Golgi Duplication in Trypanosoma brucei Requires Centrin2

See allHide authors and affiliations

Science  18 Nov 2005:
Vol. 310, Issue 5751, pp. 1196-1198
DOI: 10.1126/science.1119969


Centrins are highly conserved components of the centrosome, which in the parasitic protozoan T. brucei comprises the basal body and nucleates the flagellum used for locomotion. Here, we found TbCentrin2 in an additional bi-lobed structure near to the Golgi apparatus. One lobe was associated with the old Golgi, and the other became associated with the newly forming Golgi as the cell grew. Depletion of TbCentrin1 inhibited duplication of the basal body, whereas depletion of TbCentrin2 also inhibited duplication of the Golgi. Thus, a Centrin2-containing structure distinct from the basal body appears to mark the site for new Golgi assembly.

Organelle duplication helps to ensure propagation through successive generations. For the Golgi apparatus, a number of models have been put forward, which differ as to the role played by the old Golgi in the construction of the new (1). What has not been addressed is the mechanism that determines the location for assembly of the new Golgi. In the budding yeast, Pichia pastoris, this location appears to be random, based on the probability of components reaching a critical mass for assembly (2). In several parasitic protozoa and in other protists, however, this location seems to be defined (1, 3). In T. brucei, for example, there is a single Golgi, and a new copy is assembled at a fixed distance away (4). This new assembly site is somehow related to the new basal body (5); inhibition of basal-body segregation inhibits that of the Golgi (4) and inhibits division of the replicated kinetoplast, which contains all of the mitochondrial DNA (6). Basal bodies in particular and centrosomes in general have been implicated in the biogenesis of a number of membrane-bound organelles, in a variety of organisms (6, 7), prompting us to study further their role in Golgi duplication.

Centrins are Ca2+-binding proteins that are highly conserved and essential components of all centrosomes (8). The monoclonal antibody, 20H5, raised against Chlamydomonas reinhardtii Centrin (9), labels centrosomes in a wide range of organisms. It can also stain the basal bodies in T. brucei at different stages of the cell cycle (Fig. 1). The basal bodies are closely associated with the kinetoplast and mediate the division of the replicated kinetoplast (6).

Fig. 1.

A Centrin-containing structure associated with the Golgi. Gallery of images through the cell cycle of cells triple labeled for Golgi [anti–Golgi Reassembly Stacking Protein (GRASP), red], Centrin (20H5, green), and DNA [4′,6′-diamidino-2-phenylindole (DAPI), blue]. Basal bodies [open arrowheads (A and B)] underwent duplication (C and D) and mediated the division of the replicated kinetoplast (E). Centrin associated with the Golgi was present in a bi-lobed structure (solid arrowheads), with one lobe near to the old Golgi (A) and the other more toward the posterior of the parasite, marking the site where the new Golgi was undergoing assembly [compare (A) and (B)]. Increasing separation of the old and new Golgi (C) was accompanied by duplication of this bi-lobed structure [(C) to (E)] at about the same time as the replicated kinetoplast divided (E). (F) Just before cytokinesis, additional Golgi (arrows) appeared that were not associated with Centrin. Scale bar, 5 μm.

The 20H5 antibody stained an additional, bi-lobed structure (Fig. 1). Early in the cell cycle, the old Golgi was adjacent to one lobe (Fig. 1A), whereas the new Golgi was later seen to be adjacent to the other, more posterior lobe (Fig. 1B), suggesting that it might be marking the site for new assembly. As the new Golgi grew and increasingly separated from the old (Fig. 1C), the bi-lobed structure itself duplicated, so that one remained with the old Golgi and one with the new (Fig. 1, D and E). This occurred at about the same time as the division of the replicated kinetoplast (Fig. 1E). At later times in the cell cycle, when additional Golgi appeared (Fig. 1F, arrows) (4), they were not associated with this new structure. Though the function of these additional Golgi is unclear, the data do suggest that they are not part of the core duplication process.

There are five putative Centrins in T. brucei (10). TbCentrins1 to 4 are the most homologous, by both sequence and phylogenetic analysis, to the C. reinhardtii Centrin used to raise the 20H5 monoclonal antibody. Each of these was tagged with the yellow fluorescent protein (YFP) and stably expressed, followed by immunoblot analysis. Only TbCentrin1 and TbCentrin2 were recognized by the 20H5 antibody (fig. S1). Whereas TbCentrin1 was localized exclusively to the basal body (Fig. 2A), TbCentrin2 was also associated with the Golgi (Fig. 2, B and C). This association did not depend on the position of the YFP, because tagging the N terminus gave the same pattern at all stages of the cell cycle and was identical to that seen with 20H5 (compare Fig. 1, A to F, with Fig. 2, F to J). Endogenous TbCentrin2 could not be examined by immunofluorescence microscopy because the polyclonal antibodies proved to be too insensitive. Both alleles were thus replaced by TbCentrin2 tagged at the C terminus with the short, viral, BB2 tag (11). The cells had the same growth characteristics as the parental line, indicating that this construct was fully functional. The location of TbCentrin2-BB2 was also identical to that identified by 20H5 and both of the YFP constructs (compare Fig. 2, D and E, with Fig. 1, A to F, and Fig. 2, F to J). Thus, a pool of TbCentrin2 defines a bi-lobed, Golgi-associated structure apparently distinct from the basal bodies.

Fig. 2.

TbCentrin2 associated with the Golgi. (A to J) Cells stably expressing the indicated TbCentrin constructs (green) were labeled for tyrosinated α-tubulin (YL1/2) (red) in (A) and (B) or GRASP (red) in (C) to (J), and DNA (DAPI, blue). TbCentrin2, tagged with BB2 [(D) and (E)], replaced both of the endogenous alleles. TbCentrin1 localized to the basal body alone [open arrowheads in (A)], whereas TbCentrin2 also localized to the Golgi [solid arrowheads in (B) to (E)] throughout the cell cycle [(F) to (J)]. (K to O) Cells were labeled for Centrin (green) with the use of 20H5 [(K), (N), and (O)] or by stable expression of TbCentrin2-YFP [(L) and (M)]. Centrin associated with the Golgi (solid arrowheads) did not colocalize with the basal-body markers (red), gamma-tubulin [tagged with YFP in (K) or BB2 in (L) and (M)] or tyrosinated α-tubulin [YL1/2 in (N) and (O)]. Open arrowheads mark the basal bodies, whereas arrows mark the mitotic spindle poles. Scale bar, 5 μm.

To distinguish this structure from basal bodies in particular and centrosomes in general, we tested other markers. Gamma-tubulin is found in all centrosomes (12) and is associated with the Golgi in mammalian cells (13). None could be detected, however, in the bi-lobed structure when YFP-tagged gamma-tubulin was stably expressed. It was only found in the basal bodies (Fig. 2K). Gamma-tubulin is known to be present in other structures (14) and the YFP tag might thus have restricted its distribution. Gamma-tubulin was thus tagged with BB2 and stably expressed. Additional staining of the mitotic spindle poles during mitosis was then observed (Fig. 2, L and M), as expected (14), but there was no staining of the bi-lobed structure. There was also no staining using the YL1/2 antibody (Fig. 2, N and O), which recognizes tyrosinated α-tubulin. This is found in the basal bodies throughout the cell cycle and, during certain stages of the cell cycle, in the flagellar axoneme and the posterior third of the subpellicular microtubules that contain newly synthesized microtubules (15).

To test whether TbCentrin2 plays a functional role in Golgi duplication, the inducible and inheritable RNA interference (RNAi) system in T. brucei (16) was exploited so as to deplete either TbCentrin1 or TbCentrin2. In the presence of tetracycline, to induce expression of the double-stranded RNA, the cells stopped growing in the absence of either Centrin (Fig. 3, A and B), consistent with the effects of Centrin depletion in other organisms (1719). Polyclonal antibodies directed against either TbCentrin1 or TbCentrin2 (fig. S1) were used to monitor the levels. Depletion of TbCentrin1 had no effect on the levels of TbCentrin2 and vice versa (Fig. 3, C and D).

Fig. 3.

RNAi knockdown of TbCentrin1- or TbCentrin2-inhibited cell growth. Cells were grown in the absence or presence of tetracycline to induce RNAi, and samples were taken for (A and B) counting (results presented as mean ± SD, n = 3) and (C and D) immmunoblotting (IB) to assess levels of TbCentrin1 and TbCentrin2 in fractionated extracts. Cessation of growth occurred earlier in cells depleted of TbCentrin1, consistent with the more rapid loss of protein. Loss was specific; i.e., depletion of TbCentrin1 had no effect on levels of TbCentrin2, and vice versa.

In both cases, depletion led to an inhibition of basal-body duplication, and hence kinetoplast division, but had no effect on nuclear division. At early times, this resulted in cells containing one kinetoplast and two nuclei. At later times, these became multinucleate because cytokinesis was also inhibited. To facilitate analysis, cells containing one kinetoplast and two nuclei were counted. Cells depleted of TbCentrin1 stopped growing earlier than those depleted of TbCentrin2, the likely consequence of TbCentrin1 being depleted more rapidly than TbCentrin2 (compare Fig. 3, C and D). We thus chose to look at TbCentrin1-depleted cells at 24 hours and TbCentrin2-depleted cells at 48 hours postinduction. At these time points, the cells containing one kinetoplast and two nuclei constituted ∼10% of the total population.

Depletion of either TbCentrin1 or TbCentrin2 had a marked effect on the duplication of the basal bodies (Fig. 4, A and B). Most cells had one labeled structure and some had two, although they were not separated to the extent observed in uninduced controls at the same cell cycle stage (compare Fig. 4, A and B, with Fig. 4C). The precise nature of the inhibition remains unclear, although it clearly led to an inhibition of kinetoplast division.

Fig. 4.

RNAi knockdown of TbCentrin2 but not TbCentrin1 inhibited Golgi duplication. Cells were fixed at (A) 24 hours (TbCentrin1) or (B) 48 hours (TbCentrin2) after induction of RNAi, and triple labeled for GRASP (red), tyrosinated α-tubulin (YL1/2) (green) and DNA (DAPI, blue). (C) Uninduced cells were used as controls. (D) Cells late in the cell cycle, containing two nuclei and one kinetoplast, were imaged and the number of Golgi counted (results presented as mean percentage SD, n = 3). The depletion + of TbCentrin1 inhibited complete duplication of the basal body but not the Golgi, whereas depletion of TbCentrin2 inhibited duplication of both. Scale bar, 5 μm.

Duplication of the Golgi was dependent on the presence of TbCentrin2 but not TbCentrin1. Depletion of TbCentrin2 led to the presence of only one Golgi in ∼72% of the cells, whereas depletion of TbCentrin1 led to the opposite result: About 77% of the cells had two Golgi (Fig. 4D). Thus, TbCentrin2 is essential for the Golgi duplication process.

The simplest interpretation is that the bi-lobed structure containing TbCentrin2 helps determine the location at which the new Golgi assembles. One lobe is associated with the old Golgi, the other with the new. When TbCentrin2 is absent, only one Golgi is seen. What is not yet clear is whether the Golgi has doubled in mass but not separated, or whether the actual doubling process has also been inhibited. TbCentrin2 may be involved in either or both of these processes.

The location of the endoplasmic reticulum (ER) exit sites was also determined in relation to this bi-lobed structure. These sites are adjacent to the cis face of the Golgi stack in T. brucei and a new ER exit site grows at the same time as the new Golgi (4). The coat protein II (COPII) coat component, Tb-Sec13p, a component of ER-to-Golgi transport vesicles, was closer to the cis Golgi than the TbCentrin2-containing structure at all stages of the cell cycle (fig. S2A). Because the COPII coat is present on vesicles near to the Golgi as well as on those budding from the ER exit sites, this would place at least part of the bi-lobed structure over the ER exit sites. This was confirmed at the electron microscopic level by immunolabeling thin sections from cells expressing YFP-tagged TbCentrin2. Labeling was found over basal bodies and ER exit sites but not over other parts of the ER (fig. S2B). The density of labeling over the ER exit sites was ∼3 times as high as that over the Golgi stack and ∼6.6 times as high as that over mitochondria (fig. S2C). Comparable experiments with cells expressing TbCentrin1 showed levels of labeling over ER exit sites only slightly higher than those over the Golgi stack and mitochondria.

Thus, Centrin2 in T. brucei is additionally present in a bi-lobed structure and is needed for Golgi duplication. The morphology and location of this structure suggest that it helps define the site at which a new ER exit site and Golgi are constructed. Such an arrangement might help to ensure accurate duplication and ultimate partitioning to each daughter cell. Whether this also pertains in higher eukaryotes is the focus of present research.

Supporting Online Material

Materials and Methods

Figs. S1 and S2


References and Notes

View Abstract

Stay Connected to Science

Navigate This Article