|Home | About | Journals | Submit | Contact Us | Français|
Centrosomes are the cellular organelles that nucleate microtubules (MTs) via the activity of gamma-tubulin ring complex(s) (γTuRC) bound to the pericentriolar material of the centrosomes. BRCA1, the breast and ovarian cancer specific tumor suppressor, inhibits centrosomal MT nucleation via its ubiquitin ligase activity, and one of the known BRCA1 substrates is the key γTuRC component, γ-tubulin. We analyzed the mechanism by which BRCA1 regulates centrosome function using an in vitro reconstitution assay, which includes separately staged steps. Our results are most consistent with a model by which the BRCA1 ubiquitin ligase modifies both γ-tubulin plus a second centrosomal protein that controls localization of γTuRC to the centrosome. We suggest that this second protein is an adapter protein or protein complex that docks γ-TuRC to the centrosome. By controlling γ-TuRC localization, BRCA1 appropriately inhibits centrosome function, and loss of BRCA1 would result in centrosome hyperactivity, supernumerary centrosomes and, possibly, aneuploidy.
BRCA1 is an ovarian and breast cancer specific tumor suppressor that associates with BARD1 to form a heterodimer with powerful E3 ubiquitin ligase activity.1,2 BRCA1 and BARD1 localize to centrosomes as well as the nucleus,3,4 and BRCA1 binds to γ-tubulin.5 Suppression of BRCA1 expression in breast cells leads to centrosome amplification and hyperactivity.4,6 BRCA1/BARD1 mono-ubiquitinates γ-tubulin at lysines K48 and K344.6 Expression of mutant γ-tubulin with a substitution at lysine-48 to arginine in live cells results in a dominant phenotype of centrosome amplification and hyperactivity.4,6 Since inhibition of BRCA1 enzyme activity or the expression of a mutant substrate (γ-tubulin), produce the same phenotype, the results suggest an enzyme-substrate relationship between BRCA1 and γ-tubulin. This biochemical reaction catalyzed by BRCA1 is thought to restrain centrosomes from re-duplication. Thus, when BRCA1 function is lost, centrosome amplification occurs.7 This phenotype of centrosome amplification and hyperactivity is a common feature of early breast cancer lesions,8,9 indicating the importance of BRCA1 for the control of centrosomes.
Although it was clear from earlier studies that the enzymatic activity of BRCA1 is required for inhibition of centrosome function in vitro,10 it was not clear if this inhibition was a result of the direct ubiquitination of γ-tubulin or a result of ubiquitination of other proteins that may be required for the retention of bound γ-tubulin to the centrosome.
To define further the role and mechanism of BRCA1-dependent inhibition of centro-some MT nucleation, we refined the in vitro assay in three ways: (1) the MT nucleation assay was divided into stages, first ubiquitination of centrosomes, followed by a test for MT nucleation potential. In order to ensure that each step was discrete, the centrosomes were washed between each stage of the reaction. (2) The source of tubulin used for aster formation included either sea urchin tubulin or the crude Xenopus extracts. In this way, we could determine whether some unknown component of the Xenopus extract contributed to the results. (3) We included washes with chaotropic agents, such as 1 M potassium iodide (KI). Based on these experimental observations we propose a model emphasizing the role of BRCA1 in ubiquitinating a putative adapter protein or protein complex in regulating the amount of γ-tubulin associated with centrosomes that directly affects the efficiency of MT nucleation by centrosomes.
Human γ-tubulin cDNA clone was purchased (Invitrogen) and the full-length cDNA was amplified using the primers gtubf (5' CGG GGG CCG GCC AAA TGC CGA GGG AAA TCA TCA C 3') and gtubr (5' CGG GGG CGC GCC CTC ACT GCT CCT GGG TGC 3'). The PCR product was cloned into a retroviral vector, p-zz Babe, which encodes an amino-terminal zz (IgG binding domain of protein A) tag.
HeLa cells were cultured according to ATCC recommendations. pzz-γ-tubulin was transfected into HeLa cells using Lipofectamine 2000 (Invitrogen) according to manufacturer’s instructions. Forty-eight hours post transfection, cells were lysed and zz-γ-tubulin purified down as described below.
Anti γ-tubulin and anti α-tubulin (Sigma) were used at 1:500 dilution for immunofluorescence microscopy. Anti γ-tubulin antibody was used at 1:1000 dilution for immunoblots.
Cover slips were blocked in PBS containing 3% BSA and 0.1% Triton X-100. Primary antibodies were diluted 1:1000 in blocking buffer. FITC- or Texas Red labeled secondary antibodies were used at 1:250 dilution. Intermediate washes were carried out with 1X PBS containing 0.1% Triton X-100. Stained coverslips were mounted onto glass slides with a drop of vectorshield mounting solution (Vector laboratories, Burlingame, California).
Images were viewed using the 60X or 100X objective lens with a Nikon Eclipse TE2000-S microscope and captured using a model 2.3.1 SPOT digital camera. Images were processed using the advanced SPOT program.
Transfected cells were washed once in cold HBS (50 mM Hepes, pH 7.4, 150 mM NaCl) and lysed in 150 mM NaCl, 50 mM Hepes pH 7.4, 2 mM EDTA, 1 mM DTT, 0.5% Triton X-100, Protease Inhibitor Cocktail, 50 mM sucrose and 1 mM GTP. The lysate was bound to IgG sepharose beads at 4°C for 1.5 hours. The beads were washed twice with the lysis buffer, twice with 250 mM salt buffer (250 mM NaCl, 50 mM sucrose, 50 mM Hepes, pH 7.4) and thrice with 1X PEM buffer (5 mM PIPES, 1 mM EGTA, 1 mM MgSO4) containing 10% sucrose and 1 mM GTP.
Reactions contained 1X BRB80 (80 mM K-Pipes, 1 mM MgCl2, 1 mM EGTA), 2 mM ATP, 4 mM MgCl2, 1 mM ubiquitin, 200 nM E1, 5 µM UbcH5c-his11 and 30–40 nM BRCA1-BARD1. Centrosome fraction (~500–1000 ng) or 10–20 µl of zz-γ-tubulin bound to IgG sepharose beads in 1X PEM buffer containing 10% sucrose and 1 mM GTP, were added to the reaction and incubated at 37°C for 30 minutes. In some experiments centrosomes were treated with 1M potassium iodide (KI) for 15 minutes at room temperature. The ubiquitinated and/or KI treated centrosomes were centrifuged through 2 ml 1X PEM buffer on to glass coverslips. The centrosomes attached to the coverslips were allowed to grow asters in the presence of either 0.35 mg/ml of purified 3X cycled sea urchin tubulin in re-assembly buffer (RAB: 100 mM PIPES, pH 6.9, 1 mM EGTA, 5 mM MgSO4 and 2 mM GTP) or 10 µl xenopus extract (20 mg/ml) at 25°C for 12–15 minutes. The asters were fixed with 1% glutaraldehyde in RAB for ten minutes followed by fixation with cold methanol for ten minutes at room temperature. Excess glutaraldehyde was removed by treatment with 10% sodium borohydride in PBS for ten minutes. The asters were visualized by immunostaining with α- and γ-tubulin specific antibodies as described earlier in the immunofluorescence section.
Hi-Five cells were co-infected with the BRCA1 and BARD1 expressing baculoviruses using standard protocols and purified using M2-agarose (Sigma, St. Louis, MO) as described earlier.6
Centrosome fractions were prepared from HeLa S3 cells according to protocol previously described.12,13 3X cycled sea urchin tubulin was prepared as described in ref. 14. Xenopus extract was prepared as described in ref 15 and kindly provided by A. Groen (Harvard Medical School).
Purified asynchronous HeLa centrosomes were incubated with the ubiquitination enzymes E1, E2 and ubiquitin, with or without the E3 ubiquitin ligase BRCA1/BARD1 (the BRCA1/BARD1 heterodimer will be referred to as “BRCA1”). Asters formed in the in vitro assay were visualized by immunostaining with antibodies against α-tubulin (red) and γ-tubulin (green). In the absence of BRCA1 the centrosomes retained microtubule nucleating potential (MNP) and formed asters when treated with tubulin (Fig. 1A). In contrast, when BRCA1 was included in the ubiquitination reaction cocktail followed by the test for microtubule potential described above, centrosomes lost their MNP (Fig. 1B). Few asters were detected as a result of BRCA1 treatment, and those that were detected nucleated only a few MTs. This observation was similar to that found previously when Xenopus extract was used as a tubulin source instead of purified tubulin in similar in vitro assays.4 From these results we conclude that the loss of centrosome MNP is a direct result of BRCA1-dependent ubiquitination of centrosome proteins.
To test whether the inhibitory effect of ubiquitination was a general feature of animal cell centrosomes or specific to human centrosomes, we isolated centrosomes from an evolutionarily distant organism, Spisula solidissima, and tested for ubiquitination effects on centrosome MNP. Interestingly, the MNP of isolated clam centrosomes was also inhibited in an ubiquitin, E1 and E2 dependent manner in this assay (Fig. 1C and D). However, unlike the centrosomes from human cells, addition of BRCA1 was not required for the ubiquitin-dependent loss of MNP. Therefore, we propose that an E3 ubiquitin ligase is a stable and tightly-associated component that copurifies with Spisula centrosomes. This surprising result indicates that ubiquitin ligase activity can reside within the centrosome structure itself. Importantly, these observations indicate that that the regulation of centrosome function by ubiquitination spans evolution from invertebrates to vertebrates.
The key centrosomal protein component involved in MT nucleation is γ-tubulin, one of a number of protein components that make up a γ-tubulin ring complex (γ-TuRC). We tested whether the inhibition of centrosomal MNP was the direct consequence of ubiquitination of γ-tubulin. An epitope-tagged γ-tubulin was expressed in HeLa cells and purified (Fig. 2A). The protein complex was ubiquitinated in vitro (Fig. 2B) and incubated with purified sea urchin tubulin to test for stimulation of MT assembly. The results demonstrated that in the presence of ubiquitinated γ-tubulin MTs assembled as efficiently, if not more efficiently, than when the unmodified form of γ-tubulin was incubated with α/β tubulin (Fig. 2C andD). Based on this observation, we suggest that the inhibitory effect of BRCA1 dependent ubiquitination on centrosomal MT nucleation is not due to the direct inhibition of γ-tubulin interactions with α/β tubulin during MT nucleation.
Previously we showed that γ-tubulin is mono-ubiq-uitinated by BRCA16 and that ubiquitination of centrosomes led to decreased γ-tubulin staining at the centrosomes.4 Additionally, expression of gamma tubulin mutants that have single amino acid substitution of the ubiquitin acceptor lysines to arginine led to both centrosomal amplification as well as hyperactivity. However, we observed in cell free reactions that the quantity of ubiquitinated γ-tubulin (as observed by Western blot analysis) was low in comparison to the severity of the inhibitory effect on centrosomal MNP in the in vitro reactions.4,10 Based on these results we reasoned that although gamma tubulin ubiquitination was important in regulating centrosomal function, BRCA1 targets other centrosomal proteins, that regulate centrosomal MT nucleation activity by controlling the amount of γ-tubulin associated with the centrosome at any given time.
It has been shown that following salt-stripping of centrosomes, in addition to γ-tubulin and γ-TuRCs, other proteins present in cytoplasmic extracts of Xenopus16 and Spisula17 were required for the recovery of centrosome MNP.16 This led to the speculation that additional factors or adapters are needed to mediate the binding of γ-TuRCs to centrosomes (reviewed by ref. 18). Treatment of isolated HeLa centrosomes with 1M KI for 40 minutes at room temperature removed centrosome MNP and rendered them inactive in the aster formation assay (Fig. 3Ac). However, these salt-stripped centrosomes regained MNP when incubated in Xenopus extract (Fig. 3Ad). By comparison, intact centrosomes ubiquitinated by BRCA1 permanently lost MNP and failed to form asters, even when treated with Xenopus extract (Fig. 3Ab). This result suggests that, unlike salt-stripping, ubiquitin mediated inhibition of centrosome function cannot be reversed by treatment with Xenopus extract. BRCA1 ubiquitination of centrosomes in the in vitro assays is associated with a corresponding decrease in γ-tubulin staining at the centrosomes. Hence, inhibition of MT nucleation by the BRCA1 ubiquitin ligase is due to the physical loss of γ-tubulin/γ-TuRC proteins. Additionally, the ubiquitination of centrosomes also left a lasting mark on the centrosomes that blocked the addition of fresh γ-TuRCs for restoration of MNP. Thus, although ubiquitination by BRCA1 leads to removal of γ-TuRCs from centrosomes, this does not explain the inability of Xenopus extract to replenish the MNP of ubiquitinated centrosomes. We reasoned that the covalent modification of γ-TuRC docking proteins could prevent the reloading of fresh γ-TuRCs onto salt-stripped centrosomes.
To test whether the γ-TuRC docking proteins, “adapters”, are targets for ubiquitination, the MNP of ubiquitinated centrosomes were compared to centrosomes that were first ubiquitinated and then subsequently salt stripped with 1 M KI. If indeed the ubiquitination of docking or adapter proteins is responsible for preventing addition of unmodified γ-tubulin/γ-TuRC onto centrosomes, replacing these modified proteins with their unmodified counter-parts from Xenopus extract should allow ubiquitinated centrosomes to once again become competent to bind γ-TuRCs and thus nucleate MTs. Potassium iodide treatment of ubiquitinated centrosomes followed by incubation with Xenopus extract led to a complete reversal of the ubiquitin mediated inhibition of centrosome microtubule nucleation activity (Fig. 3Bd). This observation suggests that BRCA1 targets a key centrosomal protein adapter, or adapter complex, that when ubiquitinated blocks the binding of γ-TuRCs to centrosomes.
Based on the results from this study it is clear that the reduction of MT nucleation capacity of non-mitotic centrosomes via BRCA1-mediated ubiquitination is due to the net effect of the loss of γ-tubulin/γ-TuRCs as well as the inability of the ubiquitinated centrosome to recruit fresh un-ubiquitinated γ-tubulin to centrosomes (Fig. 4). We cannot, however, rule out the possibility that BRCA1 ubiquitinates components of γ-TuRCs resulting in the dissociation of the complex and the removal of γ-tubulin from the centrosomes. The KI treatment would be required to remove the remnants of the complex from the centrosomes before new γ-TuRCs could bind to centrosomes. Thus, either ubiquitination of a γ-TuRC component other than γ-tubulin, or ubiquitination of an adapter component that links γ-TuRCs to the centromatrix, the centrosomal structural material, must be responsible for loss of microtubule nucleation potential. We thus hypothesize that through modification of one or more such adapter/docking proteins, BRCA1 not only strips the centrosome of bound γ-TuRCs but also prevents further recruitment and turnover of centrosome γ-TuRCs in cells and in cell extracts, resulting in a down regulation of centrosome MT nucleation potential.
Several proteins including Ninein, Ninein-like protein, and NEDD1 have been described that possess γ-TuRC anchorage function. 19–22 It is not clear whether any of these proteins or some other unidentified protein contributes to BRCA1 mediated regulation of centrosome microtubule nucleation function. It is, however, clear from this study that the ubiquitination of either a γ-TuRC or adapter component inhibits γ-TuRC binding to the centrosome. This modification may thus be potentially useful as a tool for isolating and identifying components required for adapter function or γ-TuRC-adapter binding.
Through these experiments, and other earlier results, the biochemical mechanism for how BRCA1 controls centrosome function is being revealed. Since centrosomes are often aberrant in early breast cancer lesions,8,9 elucidating the biochemical basis of BRCA1 regulation of centrosome function is required for understanding the pathogenesis of the disease.
This work was supported by a Komen Foundation Fellowship (S.S.) and NCI grant CA111480 (J.D.P.) and NIH-NIGMS grant GM43264 (R.E.P.).