Hpo1 is an ARMC9-like protein, TTHERM_001276421

Tetrahymena cells have two permanent polarity axes: anterior-posterior (AP) and circumferential (C). AP polarity is reflected by an asymmetric placement of major cortical organelles: the oral apparatus (OA) near the anterior cell end, the contractile vacuole pores (CVPs) and the cytoproct near the posterior cell end (Fig 1A-1). There are about 20 longitudinal (locomotory) ciliary rows that circumvent the cell. While most of the longitudinal rows span the entire length of the cell, on the ventral side, two rows (called postoral left and right) are shorter, as they terminate posteriorly to the OA (Fig 1A-1). C axis polarity is revealed by asymmetric lateral positions of cortical organelles. While the cytoproct is located at the same longitude as the OA, the CVPs are located on the cell’s right lateral side (Fig 1A-1). During cell division, new organelles form at the same cell longitudes as the preexisting organelles. The new OA (oral primordium or OP) assembles initially as a group of ciliary basal bodies (BBs) forming at a subequatorial AP position in the middle of the cell’s ventral side, on the left side of the right postoral row (Fig 1A-2). BBs proliferate and form four oral ciliary rows (UM on the right and M1, M2 and M3 on the left) (Fig 1A-2, 1B). In the advanced stage of oral development, a fission zone bisects all somatic ciliary rows anteriorly to the OP (Fig 1A-3). The new CVPs and the new cytoproct form at the posterior ends of the anterior half-rows (Fig 1A-4). The cell completes nuclear divisions and undergoes cytokinesis (Fig 1A-4).

Download:

• PNG
  larger image
• TIFF
  original image

Fig 3. hpo1 alleles map to TTHERM_001276421 encoding an ARMC9-like protein.

(A) Mapping of hpo1-3 by the ACCA method. A linkage peak is present around 5 mB on the micronuclear chromosome 3. (B) The 3D structure of TTHERM_001276421 protein generated by AlphaFold2. (B’) The domain organization of TTHERM_001276421 protein based on InterPro. The positions of two substitutions found in strains homozygous for the four hpo1 alleles are marked. (C-E’) SR-SIM images show the cortical localization of either TTHERM_001276421-3xHA (a C-terminally tagged wild type protein) (C,C’) or TTHERM_001276421-3xHA with the hpo1-3-linked substitution, F318S (D-E’). Note that expression of the F318S variant of TTHERM_001276421 phenocopies the cortical organization (multiple and shifted OPs) of hpo1-3. After growth at 38°C for 3 hours, the cells were labeled with the anti-HA antibody (red), 20H5 anti-centrin antibody (green), and DAPI (blue).

https://doi.org/10.1371/journal.pgen.1011735.g003

Download:

• PNG
  larger image
• TIFF
  original image

Fig 4. Hpo1 forms a bidirectional gradient with a high point on the cell’s right lateral side.

(A-C”’) SR-SIM images showing two sides of cells expressing Hpo1-3xHA during the cell cycle. Cells were labeled with the anti-HA (red) and 20H5 anti-centrin (green) antibodies, and DAPI (blue). (D-D’”’) Confocal image of an apical fragment of a cell expressing Hpo1-3xHA. The cell was labeled with anti-HA (red) and 20H5 anti-centrin (green) antibodies, and DAPI (blue). Single channel gray scale images for centrin (D’,D”) and Hpo1-3xHA (D”’,D’”’) are shown. (E) The intensity plots document the gray values measured across the region marked by the yellow lines shown in panels D” and D”.” The orange arrows indicate where the measurements were started, and the purple arrow indicates the position of the dorsal discontinuity. Abbreviations: um, undulating membrane row in OA; op, oral primordium. The numbers mark the positions of ventral rows. Yellow arrows mark the Hpo1 “contrast row”, a row with high Hpo1 level next to a row with a drop off in Hpo1 level.

https://doi.org/10.1371/journal.pgen.1011735.g004

We adopted the longitudinal row numbering scheme after Cole and colleagues [15]. The right postoral (stomatogenic) row is designated as row 0 (Fig 1B). The remaining rows are numbered as +1 or higher when moving to the cell’s right side or -1 or lower when moving to the cell’s left side (Fig 1B). hpo1 (hypoangular 1) alleles disturb the positioning of oral development along the C axis [16]. While in the wild type, a single OP forms near the left side of row 0, in the hpo1 mutants, multiple OPs form near several adjacent ventral rows located to either the left (rows -1, -2…) or right (rows +1, + 2…) of the stomatogenic row 0 (Fig 2B, 2E compare to Fig 2A, 2D, 2G, 2H). One complication is that in some hpo1 mutant cells, the oral apparatus is wider than normal, which is correlated with the presence of 3 post-oral rows. In such cases we designated the middle postoral row as 0 [16]. As described by Frankel and colleagues [16], in the hpo1-3 mutants, the extra OPs formed more frequently to the right of row 0 (Fig 2B, 2E, 2H). However, when multiple OPs were present, they were located near adjacent rows, typically including row 0 or a row in the proximity of row 0. Thus, in the hpo1 mutants, the OP positions are not random. Rather, the hpo1 phenotype can be described as an inability to focus to OP position to a single ventral row. Also as described [16], the severity of the hpo1 allele phenotypes increases at 38°C as compared to the standard temperature 30°C (Fig 2G, 2H). As oral development progresses, adjacent OPs frequently fuse into one oral field and consequently most mutant cells have a single, mature OA (labeled oa in Fig 2B, 2E). Intriguingly, the hpo1 alleles also reduce the number of CVPs per cell: while most wild-type cells have two CVPs (near two adjacent rows on the right side, usually +4 and +5), hpo1 mutant cells often have only one CVP [16], (S1B Fig compare to S1A and S1H Fig).

We used comparative NGS to map genomic variants that co-segregate with each of the four available hpo1 alleles (hpo1-1, hpo1-2, hpo1-3 and hpo1-4). Linkage peaks were consistently detected on the micronuclear chr3 around ~5 Mb (see the data for hpo1-3 in Fig 3A), in agreement with the previous assignment of hpo1-1 to chr3 using complementation tests in crosses between mutant homozygotes and micronuclear nullisomic strains [16]. Within the linkage region, we located two homozygous variants: chr3:5363773 G/A in hpo1-1 and hpo1-4 and chr3:5364108 T/C in hpo1-2 and hpo1-3 genomes, respectively. Both variants are located within the same protein-coding gene, TTHERM_001276421. BlastP searches identified TTHERM_001276421 homologs in diverse ciliate species including another oligohymenophoran, Paramecium tetraurelia (GSPATG00001303001 and 6 paralogs), and more evolutionarily-distant ciliates: the hypotrich Oxytricha fallax (g7224, g191167 and g3871) and the heterotrich Stentor coeruleus (SteCoe_23425, SteCoe_24205). While BlastP searches failed to identify proteins with significant homology outside of the ciliate phylum, the domain organization of TTHERM_001276421 resembles that of the conserved ciliary tip protein ARMC9, whose mutations cause Joubert syndrome [17–19]. Both ARMC9 and TTHERM_001276421 have the same set of protein domains arranged in the same order (Fig 3B-B’). The N-terminal one-third of TTHERM_001276421 is classified by Interpro as the ARMC9 domain (IPRO040369), within which there is a LisH domain (IPR006594), and two short coiled-coil regions (Fig 3B-B’). The C-terminal region of TTHERM_001276421 contains an ARM-like domain (IPRO11989) (Fig 3B-B’). Thus, it appears that TTHERM_001276421 is a ciliate lineage-specific ARMC9-like protein. The chr3:5363773 G/A variant (hpo1-1, hpo1-4) results in the S236N amino acid substitution, while the chr3:5364108 T/C variant (hpo1-2, hpo1-3) results in the F318S substitution, both within the ARM-like domain. In agreement with our findings, homozygotes for hpo1-2 and hpo1-3 were reported to have a similar temperature-sensitive phenotype, less severe than that of the hpo1-1 and hpo1-4 homozygotes ([16], and information deposited at the Tetrahymena Stock Center (https://tetrahymena.vet.cornell.edu/display.php?stockid=SD01466).

To test whether TTHERM_001276421 is the locus of hpo1 alleles, we used homologous DNA recombination to edit TTHERM_001276421 (in the wild-type background) to encode the hpo1-2(3) linked F318S variant with a C-terminal 3xHA tag. Homozygotes expressing TTHERM_001276421-F318S-3xHA had the cortical phenotype similar to the one observed in the original hpo1-3 mutants: multiple early-stage OPs forming near adjacent ventral rows (Fig 3D-D’ compare to Fig 3C-C’) or a single compound primordium (resulting from fusion of adjacent primordia) at a later cell division stage (Fig 3E-E’). While the wild-type TTHERM_001276421-3xHA protein was enriched near the anterior BBs along a subset of ciliary rows on the cell’s right side (Fig 3C-C’) (see below for a detailed analysis of the localization pattern), the F318S Hpo1 variant was either not-detectable above the background (Fig 3D-D’) or greatly diminished but still biased to the cell’s right side (Fig 3E-E’ compare to Fig 3C-C’). Thus, F318S may reduce the targeting of TTHERM_001276421 to the cell cortex or decrease its stability. Next, we created a germ-line based strain with a deletion of TTHERM_001276421. The TTHERM_001276421-KO macronuclear homozygotes showed a phenotype similar to the original hpo1 mutants with a subtle difference: the knockout cells had a higher number of OPs per cell as compared to the hpo1-3 homozygotes (Fig 2C, 2F, 2G). Furthermore, while in the hpo1-3 homozygotes, the OPs were predominantly shifted to the cell’s right side, in the TTHERM_001276421-KO cells the OPs were more frequently shifted to the cell’s left side at both 30 and 38°C (Fig 2C, 2F compare to Fig 2B, 2E, Fig 2H). A leftward shift was reported to occur in the hpo1-2 homozygotes grown at 39°C for 24 hr [16]. Thus, the direction of the shift in the stomatogenic row position appears to depend on the degree of loss of function of Hpo1 and the original hpo1 alleles are likely hypomorphs. Overall the phenotypes conferred by the engineered alleles are remarkably similar to the phenotypes observed in the original hpo1 mutants [16] and therefore we concluded that TTHERM_001276421 is the sought HPO1 gene.

Hpo1 forms a circumferential cortical gradient with a high point on the cell’s right side

During interphase, Hpo1-3xHA was enriched near the most anterior BBs of ~6 somatic ciliary rows on the cell’s right side (Fig 4A-A”’). Along these rows, the levels of Hpo1-3xHA were high within the first most anterior 8-10 BBs and decreased along row length (Fig 4A-A”’). In addition, Hpo1 marked the BBs of the UM row positioned on the right side of the OA (labeled um in Fig 4A-A’). On the ventral side, there was an abrupt drop in the Hpo1 level between row + 1 and row 0 (Fig 4A-A’). This seems important because the OP starts to form in the immediate proximity of BBs of row 0 (likely by nucleation from somatic BBs of row 0 serving as templates [20]). In the early divider with a young OP, the Hpo1-3xHA pattern was unchanged, with a drop-off between rows +1 and 0 and the presence in UM (Fig 4B-B”’). In dividers with a fission line, the pattern of Hpo1-3xHA was duplicated along the AP axis (Fig 4C-C”’). Namely, Hpo1-3xHA appeared along the anterior BBs of a subset of posterior ciliary half-rows on the right side of the cell, mirroring the positions of the pre-existing Hpo1-3xHA in the anterior hemi-cell (Fig 4C-C”’). Thus, the C pattern of Hpo1 is faithfully replicated during the cell cycle. Observations of live cells expressing GFP-Hpo1 revealed a pattern of Hpo1 distribution consistent with the data obtained in fixed cells (S2B, S2C Fig). In addition to the localization at the BBs, Hpo1 weakly localized at positions consistent with the microtubule bundles associated with the BBs (transverse and longitudinal microtubule bundles, S2A–S2C Fig). We did not detect displacements of GFP-Hpo1 foci in live cells.

A close inspection of the levels of Hpo1-3xHA around the cell’s circumference revealed a second drop-off on the cell’s dorsal side (Fig 4A”-A”’, 4B”-B”’). In the cells stained for fenestrin, a marker of CVPs [21,22], the dorsal drop-off was apparent between the second and third row past the CVP row counting clockwise (if the cell was viewed from the apical end) (Fig 5C’). To map the position of the dorsal Hpo1 drop-off more accurately, we imaged the cell’s apical region in cell fragments that offered a “polar” view of the circumference (Fig 4D, 4E). The centrin signal was nearly uniform around the cell circumference (Fig 4D, 4E). On the dorsal surface, the Hpo1-3xHA signal was highest at row + 4 and decreased to almost the background level in two steps: between row + 5/ + 6 and +7/ + 8. (Fig 4D, 4E).

Download:

• PNG
  larger image
• TIFF
  original image

Fig 5. The dorsal discontinuity in the Hpo1 gradient marks the cryptic position for oral development expressed in the Janus C mutant.

(A-B’) Pairs of confocal images showing two sides of janC-4 homozygote cells that have either a single OA (A’A’) or two OAs (B-B’) (note that the janC-4 phenotype is not fully penetrant). Cells were labeled with the anti-HA (red) and 20H5 anti-centrin (green) antibodies, and DAPI (blue). (C-D’) SR-SIM image pairs of cells that are either wild-type (C, C’) or janC-4 homozygous (D, D’) and express Hpo1-3xHA. The cells were decorated with the anti-HA (red), anti-fenestrin (green) antibodies, and DAPI (blue). (E, F) Graphs reveal an increase in the number of rows with enriched Hpo1-3xHA in the janC-4 background as compared to the wild type (E) but the fraction of the circumference occupied by high Hpo1-3xHA remains unchanged when the increase in the total number of rows is taken into account (F). Mean + /- SD. Two samples of cells were analyzed for each genotype and the data were combined. The number of analyzed cells is indicated (n=). A two-tailed unpaired t-test was executed for statistical significance (**: P < 0.01; ns: not significant). Abbreviations: pOA, primary oral apparatus; sOA, secondary (dorsal and usually partially assembled) OA; cvps, contractile vacuole pores.

https://doi.org/10.1371/journal.pgen.1011735.g005

Overexpression of Hpo1-3xHA using the cadmium-dependent promoter MTT1 [23], resulted in strong accumulation in the cell body. However, the right-side- and anterior-biased localization of Hpo1 was still visible and the cortical pattern was not disturbed (S3 Fig). Likely, the biased localization of Hpo1 requires its binding to another spatially-biased cortical factor, and Hpo1, while required, is not rate-limiting for precise positioning of oral development.

The dorsal Hpo1 drop-off marks the position where an extra oral apparatus assembles in the janus C (janC-4) mutant

In the janus mutants (janA, janB and janC allelic groups), ventral structures are abnormally duplicated on the dorsal cell surface [24–27]. The dorsal (secondary) OA is underdeveloped and often inverted in its left-right internal polarity (sOA in Figs 5B and S4A). The janus phenotype was interpreted as a global mirror-image circumferential pattern duplication [24,25,27]. In the janC-4 homozygotes, the distribution of Hpo1-3xHA was similar to that in the wild type, including an enrichment of Hpo1 on the cell’s right side (Fig 5A-B’ compare to Fig 4). The average number of rows with high Hpo1-3xHA (rows between the two drop-off positions) was elevated from 7 in the wild type to 8 in the janC-4 homozygotes (Fig 5E), but janC-4 cells had more total rows per cell and therefore the ratio of Hpo1-enriched rows to the total cell circumference was unchanged (Fig 5F). The expression of the sOA in janus mutants is partially-penetrant [25–27]. Both janC-4 mutant cells with and without a sOA had a similar number of Hpo1-enriched rows (Fig 5B-B’ compare to Fig 5A-A’). Strikingly, in cells with a fully expressed Janus phenotype, the sOA was located at the position of the dorsal discontinuity of Hpo1 (next to row + 7 in Fig 5B-B’; S4 Fig). Thus, in the janus background, Hpo1 is a bilateral marker for oral development.

The janus mutants also have an increased number of CVPs that often are arranged as two sets separated by 1-2 rows [24,25]. Thus, there is a correlation between the widening of the Hpo1-enriched domain and the increased number of CVPs in the janC-4 homozygotes. The mid-point of the CVP domain was located along or close to the row with the peak level of Hpo1 in both the wild type (Fig 5C’) and janC-4 cells (Fig 5D’).

To summarize, OA development occurs at positions where Hpo1 levels drop off, while CVPs form at the peak level of the Hpo1 circumferential gradient.

Hpo1 acts bidirectionally to restrict circumferential positions competent for oral development

Based on its circumferential distribution pattern in the janC-4 homozygotes, Hpo1 may act bidirectionally to exclude OPs from the region of its enrichment on the cell’s right lateral side. To test this idea, we analyzed double mutant homozygotes, janC-1;hpo1-3. The double mutants presented a highly penetrant phenotype: two mature OAs were located side by side near the anterior cell end, sometimes separated by a gap of one or more ciliary rows (Fig 6D compare to Fig 6A–6C). While such cells were rare in the single janC-1 or hpo1-3 mutants, the frequency reached 70% in the double mutants (Fig 6I, 6J). Given its high frequency, the “twin OA” configuration appears relatively stable and is likely propagated during division. Some twin OAs appeared to integrate into a single, compound organelle (Fig 6E). The twin OAs likely result from the “cortical slippage”: lateral shifts in the OP positions in relation to the positions of the parental OAs. Likely both the primary and secondary OAs undergo cortical slippage and eventually posterior cells are produced in which two OAs collide (a situation predicted in [14]) (Fig 6K). The high frequency of twin OAs suggests that the circumferential direction of slippage is opposite for the primary and secondary OAs. We observed dividing janC-1;hpo1-3 cells in the course of “cortical slippage” where the distance between the OAs decreased within one generation; the dividing cells in Fig 6F–6H have two old OAs still separated by a gap, and multiple immediately adjacent OPs. When the parental OAs are relatively close, they are mostly correct in structure suggesting the secondary OP almost invariably underdeveloped (an often inverted) in the single janus mutants, normalizes when it moves toward the primary OA in the double mutants. It appears that the predominant direction of cortical slippage is opposite: toward the cell right for the primary OP, and to the cell left of the secondary OP as depicted in Fig 6K. However, we can not exclude the possibility that the two OAs undergo slippage in the opposite directions (primary OP to the left and secondary OA to the right) or that the slippage directions are variable. Overall, these data suggest that Hpo1 acts to separate the two regions competent for oral development, one of which (dorsal) is repressed in the wild type through action of the Janus gene products. Furthermore, these data correlate the bilateral gradient pattern of Hpo1 (described above) with its bilateral (OP excluding) activity.

Download:

• PNG
  larger image
• TIFF
  original image

Fig 6. Hpo1 acts bidirectionally to separate the two oral apparatuses in the janus C mutant.

Confocal (A-D) and SR-SIM (E-H) images cells with indicated genotypes labeled with the 20H5 anti-centrin antibody (red), and DAPI (blue). Cells shown in all panels except H were incubated for 4 hours at 38°C to enhance the hpo1-3 phenotype. (I-J) The graphs quantifies the frequencies of cells with two mature OAs (I) and two OAs in close proximity (J). Mean + /- SD. N = 4 experiments. Between 6-27 cells were analyzed per experiment and per genotype. A two-tailed unpaired t-test was executed for statistical significance. Stars indicate a statistically significant difference (*: P < 0.05, **: P < 0.01, ***: P < 0.001, ns: not significant). The total number of cells per genotype is indicated (n=). (K) The drawing illustrates the likely outcome of the genetic interaction between janC-1 and hpo1-3. The yellow ovals are OPs. The green arrows shows the proposed predominant direction of shifts in the OP positions. Abbreviations: pOA primary OA; sOA secondary OA; op, oral primordium.

https://doi.org/10.1371/journal.pgen.1011735.g006

Hpo1 interacts genetically with Bcd1, a left side-enriched OP positioning factor

Bcd1 is a conserved Beige-BEACH domain protein [8] whose loss of function in Tetrahymena confers a phenotype similar to that of hpo1: multiple adjacent oral primordia [15]. However, while the hpo1 alleles reduce (S1A–S1C, S1H Fig), bcd1 alleles increase the number of CVPs per cell, respectively [8,15] and (S1F, S1H Fig). Importantly, while Hpo1 is enriched on the cells’ right side, Bcd1 is enriched on the cell’s left side, where it appears to also form a bidirectional circumferential gradient when viewed toward the apical cell end (See Fig 5H in [8], and S5A–B”’ Fig). The AP distribution of Bcd1 protein resembles that of Hpo1, high near the most anterior row ends and fading away toward the posterior row ends ([8] and S5A–B”’ Fig).

The similarity of the hpo1 and bcd1 OP phenotypes and the enrichment of Hpo1 and Bcd1 proteins on the opposite sides of the stomatogenic ciliary row suggest that the two proteins act collectively to restrict the OP formation from spreading to either the cell’s right (Hpo1) or cell’s left (Bcd1) of row 0. To look for interactions between Hpo1 and Bcd1, we first examined whether a deficiency of Bcd1 (bcd1-2) affects the pattern of Hpo1-3xHA (Figs 7, 8). In the bcd1-2 homozygotes, the average number of Hpo1-enriched rows (rows between the discontinuities) was increased to 9 (Fig 8F) but the total number of rows per cell also increased and therefore Hpo1-enriched domain was proportionally unchanged (Fig 8G). As in the wild type, in the bcd1-2 homozygotes, Hpo1-3xHA showed a drop-off in concentration both ventrally and dorsally (Fig 7B-B’ compare to Fig 7A-A’; top view in Fig 7D, 7E compare to Fig 4D, 4E). However, the bcd1-2 background altered the AP distribution of Hpo1-3xHA. In the wild type, the anterior-posterior Hpo1-3xHA gradient is steep, dropping to half signal intensity within several BBs of the anterior cell end. In bcd1-2 homozygotes, the AP distribution of Hpo1-3xHA was more uniformed, only dropping to ~75% signal intensity over the entire row length (Fig 7A”, B”, C).

Download:

• PNG
  larger image
• TIFF
  original image

Fig 7. A Bcd1 deficiency changes the A/P distribution of Hpo1.

(A-B”) SR-SIM images of Hpo1-3xHA-expressing cells that are either otherwise wild-type (A-A”) or bcd1-2 homozygous (B-B”). The cells were labeled with the anti-HA (red) and 20H5 anti-centrin (green) antibodies, and DAPI (blue). The grey scale images are the single channel signals of Hpo1-3xHA. The yellow lines (A” and B”) cover the region whose gray values were measured along the A/P axis, and the light blue arrows indicates the initial measurement. The corresponding A/P grey value intensity plots for Hpo1-3xHA in the two genetic backgrounds are shown in the graph in panel C. (D-D”) An SR-SIM image of the apical region of a cell expressing Hpo1-3xHA (red) in the bcd1-2 background, also stained with anti-centrin antibody (green). The single channel grey scale images for centrin (D’) and Hpo1-3xHA (D”) were used for measurements of the circumferential signal intensities. The measured regions are marked by the yellow lines and the orange arrows orient the measurements. The resulting signal intensity plots are shown in the graph in panel E. Compare to the similar images of the wild-type apical region in Fig 4D, 4E.

https://doi.org/10.1371/journal.pgen.1011735.g007

Download:

• PNG
  larger image
• TIFF
  original image

Fig 8. Bcd1 influences the circumferential pattern of Hpo1-3xHA.

Confocal (A-D’) and SR-SIM (E,E’) images of cells expressing Hpo1-3xHA that are either wild-type (A-B’) or homozygous for bcd1-2 (C-E’). After overnight incubation at 30^oC, the cells were labeled with the anti-HA (red), 20H5 anti-centrin (green) antibodies, and DAPI (blue). The numbers mark the ventral row positions. Yellow arrows mark the Hpo1 “contrast row”, a row with high Hpo1 level next to a row with a drop off in Hpo1 level. Note the presence of two adjacent contrast rows on the ventral surface in the bcd1-2 homozygotes (C-E, compare to A,B). (F, G) Graphs quantify the number of rows enriched in Hpo1-3xHA (F), and the ratio of Hpo1 enriched rows to the total number of rows per cell (G). Mean + /- SD. Cells were analyzed in 3 experiments and the data were combined. A two-tailed unpaired t-test was executed for statistical significance. Stars indicate a statistically significant difference (***: P < 0.001, ns: not significant). The total number of cells per genotype is indicated (n=).

https://doi.org/10.1371/journal.pgen.1011735.g008

Next, we examined how the position of the ventral discontinuity of Hpo1 (that marks the position of normal oral development) is affected by bcd1-2. In wild-type cells, the ventral Hpo1 drop-off occurs between row + 1 (contrast row) and row 0 (Fig 8A-B’). In the dividing bcd1-2 cells in which the OP fields were shifted laterally, the contrast row position was consistently shifted to the row on the right side of the right-most OP. For example, in the dividing cell shown in Fig 8D-D’, there are two early OPs at positions 0 and +1 and the contrast row is located to the right of the OP pair at row + 2. While we have not observed clear cases of a shifted contrast row in interphase cells, there were instances of apparent ambiguity in position of the contrast row. In the bcd1-2 mutant cell shown in Fig 8C-C’, within the anterior 1/3 of the cell the contrast row is + 1. However, in the posterior 2/3 of the cell the contrast row is + 2 (yellow arrows). This cell has 3 postoral rows, a feature common to both bcd1 and hpo1 mutants that correlates with the increased width of the mature OA. However, another interphase bcd1-2 cell shows the same “ambiguous contrast row” phenotype in the presence of two postoral rows (yellow arrows in Fig 8E-E’). Taken together, these data reveal that the Bcd1 deficiency modifies both the AP and C distribution of Hpo1. These observations suggest that there is a cross-talk between Hpo1 and Bcd1, possibly as a boundary interaction between the right and left ventral cortical region. No obvious effect of expression of hpo1-3 on the distribution of Bcd1-GFP was observed but the signal of Bcd1-GFP was weak even in the wild-type background making the evaluation difficult (S5 Fig).

We next asked whether there is a genetic interaction between Hpo1 and Bcd1 by analyzing the phenotypes of double mutants homozygous for bcd1-2 and hpo1-KO alleles. Surprisingly, the bcd1-2;hpo1-KO homozygotes were not viable. To determine the penetrance of synthetic lethality, we attempted to rescue the progeny of mating (double mutant) heterokaryons by biolistic introduction of a transgene expressing Hpo1-3xHA in an unrelated locus. A total of about 5 x 10⁶ mating double mutant heterokaryon cells were either subjected to biolistic bombardment with a Hpo1-3xHA transgene or mock-transformed without plasmid DNA and selected with paromomycin (conferred by the neo gene embedded in the hpo1-KO allele). Eighteen drug-resistant clones were isolated from the population bombarded with the Hpo1-3xHA transgene and one clone was isolated from the mock-transformed population. Six clones from the plasmid bombarded cells were analyzed by immunofluorescence and all were positive for Hpo1-HA (S6B-C’ Fig). The single clone selected in the mock-transformed population was negative for Hpo1-HA as expected (S6A-A’ Fig). A PCR amplification detected a wild-type Hpo1 sequence corresponding to the portion that was deleted in the hpo1-KO allele in the “escapee” clone (S6D-D’ Fig). Likely, in this single clone there was a transfer of the wild-type HPO1 gene from the parental macronucleus to the newly developing macronucleus that rescued the lethality. These observations confirm that the loss of function of Bcd1 and Hpo1 is lethal and 100% penetrant (n = 5 x 10⁶). Next, we evaluated how the cortical pattern changes during the transition from the wild-type to double mutant phenotype in the mass-selected progeny of mating heterokaryons. To this end, heterokaryon strains with complementary mating types were allowed to mate, refed and selected with paromomycin to kill the non-mating (phenotypically wild-type) parental cells. After 24 hr of drug selection, the double mutant progeny cells looked nearly normal except for the frequent OAs with abnormally wide oral M rows (white arrows in Fig 9A). After 30–46 hr, dividing cells frequently had posteriorly displaced primordia (yellow arrows in Fig 9B, 9E–9G). In the cells with a posteriorly shifted OP, the division boundary position was variable, either also shifted to the posterior (asterisks in Fig 9F) or equatorial (asterisks in Fig 9E) or shifted anteriorly (asterisks Fig 9G). Cells with a posteriorly shifted OP and an anteriorly shifted division plane may produce a posterior daughter cell with mature OA near the posterior cell end (Fig 9H). At ~46 hr the cells were uniformly arrested in cell proliferation and had an abnormally elongated and curved morphology with OAs that had curved (pink arrows in Fig 9C, 9I) or fragmented M rows (green arrows Fig 9C). The somatic ciliary rows appeared excessively long, and sometimes twisted (Fig 9D, 9H, 9I). SR-SIM imaging revealed additional patterning defects including OAs that lacked the UM (Fig 9G, 9I) and OAs with M rows arranged with orthogonal orientations (small arrows in Fig 9J, 9K). In addition, there were defects in the organization of the buccal cavity including a missing or fragmented ribbed wall of microtubules (Fig 9G, 9I–9K). The bcd1-2;hpo1-KO heterokaryon progeny were not viable on the specialized culture medium (MEPP) that supports proliferation of mutants lacking a functional oral apparatus [28]. Thus, the lethality is not solely caused by a lack phagocytotic activity that occurs exclusively at the buccal cavity of the oral apparatus. We conclude either Hpo1 or Bcd1 are required for survival and that both proteins contribute to patterning in the ways that extend beyond positioning of the OP on the cell-wide polarity axes, including shaping the internal organization of the OA and controlling the longitudinal expansion of the somatic ciliary rows.

Download:

• PNG
  larger image
• TIFF
  original image

Fig 9. A loss of both the left-side enriched Bcd1 and the right-side enriched Hpo1 severely disturbs cortical patterning.

Confocal (A-C) and SR-SIM (D-K) images of bcd1-2;hpo1-KO double homozygotes obtained as progeny of mating heterokaryons. The times shown in panels A-C refer to the period after refeeding of the mating heterokaryons (after 18 hr of mating in an inorganic medium at 30^oC). Cells were labeled with the 20H5 anti-centrin antibody (red), and DAPI (blue). Abbreviations: m, M oral row; um, UM oral row; rw, ribbed wall of microtubules; ac, apical crown; cs, cortical subdivision (fission line). The stars show the gaps in the longitudinal rows that indicate the position of the “cortical subdivision” (a set of gaps in rows at the division plane). Arrows of multiple colors indicate types of cortical defects: white, OAs with long oral rows; yellow, OP displaced posteriorly; green, OA with disorganized or fragmented rows; pink, OA with circular rows.

https://doi.org/10.1371/journal.pgen.1011735.g009

Strains and culture

All strains used were obtained from the Tetrahymena Stock Center (Cornell University, Ithaca NY, currently housed at Washington University, St Louis, MO USA; https://sites.wustl.edu/tetrahymena/). Cultures were grown in the SPPA medium (1% proteose-peptone, 0.2% dextrose, 0.1% yeast extract and 0.003% EDTA:ferric sodium salt) supplemented with antibiotics (SPPA medium) [59,60]. CU428 mpr1-1/mpr1-1 (MPR1; VII)(TSC_SD00178) and B2086 mat1-2/mat1-2 (mat1-2; II) (TSC_SD00709) were used as wild type controls. CU427 chx1-1/chx1-1 (CHX1; VI) (TSC_SD00715) was used for outcrosses to map the hpo1 mutations. B*VII (TSC_SD00023) was used for self-crosses. The following mutant strains were used: IA393 hpo1-1/hpo1-1; eja1-1/eja1-1 (hpo1-1, eja1-1; II) (TSC_SD01463), IA418 hpo1-2/hpo1-2 (hpo1-2, VI) (TSC_SD_01465), IA443 hpo1-3/hpo1-3 (hpo1-3; II) (TSC_SD01455), IA480 hpo1-4/hpo1-4 (hpo1-4, II) (TSC_SD01466), IA359 janC-1/janC-1 (janC-1, IV) (TSC_SD00637), IA479 janC-4/janC-4 (janC-4; VII) (TSC_SD01520), IA378 bcd1-2/bcd1-2 (bcd1-2, V) (TSC_SD00641), IA437 janC-1/janC-1 and hpo1-2/hpo1-2 (janC-1, hpo1-2(3); IV) (TSC_SD01576). In addition, the following strains were made by editing the micronucleus by homologous DNA recombination (see below): UG20 hpo1-F318S-3xHA-neo4/ hpo1-F318S-3xHA-neo4 (hpo1-F318S-3xHA-neo4), UG21 hpo1::neo2/hpo1::neo2 (hpo1::neo2), and heterokaryon strains UG22 bcd1-2/bcd1-2 hpo1::neo2/hpo1::neo2 (BCD1, HPO1, pm-s), UG23 bcd1-2/bcd1-2 hpo1::neo2/hpo1::neo2 (BCD1, HPO1, pm-s, mates with UG22). To obtain progeny cells with the (terminally lethal) macronuclear genotype hpo1::neo2, bcd1-2, UG22 and UG23 heterokaryon strains were starved and allowed to mate for 24 hr at 30^oC, the cell population was incubated in the SPPA medium for 6 hr followed by mass selection with paromomycin 100 μg/ml in SPPA to kill the parental cells.

Mapping of hpo1 alleles and protein structure prediction

We applied the ACCA method [31] to map the causal mutation for hpo1-3 as described below (the same strategy was used to map the remaining three hpo1 alleles using appropriate homozygous strains). Strain IA443 was crossed to CU427 (homozygous for the cycloheximide (cy)-resistant allele chx1-1 in micronucleus). The heterozygous F1 progeny (hpo1-3/HPO1; chx1-1/CHX1) was assorted in SPPA to cycloheximide (cy) sensitivity. A cy-sensitive F1 was mated to B*VII to produce F2 homozygotes using uniparental cytogamy [61]. F2 clones were selected with 15 µg/ml cy and their cortical phenotype was evaluated by immunofluorescence. Twenty-three phenotypically wild-type or mutant F2 clones were pooled, cultured in 25ml of SPPA overnight, and subjected to starvation in 60 mM Tris-HCl (pH 7.5) for 2 days at 30°C. Total genomic DNA was extracted from the starved F2 pools and used to construct genomic libraries using Illumina Truseq primer adapters. The libraries were sequenced on Illumina HiSeq X to obtain paired-end 150 bp reads at 90x coverage. MiModD in the ACCA mode was used to identify variants linked to the hpo1 phenotype as described in detail in [29,31]. A 3D model of Hpo1 was retrieved from the AlphaFold protein structure Database (https://alphafold.com/). Protein domains were identified by InterPro (https://www.ebi.ac.uk/interpro/).

Gene editing in T. thermophila

To construct a plasmid for a genomic knockout of TTHERM_001276421/HPO1, fragments were amplified from the genomic DNA of the wild-type strain CU428 and subcloned on the sides of the neo4 selectable marker using primer pairs: 5’-AATTCCGCGGCGAACTTC TGAGTCATCATTG-3’, 5’-AATTCTGCAGCTTAAAGGCGTCTACCATTTTATTC-3’ and 5’-ATTCCCGGGTCAAGTATTCAACTCCTCTAAGTG-3’, 5’-AATTGGGCCCGAATACTCTGCTCGTGATGTCGA-3’. The resulting plasmid pcand2-KO-neo4 was design to target for deletion a portion of the coding region composed of 1383 bp starting at codon 250 and including 371 bp of the 3’ UTR. Another targeting plasmid, pIA443-3HA was constructed to introduce the F318S substitution (hpo1-3) into the wild type background and simultaneously add a 3xHA epitope tag sequence at the 3’ end of the coding region. To this end, HPO1 fragments were amplified using the total genomic DNA from strain IA443 (homozygous for hpo1-3) using primers pairs: 5’-AATTCTCGAGGAACTGTATTAGGAGTGATTTCAC-3’, 5’-TTAACTGCAGCATAAAATATCCAATTAATTTCAAATATCCACAAATATTATC-3’ and 5’-AATTGGATCCTTAACTAACTTCATCCTAGAAGCATTC-3’, 5’-AAATACGCGTTTAGTTCAACACTTAGAGGAGTAGAATAC-3’ and used to replace the gene targeting parts of the plasmid pIFT54-3HA-native-neo4 [62]. The targeting fragments of the above plasmids were released using restriction enzymes cleaving near the ends of flanking homologous sequences and the digested DNA was used for biolistic bombardment of mating CU428 and B2086 cells at an early stage of conjugation optimal for targeting genes in the micronucleus, followed by standard crosses to make mutant heterokaryons and homokaryons [63].

To prepare a strain for live imaging of Hpo1, the coding region of TTHERM_001276421/HPO1 was cloned into the plasmid pGFP_PLK2-BTU1ov, downstream of the MTT1 promoter and the GFP sequence. The primer pairs used to amplify HPO1 were 5’-CTATACAAACGCGTGATGTAAAACTTACCTGATTGC-3’, and 5’-GTTCGCTTACGGATCCTCATTAACTAACTTCATCCTAGAAGC-3’. The transgene was placed between the UTR sequences of BTU1 for targeting to this locus by homologous DNA recombination. The plasmid was digested by BamHI and MluI, biolistically introduced into the CU428 strain and transformants were selected with 100 μg/ml paromomycin.

To overexpress Hpo1-HA, a plasmid was made (pMTT1_Hpo1_HA) with the following sequence elements cloned between the UTR sequences of the BTU1 gene: Bsr selectable marker, MTT1 promoter, coding region of HPO1 amplified with primers: 5′-TAAAATAATGGCCAAGTCGACAATGTAAAACTTACCTGATTGCG-3′ and 5′-AACATCATAAGGATAAGCACCGGATCCTTAACTAACTTCATCCTAGAAGCATTC-3′, HA epitope tag sequence. The targeting portion of the plasmid was released with SacI and BamHI, introduced biolistically into the BTU1 locus (of CU428 strain) and transformants were selected with 60 μg/ml blastidicin S. To induce overproduction, the transgene-carrying cells were exposed to 2.5 μg/ml cadmium chloride for 6 hr.

Microscopic imaging

T. thermophila cells were fixed and stained by immunofluorescence as described [29,60]. The primary antibodies used were: polyclonal anti-GFP (Rockland Immunochemicals, #600-401-215; 1:800 dilution), monoclonal anti-HA 16B12 (Covance; 1:300), polyclonal anti-HA (Proteintech, 51064-2-AP; 1:300) and monoclonal anti-centrin 20H5 (EMD Millipore; 1:200-300; [64]). The secondary antibodies were conjugated to either Cy3 or FITC (Jackson ImmunoResearch, 115-095-146 and 111-165-003; 1:300). The nuclei were stained with DAPI (Sigma-Aldrich). The labeled cells were mounted in 90% glycerol, 10% PBS supplemented with 100 mg/ml DABCO (Sigma-Aldrich). To image the apical surface (top view), cell fragments were obtained as follows: 1.5 ml of cell culture was concentrated at 2,800 rpm for 3 min, and washed with 1 ml of the nuclear isolation medium A (0.1M sucrose, 4% gum arabic, 0.0015M MgCl₂, 0.01% spermidine-HCl, pH 6.75) [65]. Cells were concentrated by centrifugation to 150 µl and combined with 160 μl of 1% paraformaldehyde/0.25% Triton X-100, followed immediately by addition of 1.9 μl of octyl alcohol. The mixture was vortexed for 10–60 seconds and 20 μl of the sample was air-dried at the room temperature on a cover glass. Next, immunofluorescence was conducted as described above. Images were collected on a Zeiss LSM 710 confocal microscope with a Plan-Apochromat 63 × /1.40 oil DIC M27 objective and on an ELYRA S1 SR-SIM microscope equipped with a 63 × NA 1.4 Oil Plan-Apochromat DIC. TIRF microscopy was executed as described [66] except that partial immobilization of cells was achieved by entrapment in a small volume of culture medium.

Statistical analysis

Using the GraphPad PRISM software, we executed two-tailed unpaired t-tests to evaluate differences. P < 0.05 was deemed statistically significant.