Autoubiquitination of BCA2 RING E3 ligase regulates its own stability and affects cell migration

Abstract

Accumulating evidence suggests that ubiquitination plays a role in cancer by changing the function of key cellular proteins. Previously, we isolated BCA2 gene from a library enriched for breast tumor mRNAs. The BCA2 protein is a RING-type E3 ubiquitin ligase and is overexpressed in human breast tumors. In order to deduce the biochemical and biological function of BCA2, we searched for BCA2-binding partners using human breast and fetal brain cDNA libraries and BacterioMatch two-hybrid system. We identified 62 interacting partners, the majority of which were found to encode ubiquitin precursor proteins including ubiquitin C and ubiquitin A-52. Using several deletion and point mutants, we found that the BCA2 zinc finger (BZF) domain at the NH(2) terminus specifically binds ubiquitin and ubiquitinated proteins. The autoubiquitination activity of BCA2, RING-H2 mutant, BZF mutant, and various lysine mutants of BCA2 were investigated. Our results indicate that the BCA2 protein is strongly ubiquitinated and no ubiquitination is detected with the BCA2 RING-H2 mutant, indicating that the RING domain is essential for autoubiquitination. Mutation of the K26 and K32 lysines in the BZF domain also abrogated autoubiquitination activity. Interestingly, mutation of the K232 and K260 lysines in and near the RING domain resulted in an increase in autoubiquitination activity. Additionally, in cellular migration assays, BCA2 mutants showed altered cell motility compared with wild-type BCA2. On the basis of these findings, we propose that BCA2 might be an important factor regulating breast cancer cell migration/metastasis. We put forward a novel model for BCA2 E3 ligase-mediated cell regulation.

Figure 1

A) Schematic diagram of BCA2. The position of the six lysine residues, K26 and K32 (N-terminal lysines), K206 and K208 (internal lysines) and K232 and K260 (C-terminal lysines), are shown as gray letters. Conserved cysteine (C22, C25, C38 and C41) residues in the BCA2 zinc finger (BZF), which we show is the functional ubiquitin binding domain and cysteine and histidine (C228, C231, C236, H238, H241, C244, C255 and C258) residues in the RING domain are shown on black background. The AKT phosphorylation sites (S132 and S133) are shown on gray background. B) Gene names and accession numbers of positive clones recovered from the bacterial two-hybrid tertiary screen of a human fetal brain library with the BCA2 wild type protein as bait are shown. C) Recombinant BCA2 binds ubiquitin. GST-Ub(n) (1, 2, 4, or 6 ubiquitin tandem repeats) immobilized on glutathione sepharose beads were incubated with recombinant His-tagged BCA2. The pulldowns were immunoblotted with anti-His antibody (upper panel). The coomassie brilliant blue R-250 staining of SDS-PAGE gel shows the amount of GST-Ub(n) fusion proteins used in each pulldown assay (lower panel).

Figure 2

BCA2 interacts with ubiquitin via BZF domain. (A) Sequence comparison of the BZF domain and various zinc finger type ubiquitin binding domains. Amino acid sequence were taken from PAZ (25), NZF (26), UBZ (27) and A20 (28). Gray shaded letters denote conserved cysteine and histidine residues for zinc finger domains. Open boxes show conserved amino acids within BZF domains of BCA2 and RNF126. A dash indicates a gap inserted to maximize conserved amino-acid alignment. (B) Schematic diagram of BCA2 and its deletion mutants used for directed bacterial two hybrid assay. (C) Bacterial cells expressing the BCA2 wild type or N-terminal fragment (1-225) of BCA2 interact with both 6 tandem ubiquitin and UbA-52 ribosomal fusion protein grown in selective plate containing 3-AT and streptomycin while pBT empty vector, BCA2 internal fragment (47-225) and C-terminal fragment (226-304) failed to interact with Ubiquitin. (D) Schematic diagram of BCA2 and its point mutants used for pull down assay in (E). (E) Recombinant his-tagged BCA2 and its BZF, AKT and RING mutants were subjected to pull-down assays with GST-Ub(4) immobilized on glutathione sepharose beads. GST-Ub(4) bound to BCA2 wild type, AKT and RING mutants but not to BCA2 BZF mutant.

Figure 3

Quantification of BZF domain affinity for ubiquitin. (A) Determination of the dissociation constant by measuring the amount of ubiquitin bound to BZF domain of BCA2 using SPR. Ubiquitin protein was injected in triplicate at concentrations of 0-1000 mM over His tagged BCA2 immobilized on CM5 sensor chip (squares, Kd = 29.6+/-3.2mM). Much weaker binding (Kd) response was obtained when the same concentrations of ubiquitin was injected over BCA2 treated with zinc ejecting compound, disulfiram (triangles with dotted line). Fitted curves are shown. Error bars indicate standard deviation based on three independent experiments. (B) Comparison of binding affinity of BCA2-BZF domain and various distinct ubiquitin binding domains, UEV (34), UBA (35), UIM (36), NZF (26) and A20 (28) All binding affinities were measured by SPR.

Figure 4

Autoubiquitination of the BCA2 proteins. BCA2 and its mutants as indicated above each lane were expressed using the pCMV-tag2B-Flag vector or pEBG-GST vector in HEK293T cells. Flag-tagged proteins were immunoprecipitated (IP) with monoclonal anti-Flag antibodies and incubated with protein A sepharose beads. GST-tagged proteins were pull-downed with Glutathione Sephaose 4B. Washed beads were incubated with ubiquitin, ATP, mouse E1, and with (+) or without (-) the bacterial E2 UbcH5b. Left panel shows anti-Flag immunoblot of autoubiquitination assay for Flag-tagged BCA2. Right panel shows anti-GST immunoblot of autoubiquitination assay for GST-tagged BCA2.

Figure 5

Time course for the detection of BCA2 and its mutants autoubiquitination. Time course of in vitro ubiquitination reactions (15 min, 30 min, 60 min and 90 min) conducted in the presence (+) or absence (-) of bacterial extract containing the E2-conjugating enzyme UbcH5b and purified ubiquitin with E1 protein, using bacterially expressed (A) His-BCA2, (B) His-BCA2 AKT mutant, (C) His-BCA2 BZF mutant, (D) His-BCA2 Ring mutant, (E) His-BCA2 N-terminal lysine (K26R, K32R) mutant, (F) His-BCA2 K26R mutant, (G) His-BCA2 K32R, (H) His-BCA2 internal lysine (K206R, K208R) mutant, (I) His-BCA2 C-terminal (K232R, K260R) mutant, (J) His-BCA2 K232R mutant, (K) His-BCA2 K260R mutant, (L) and His-BCA2 BZF + C-terminal (K232R, K260R) mutant. Autouiquitination assay conditions are described in Material and Methods

Figure 5

Time course for the detection of BCA2 and its mutants autoubiquitination. Time course of in vitro ubiquitination reactions (15 min, 30 min, 60 min and 90 min) conducted in the presence (+) or absence (-) of bacterial extract containing the E2-conjugating enzyme UbcH5b and purified ubiquitin with E1 protein, using bacterially expressed (A) His-BCA2, (B) His-BCA2 AKT mutant, (C) His-BCA2 BZF mutant, (D) His-BCA2 Ring mutant, (E) His-BCA2 N-terminal lysine (K26R, K32R) mutant, (F) His-BCA2 K26R mutant, (G) His-BCA2 K32R, (H) His-BCA2 internal lysine (K206R, K208R) mutant, (I) His-BCA2 C-terminal (K232R, K260R) mutant, (J) His-BCA2 K232R mutant, (K) His-BCA2 K260R mutant, (L) and His-BCA2 BZF + C-terminal (K232R, K260R) mutant. Autouiquitination assay conditions are described in Material and Methods

Figure 6

Comparison of E3 ligase activity of BCA2 and its mutants. The optical density (OD) of individual band of each Western blot (Fig. 6 A - L) was quantitated using the Quantity One software (ver. 4.3.1, Bio-Rad Laboratories, Segrate, Milan, Italy) and normalized by subtracting the OD of an area of identical size from the blank lane of the each gel.

Figure 7

Stability of transiently expressed BCA2 wild type and its mutants. Equal amounts of His-tagged BCA2 wild type and its mutant vector DNA were transfected into HEK293T cells. The total transfection period was 48 hours. 10mM of MG-132 (+) or DMSO (-) were added 40 hours after transfection, resulting in drug exposure of 8 hours. Whole cell lysate were immunoblotted with anti-His (upper panel) or β-actin was used as a loading control (lower panel). Bar graph shows normalized fold protein level of BCA2 and its mutants relative to BCA2 wild type MG-132 (-) using β-actin as internal control.

Figure 8

BCA2 and its variants expression in MCF7 cells. MCF7 cell lines that stably express BCA2 or selected mutant variants were examined by immunofluorescence as described in Material and Methods. Fixed cells were probed with the monoclonal anti-polyhistidine (AD1.1.10) antibody for 8 hours followed by anti-mouse FITC-secondary fluorescent antibody. Cell nuclei were stained with DAPI. In situ immunoflourecence reveals different expression patterns for the BCA2 protein variants. (A) MCF7 cells expressing BCA2 wild type protein. (B) MCF7 cells expressing BCA2 N-terminal lysine mutant protein. (C) MCF7 cells expressing BCA2 C-terminal lysine mutant protein. (D) MCF7 cells expressing BCA2 RING mutant protein. (E) MCF7 cells expressing BCA2 BZF mutant protein. (F) MCF7 cells expressing His tag alone (Empty vector). (G) Negative control in which only secondary FITC antibody was incubated with MCF7 cells expressing the His tag from the empty vector.

Figure 9

Effects of BCA2 and mutants on MCF7 cell migration. (A) anti-His immunoblot of MCF7 cells stably transfected with pEF1-His-BCA2 wild type and its mutants (upper panel). Non-specific band obtained by anti-His was used as a loading control (lower panel). (B) MCF7 cells stably expressing BCA2 or its mutants were serum starved for 18 h, scratched with a sterile pipette tip, and treated with DMEM medium containing 10% FBS and 10mM Hydroxyurea. Cells were allowed to migrate into the denuded area, and phase micrographs taken at 0, 24, and 48 h. (C) Graph shows the percentage of recovered wound area from the initial wound area after 24 and 48 h of migration. Error bars indicate S.D. based on three independent experiments.

Figure 10

Proposed Model for BCA2 Regulation