| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
Previous Article | Next Article 
Eukaryotic Cell, March 2006, p. 555-567, Vol. 5, No. 3
1535-9778/06/$08.00+0 doi:10.1128/EC.5.3.555-567.2006
Copyright © 2006, American Society for Microbiology. All Rights Reserved.
Department of Biological Sciences, University of Iowa, Iowa City, Iowa,1 Department of Biology, University of Rochester, Rochester, New York2
Received 20 September 2005/ Accepted 2 December 2005
 |
ABSTRACT |
|---|
 Top Abstract IntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
mutant cells, ultrastructural abnormalities of outer doublet microtubules were present in some of the axonemes. Nonetheless, these cells were still able to assemble cilia after deciliation. The nearly paralyzed ACT1 cells completed cleavage furrowing normally, but the presumptive daughter cells often failed to separate from one another and later became reintegrated. Clonal analysis revealed that the cell cycle length of the ACT1 cells was approximately double that of wild-type controls. Clones could nonetheless be maintained for up to 15 successive fissions, suggesting that the ACT1 gene is not essential for cell viability or growth. Examination of the cell cortex with monoclonal antibodies revealed that whereas elongation of ciliary rows and formation of oral structures were normal, the ciliary rows of reintegrated daughter cells became laterally displaced and sometimes rejoined indiscriminately across the former division furrow. We conclude that Act1p is required in Tetrahymena thermophila primarily for normal ciliary motility and for phagocytosis and secondarily for the final separation of daughter cells. |
INTRODUCTION |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
Perhaps the best-established function of ciliate actin is associated with phagocytosis, as cytochalasins and other drugs that affect actin polymerization (9) inhibit formation of food vacuoles in Tetrahymena pyriformis (49, 50), T. vorax (23), T. thermophila (72), and Paramecium tetraurelia (8, 14). Cytochalasin B also binds to isolated oral apparatuses of T. thermophila (21), and inhibits the formation of this organelle, but does so at a much higher concentration than is sufficient to inhibit feeding (20). A recent study also has shown that latrunculin, an inhibitor of actin polymerization, interferes with the posteriorly directed movement of phagosomes within this cell (31).
From past work it is unclear whether perturbing actin function interferes with cell division. Cytochalasin B did not prevent cell division of synchronized Tetrahymena pyriformis cells at concentrations that inhibit food vacuole formation (50) and also failed to prevent synchronized division of T. thermophila (strain WH-6) when added at a high concentration to cells after (predivision) stage 4 of oral development (20). However, two novel methods of interference with actin function did succeed in bringing about division arrest. First, microinjection of skeletal muscle actin into T. thermophila cells just prior to the onset of constriction prevented cytokinesis but had no effect on macronuclear division (53). Second, green fluorescent protein (GFP)-tagged actin expressed at a high level from an rRNA gene-based replicative vector blocked both macronuclear elongation and cytokinesis in some T. thermophila cells (32).
None of the reports on effects of cytochalasins on ciliates mention changes in cell motility, suggesting that it was not impaired. Likewise, a study of the effects of cytochalasin D on flagellar microtubule dynamics in the alga Chlamydomonas reinhardtii stated that "the cells were able to swim normally" in this drug, and also were capable of nearly normal flagellar regeneration after amputation (11). However, a null mutation of the IDA5 gene, later shown to encode the sole conventional actin of this species (39), reduced swimming speed to less than half that of wild-type Chlamydomonas cells (38) without affecting cell growth or division.
The actin gene, ACT1, was cloned from Tetrahymena thermophila by Cupples and Pearlman (10), and a close ortholog was cloned from T. pyriformis (26). The predicted amino acid sequence of the protein encoded by the T. thermophila ACT1 gene was found to be highly divergent relative to actins in other organisms, showing, for example, only 77% identity with the most similar human actins, compared to 87 to 89% identity between Saccharomyces cerevisiae actin (GenBank accession number CAA24598 [GenBank] ) and human actins.
A search of the latest assembly of the T. thermophila macronuclear genome from The Institute for Genomic Research (http://www.tigr.org/tdb/e2k1/ttg/) and the Tetrahymena Genome Database (http://www.ciliate.org) revealed that ACT1 is one of four genes for actins in this organism. The predicted amino acid sequences of the proteins encoded by the other three actin genes were found to be even more divergent than those encoded by ACT1, and these proteins also are less highly expressed (C.-C. Tsao and M. A. Gorovsky, unpublished observations). Therefore, despite its divergent amino acid sequence and its atypical properties (27), we consider the protein encoded by the ACT1 gene to be the "classical" Tetrahymena actin, and have named it Act1p.
As a vital step in elucidating the role of actins in T. thermophila, we disrupted the ACT1 gene of T. thermophila with the neo3 inactivation cassette. In spite of the multiplicity of actin genes, we obtained a clear mutant phenotype from this disruption. The most obvious direct effect was the presence of almost completely nonmotile cilia. Food vacuole formation was also inhibited. However, macronuclear division and cytoplasmic cleavage furrowing were unimpaired; also, the ACT1 gene apparently was not essential for viability. As reported earlier (3, 5), arrest of ciliary motility brought about failure of the final step of separation of presumptive daughter cells. This failure impeded clonal growth and also generated a unique cellular reintegration phenotype.
|
MATERIALS AND METHODS |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
Cloning of ACT1. The complete coding sequence of T. thermophila ACT1, along with 250 bp of 5' untranslated region (UTR) (10), was amplified by PCR using both a cDNA library (courtesy of Aaron Turkewitz, University of Chicago) and genomic DNA as templates. No introns were found. The forward primer was 5'-AGATCTCATCAAACAATTA-3' and the reverse primer was 5'-TCAGAAGCACTTTCTGTGGA-3'. The 1,382-bp PCR product was cloned using the Invitrogen Topo TA cloning kit; the final construct is pAct-35-TOPO.
Somatic disruption of ACT1. The targeting construct (Fig. 1A) was created by insertion of the neo3 disruption cassette (58), which confers resistance to paromomycin, into pAct-35-TOPO at a unique SfuI site found at position 625 of the ACT1 gene. Plasmid pMNBL containing the neo3 disruption cassette was a gift from Martin Gorovsky, University of Rochester. The neo3 cassette consists of the cadmium-inducible MTT1 promoter, the neomycin coding region, and the BTU2 3'-flanking region.
|
Four somatically transformed clones were analyzed 200 generations after transformation by whole-cell PCR to detect the presence of ACT1 genes with and without the neo3 cassette (43). For each sample, cells in 150 µl of medium were concentrated by centrifugation, incubated for 1 h in 100 µl buffer K (0.03 mM proteinase K in 1x PCR buffer from Promega) at 55°C, and then boiled for 20 min. Two aliquots from each lysate were subjected to PCR as described above. Using the two ACT1 primers described above, a PCR product of 1.38 kb is obtained if the wild-type ACT1 gene is present. PCR using the ACT1 forward primer plus a reverse primer that anneals to the neo3 internal sequence, 5'-AGAGTCCTTGGTCTTAACACT-3', gives a 1.04-kb product when the transgene is present.
Germ line disruption of ACT1.
The pAct-35-TOPO-neo3 construct (Fig. 1A) digested with BstXI was introduced into conjugating CU428 and B2086 cells by biolistic particle bombardment 3 h after mixing (7). The germ line transformants were selected, assorted, and made homozygous for the knockout allele (Fig. 1A) in the micronucleus as described (25). ACT1F10ns and ACT1C3ns, two heterokaryon lines with different mating types, were obtained. These cells each have the ACT1 gene with the drug resistance disruption cassette (the knockout allele in Fig. 1A) in their micronuclear genome and have phenotypically assorted their macronuclear genome to wild-type ACT1 alleles.
ACT1F10ns and ACT1C3ns heterokaryon cells were mated and individual conjugating pairs were isolated between 6 and 10 h postmixing and transferred into SPP medium drops to finish conjugation. To determine if progeny exhibited the same phenotypes as the somatic knockout cells, these progeny were examined with an Olympus SZH-ILLD dissecting microscope every 6 to 8 h. After 2 days, progeny cells were transferred to microtiter plates and further selected with 120 µg/ml paromomycin and MEPP medium with 1 µg/ml CdCl2 to eliminate the paromomycin-sensitive cells derived from aborted conjugants. The true progeny cells (homozygous ACT1 homokaryons) were then maintained in rich medium (PPYGFe or MEPP) in Erlenmeyer flasks with vigorous shaking.
For analysis of the genotype of the homozygous ACT1 homokaryons, genomic DNA extracted from CU428 and ACT1 cells was analyzed by PCR (43) using primers (B1-ACT1-F10, 5'-ATAGGATCCATGGCTGAAAGTGAATCCCCCGCT-3', and A1-ACT1-R10, 5'-TAAGGCGCGCCGAAGCACTTT CTGTGGACGATG-3') designed to amplify the whole ACT1 open reading frame (ORF) (Fig. 1A). PCR was performed using TripleMaster mix (Eppendorf).
For reverse transcription (RT)-PCR analysis to assess ACT1 expression, total RNA from CU428 and ACT1 cells in vegetative growth was extracted using Trizol reagent (Invitrogen) following the manufacturer's instructions; 2 µg of total RNA was reverse-transcribed using SuperScript III (Invitrogen) with oligo(dT) primers at 50°C for 60 min following the manufacturer's instructions.
Multiplex PCR (94°C, 30 seconds; 50°C, 30 seconds; 72°C, 1 min for 23 cycles) was performed using the B1-ACT1-F10 and A1-ACT1-R10 primers to detect the ACT1 message. A pair of specific primers (L21Fw: 5'-AAGTTGGTTATCAACTGTTGCGTT-3', and L21Rv: 5'-GGGTCTTTCAAGGACGACGTA-3'), which flank the first two introns of the ribosomal protein RPL21 gene and amplify a 0.36-kb region from the cDNA or a 0.84-kb region from the genomic DNA template, respectively, were used in the same reaction as an internal control.
Rescue experiment. The whole ACT1 ORF was obtained by PCR using B1-ACT1-F10 and A1-ACT1-R10 primers and CU428 genomic DNA template. The PCR product was directionally ligated between a 1.6-kb 5'-flanking sequence and a 0.5-kb 3'-flanking sequence of the MTT1 gene (58) and cloned into pBlueScript vector (Stratagene).
For the MTT-ACT1-3xHA construct, a triple hemagglutinin (3xHA) oligonucleotide was inserted in frame downstream of the initiator ATG of MTT1 and upstream of the 5' end of the ACT1 ORF to add a 3xHA tag at the amino terminus of Act1p. For the MTT-ACT1-FLAG construct, an oligonucleotide carrying a FLAG epitope was incorporated in frame at the 3' end of ACT1 ORF to add a single FLAG tag at the carboxyl terminus of Act1p. To retransform the ACT1 cells, the MTT-ACT1-3xHA and MTT-ACT1-FLAG rescue constructs were digested with SacII and XhoI and introduced into vegetative cells by biolistic bombardment. For a negative control, cells were bombarded with the empty vector. Cells were incubated in SPP medium with 1 µg/ml of CdCl2 for 4 h at 30°C and then aliquoted to microtiter plates. Rescued swimming cells were observed after about 2 to 3 days.
Phenotypic analyses of living cells.
Cells were examined using an Olympus CK2 inverted microscope, and photographed using a Leica inverted model DMIRBE bright-field microscope equipped with a Hamamatsu Orca digital camera, model C4742-5. Living homozygous ACT1 homokaryons were observed in drops on a petri dish kept in a moist chamber. More extensive microscopic observations on individual cells and their vegetative progeny were made at intervals after isolation of single cells into 2- to 5-µl drops of culture medium plated on plastic petri plates (Fisher) and then covered with a thin layer of light mineral oil (Fisher, 021-1) (17, 44). These microdrop cultures were maintained at room temperature (22 to 24°C) and periodically examined using a Zeiss Stereomicroscope-II and an Olympus CK2 inverted microscope.
Preliminary low-power observations of food vacuole uptake by live cells were carried out using carmine (Alum Lake, Fisher) suspended in culture medium at a concentration of 0.1 µg/ml. In a subsequent more detailed analysis, live cells were washed in 10 mM Tris (pH 7.5) and incubated in Higgin's ink (concentration of about 0.5 µl ink per 1 ml of cells) at 30°C for 1 h. Cells were pelleted at 365 x g for 2 min and then fixed for 30 min with 2% paraformaldehyde in PHEM buffer (60). After fixation, cells were washed and resuspended in phosphate-buffered saline (60), and wet-mount samples were observed directly by interference microscopy.
Cilium regeneration and measurement of cilium length. Cells were deciliated and then allowed to regenerate cilia for 2 h, as previously described (6, 55). Measurements were made either on cilia still attached to cells or on isolated axonemes taken from the supernatant from cells that had regenerated cilia; both were fixed with 2% paraformaldehyde in PHEM buffer with 0.05% Triton X-100. Cells or axonemes were stained with rabbit polyclonal antibody raised against Tetrahymena tubulin (1:500) (66) and images were taken with the Leica TCS SP confocal microscope. The length of axonemes was measured by ImageJ software obtained from http://rsb.info.nih.gov/ij.
Immunofluorescent studies. The axonemes from MTT-ACT1-3xHA rescued cells were prepared as described (24). Cells were fixed in 2% paraformaldehyde in PHEM or in Triton-ethanol (68) followed by staining with anticentrin monoclonal antibody 20H5 (1:500), kindly provided by Jeffrey Salisbury of the Mayo Clinic. In some preparations, cells were doubly stained with anticentrin (1:500) and with the monoclonal antibody 9A9 (1:5) that stains contractile vacuole pores (33). Cilia and cortical microtubule bands were visualized with polyclonal antibodies raised against T. pyriformis tubulins (66, 67) and secondary goat anti-rabbit immunoglobulin G fluorescein isothiocyanate conjugate (1:500, Sigma-Aldrich). 3xHA-tagged Act1p was stained with mouse anti-HA antibody (16B12, 1:500, Sigma-Aldrich) and secondary Alexa Fluor 568 goat anti-mouse immunoglobulin G (1:500, Invitrogen). Nuclei were stained with 4',6'-diamidino-2-phenylindole (DAPI) as described (60), and stained cells were examined in the DMIRBE inverted microscope equipped with a Chroma fluorescent cube. Fixed and stained cells were observed with a Leica TCS SP confocal microscope (PL Apo 100x lens, N.A. 1.4) (Fig. 3A and B and 4) or with an Olympus BH-2 fluorescent microscope (Fig. 3C and 9).
|
|
RESULTS |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
cells were obtained by transformation of macronuclei with the ACT1 disruption plasmid construct (pAct-35-TOPO-neo3) followed by selection in paromomycin with cadmium chloride present. Genes in the ciliate macronucleus are borne by DNA molecules that lack centromeres and segregate randomly during macronuclear division, bringing about the phenomenon known as phenotypic assortment (47, 70). In time, genes being selected against will no longer be represented among the progeny. Increasing concentrations of paromomycin impose a strong selection favoring macronuclear chromosomes containing the neo3 disruption cassette such that all cells should eventually have macronuclei that are fully assorted for the knockout allele, unless there is countervailing selection for maintenance of copies of the wild-type allele.
The persistence of phenotypic heterogeneity as revealed by clonal analysis (see below) as well as the strong signals obtained in PCR-based assays with both nondisruption and somatic disruption primer sets (see Materials and Methods) together strongly suggest that the somatic ACT1 transformants maintained some wild-type alleles in the macronucleus. We conclude that the ACT1 knockout allele probably could not be driven to fixation in growing populations following somatic disruption of the wild-type ACT1 allele, although such fixation may have occurred in some individual cells within these populations (see below).
Using the pAct-35-TOPO-neo3 disruption construct, homozygous ACT1 homokaryon cells were created by germ line transformation (25). To examine the genotype of these homozygous ACT1 homokaryon cells, genomic DNA was extracted and subjected to PCR analysis. As shown in Fig. 1B, the ACT1 primers used specifically amplified the 1.2-kb ACT1 coding region using DNA from wild-type cells. When the same primers were used with ACT1 DNA, however, a 4.2-kb band corresponding to the disrupted ACT1 allele was obtained and no 1.2-kb band was visible, confirming that all of the ACT1 genes had been disrupted in both the macronucleus and micronucleus.
In wild-type cells, a 1.2-kb band (Fig. 1C, arrowhead) was amplified by RT-PCR using ACT1-specific primers. However, RT-PCR failed to amplify an ACT1 transcript from ACT1 cells under the same conditions, while the control primers specific to ribosomal protein RPL21 amplified the 0.36-kb cDNA band (Fig. 1C, asterisk) in the same multiplex PCR. This result showed that ACT1 message was not detected in the ACT1 cells.
Phenotypes of living cells with disrupted ACT1 genes. The most obvious effect of the somatic ACT1 disruption was the appearance of many cells that seemed to be blocked in fission after at least 50 generations in progressively increasing concentrations of paromomycin (Fig. 2A). Examination by light microscopy of living cells (e.g., Fig. 2F) and by DAPI staining indicated that the macronuclei of these fission-blocked cells were completely divided (see also Fig. 9B).
|
|
homokaryons exhibited the same phenotype, but much more uniformly. Individual conjugating pairs of ACT1 heterokaryons were isolated into drops of SPP medium on a petri dish and examined with a dissecting microscope. Initially the exconjugants could swim and divide after conjugation was complete. After three to four divisions, however, cells started to lose their motility and cells that had failed to complete cytokinesis could easily be identified. After five or six divisions, cells became severely paralyzed and lost their normal motility; the motility that remained consisted of a slow and spasmodic oscillatory motion (see films A and B in the supplemental material to compare locomotion in wild-type cells and in ACT1 cells). Tandem chains of two to four cells that had failed to separate (similar to those shown in Fig. 2) also became more common. These observations indicated that after five to six rounds of divisions from a single pair, the remaining wild-type Act1p from the parental pool was no longer sufficient to support motility in the progeny cells. Such nearly paralyzed cells were used for a more detailed clonal analysis (see below).
Close observation of dividing ACT1 cells indicated that all such cells developed an advanced cleavage furrow. Cell separation was delayed or arrested at a terminal stage of cytokinesis, when the connection between the two daughters was extremely narrow. There was no evidence for blockage at earlier stages of cytokinesis, strongly suggesting that the various apparent fission-arrested phenotypes seen in Fig. 2A were consequences of progressive reintegration of cells that had succeeded in furrowing but not in separating.
Figures 2B to D, albeit taken of different cells, illustrate the most likely course of division in the affected cells (see also film B in the supplemental material). Since all living cells observed in division completed constriction but failed to separate prior to arrest, the configurations seen in Fig. 2E and F as well as in Fig. 2A can most easily be interpreted as stages of regression of previously completed furrows. Cells may then persist with regressed furrows, form new furrows while retaining a linear order (Fig. 2G), undergo an anterior telescoping of the posterior presumptive daughter cell (cell 5 in Fig. 2A), or carry out other spatial readjustments (Fig. 2 H and I) ultimately leading to the formation of irregular monsters (Fig. 7D and 9I).
|
cells. The ACT1 cells were transformed with the MTT-ACT1-3xHA construct or empty vector and maintained in microtiter plates in SPP with 1 µg/ml of CdCl2 to induce expression of the tagged Act1p. After 3 days, swimming rescued cells could be found in more than two-thirds of total microtiter wells, while in the negative control plates transformed with vector DNA no cell gained normal motility. The rescued cells looked indistinguishable from wild-type cells. Thus, restoration of normal motility in the mutant cells demonstrated that knockout of ACT1 indeed caused the ciliary defect. A similar rescue was achieved with the MTT-ACT1-FLAG construct.
Analysis of the ciliary defect.
Examination of nearly paralyzed cells both among the somatic ACT1 transformants and in the homozygous ACT1 homokaryons indicated that full-length cilia were present (Fig. 3 and 6), suggesting that in this case, unlike the previously described T. thermophila mutants with paralysis and subsequent failure of cell separation (3, 5), there was no loss of cilia. Cilia could also regenerate nearly normally following deciliation. First, both wild-type cells and homozygous ACT1 homokaryons regenerated cilia with similar kinetics in the recovery buffer (data not shown). Second, regenerated cilia measured on intact ACT1 cells averaged 4.8 ± 0.3 µm (n = 36) in length compared to 5.4 ± 0.5 µm (n = 37) in parallel CU428 cells; when isolated, regenerated cilia were measured, the corresponding lengths were 4.8 ± 0.6 µm (n = 83) for the ACT1 cells, 4.9 ± 0.8 µm (n = 84) for CU428 cells, and 5.6 ± 0.8 µm (n = 77) for the MTT-ACT1-3xHA rescued cells.
|
cells (significantly so for cilia measured on intact cells), it was nonetheless clear that the disruption of the ACT1 gene imposed no major impairment on the formation of new cilia. Close observations of ACT1 and wild-type cells under phase- contrast microscopy revealed that the cilia in mutant cells beat at a much lower frequency than those in wild-type cells, and the beating was not metachronal (see film B in the supplemental material). Unlike the constant and fast ciliary beating in wild-type cells, the mutant cilia made one or several strokes, paused, and then made another one or several strokes. All of these observations argue that ACT1 plays a role in ciliary motility but is not required for the assembly of an axoneme.
Our evidence for a functional deficit in cilia of ACT1 cells made it desirable to confirm an earlier report (46) that Act1p is present in ciliary axonemes of T. thermophila. This was demonstrated in homozygous ACT1 homokaryons rescued by the MTT-ACT1-3xHA construct (Fig. 4). Comparison of the slightly offset antitubulin and anti-HA images (Fig. 4B) indicated that Act1p and tubulin were largely colocalized in the isolated cilia. There were very few regions where Act1p was visualized in the absence of coincident tubulin, but more places where tubulin immunofluorescence was seen without corresponding anti-HA fluorescence. This was most likely due to the presence of an additional ATG in the MTT-ACT1-3xHA construct, located between the 3xHA epitope tag and the ACT1 gene. The existence of a second translational start signal downstream from the sequence encoding the 3xHA tag implies that some of the Act1p present in the cilia might not possess the epitope tag. Intracellular immunostaining of cells transformed with this construct was variable; however, in no case did we detect localization of the 3xHA-Act1p in the division furrow.
|
transformants and homozygous ACT1 homokaryons, were studied by transmission electron microscopy. The majority of ciliary cross sections were similar to those of wild-type cells (as shown, for example, in Fig. 8 of reference 69), but 1 to 5% had relatively subtle but distinct abnormalities, notably incomplete doublet microtubules, most commonly the B tubule (Fig. 5A, B, and D), but occasionally the A tubule (Fig. 5C). Very few cilia were grossly abnormal (not shown). This strongly suggests that the ciliary defect is related to an inability to assemble fully normal axonemes.
|
|
homokaryons failed to take up carmine, and only a minority of about 10% took up a few of the smallest particles over a period of hours. A more detailed examination using black ink indicated that while the wild-type cells took up numerous large ink particles that became concentrated near the rear end of the cells (Fig. 6A and B), the great majority of the ACT1 cells took up only a small amount of ink, and that was almost invariably wedged in the oral apparatus (Fig. 6C and D). These results are reminiscent of those obtained earlier (49) for cells of T. pyriformis exposed to cytochalasin B, except that the amount of ink trapped in the oral region was far less in the ACT1 cells of T. thermophila than in the cytochalasin-inhibited cells, which were not reported to have impaired motility. These results strongly suggest that Act1p is directly as well as indirectly required for successful phagocytosis (see Discussion).
Clonal analysis.
Clones of homozygous ACT1 homokaryons were followed for extended periods in microdrops of culture medium to determine the effects of absence of the ACT1 gene on cell division, the cell cycle, and long-term survival.
About 2 weeks after the creation of the ACT1 homokaryons, 50 normal-appearing cells were isolated into microdrops, 25 into SPP and 25 into PPYGFe, and 24 control cells of the ACT1 heterokaryon parental strain were isolated into SPP. Control isolates all generated vigorously growing populations of cells swimming rapidly throughout the microdrop (Fig. 7A). In contrast, cells isolated from ACT1 cultures all crept extremely slowly on the plastic substratum of the drop, and each generated variable numbers of nearly paralyzed progeny (Fig. 7B to D). No ACT1 cell ever recovered normal or near-normal motility.
Among ACT1 cells that were followed closely through their first two division cycles, all progressed to the late constriction stage. The proportion of success of separation in the first two attempts at division was 43% in SPP (n = 40) and 69% in PPYGFe (n = 48), a statistically significant difference (
2 = 6.12, 0.02 > P > 0.01). We have also observed that this difficulty in separation is associated with a defect in the final twisting apart of the nearly separated daughter cells, a process that was named rotokinesis by its discoverers (4).
The average duration of the first cell cycle, measured as the time between successive onsets of furrowing, was 6.6 h (range, 5.0 to 9.5 h, n = 14) in SPP, and 5.9 h (range, 4.25 to 7.5 h, n = 15) in PPYGFe, both at 23 ± 1°C. The average interfission interval in the control heterokaryon clones was 3.1 h (range, 2.5 to 3.75 h, n = 44).
While the control heterokaryon clones reached stationary phase with 200 or more cells per microdrop within 2 days, ACT1 clones grew slowly and variably, probably as consequences both of frequent failure to complete constriction and of lengthened cell cycles. The largest clones (Fig. 7B) included many nearly normal cells and went through as many as 15 cell cycles after inception. Average clones, consisting of 20 to 30 cells, generally included a variety of abnormal reintegrated cells, irregular monsters, and tiny cells that may have budded off from abnormal larger cells (Fig. 7C). The smallest clones were generally those that persistently failed in cell separation; they often included one or a few giant irregular monsters, derived from repeated unsuccessful attempts at cell separation (Fig. 7D).
Figure 8 gives a summary of the quantitative subclonal data. A comparison of histograms A and B indicates that the maximum clone sizes of ACT1 cells were greater in the richer PPYGFe medium than in SPP, but there was no statistically significant difference in the average clone sizes (Mann-Whitney U test). Histogram C indicates that differences in clone size and proliferation rate are not heritable. Two subclonal sets, initiated with normal appearing cells from clones (a and c in Fig. 8B) of very different sizes, both reiterated the range of population sizes seen in the original clones, with no significant differences between them and no consistent loss of proliferative vigor. The subclonal set derived from clone b was strikingly different; there were few normal-appearing cells left in clone b at the time of subcloning and the subclones were uniformly small and mostly died before the end of the experiment.
A clonal analysis was also carried out (in SPP) for cells selected from mass cultures of somatic ACT1 transformants. In the majority (50 out of 60) of such clones, cells swam and divided normally, and produced vigorous populations much like those of the heterokaryon controls described above. However, 10 of the 60 selected cells gave rise to clones very similar to those described above for the homozygous ACT1 homokaryons. In these clones, speed of swimming and frequency of division slowed dramatically within a day after the selection, and many cells underwent a late-fission arrest. These results led us to suspect that the culture was heterogeneous; whereas most somatic ACT1 transformants had not fully assorted, those cells that produced the "slow" clones probably had assorted to fixation or near-fixation of macronuclear chromosomes with disrupted ACT1 alleles. This interpretation was supported by the uniform results (described above) that were later obtained with the homozygous ACT1 homokaryons.
These clonal analyses thus indicate that (i) the division arrest phenotype of the ACT1 cells is due solely to failure of fully furrowed cells to complete their separation, (ii) this failure of separation is invariably associated with loss of cell motility, (iii) cell cycles of the ACT1 cells are substantially lengthened but not halted, (iv) an abnormal cellular geometry resulting from a failure of cell separation impedes but does not totally prevent subsequent cell division, and (v) T. thermophila. cells have the potential to live, grow and divide for a long time without the ACT1 gene, although that potential is not always realized.
Observations of immunostained cells with disrupted ACT1 genes.
Cortical configurations were examined in somatic ACT1 transformants, primarily using the 20H5 anti-centrin monoclonal antibody, which detects basal bodies, oral crescents, and the filament ring underlying the anterior crown of basal body couplets (33, 59). Oral crescents are absent at the onset of cell division (Fig. 9A) and develop in both sets of oral structures during cell division (12, 18, 34). Therefore, cells that have well-developed oral crescents (OC in Fig. 9C, D, and F) are likely to have undergone cell division.
Examination of at least 200 somatic ACT1 transformants fixed at all stages of cortical development (staging from reference 1) up to the onset of division furrowing revealed no abnormalities in cortical patterns visible with anticentrin and antitubulin immunostaining. Most relevant, the oral primordium was always seen at its normal equatorial position along the same cell longitude as the anterior oral apparatus. This held true even at the final stage before the onset of cytokinesis (Fig. 9A), when the fission zone has already formed as a complete equatorial ring of gaps in the longitudinal ciliary rows.
Cells that had undergone previous failure of cell separation almost invariably had oral structures with normal compound ciliary structures (membranelles and undulating membrane) and normal oral crescents; however, they were abnormally positioned, along different cell longitudes in the two presumptive daughter cells (Fig. 9B). In cells that were at advanced stages of regression of the division furrow, the posterior oral apparatus was virtually always on a different cell longitude than the anterior oral apparatus (Fig. 9C, D, and F). In one tally, 46 out of 50 reintegrated cells, diagnosed as such by their mature oral apparatuses, had their two sets of oral structures on different cell longitudes. Since this condition was never seen in predividing cells, it was presumably generated by a lateral displacement of presumptive daughter cells during the aborted cell separation.
A consequence of this lateral displacement was that different ciliary rows abutted one another across the division furrow. Often, these rows did not rejoin, and polar structures were sometimes formed (Fig. 9E, AC2). Disorder and displacement of ciliary rows in the furrow region were commonly observed (Fig. 9F, arrow, and 9G, rows a and b).
Surprisingly, some ciliary rows did rejoin across the furrow plane. This is seen for a single row (arrowhead) in Fig. 9D, and for three rows (a, b, and c) in Fig. 9F. This clearly was not a rejoining of the same rows that had been together prior to division. Note, in Fig. 9F, the major lateral displacement of the posterior relative to the anterior oral structures, as well as the circumstance that three ciliary rows are interposed between the rejoined ciliary rows labeled b and c posterior to OA2, with only one such row interposed anterior to OA2. The microtubular assemblies (longitudinal microtubule bands) associated with the basal bodies sometimes also rejoined (Fig. 9G, rows c and d).
When the presumptive daughter cells retained their linear order, they could attempt to divide again, probably along newly formed fission zones (Fig. 9H). Eventually, many such cells formed irregular monsters in which the oral apparatuses were chaotically distributed albeit still internally normal (Fig. 9I, arrow).
|
DISCUSSION |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
|
An association of actin with inner-arm dynein chains was demonstrated in C. reinhardtii. (39, 71) and in T. thermophila (46) by fractionation and Western blotting. Because Tetrahymena actin differs significantly from other actins, the association of actin with inner-arm dynein may be a conserved property of actin (46). The association was also visualized ultrastructurally in Chlamydomonas reinhardtii by showing a subtle but definite alteration of the inner arms in the actin-minus ida5 mutant (52). For this reason, we examined the ultrastructure of cilia in T. thermophila cells with disrupted ACT1 genes. Although we were unable to assess whether there was a subtle change in the dynein arms, we did find a low proportion of ciliary cross sections with incomplete microtubule doublets (Fig. 5).
The images encountered most frequently suggest that the ACT1 cells experience difficulty in the proper attachment of the B tubule to the A tubule. This suggests that Act1p may be involved, either directly or indirectly, in the assembly of axonemal microtubules in Tetrahymena cells. Interestingly, an identical defect of incomplete B tubules was recently observed in a T. thermophila mutant lacking sites of ß-tubulin polyglycylation (57), a posttranslational modification found mainly in cilia (2). It has been suggested that polyglycylation regulates binding of axonemal microtubule-associated proteins (63). We can speculate that ß-tubulin polyglycylation is required for proper binding or activity of Act1p on axonemal microtubules. Alternatively, both ß-tubulin polyglycylation and Act1p could be cooperatively required for another component(s) which stabilizes the B tubule.
A second effect of ACT1 gene disruption (Table 1) was a nearly complete absence of food vacuole formation in cells lacking normal ACT1 genes. The simultaneous inhibition of cell motility complicates our interpretation of this result, since the absence of food vacuoles might be caused indirectly (by failure to sweep particles into the buccal cavity) or directly (by failure to form normal food vacuoles). A comparison of our results with an earlier observation (49) leads us to conclude that both causes operate simultaneously. Exposure of Tetrahymena pyriformis to cytochalasin B caused a large amount of a marker (carmine) to accumulate in the food vacuole-forming region, while very little got into the cell proper (49). We observed a much smaller accumulation of our marker (ink) in the same region (Fig. 6). This reduced uptake is most likely due to weaker ciliary currents in the oral region, but the identical site of accumulation suggests a defect in internalization of food vacuoles, in this case brought about genetically rather than pharmacologically. This interpretation is consistent with many reports that actin is required for phagocytosis in ciliates.
Our initial observations suggested that Act1p might also be involved in division furrow constriction, since the most obvious phenotype emerging after disruption of the normal ACT1 gene was the appearance of "chains" of connected cells. However, this phenotype was found to be a consequence of cellular reintegration following a failure of separation of cells in which cleavage furrowing had been completed, and it invariably occurred in cells that were virtually immobile and therefore hampered in the final twisting apart (rotokinesis) that is necessary for separation of most cells (4). The observed course of fission and reintegration seen in living ACT1-disrupted cells was identical to that recorded after disruption of two genes (KIN1 and KIN2) encoding the Tetrahymena homologs of kinesin-2 (see Fig. 9 in reference 5). More recently, disruption of another gene required for assembly of cilia, IFT52, was shown to also lead to arrest of cytokinesis (3). Thus, cellular immobility, whether caused by loss of cilia (3, 5) or by impairment of ciliary function as a consequence of ACT1 disruption (this study), prevents or retards rotokinesis (4) and thereby impairs cell separation.
We also observed that even when cell separation failed and various bizarre cortical configurations arose as a consequence, new oral apparatuses remained normal, as did the general architecture of the ciliary rows. Actin has been reported in these structures in T. thermophila. (29) and in P. tetraurelia (41), but there is no evidence available yet that it is essential for their formation. It is possible that one of the other actins might be necessary and sufficient for formation of basal bodies and of the oral apparatus in the absence of Act1p.
We conclude, therefore, that Act1p is required directly for cellular locomotion and for food vacuole formation in Tetrahymena spp. and is required indirectly for normal cell separation (Table 1). Act1p is not essential for the constriction of the division furrow or for the formation and arrangement of cortical structures made of basal bodies and associated cytoskeletal components. The presence of Act1p in the division furrow had been strongly suggested by its immunostaining with an antibody against an N-terminal peptide specific to Act1p (22). We failed to confirm this observation, but since we also failed for technical reasons to get consistent internal immunostaining with our MTT-ACT1-3xHA construct, our observations do not conclusively indicate that Act1p is absent from the division furrow, only that if it is present there, it is not essential for furrowing.
The contribution of the other genes in the actin gene family to these various activities remains to be determined. In our preliminary investigation of the mRNA levels of other actin genes in the homozygous ACT1 homokaryons, expression of at least one of the other actin genes was greatly induced when ACT1 was deleted (C.-C. Tsao and M. A. Gorovsky, unpublished data). We suggest that Act1p and other actin isotypes might share partially, but not completely, redundant functions.
The multiplicity of Tetrahymena actins might also help to account for the differences between the results of ACT1 disruption reported here and the effects of injection of skeletal muscle actin (53) and of amplification of GFP-actin (32). In our study, we specifically examined the physiological role of Act1p; in both of the previous studies, the addition of a perturbing agent may have had dominant-negative effects that impinge upon several different actins. Injected skeletal muscle actin might form complexes with any (or all) of the several Tetrahymena actins. For GFP-actin (32), this possibility is less obvious, since the GFP coding sequence was fused to the same ACT1 gene that was disrupted in our study. However, this gene was carried on an rRNA gene-based processing vector (37) that was highly amplified; hence, it is likely to be even more highly overexpressed than the ACT1 gene under control of the MTT1 promoter. If different actins compete for common ligands, an abundant GFP-Act1p could interfere not only with the function of native Act1p but also with the functions of other actins by titrating shared factors; it might also generate abnormalities on its own, as is known, for example, for GFP-centrin constructs (59).
The inability of ACT1 cells to take up food vacuoles limits a major route of entry of nutrients and therefore might be expected to retard cell growth and impair long-term viability. However, the availability of alternative routes of nutrient entry such as clathrin-mediated endocytosis (13, 51) and carrier-mediated uptake (30) allows survival and growth in the absence of food vacuole formation in certain media (54). We propose that the doubling of generation time that we observed in ACT1 cells was not a direct consequence of loss of the ACT1 gene but rather was an indirect consequence of forcing the cells to use alternative, less efficient systems of nutrient entry. A similar increase in generation time was seen in cells that were prevented from feeding normally by disruption of the ciliary assembly gene IFT52 (D. Dave and J. Gaertig, personal communication). Likewise, our cloning experiments suggested that cells lacking a normal ACT1 gene have the potential to survive indefinitely since some (but not all) of the subclones of ACT1 cells maintained undiminished vigor when the experiment was terminated 6 weeks after the loss of the normal ACT1 gene.
Laterally displaced cortical reintegration: a serendipitous experiment. The cortical reintegration phenotype described here, in which fusion of different ciliary rows can take place after a lateral displacement of the presumptive daughter cells relative to each other, has no precedent in the great variety of fission-block mutants collected over the course of two decades (17, 19, 35, 36, 42; Frankel, unpublished observations). This novel phenotype arises because, unlike any of the conditional fission arrest mutants, every reintegrated divider bearing an ACT1 disruption cassette had at some previous time normally completed cytokinesis, with the two daughters connected only by a thin strand of cytoplasm, which in normal cells gets broken by rotary twisting (rotokinesis). If that twisting should be insufficient to separate the daughters, they will then reintegrate at whatever alignment they happen to be in when the effort at separation by rotation finally ends (Fig. 9B; notice that OA1 and OA2 are not aligned). This seemingly random reintegration puts different ciliary rows of the two daughter cells into close juxtaposition, after which they can either ectopically heal together or remain separate. Cells lacking KIN1 and KIN2 appear to undergo a similar lateral displacement during rotokinesis (Fig. 3E in reference 5), followed by rejoining of ciliary rows (Fig. 1C in reference 4) or anterior shifting of the presumptive posterior daughter cell after reintegration (Fig. 3F in reference 5).
A reintegration of the longitudinal microtubule bands (Fig. 9G) does not necessitate a rejoining of severed microtubules, because these bands consist of short microtubules that are tilted relative to the longitudinal axis of the cell, so that each individual microtubule of the longitudinal band is much shorter than the cell (56).
If our interpretation of the origin of the lateral displacements observed in the reintegrated cells is correct, then the loss of motility brought about by disruption of the ACT1 gene caused Tetrahymena thermophila to carry out a type of experiment that was performed microsurgically on the large ciliate Stentor coeruleus (61, 65). These experiments involved first a horizontal severing of anterior and posterior portions of the same cell, then a rotation of the two subcells relative to one another, and finally a rejoining with a circumferential misalignment. The cell's response could be assessed by observing the severed longitudinal pigment stripes, which in Stentor coeruleus are located between the less visible ciliary rows. In the control experiment, in which the anterior and posterior subcells were rejoined without any lateral displacement, the pigment stripes, and presumably the ciliary rows between them, healed. However, if there was a misalignment, the stripes did not heal, and the cortex of either the basal or apical subcell subsequently grew anteriorly or posteriorly, replacing the other subcell; which of these two misaligned subcells prevailed depended on the level at which the severing and rejoining took place, with a tendency for the posterior subcell to dominate (65). These results can be interpreted to reflect a structural "incompatibility gradient" (see reference 16, p. 126-127) as a correlate of the visible circumferential gradient in pigment stripe widths that characterizes this organism (62); the different longitudes in this organism appear to be different in some way that makes mutual adjustment impossible.
A similar rotation and rejoining, accomplished in Tetrahymena thermophila by aborted rotokinesis, appeared to generate no such radical readjustments; at least some ciliary rows rejoined while the posterior oral apparatus remained at its normal equatorial position (Fig. 9F and G). The absence of a circumferential incompatibility gradient is in accord with the absence in Tetrahymena spp. of the visible structural gradation that is obvious around the circumference of Stentor coeruleus. This also indicates that in some respects Tetrahymena may exceed even Stentor's legendary capacity for regulative reorganization.
|
ACKNOWLEDGMENTS |
|---|
This research was supported in part by a grant from the U.S. National Institutes of Health (GM-26973) to Martin A. Gorovsky of the University of Rochester.
|
FOOTNOTES |
|---|
Supplemental material for this article may be found at http://ec.asm.org/.
|
REFERENCES |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
