Depletion of Mitochondrial Acyl Carrier Protein in Bloodstream-Form Trypanosoma brucei Causes a Kinetoplast Segregation Defect

Construction of the BSF T. brucei conditional knockout cell line for ACP was reported previously (26). Since deletion of both genomic ACP alleles was lethal, an ectopic copy of the ACP gene controlled by the tetracycline repressor was expressed during construction of the homozygous knockout. These cells were maintained at 37°C in 5% CO₂ and HMI-9 medium (11) containing 10% fetal bovine serum and 10% Serum Plus (JRH Biosciences). The antibiotic G418 (2.5 μg/ml) was used to maintain the T7 polymerase and tet repressor, while phleomycin (2.5 μg/ml) maintained the presence of the ectopic copy of ACP and tetracycline (Tet; 1 μg/ml) or doxycycline (1 μg/ml) allowed its expression.

ACP conditional knockout cells (∼1 × 10⁸ to 5 × 10⁸) grown in the presence or absence of tetracycline for up to 6 days were harvested and extracted according to the method of Bligh and Dyer (3), dried under N₂, and stored at 4°C until analyzed by electrospray mass spectrometry (ES-MS) and ES-tandem MS (ES-MS/MS), as previously described (10).

For staining of DNA in cells, 2 μg/ml of DAPI (4′,6-diamidino-2-phenylindole) was added to fixed cells, and the slides were mounted in Vectashield (Vector Laboratories). Following fixation, fluorescence in situ hybridization (FISH) was performed as described previously (7, 15). All imaging was done with a Zeiss Axioskop microscope equipped with a Retiga Exi charge-coupled-device (CCD) camera (QImaging Corp.).

Total DNA isolation and Southern blotting were described previously (28). For analysis of free minicircle replication intermediates, total DNA was run directly in an agarose gel in the presence of ethidium bromide to resolve gapped and covalently closed species. To analyze total minicircle and maxicircle content, total DNA was digested with HindIII and XbaI prior to electrophoresis. kDNA species were detected on a Southern blot with ³²P-labeled probes for minicircles and maxicircles. The trypanosome hexose transporter (THT) gene was probed as a loading control.

kDNA networks were isolated as described previously (20), by centrifugation of cell lysates through a 20% sucrose cushion. For DAPI staining, 1 μl of isolated network solution was added to 19 μl of 1× phosphate-buffered saline (PBS). This solution was then spotted into one well of a 6-mm, 8-well slide coated with 0.01% poly-l-lysine. Networks were allowed to adhere to slides for 1 h, followed by washing with PBS. Finally, 2 μg/ml DAPI in PBS was added to the slide and incubated for 5 min. Slides were washed briefly with PBS and then mounted in Vectashield (Vector Laboratories) and examined by fluorescence microscopy as described above. IPLab software (Scanalytics, Inc.) was used for imaging and for network surface area measurements (6). For electron microscopy (EM), isolated networks were spread on nitrocellulose-coated grids by the formamide method (21). The grids were then rotary shadowed with Pt-Pd, using a Denton vacuum evaporator (DV502-A). DNA was imaged with a Phillips (CM-120) transmission electron microscope operating at 80 kV, recorded with an Orius SC 1004 2k×2× CCD camera, and digitized with a Gatan v. 3.11.0 digital micrograph. Brightness and contrast of some EM and fluorescence images were adjusted uniformly by using Adobe Photoshop on a Macintosh computer.

ACP knockout cells were grown in the presence of 1 μg/ml doxycycline (with similar effects to those with tetracycline) or in the absence of doxycycline for 3 days. Cells were fixed in 2.5% glutaraldehyde, postfixed in osmium tetroxide, and embedded in Epon (2). Thin sections were examined on an FEI Tecnai 12 transmission electron microscope.

Our previous study on PCF trypanosomes revealed that RNAi knockdown of ACP, and thus the mitochondrial FAS pathway, decreased phosphatidylinositol (PI) and phosphatidylethanolamine (PE) levels. This observation established the importance of mitochondrial FAS in the maintenance of phospholipid content (10). To investigate whether mitochondrial FAS has the same role in BSFs, we analyzed phospholipid content in cells with a conditional knockout of ACP (26). In these cells, we replaced both ACP alleles with drug resistance genes, but since ACP is essential for growth, we also introduced an ectopic copy of the ACP gene whose expression depended on the addition of Tet or doxycycline to the culture medium. As in our experiments with PCFs (10), we then used ES-MS to compare the phospholipid content of whole-cell extracts of BSF cells expressing ACP (measured at day 0, prior to removal of Tet) with that of similar extracts of BSF cells grown in the absence of Tet for 4 or 6 days. To search for major changes in levels of various phospholipids, we first conducted survey scans in negative-ion mode, for detection of PI, PE, phosphatidylserine (PS), phosphatidylglycerol (PG), and phosphatidic acid (PA) (Fig. 1A), and in positive-ion mode, mainly for detection of phosphatidylcholine (PC) and sphingomyelin (SM) (Fig. 1B). The molecular species for the above phospholipid classes have previously been characterized for T. brucei BSFs (22).

From the negative-ion survey scan (Fig. 1A), we observed that the ratios of the major PS (772 m/z; alkyl-C_18:0 and C_18:2) and PI (912 m/z; C_18:0 and C_22:5) species remained relatively constant, while the major PE peak (726 m/z; alkyl-C_18:1 and C_18:2) decreased by day 4 of ACP depletion (Fig. 1A, day 0; these 3 phospholipids are marked with asterisks). This observation was supported by further investigation of the collision-induced daughter ion characteristic of PE phospholipids (parents of 196 m/z) by ES-MS/MS, where a decrease in the major plasmenyl-PE was observed by day 6 (Fig. 2A, 726 m/z). The ion at 659 m/z, previously identified as ethanolamine-phosphorylceramide (27), changed little; its apparent increase at day 6 was due to the significant decrease in PE (726 m/z). This decrease in PE upon ACP loss paralleled relative increases in the phospholipids PA (737 m/z; alkyl-C_18:0 and C_22:4) and PG (805 m/z [C_18:0 and C_20:0], 821 m/z [C_18:0 and C_22:5], and 869 m/z [C_22:4 and C_22:5]) (Fig. 1A, day 6 [marked with asterisks]). Further investigation of the PS species by collision-induced fragmentation (neutral loss of 87 m/z) showed only a slight decrease in diacyl-PS, while PI species (parents of 241 m/z) showed insignificant fluctuations in level after ACP depletion (data not shown).

We detected minor changes in the positive-ion survey scan at day 6 compared with day 0 (Fig. 1B). We further investigated these potential differences by collision-induced parent ion scanning of 184 m/z, which is characteristic of PC and SM phospholipids (Fig. 2B). Changes in the PC and SM phospholipid content after 4 days of ACP depletion became more severe by day 6. These differences included relative increases in plasmenyl-PC species at 743 m/z (C_34:1) and 771 m/z (C_36:1 and alkyl-C_18:0/18:2), with corresponding decreases in diacyl-PC species at 807 m/z (C_18:0/20:4 and C_16:0/22:4) and 835 m/z (C_18:0 and C_22:5) (Fig. 2B, day 6 [marked with asterisks]). Thus, we concluded that ACP depletion alters the phospholipid content of BSF trypanosomes by decreasing PE, increasing PA and PG, and shifting to more plasmenyl-PC at the expense of diacyl-PC.

Again using the ACP conditional knockout, we found that BSF cells had significant changes in kinetoplast size that began prior to growth arrest (Fig. 3A and B). This observation was unexpected, as we had detected no effect on kDNA following RNAi-mediated knockdown of ACP in PCF trypanosomes (10, 26). Using fluorescence analysis of DAPI-stained cells, we evaluated the kinetics of kinetoplast size change during the days following Tet removal (Fig. 3B shows DAPI-stained examples of kinetoplasts of various sizes). After 4 days, we observed that approximately 25% of cells had no kinetoplast, 6% had kinetoplasts that were smaller than normal, and 19% had kDNA that appeared larger than normal (Fig. 3B).

We next used Southern blot analysis of minicircles and maxicircles as an alternative approach for evaluating kDNA levels. Surprisingly, we observed no significant change in level of either minicircles or maxicircles following Tet removal (Fig. 3C). Since these Southern blots reflect the kDNA content of the entire population after ACP removal, these data suggest that the DAPI-stained cells in Fig. 3B that appear to have less kDNA (small kinetoplasts or none at all) and those with more kDNA (large kinetoplasts) cancel each other out, leaving little, if any, net change in the level of minicircles or maxicircles (Fig. 3C).

Since kDNA loss could be due to a replication defect, we then looked for changes in free minicircle replication intermediates. kDNA replication (reviewed in references 17 and 25) involves release of monomeric covalently closed minicircles from the network, most likely by a topoisomerase II, followed by replication of the free minicircles to produce gapped progeny. These reattach to the network in another topoisomerase reaction. These closed and gapped free minicircles are easily separated by agarose gel electrophoresis in the presence of ethidium bromide. Autoradiography of a Southern blot showed that in the days following tetracycline removal, there were no major changes in these free minicircle species, except for a slight increase in oligomeric forms (designated “oligomers” in Fig. 3D). These experiments provided strong evidence that ACP depletion had little effect on minicircle replication.

To examine more closely the structure of kDNA networks following ACP depletion, we isolated networks from cells before (day 0) and after (days 2 and 4) ACP depletion (by Tet removal), stained them with DAPI, examined them by fluorescence microscopy, and measured the surface area of about 130 networks at each time point (Fig. 4A). The size distributions of networks isolated from cells expressing ACP (day 0) and from those after ACP depletion for 2 days centered around unit-size networks (∼5 μm²). However, after 4 days of ACP depletion, the distribution of network surface areas became much more heterogeneous, with more networks being smaller or larger than unit size (Fig. 4A). These measurements were consistent with our in-cell observations and provided support for the hypothesis that most ACP-depleted cells contained networks that were small, normal, or large. The larger networks rarely exceeded a doubled size, which is the size of a network after replication and prior to segregation.

We also considered the possibility that network size changes reflected alterations in network topology. Minicircles in the kDNA network are each concatenated to several neighboring minicircles, and the number of neighbors varies according to the stage of kDNA replication (4). The number of interlocks per minicircle determines the minicircle density and therefore can affect network size. For example, if the minicircle density decreases due to a reduction in the average number of neighbors of each minicircle, the network expands in terms of surface area (4). Similarly, an increase in minicircle density would shrink the network surface area. Finally, if networks develop holes, as occurs when monomeric minicircles are released for replication, they also enlarge significantly (16); these holes are observed following RNAi of the mitochondrial topoisomerase II. To evaluate these possibilities, we employed EM (Fig. 4B) to examine isolated networks following ACP depletion. These experiments revealed neither holes nor obvious changes in minicircle density. However, as shown by the examples in Fig. 4B, we did observe considerable variation in network size at day 4. Compared to the normally sized network at day 0 (upper left), day 4 networks were either normal (lower left), large (upper right), or small (lower right), consistent with the histograms in Fig. 4A. Regardless of their variation in size, most networks at day 4 appeared to have a relatively normal shape.

kDNA replication produces a double-size network that splits into two equal-size progeny, which then segregate into daughter cells during cytokinesis. The fact that some ACP-depleted cells had large kDNA networks while others had small kDNA networks or none at all raised the possibility of a defect in kDNA segregation. A complete block in segregation would produce a cell with a double-size network and a sister cell with none at all, consistent with our observations. In another, less severe type of segregation defect, the double-size product of replication divides asymmetrically, leading to one progeny network that is larger than normal and one that is smaller, again consistent with our observations. To explore the possibility of asymmetric segregation, we DAPI-stained cells following depletion of ACP. Fluorescence microscopy revealed that by day 6, about 18% of all cells had asymmetrically dividing kinetoplasts (Fig. 5A). We also detected cells with small DAPI-staining structures that appeared to be positioned distant from the kinetoplast but presumably within the mitochondrial matrix (Fig. 5B). Such structures, known as ancillary kDNA, have been observed previously in wild-type trypanosomatids (mainly Trypanosoma cruzi) (18) and in RNAi cell lines (e.g., following knockdown of DNA polymerase IC [12]). These are thought to be small network fragments in abnormal positions within the mitochondrial matrix. Examination of cells 4 days after ACP depletion revealed that 13% (15/116 cells) contained ancillary kDNA (Fig. 5B). To determine if these small DAPI-stained structures were actually ancillary kDNA and not, for example, nuclear DNA fragments, we used FISH with minicircle and maxicircle probes. Ancillary kDNA was detectable with the minicircle probe, proving that these structures were in fact small fragments of kDNA (Fig. 5B). We could not detect maxicircles in ancillary kDNA, either because they were not present or because their level was too low to detect. Also, if they were covalently closed, they would be nondenaturable and therefore unable to hybridize with the probe. This result suggests that ACP loss may promote network fragmentation and that the fragments may disconnect from the TAC.

Electron microscopy of PCFs following ACP depletion revealed mitochondrial enlargement and abnormal ultrastructural features, such as membranous bodies thought to be disrupted cristae (10). We performed thin-section EM on ACP-depleted BSF trypanosomes 3 days after Tet removal to assess whether structural abnormalities of the mitochondrion could help to explain the segregation defect in this life cycle stage. Control cells prior to tetracycline removal displayed unit-size kDNA networks surrounded by a tightly fitting mitochondrial membrane (Fig. 5C, day 0 [K]). In day 3 cells, basal bodies appeared properly aligned with the kDNA disk (Fig. 5C, black arrowheads), and flagellar pockets were properly formed (white arrowheads). We observed large and possibly abnormally dividing kinetoplasts (K) after 3 days of ACP depletion. Overall, the mitochondrial structure appeared normal in size and density, but membranous bodies (white arrows) were present within some ACP-depleted BSF mitochondria.