The Hook family of proteins has a conserved NH₂-terminal domain and central coiled coil with a divergent COOH-terminal domain. (A) Hook proteins are schematically aligned, and percent identities are indicated for the NH₂- and COOH-terminal globular domains and the central coiled coil (gray) compared with hHK3. The sequence of the hHK3 protein was deposited in Genbank (sequence data available from GenBank/EMBL/DDBJ under accession number AF241830). (B) Endogenous Hook proteins in HEK293 cell lysates were detected with affinity-purified antisera. Proteins close to the predicted size for hHK1 (lane 1), hHK2 (lane 3), and hHK3 (lane 5) were specifically recognized. Preabsorption of antibodies with 10 μg/ml of a corresponding His₆ fusion protein abolished detection (lanes 2, 4, and 6). (C) Expression of Hook3 protein was detected in all murine tissues examined, 293 cells, and all human and monkey cell lines tested (data not shown). We do not know whether the bands of ∼36 kD detected in liver and kidney cells correspond to degradation products, alternative splice forms, or cross-reacting proteins. Int, intestine; Liv, liver; Kid, kidney; Spl, spleen; Hrt, heart; Lng, lung; Skm, skeletal muscle, Brn, brain; Stm, stomach; 293, HEK293 cells.

Endogenous human Hook proteins function in distinct complexes. HEK293 postnuclear supernatants were subjected to immunoprecipitation using the indicated hHK antisera (IP αhHK), either directly (−) or after cross-linking with 100 μM BS³ (+). Precipitated Hook proteins were detected with the respective hHK antiserum by Western blotting. Each hHK antiserum is able to specifically immunoprecipitate its respective endogenous protein from noncross-linked samples (hHK1, lane 1; hHK2, lane 9; hHK3, lane 17). In each of the cross-linked samples, high molecular weight complexes are detected with the specific anti-hHK antibodies used for their immunoprecipitation (hHK1, lane 2; hHK2, lane 10; hHK3, lane 18) but not the other two anti-hHK antibodies (lanes 4, 6, 8, 12, 14, and 16). The identity of the bands at ∼85 kD in lanes 3 and 5 has not been determined.

Endogenous hHK 3 localizes to the cis-face of the Golgi complex. (A–C) The localization of endogenous Hook3 in Vero cells was analyzed by preimbedding immunogold electron microscopy. The panels show three representative examples of the range of staining observed. In the Golgi area, the labeling was polarized to one face of the Golgi complex (A and B, arrows), which, based on the electron micrographs and the immunofluorescence colocalization experiments (D–O), was identified as the cis-face of the Golgi complex. Labeling of the TGN was negligible in all cells observed. Outside the Golgi complex, vesicles of unknown identity were labeled (arrowheads). (D–O) Localization of hHK3 was compared with several Golgi markers by double immunofluorescence staining in Vero cells: the cis-Golgi marker GM130 (D–F), the intermediate compartment marker syntaxin5 (G–I), the trans-Golgi markers TGN46 (J–L), and clathrin (M–O). The insets show details of the staining in higher magnification. In the merged images (F, I, L, and O), hHK3 staining is shown in red; and the Golgi markers, in green. N, nucleus. Bar, 0.5 μm.

Association of hHK3 with microtubules in vivo. hHK3 localization was compared with α-tubulin (B, H, K, and N) or γ-tubulin (E) in Vero cells that were either treated with BfA (A–F), visualized during mitosis (G-L), or in which microtubules were stabilized with taxol (M–O). During the early stages of mitosis (prophase), hHK3 (G–I) is found associating with the microtubule asters (arrowheads). Later in mitosis (metaphase), hHK3 is found in a more diffuse cytoplasmic pattern (J–L). During cytokinesis, hHK3 is relocalized in a tight juxtanuclear focus (G–I, arrow). It is important to notice that images acquired from rounded mitotic cells lie above the focal plane of the surrounding, flattened interphase cells. As a result, fluorescence intensities are not directly comparable between these two populations of cells. When depolymerized microtubules were bundled in vivo in the presence of taxol (40 μM), hHK3 protein (M) was enriched near the ends of microtubule bundles (M–O, arrows).

hHK3 binds Golgi membranes through a COOH-terminal domain. Purified Golgi membranes (Ktistakis et al. 1996) were incubated with cytosol from HEK293 or HEK293:ΔC-hHK3_1–555 cells. Bound proteins were collected by centrifugation through a 20% sucrose cushion, and pellets were probed for the presence of hHK1, hHK2, and hHK3 proteins by immunoblotting. None of the Hook proteins was detected in samples without added cytosol (lanes 1 and 2). In samples combining HEK293 cytosol and purified Golgi membranes, hHK3, but not hHK1 or hHK2, was found in the Golgi-containing pellets (lanes 4 and 6). When HEK293:ΔC-hHK3_1–555 cytosol was used, endogenous full-length hHK3 was found in the pellet, but ΔC-hHK3_1–555 was not foundin the pellet (lane 6), and it instead was found entirely in the supernatant (lane 5).

Juxtanuclear localization of hHK3 after BfA treatment. Localization of endogenous Hook3 in Vero (A–L) and HeLa cells (M–O) was compared with that of β-COP (B and N), TGN46 (E), and GM130 (H and K), after treatment with BfA (10 μg/ml) for 1 h (A–I) or BfA (10 μg/ml) for 1 h, and then BfA and NZ (10 μg/ml) for 1 h (J–L) or NZ (10 μg/ml) alone for 1 h (M–O). After 1 h of BfA treatment, β-COP protein was cytosolic (B), but the majority of hHk3 protein remained in a juxtanuclear position (A, D, and G). In BfA-treated cells, a fraction of hHK3 protein colocalized with the cis-Golgi matrix protein GM130 in distinct punctae (G–I, arrows). The majority of hHK3 protein was released from its juxtanuclear position after NZ was added to the BfA-treated cells; then, the majority of hHK3 localized to the GM130-positive punctae (J–L, arrows). After NZ treatment of HeLa cells, hHK3 (M) was associated with the resultant Golgi fragments (arrowheads), which were identified by β-COP labeling (N). In the merged images, hHK3 is visualized in red; and Golgi markers, in green.

Endogenous hHK3 localizes to the Golgi complex. The localization of endogenous Hook proteins in Hep2 cells (A, D, G, J, and M) was compared with the Golgi marker FTCD (B, E, and K), microtubules (H), or ERGIC-53 (N). hHK1 localized to discrete unidentified subcellular structures (A) that do not significantly overlap with the Golgi complex (B). hHK2 localized to discrete subcellular structures that were often observed in linear tracks (D and G), which colocalized with microtubules (H and I). Much of hHK3 staining (J) precisely colocalized with a marker of the cis- and/or medial-Golgi (K), FTCD (Bashour and Bloom 1998). hHK3 labeling detected outside the Golgi (M–O, inset) did not colocalize with ERGIC-53, a marker for ER-to-Golgi intermediates (M–O). In the merged images (C, F, I, L, and O), staining for FTCD, microtubules, or ERGIC-53 is shown in green; and the Hook proteins, in red.

Hook proteins bind microtubules through their conserved NH₂-terminal domains. (A) Microtubule spin-down assays were performed using cytosol from HEK293 cells containing endogenous hHK1, hHK2, and hHK3 and expressing the COOH-terminal truncations ΔC-hHK1_1–555, ΔC-hHK2_1–548, and ΔC-hHK3_1–555. Samples from these assays were analyzed by Western blots using the indicated anti-Hook antibodies. hHK1, hHK2, and hHK3 are detected in the top, middle, and bottom blots, respectively. None of the human Hook proteins was detected among the purified microtubules (lanes 1 and 2). When no microtubules were added, all human Hook proteins were found in the supernatant (lane 5) but not in the pellet (lane 6). When cytosol and microtubules were combined, all full-length endogenous human Hook proteins as well as COOH-terminal truncations of each Hook protein were found copelleted with microtubules (lane 4). (B) To test for direct binding of hHK3 to microtubules, agarose beads coupled to the purified fusion proteins His₆–hHK3N1_1–164, His₆–hHK3N2_1–224, or His₆–hHK3CC_423–630 were tested for binding to stabilized microtubules. Excess tubulin is found in the supernatant of all samples (lane 1, 3, and 5). Fusion proteins containing the NH₂-terminal 164 aa of hHK3 can bind microtubules (lanes 4 and 6), whereas a fusion protein from the hHK3 coiled coil does not (lane 2).

The dHKs bind microtubules. (A) Cytosol from Drosophila tissue culture cells was used in microtubule spin-down assays to test the ability of dHK to associate with bovine microtubules. dHK was found in the pellet associating with microtubules (lane 4) only when Drosophila cytosol and taxol-stabilized microtubules were combined. (B) dHK protein also bound to endogenous Drosophila microtubules when they were stabilized by the addition of GTP–taxol. Tubulin and dHK were pelleted in the presence of taxol–GTP (lane 4) but not in their absence (lane 2).

Overexpression of hHK3 proteins disrupts Golgi morphology. Transfected cells overexpressing ΔC-hHK3_1–555 (A–C, Q, and R), full-length hHK3 (D–I), a dHK–hHK3C_589–718 chimera (J–L), or a dHK_6–562 truncation (M–O) were detected by the high levels of Hook proteins. Cells overexpressing ΔC-hHK3_1–555 (A) or hHK3 (D and G) exhibited a severely disrupted Golgi morphology, as indicated by the KDEL receptor (B and E) or GM130 (H) localizations. Arrowheads in A–F point to the well-defined Golgi complexes of untransfected cells. The COOH-terminal domain of hHk3 was sufficient to direct the dHK–hHK3C_589–718 chimera to the Golgi complex (J–L) but did not noticeably disrupt Golgi morphology (K). By itself, the dHK_6–562 part of the chimera was evenly distributed in the cytoplasma (M–O). All Hook chimeras and truncations listed in S, but not an unrelated control plasmid, caused a disruption of the microtubular network. Microtubules in transfected cells appeared noticeably disordered (R), often crossing in the cell periphery, a rare occurrence in untransfected cells (P). The disruption of the microtubular network paralleled an increase in the number of cells with multiple or multilobed nuclei in transfected cells (S; e.g., two nuclei in A are labeled N). In S, the gray bars indicate the dHK portion of chimeras.