intracellular transport between membrane-bound compartments in eukaryotic cells (1). The unit structure of the clathrin coat is the triskelion which contains three heavy and three light chains of two types, LC,,' and Lcb (2). Investigations into the assembly of clathrin in vitro have suggested that the light chains function in a structure-stabilizing or regulatory role (3). In this regard, phosphorylation of Lcb and not of LC. or the clathrin heavy chain by coated vesicle-associated casein kinase I1 may be of particular functional significance (4).
The phosphorylation of Lcb has been reported in both in vitro and in vivo systems (5)(6)(7)(8) and is known to occur on serine residues (6). Our previous analysis of the primary * This research was supported by grants from the American Cancer Society (to K. D. and P. P.), the National Science Foundation (to F. M. B.), and the National Cancer Institute, National Institutes of Health (to B. L. H.). 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.
The abbreviations used are: LC., clathrin light chain a; LCb, clathrin light chain b; HPLC, high performance liquid chromatography; SDS-PAGE, sodium dodecyl sulfate-polyacrylamide gel electrophoresis.
structures of light chains showed that at positions 9-13 Lcb has the sequence Ser-Ser-Ser-Glu-Ser whereas the corresponding region of LC, has the sequence Gly-Gly-Pro-Ala-Leu (9). This amino-terminal region provided a possible site for LCb-specific phosphorylation. Here we present experiments demonstrating that this hypothesis is correct.

EXPERIMENTAL PROCEDURES
Phosphorylation and Purification of Clathrin Light Chains-Clathrin-coated vesicles from three bovine brains were purified and phosphorylated as previously described (5) with the following modifications of the phosphorylation reaction: 2 mg/ml clathrin-coated vesicles, 100 p~ [y-32P]ATP (100 mCi/mmol), 50 pg/ml polylysine ( M , = 40,000). Light chains were isolated from Tris extracts of the coated vesicles by boiling, dialysis against water, and high speed centrifugation as previously described (10). The light chain-containing supernatant was concentrated by ultrafiltration. The light chains were separated from low molecular weight contaminants by HPLC on a reverse phase C3 column (Beckman Instruments), eluting with a gradient of acetonitrile in 0.1% trifluoroacetic acid. Aliquots of 32Pcontaining HPLC fractions were analyzed by SDS-PAGE followed by autoradiography.
Proteolysis of Phosphorylated LC,-HPLC fractions containing specifically phosphorylated Lcb were pooled and dialyzed against water to prepare for cleavage with clostripain or chymotrypsin using a 1:50 (w/w) ratio of enzyme to light chain. Conditions for digestion with clostripain were 0.1 M Tris-HCI, pH 7.4, 1 mM CaC12, 1 mM MgCl,, and 1 mM dithiothreitol. After 12 h a t 37 "C, clostripain was inactivated by adding iodoacetamide to 1 mM. Conditions for digestion with chymotrypsin were 0.1 M NH4HC03, pH 8.1, for 4.5 h a t 37 "C. Chymotrypsin was inactivated by adding trifluoroacetic acid to 0.1%. After inactivation of protease, samples were acidified with trifluoroacetic acid to 0.1%, centrifuged for 10 min in an Eppendorf microcentrifuge, and applied to the reverse phase columns.
Amino Acid Sequence Determination and Identification of Phosphorylated Residues-Amino acid sequences of chymotryptic peptides were determined using a pulsed-liquid sequenator. Radioactivity in each sequencing cycle was quantitated by liquid scintillation counting.
Inhibition of Phosphorylation by Monoclonal Antibodies-Coated vesicles (30 pg) and monoclonal antibodies (11) a t varying concentrations were preincubated in phosphorylation buffer ((5) 45 pl total volume) for 1 h at 4 "C. The phosphorylation reaction (50 p1 total volume) was initiated by adding polylysine to 50 pg/ml and [32P]ATP to 20 p~ (100 mCi/mmol), and the reaction continued for 5 min a t 30 "C. The reaction was stopped by the addition of 50 pl of 2 X SDS-PAGE sample buffer and boiling for 3 min; 10 pl was analyzed by SDS-PAGE (12). Gels were stained, dried, and autoradiographed for 16-20 h.

RESULTS AND DISCUSSION
Incubation of purified coated vesicles with [T-~'P]ATP resulted in specific phosphorylation of Lcb and not of LC, or the clathrin heavy chain ( 5 ) . After purification from the coated vesicles, 32P-labeled Lcb was first digested with chymotrypsin and the peptides analyzed by reverse phase HPLC. A single peak of radioactivity was detected which eluted with -35% acetonitrile (Fig. la). Sequence analysis of the 32Pcontaining fractions indicated the presence of at least two peptides. Inspection of the mixed sequence suggested that one peptide corresponded to a region around residues 9-13 and that another corresponded to the sequence around residue 197. Previously we have shown that these two sequences, both of which contain serine, are found in different clostripain fragments of LCb (9). Thus, by first isolating 32P-labeled clostripain fragments and subsequently digesting them with chymotrypsin, we were able to obtain pure 32P-labeled peptides of Lcb.
Reverse phase HPLC analysis of the clostripain digest of LCb showed three 32P-containing peaks, designated A, B, and C (Fig. lb). Fractions corresponding to each of these peaks were digested with chymotrypsin and analyzed separately. Chymotryptic digestion of each of A, B, and C produced one major 32P-containing peak (Fig. IC). Sequencing of each of the 32P-labeled chymotryptic fragments showed they all derived from the amino-terminal region of Lcb and that the multiple peaks resulted from heterogeneous cleavage of Lcb by protease. Each of the 32P-labeled chymotryptic peptides showed two phosphorylation sites occurring at serine residues 11 and 13 located within a cluster of 4 serine residues near the amino terminus (Fig. 2). The major chymotryptic cleavage products of A, B, and C were identical except for the variable presence of 2 amino acid residues at either terminus. Two minor peaks resulting from cleavage of A (not shown) corresponded to the major cleavage products of B and C. A single minor peak resulting from cleavage of C (Fig. IC) corresponded to the major cleavage product of B.
These results confirm our prediction that the serine-rich sequence between positions 9 and 13 is the site of LCb-specific phosphorylation. The corresponding region of LC. contains u" 600 -> u \ r:  (19,20)) and dephosphorylated aaZ casein (Casein as2 (21)) are shown. The phosphorylated serines are underlined. Although more than 20,000 cpm were present in the peptide sample sequenced, anilinothiazolinone-phosphoserine is poorly extracted by butyl chloride during sequencing resulting in a minimal recovery of the radioactivity. no serine residues, thus explaining the differential phosphorylation of the two clathrin light chains. It is also highly likely that this region is the only site of phosphorylation as no evidence for 32P labeling of other sequences has been obtained using conditions of HPLC analysis that resolve intact LC, and all proteolytic fragments.

PK-A.reg
The phosphorylated serines and surrounding sequences of Lcb are conserved between species (9,13). When compared with other known target sequences of casein kinase 11, the phosphorylation site of Lcb is similar, having the requisite acidic residues carboxyl terminal to the phosphorylation site ((14) Fig. 2). As previously reported for casein kinase 11, heparin and calcium inhibited the phosphorylation of LC, while polylysine stimulated Lcb phosphorylation ( ( 5 ) not shown).
Recently, Kohtz et al. (15) reported two monoclonal antibodies, specific for Lcb, whose binding to coated vesicles inhibited the phosphorylation of Lcb. We found that 4 of 14 monoclonal antibodies to Lcb inhibit the phosphorylation of LC,,. Two of these antibodies are specific for Lc1, (LCB.3 and X44), and two bind both LCb and LC, (LCB.l and X43). LCB.l and LCB.3 were the most potent, showing inhibitory activity at a concentration of 100 pg/ml (Fig. 3b). X43 and X44 became inhibitory at 200 Fg/ml (Fig. 3a). In each case, the inhibition is dose-dependent and specific; the phosphorylation of pp50 and tubulin, the other major phosphorylated components of the coated vesicle preparation, is not inhibited. Previously, the LCB.3 and X43 epitopes were mapped to epitopes present on two different carboxyl-terminal fragments of LCb, while the locations of the LCB.l and X44 epitopes were undetermined (11). In the present study, none of these antibodies bound to a synthetic peptide encompassing the phosphorylation site of LCb and corresponding to residues 2-21 (not shown). Therefore, these antibodies are unlikely to compete with casein kinase I1 at the phosphorylation site. These data are consistent with either of two explanations. Casein kinase 11 may interact with an extensive region of Lcb, carboxyl-terminal to the phosphorylation site, such that binding of antibodies within these regions interferes with phosphorylation. Alternatively, carboxyl-terminal regions of LCI, may be sufficiently close to the phosphorylation site that Both X43 and X44 have been mapped to epitopes that are cryptic in assembled coated vesicles (11), results which are consistent with a lower affinity of these antibodies for assembled, as compared to free, Lch. In the present study, this affinity limitation is lessened, since the antibody concentration is constant during the course of the phosphorylation reaction. Furthermore, breathing of the clathrin structure under these conditions could allow binding to epitopes that are only transiently accessible.
Immunoelectron microscopy studies have shown that clathrin light chains extend along the length of the proximal arm of the triskelion with the carboxyl terminus of the light chain oriented toward the triskelion vertex (16,17). This orientation places the amino-terminal phosphorylated serines of Lch near the triskelion elbow, a location where they may influence regional interactions between clathrin heavy and light chains (18). Serines 11 and 13 are only 9 and 7 amino acid residues amino-terminal to a sequence (residues 20-41) that is conserved absolutely between LC. and Lch and between light chains of rat and bovine origin (9,13). Phosphorylation of serines 11 and 13 provides a potential mechanism to modify differentially the functions of this region of the light chains.