Human 76p: A new member of the gamma-tubulin-associated protein family

Abstract

The role of the centrosomes in microtubule nucleation remains largely unknown at the molecular level. gamma-Tubulin and the two associated proteins h103p (hGCP2) and h104p (hGCP3) are essential. These proteins are also present in soluble complexes containing additional polypeptides. Partial sequencing of a 76- kD polypeptide band from these complexes allowed the isolation of a cDNA encoding for a new protein (h76p = hGCP4) expressed ubiquitously in mammalian tissues. Orthologues of h76p have been characterized in Drosophila and in the higher plant Medicago. Several pieces of evidence indicate that h76p is involved in microtubule nucleation. (1) h76p is localized at the centrosome as demonstrated by immunofluorescence. (2) h76p and gamma-tubulin are associated in the gamma-tubulin complexes. (3) gamma-tubulin complexes containing h76p bind to microtubules. (4) h76p is recruited to the spindle poles and to Xenopus sperm basal bodies. (5) h76p is necessary for aster nucleation by sperm basal bodies and recombinant h76p partially replaces endogenous 76p in oocyte extracts. Surprisingly, h76p shares partial sequence identity with human centrosomal proteins h103p and h104p, suggesting a common protein core. Hence, human gamma-tubulin appears associated with at least three evolutionary related centrosomal proteins, raising new questions about their functions at the molecular level.

Figure 1

Comparison of the amino acid sequence of human 76p (h76p) with its Drosophila (d75p) and Medicago (m85p) orthologues. The alignments have been performed using both amino acid similarities (L/I/V/W/M, K/R/H, A/T/S/P/G, D/E/Q/N, and F/Y) and the conservation of the expected structural motifs ( Callebaut et al. 1997). Similarities are shown by shaded areas, while bold letters indicate identical amino acids. Consensus amino acids between the three orthologue proteins are underlined. The deduced nucleotide sequences of h76p, d75p, and m85p have been deposited in the GenBank/EMBL/DDBJ databases under accession numbers AJ249677, AJ249678, and AJ249679, respectively.

Figure 2

Characterization of 76p in the γ-tubulin complexes and in cell extracts. (A) Antibodies against h76p labeled the 76-kD polypeptide band. Pig brain γ-tubulin complexes submitted to SDS-PAGE were either stained with Coomassie blue (15 μg of γ-tubulin) or immunoblotted (50 ng of γ-tubulin) with antibodies R72, R801, R190, and R29/30. All antibodies labeled the polypeptide at ≈76 kD. The specificity of the labeling was controlled by preincubation with the immunizing peptide (+P). (B) Presence of 76p and γ-tubulin in the same protein complexes. Pig brain γ-tubulin complexes (250 ng of γ-tubulin) were immunoprecipitated with immune or preimmune h76p antibodies (R801). The immunoprecipitates (protein A–Sepharose beads, lanes b) and the corresponding supernatants (lanes s, one fifth of the initial volume) were analyzed by SDS-PAGE and immunoblotted with γ-tubulin antibodies (C3). A fraction of γ-tubulin specifically coimmunoprecipitated with 76p. (C and D) Detection and determination of the amount of 76p and γ-tubulin in PtK2 cell extracts (C) and Xenopus oocyte extracts (D). Total extracts (50 μg of proteins) were submitted to SDS-PAGE and immunoblotted with antibodies against 76p (R190 and R801) or against γ-tubulin (C3). The specificity was controlled in the presence of the immunizing peptide (+P) and the amounts of 76p (R190) and γ-tubulin (C3) were assessed by comparison with a range of purified recombinant h76p and Xenopus γ-tubulin. Both extracts contained equivalent amounts of 76p and γ-tubulin.

Figure 3

Biochemical properties of 76p. (A) 76p cosediment with γ-tubulin in Xenopus γ-TuRCs. An oocyte extract (1 mg of protein) was submitted to a sedimentation in a 5-ml linear 5–40% sucrose gradient (160,000 g for 4 h and 15 min at 0°C). The fractions (0.5 ml) were analyzed by SDS-PAGE and immunoblotting with γ-tubulin (C3) and 76p (R801) antibodies. Both proteins cosediment with an apparent mass corresponding to the γ-TuRCs (the arrow corresponds to the sedimentation of thyroglobulin at 669 kD or 19 S). (B) Presence of 76p in Xenopus γ-TuRCs. The presence of 76p in γ-TuRCs was assessed by the immunoprecipitation of γ-tubulin with x76p. Fractions 4–7 of the sucrose gradient (A) corresponding to γ-TuRCs were immunoprecipitated by immune or preimmune 76p antibodies (R190). The protein A–Sepharose beads (b) and a fraction of the supernatants were analyzed by SDS-PAGE and immunoblotted with γ-tubulin antibodies (C3). Protein A–Sepharose beads carrying preimmune antibodies failed to immunoprecipitate γ-tubulin, while a small fraction of γ-tubulin was recovered linked to the beads carrying the immune antibodies. (C) Determination of the amount of 76p in purified mammalian γ-tubulin complexes. Pig brain γ-tubulin complexes, submitted to SDS-PAGE, were immunoblotted with h76p antibodies (R801), and the amount of 76p was determined by comparison with a range of recombinant h76p. (D) Immunodepletion of a small fraction of γ-tubulin complexes by h76p antibodies. High molecular masses purified pig brain γ-tubulin complexes (M), obtained by centrifugation through a 40% sucrose cushion (≈500 ng of γ-tubulin) and a Xenopus oocyte extract (X, ≈2 mg of protein), were submitted to three successive cycles of immunodepletion with h76p antibodies (beads with R801 and R190 for purified mammalian γ-tubulin complexes and beads with R190 for Xenopus extract). Then the successive protein A–Sepharose beads (lanes b₁–b₃) and supernatant fractions (lanes s₁–s₃, 1/10 and 1/25 of the initial volume for mammalian γ-tubulin complexes and Xenopus extracts, respectively) were analyzed by SDS-PAGE and immunoblotting with γ-tubulin antibodies (C3).

Figure 4

Centrosomal localization of 76p during the cell cycle. (Top) Localization of 76p. PtK2 cells were labeled with antibodies against 76p (R801) and γ-tubulin (C3). (A and B) Double immunofluorescent labeling showing the colocalization of γ-tubulin (A) and 76p (B). (C) Specificity of the interphase labeling checked by preincubating h76p antibodies with the immunizing peptide. (D) Localization of 76p to the interphasic centrosome after a 2-h treatment in the presence of 2 μM colcemid that disassembled all the microtubule cytoskeleton. (E and F) Localization of 76p to the two separating centrosomes in prophase before and after the disappearance of the nuclear envelope, respectively. (G) Localization of 76p to the poles of the metaphase spindle. (H) Localization of 76p to the two centrosomes (arrows) in colcemid-treated mitosis. (I and J) Localization of 76p to the spindle poles in anaphase (I) and telophase (J). Note the relocalization of 76p between the two chromosomal masses (I and J, arrows). (K and L) Double immunofluorescent labeling of a cell undergoing cytokinesis; γ-tubulin (K) and 76p (L) are present at the centrosomes, but in contrast to γ-tubulin (K, arrows), 76p is absent at the two minus extremities of the midbody. (Bottom) Transient mitotic recruitment of 76p to the spindle poles. The maximal intensity of fluorescence (arbitrary units) raised by R801 antibodies over the centrosome in a single preparation of PtK2 cells was determined in interphase (I) for each of the two dots constituting the diplosome and in mitosis for each of the two mitotic poles (P, prophase; PM, prometaphase; M, metaphase; A, anaphase; T, telophase; and late T, late telophase). Average values are indicated by horizontal bars and experimental values are indicated by circles.

Figure 4

Centrosomal localization of 76p during the cell cycle. (Top) Localization of 76p. PtK2 cells were labeled with antibodies against 76p (R801) and γ-tubulin (C3). (A and B) Double immunofluorescent labeling showing the colocalization of γ-tubulin (A) and 76p (B). (C) Specificity of the interphase labeling checked by preincubating h76p antibodies with the immunizing peptide. (D) Localization of 76p to the interphasic centrosome after a 2-h treatment in the presence of 2 μM colcemid that disassembled all the microtubule cytoskeleton. (E and F) Localization of 76p to the two separating centrosomes in prophase before and after the disappearance of the nuclear envelope, respectively. (G) Localization of 76p to the poles of the metaphase spindle. (H) Localization of 76p to the two centrosomes (arrows) in colcemid-treated mitosis. (I and J) Localization of 76p to the spindle poles in anaphase (I) and telophase (J). Note the relocalization of 76p between the two chromosomal masses (I and J, arrows). (K and L) Double immunofluorescent labeling of a cell undergoing cytokinesis; γ-tubulin (K) and 76p (L) are present at the centrosomes, but in contrast to γ-tubulin (K, arrows), 76p is absent at the two minus extremities of the midbody. (Bottom) Transient mitotic recruitment of 76p to the spindle poles. The maximal intensity of fluorescence (arbitrary units) raised by R801 antibodies over the centrosome in a single preparation of PtK2 cells was determined in interphase (I) for each of the two dots constituting the diplosome and in mitosis for each of the two mitotic poles (P, prophase; PM, prometaphase; M, metaphase; A, anaphase; T, telophase; and late T, late telophase). Average values are indicated by horizontal bars and experimental values are indicated by circles.

Figure 5

Centrosomal localization of GFP-h76p in mammalian cells. Cells were observed 65 h after transfection with a plasmid expressing the fusion protein GFP-h76p. (A) Presence of GFP-h76p in COS-transfected cells. In control (C) cell extracts (20 μg), no polypeptide bands were detected both by anti–GFP (GFP) and anti–76p antibodies (R801). In transfected cells (T) both antibodies labeled the 103-kD fusion protein. The average expression of GFP-h76p in transfected cell extract (T cells) was determined with R801 antibodies by comparison with a range of recombinant h76p. (B–D) Centrosomal localization of GFP-h76p in COS-transfected cells. In mitotic cells, the GFP-h76p (B and C, upper row) localized at the spindle poles as shown by a coimmunofluorescent staining with α-tubulin antibodies (B, lower row), and γ-tubulin antibodies (C, lower row). In interphase cells, the GFP-h76p (D, upper row) colocalized with γ-tubulin (D, lower row) although the fluorescence raised by GFP-h76p appeared broader than the γ-tubulin staining. The transformed cells exhibited numerous multipolar spindles and showed frequently two diplosomes localized in close proximity in interphase, but controls performed with the vector expressing GFP alone suggested that the increase in centrosome number was a consequence of the transfection treatment rather than because of h76p overexpression. (E–F) Appearance of apoptotic figures in transfected HeLa cells. The transfected cells expressing GFP-h76p (E) exhibited a nuclear fragmentation shown by DAPI staining (F).

Figure 6

The 76p complexes are necessary for aster formation by permeabilized Xenopus spermatozoa. (A) Presence of 76p in sperm basal bodies after incubation in an oocyte extract. (1) Localization of 76p, labeled by antibodies R801 (black spot shown by the arrow), at the extremity of the sperm nuclei stained with DAPI (white). (2) Colocalization of 76p and (3) γ-tubulin at the extremity of the sperm basal body (arrows) (R801 and R75 antibodies, respectively). (B) Kinetics of aster nucleation by the sperm basal bodies. Depletion of 76p (left) and γ-tubulin (right) were conducted in the same experiment. (diamond) Control, spermatozoa incubated in an untreated Xenopus oocyte extract; (open circle) spermatozoa incubated in extract depleted with h76p (R801) or γ-tubulin (R74) antibodies; (open triangle) spermatozoa incubated in extract depleted with h76p or γ-tubulin antibodies in the presence of the corresponding immunizing peptide (500 μg/ml); (closed circle) spermatozoa incubated in extract depleted with h76p and γ-tubulin preimmune antibodies; (inverted triangle) spermatozoa incubated with antibodies unrelated to Xenopus centrosomal proteins (R82 directed against the carboxy-terminal region of Arabidopsis γ-tubulin); (square) spermatozoa incubated in h76p- or γ-tubulin–depleted extract further complemented with purified pig brain γ-tubulin complexes (corresponding to 2 ng of γ-tubulin and 0.4 ng of 76p), and (closed triangle) spermatozoa incubated in a 76p- or γ-tubulin–depleted extract complemented with an extract containing recombinant h76p (1.5 μg of recombinant h76p). No effects were observed when an extract devoid of recombinant protein was used (not shown). (C) Recruitment kinetics of 76p and γ-tubulin to the sperm basal bodies. Spermatozoa (Sp, 40 μg) were incubated in the oocyte extract in the presence of 5 μM nocodazole. Samples (Sp + E) taken at various times were centrifuged at 4°C for 5 min. 76p and γ-tubulin, recruited to spermatozoa, were revealed after SDS-PAGE by immunoblotting (R801 and R74 antibodies, respectively) and compared with untreated spermatozoa (Sp). (D) Specificity of the recruitment of the 76p and γ-tubulin to the sperm basal bodies. γ-tubulin and 76p were detected by immunoblotting with R74 and R801 antibodies, whereas the polypeptide band at 33 kD, observed after Coomassie blue staining, was used as a standard to check the recovery of sperm heads. (control, lane 1, Sp) Nonincubated spermatozoa (Sp, 40 μg), (lane 2, Sp + E) spermatozoa incubated for 20 min in the oocyte extract (E), (lane 3, Sp + E + IS) spermatozoa incubated in oocyte extract depleted with R801 immune serum (IS), (lane 4, Sp + E + IS + p) spermatozoa incubated in oocyte extract depleted with the R801 immune serum in the presence of the immunizing peptide (p), (lane 5, Sp + E + pIS) spermatozoa incubated in the oocyte extract depleted with R801 preimmune serum (pIS), and (lane 6, E) oocyte extract incubated alone showing that the recruitment is linked to the presence of spermatozoa. (E) Quantification of the amount of 76p in immature and mature sperm basal bodies. Spermatozoa (Sp, 160 μg) and spermatozoa incubated for 20 min in the oocyte extract in the presence of 5 μM nocodazole (Sp + E, 40 μg) were immunoblotted with R801 antibodies and the amount of 76p was assessed by comparison with a range of purified recombinant h76p.

Figure 7

Comparison of the amino acid sequences of human 76p, 103p (h97p) and 104p (h98p). (A) Regions of homology of the three proteins (shaded area). (B) Sequence comparison in the regions of homology. Identical and homologue amino acids are shown by bold letters and by shaded areas, respectively. Consensus amino acids between h76p, h103p and h104p are underlined. Comparison between Fig. 1 and Fig. 7 shows that the consensus amino acids between h76p, d75p, and m85p and the consensus amino acids between h76p, h103p, and h104p are 18% identical. The alignments have been performed using both the similarities of sequences and the conservation of the expected structural motifs ( Callebaut et al. 1997).