Snf2 Proteins Are Required to Generate Gamete Pronuclei in Tetrahymena thermophila

During sexual reproduction/conjugation of the ciliate Tetrahymena thermophila, the germinal micronucleus undergoes meiosis resulting in four haploid micronuclei (hMICs). All hMICs undergo post-meiotic DNA double-strand break (PM-DSB) formation, cleaving their genome. DNA lesions are subsequently repaired in only one ‘selected’ hMIC, which eventually produces gametic pronuclei. DNA repair in the selected hMIC involves chromatin remodeling by switching from the heterochromatic to the euchromatic state of its genome. Here, we demonstrate that, among the 15 Tetrahymena Snf2 family proteins, a core of the ATP-dependent chromatin remodeling complex in Tetrahymena, the germline nucleus specific Iswi in Tetrahymena IswiGTt and Rad5Tt is crucial for the generation of gametic pronuclei. In either gene knockout, the selected hMIC which shows euchromatin markers such as lysine-acetylated histone H3 does not appear, but all hMICs in which markers for DNA lesions persist are degraded, indicating that both IswiGTt and Rad5Tt have important roles in repairing PM-DSB DNA lesions and remodeling chromatin for the euchromatic state in the selected hMIC.


Introduction
The model unicellular eukaryotic ciliate protozoan Tetrahymena thermophila (hereafter referred to as Tetrahymena) stably maintains two morphologically and functionally differentiated nuclei within a single cell [1,2]. The large, transcriptionally active macronucleus (MAC) contains the somatic genome, whereas the small, diploid micronucleus (MIC), which is mostly transcriptionally inert, contains the germline genome. The phenotype of a cell depends on the genetic constitution of its MAC, whereas only the MIC genome is transmitted to progeny MACs and MICs during sexual reproduction called conjugation (Supplemental Figure S1). When conjugation initiates between two cells of different sexes/mating types, their MICs undergo synchronous meiosis. During meiotic prophase, the MICs elongate to form bivalent chromosomes without synaptonemal complex formation [3]. Thereafter, two continuous nuclear divisions take place to form four identical haploid MICs (hMICs) that are in the G2 phase of the cell cycle, which is concomitant with anaphase II [4]. All four hMICs undergo post-meiotic DNA double-strand break (PM-DSB) formation, which is induced by a timely and functionally different manner in the meiotic DSB formation [3]. DNA lesions in hMICs correlate with the appearance of γ-H2AX foci [5], which are markers of DSBs [6]. The γ-H2AX foci disappear only from one hMIC, and this occurs at the same time as the specific localization of the DNA repair proteins; DNAPKcs are involved in DNA repair by non-homologous end-joining and Rad51 involved in recombinational repair to the nucleus [5]. Only this hMIC is selected to form the gamete pronucleus, whereas the three unselected hMICs with persistent γ-H2AX foci are degraded through autophagy mediated by autophagy-related genes (ATGs) [7,8]. Attenuated PM-DSB formation by loss of Microorganisms 2022, 10, 2426 2 of 18 topoisomerase II orthologs culminates in autophagy for all hMICs [5], strongly suggesting that hMIC selection involves self-inflicted DNA damage in all hMICs followed by DNA repair in only one. The selected hMIC undergoes an additional nuclear division, known as gametogenic mitosis, to produce gametic pronuclei [5]. One of the pronuclei migrates to the partner cell to fertilize its stationary pronucleus, whereas the other becomes fertilized by the migratory pronucleus of the partner cell. This reciprocal pronuclear exchange takes place at the conjugation junction where the plasma membranes of conjugating cells are fused [9] and lead to the formation of zygotic nuclei from which both the progeny MACs and MICs differentiate in both partners.
Once selection takes place, the selected hMIC acquires special properties that are not seen in the other three unselected hMICs. For example, the periphery of the selected hMIC is marked by the transmembrane protein Semi1, which mediates the attachment of the nucleus to the conjugation junction [10]. In addition, Semi1 recruits a zinc finger protein Zfr3 to the selected hMIC, which is required for the reciprocal exchange of pronuclei [10]. When it comes to the chromosomes, the disappearance of the γ-H2AX foci from the selected hMIC occurs simultaneously with histone H3 acetylation at lysine 56 (H3K56ac) [5], which is an epigenetic marker of reconstituted chromatin on nascent DNA [11], and the histone H3-H4 chaperone Asf1 specifically localizes to the selected hMIC [5]. Histone H3 also becomes acetylated at lysine 18 (H3K18ac) and 27 (H3K27ac) [5], which are strongly enriched in euchromatin [12,13]. Additionally, a protein containing a high mobility group (HMG) box domain which decreases the compactness of the chromatin fiber [14,15] is abundantly expressed in the nucleus [16]. These findings suggest that the repair of PM-DSBs allows changes in the nucleosome composition to produce fertile gametes. A recent study shows that Nrp1, a NASP-related protein in Tetrahymena, interacts with Asf1 and core histones [17,18]. Knockdown of NRP1 leads to post-meiotic arrest where the γ-H2AX foci persist in all four hMICs [17]. Canonical histones around PM-DSB lesions may be removed either alone or together with chromatin remodelers or histone chaperones [19] to allow access by DNA repair machinery.
The Snf2 family proteins are core to forming chromatin remodelers which have the capacity to add, slide, or eject nucleosomes. The Snf2 family proteins produce energy by hydrolysis of ATP and the remodelers utilize the energy for remodeling activities [20]. The Snf2 family proteins are classified into six groups comprising 24 subfamilies that correlate with functional characteristics and evolutionary phylogenetic relationships [21,22]. The Tetrahymena genome possesses multiple genes encoding the Snf2 family proteins [23,24]. Among them, Brg1 Tt , Chd3 Tt , and Chd7 Tt have been investigated [23][24][25]. These three Snf2 family proteins are abundantly and exclusively expressed in the MAC in vegetative cells. Their genes are essential for cell viability, and these proteins can interact with modified histone proteins directly and indirectly [23][24][25]. They have therefore been regarded as participating in transcription regulation for cell maintenance. Brg1 Tt , Chd3 Tt , and Chd7 Tt also appeared in the progeny MAC Anlagen during conjugation and have been suggested to play roles in the genome rearrangement for progeny MAC formation [23,24]. However, it is yet to be indicated which Snf2 family protein appears in the selected hMIC, and it is yet to be elucidated whether these Snf2 family proteins are involved with the hMIC selection involving epigenetic modifications.
Here, we focused on Tetrahymena Snf2 family proteins that are expressed during the early to mid-phase of conjugation and whose subcellular localization has not previously been elucidated. Our investigation revealed that four Snf2 family proteins (one of two Iswi; IswiG Tt , Rad54 Tt , Rad5 Tt , and Lodestar Tt ) appeared in the selected hMIC and that both IswiG Tt and Rad5 Tt were required for the hMIC selection followed by gametic pronuclei generation. In the ISWIG Tt and RAD5 Tt knockout cells, γ-H2AX foci persisted in all hMICs, and none of the epigenetic markers for euchromatin formation appeared in any hMICs. These findings provide keys to understanding the mechanism of hMIC selection in Tetrahymena.

Database Searching and Motif Scanning
The Snf2 family proteins are characterized by harboring both the SNF2_N domain (PF00176) and HELICc domain (SM00490) [21,22]. To find Tetrahymena orthologs to Snf2 family proteins, we used amino-acid sequences of the domains from representative mammalian and yeast Snf2 family proteins as query sequences in the BLASTp search in the Tetrahymena Genome Database (TGD; http://ciliate.org). We retrieved 15 proteins. The NCBI Conserved Domain Search (https://www.ncbi.nlm.nih.gov/Structure/cdd/wrpsb.cgi) and the MOTIF Search (https://www.genome.jp/tools/motif/) were utilized to predict conventional domains/motifs.

Phylogenetic Analysis
The collected sequences were aligned using the multiple sequence alignment program MAFFT v7.294b [27] with the global pair and maxiterate options. The aligned amino-acid sequences corresponding to conserved SNF2_N and HELICc domains were cropped, and all gaps were eliminated. The remaining 329 sites with 126 taxa were utilized for maximum likelihood (ML) phylogenetic tree reconstruction. For ML calculation, the best substitution model and optional parameters were evaluated using Aminosan [28], and LG+FC+I+G was suggested as the best setting. The maximum likelihood ML phylogenetic relationships were calculated using RAxML-NG v. 1.1 [29], and 100 replicated trees were reconstructed from the same model to evaluate the bootstrapping value.

Construction of C-Terminal EGFP-Tagging and N-Terminal mCherry-Tagging Vectors
For the expression of IswiS Tt -EGFP, IswiG Tt -EGFP, Mot1 Tt -EGFP, Rad5 Tt -EGFP, Rad16A Tt -EGFP, Rad54 Tt -EGFP, Lodestar Tt -EGFP, and pSmarcal1 Tt -EGFP, C-terminal tagging of endogenous proteins was conducted using a knock-in strategy (Supplemental Figure S2A) [30]. Approximately 2 kb from the 3 end of the coding sequence and 2 kb from a downstream region were amplified from genomic DNA of wild-type CU427 cells with PrimeSTAR Max DNA Polymerase (TaKaRa Bio Inc., Shiga, Japan) and the primer sets listed in Supplemental Table S2. Amplified fragments were cloned into the pEGFP-NEO4 (GenBank: AB570109.1) plasmid using the NEBuilder HiFi DNA Assembly Cloning Kit (New England Biolabs Japan Inc., Tokyo, Japan).
The procedure for DNA transfection into the MAC in starved cells is performed in a similar manner to the DNA transfection into conjugating cells [31,32]. The details of the procedure are described in Supplemental Figure S3. It should be noted that, in the DNA transfection targeting the MAC locus with electroporation, the transformation efficiency likely increases if the length of the flanking regions for homologous recombination is longer than used with the gene gun (Dr. Masaaki Iwamoto, personal communication). In addition, transformation efficiency seems to vary considerably depending on the targeted loci.
Initially, C-terminal tagging of endogenous proteins was carried out for all Snf2 family proteins that were examined for subcellular localization analysis. Since C-terminal EGFP-tagged proteins were not expressed in two genes, SWR1 Tt and INO80 Tt , we used N-terminal mCherry tagging modules for the genes. For the expression of mCherry-Ino80 Tt and mCherry-Swr1 Tt , N-terminal tagging of exogenous proteins was conducted as demonstrated previously for the expression of EGFP-DNAPKcs (Supplemental Figure  S2B) [5]. Approximately 2 kb from an upstream region and the 5 end of the coding sequences were amplified from genomic DNA of wild-type CU427 cells with PrimeSTAR MAX DNA polymerase and the primer sets listed in Supplemental Table S2. Amplified forward and reverse target fragments were cloned into the SacI-SalI and SpeI-KpnI sites, respectively, of the backbone plasmid pmCherry-PAC [10] using the NEBuilder HiFi DNA Assembly Cloning Kit. We replaced the PAC-based drug resistance markers which confer cells resistance to puromycin in the plasmids with a paromomycin resistance marker (NEO5), excised from pBNMB1-EGFP (a gift from Kazufumi Mochizuki, Institute of Human Genetics, Montpellier, France) with SalI plus XmaI (New England Biolabs Japan Inc., Tokyo, Japan), using T4 DNA ligase (New England Biolabs Japan Inc., Tokyo, Japan). For DNA transfection, the N-terminal mCherry tagging modules for Tetrahymena were amplified from the constructed plasmids (pmCherry-INO80 Tt -NEO5 and pmCherry-SWR1 Tt -NEO5) with PrimeSTAR Max DNA Polymerase and M13 primers. The resulting PCR products were purified and used for electroporation as described above. Resistant cells appeared within 3 days without adding CdCl 2 and were grown in SPP medium containing an increasing concentration of paromomycin sulfate from 100 to 10,000 µg/mL to allow phenotypic assortment [33]. Protein expression was induced in cells by adding 0.075 µg/mL CdCl 2 to starved cells.

Fluorescence Microscopy of Living Cells
Sixty minutes before observation, Hoechst33342 solution (H342, Cellstain ® , DOJINDO Laboratories, Kumamoto, Japan) was added to conjugating cells to reach 50 ng/mL. The cells were concentrated by centrifugation and resuspended in 3% polyethylene oxide (Sigma-Aldrich Japan K.K., Tokyo, Japan) to increase the viscosity of the medium. Seven microliters of the cell suspension were placed onto a glass slide, and a 22 mm square coverslip was gently placed on the slide. Prepared slides were immediately applied for microscopic observation (Olympus BX50 equipped with UPLFLN60X Objective lens: NA = 0.9 and DP71 Digital Camera; Olympus, Tokyo, Japan).

Indirect Immunofluorescence
Cells were fixed with methanol at −20 • C for 1 h. After removal of methanol by centrifugation (3000× g, 1 min), the cell pellet was postfixed in 1% formaldehyde (Nacalai Tesque Inc., Kyoto, Japan) in PBS pH 7.5 at 4 • C for 1 h. After removal of formaldehyde by centrifugation (3000× g, 1 min), the pellet was washed with PBS three times and incubated for 1 h at room temperature or overnight at 4 • C with primary antibodies. After washing with PBS, cells were incubated with secondary antibodies for 1 h at room temperature in the dark. After washing with PBS, cells were resuspended in PBS containing 1 µg/mL DAPI (Dojindo Laboratories, Kumamoto, Japan), dropped onto a slide, and mounted under a coverslip. Details for antibodies used are presented in Supplemental Table S3.

Gene Knockout Constructions for the Snf2 Family Proteins Localizing to the Selected hMIC
Approximately 2 kb sequences upstream (5 ) and downstream (3 ) of the macronuclear LODESTAR Tt (TTHERM_00313280), RAD5 Tt (TTHERM_00037210), and RAD54 Tt (TTHERM_00237490) loci were amplified from CU427 genomic DNA with PrimeSTAR Max DNA Polymerase and the primer sets listed in Supplemental Table S4. The amplified PCR products were cloned into the pNEO4 (GenBank: EU606202.1) plasmid [34] using the NEBuilder HiFi DNA Assembly Cloning Kit. For DNA transfection, the modules that contain the NEO4 cassette and the 5 and 3 portions of the LODESTAR Tt , RAD5 Tt , or RAD54 Tt genomic locus for homologous recombination were amplified from the constructed plasmids (pKoLODESTAR Tt -NEO4, pKoRAD5 Tt -NEO4, and pKoRAD54 Tt -NEO4) with PrimeSTAR Max DNA Polymerase and M13 primers. The resulting PCR products were purified and used for electroporation as previously described. Resistant cells appeared within 3 days and grew in the SPP medium containing an increasing concentration from 100 to 100,000 µg/mL of paromomycin sulfate to allow phenotypic assortment [33]. Replacement of the target loci from the MAC was confirmed by PCR using the primer sets listed in Supplemental  Table S5.

Acetic Orcein Stain
Ten microliter cell suspension was pipetted onto a glass slide and air dried. The glass slide was fixed in 3:1 methanol: acetic acid for 5 min, incubated in 5 N HCl for 5 min to degrade RNA, and then rinsed in distilled water for 10 sec. Acetic orcein solution (Sigma-Aldrich Japan K.K.) was applied to the sample, and stained nuclei were observed under light microscopy.

Tetrahymena Snf2 Family Proteins
Fillingham et al. indicated that Tetrahymena has 16 genes encoding Snf2 family proteins [24]. Since that study, the genome annotation of the Tetrahymena Genome Database (TGD; www.ciliate.org) has been updated several times [35][36][37][38][39][40]. Using the most recently updated genomic data, we searched for the Tetrahymena genes encoding Snf2 family proteins. Snf2 family proteins are characterized by the SNF_N domain in the N-terminal region followed by a HELICc domain in the C-terminal region [21,22]. We explored the TGD for genes encoding proteins containing these two consecutive domains and found 15 genes (Supplemental Table S1).
The Snf2 family proteins are classified into six groups comprising 24 subfamilies that correlate with functional characteristics and evolutionary phylogenetic relationships [21,22]. THERM_01245640, TTHERM_00193800, and TTHERM_0049310 were previously identified as BRG1 Tt , CHD3 Tt , and CHD7 Tt , respectively [23,24]. However, for the Snf2 family proteins encoded by other genes, their groups and subfamilies had not been identified. To identify these, we conducted a molecular phylogenetic analysis using the conserved amino acid sequences corresponding to the SNF_N and HELIC_C domains of the Tetrahymena Snf2 family protein genes together with various Snf2 family proteins in evolutionarily diverse organisms. The reconstructed ML phylogenetic tree is shown in Figure 1A. Except for the Rad5/16 group, the subfamilies within each group appeared as robustly supported monophyletic groups (Supplemental Figure S4). Five Snf2 family proteins, including Brg1 Tt , Chd3 T t, and Chd7 Tt , were classified into the Snf2-like group. Brg1 Tt , Chd3 Tt , and Chd7 Tt fell into the Snf2 subfamily clade, Mi-2 subfamily clade, and Chd7 subfamily clade, respectively. These results were consistent with the previous classification based on their domain structures, strongly indicating that the reconstructed ML tree was valid. The remaining two genes in the Snf2-like group, TTHERM_00137610 and TTHERM_00388250, were identified as the genes encoding Iswi subfamily proteins. Because the proteins encoded by THERM_00137610 and TTHERM_00585520 exclusively localized to the MAC and the MIC, respectively, we refer to the Iswi Tt encoded by THERM_00137610 as IswiS Tt ; Somatic MAC Iswi Tt , and the Iswi Tt encoded by TTHERM_00585520 as IswiG Tt ; Germline MIC Iswi Tt (Figure 2 and Supplemental Figures S5 and S6). Proteins encoded by TTHERM_00343570 and TTHERM_01546860 belonged to the Swr1-like group. The former was identified as the gene encoding the Ino80 subfamily protein, and the latter was identified as the gene encoding the Swr1 subfamily protein. TTHERM_00237490 was the only gene encoding the Rad54 subfamily. The Snf2 family protein encoded by TTHERM_00313280 was identified as a member of the Mot1 subfamily in the SSO1653-like group. Proteins encoded by TTHERM_00313280, TTHERM_00420480, TTHERM_0037210, and TTHERM_00933250 fell into the Rad5/16-like group. The Rad5/16-like group comprises four subfamilies; the Rad5/16 subfamily, the Ris1 subfamily, the Lodestar subfamily, and the Shprh subfamily. Except for the last subfamily, no supported monophyletic groups representing subfamilies were reconstructed in the ML phylogenetic tree (Supplemental Figure S4). We could therefore not identify the subfamily of four proteins assigned to the Rad5/16-like group from the phylogenetic analysis. The remaining two genes, TTHERM_01080500 and TTHERM_00627150, were identified as the genes encoding the Smarcal1 subfamily proteins within the distant group. In this study, we refer to the Smarcal1 Tt encoded by TTHERM_01080500 as Smarcal1A Tt , and the Smarcal1 Tt encoded by TTHERM_00627150 as Smarcal1B Tt .
The Snf2 family proteins have subfamily-specific accessory domains/motifs, and these accessory domains/motifs are responsible for the functional characteristics of each subfamily [21,22]. We applied two different motif scan tools to predict the accessory domains/motifs of the 15 Snf2 family proteins found in Tetrahymena. Figure 1B shows the structure of the predicted domains/motifs. The accessory domains/motifs detected in Brg1 Tt , Chd3 Tt , and Chd7 Tt were consistent with the structures identified in previous studies. For two Iswi Tt , Ino80 Tt , Swr1 Tt , and Rad54 Tt , accessory domains/motifs unique to each subfamily were identified (e.g., Iswi subfamily proteins possessed SANT and SLIDE domains [41]. Ino80 harbored the DBINO domain at the N-terminal end of the SNF2_N domain [42]). These results were consistent with the classification of these Snf2 family proteins based on the molecular phylogenetic analysis. The four Snf2 family proteins belonging to the Rad5/16-like group were identified as subfamilies based on the detected accessory domains/motifs. Proteins encoded in TTHERM_00420480 and TTHERM_00933250 were classified as Rad16 Tt in the Rad5/16 subfamily because both proteins harbor a ring finger motif between the SNF2_N domain and the HELICc domain [43,44]. In this study, we refer to the Rad16 Tt encoded by TTHERM_00420480 as Rad16A Tt , and the Rad16 Tt encoded by TTHERM_00933250 as Rad16B Tt . TTHERM_00037210 encodes Rad5 Tt in the Rad5/16 subfamily because the HIRAN motif was found in the N-terminal region in addition to the Ring finger motif [45]. The remaining gene TTHERM_00313280 was identified as the gene encoding the Lodestar subfamily protein because the protein lacked both the Ring finger and HIRAN motifs.
we refer to the Rad16 Tt encoded by TTHERM_00420480 as Rad16A Tt , and the Rad16 Tt encoded by TTHERM_00933250 as Rad16B Tt . TTHERM_00037210 encodes Rad5 Tt in the Rad5/16 subfamily because the HIRAN motif was found in the N-terminal region in addition to the Ring finger motif [45]. The remaining gene TTHERM_00313280 was identified as the gene encoding the Lodestar subfamily protein because the protein lacked both the Ring finger and HIRAN motifs.

Snf2 Family Proteins Appearing in the Selected hMIC
To identify Snf2 family proteins that are involved in hMIC selection, the EGFP-or mCherry-tagged constructs were introduced into WT cells to analyze their subcellular localization. Each fusion protein with EGFP or mCherry was expressed from its endogenous MAC locus or from the cadmium inducible MTT1 promoter [46]. Based on DNA repair and euchromatin formation, the indexes for hMIC selection exclusively appear in one meiotic product that gives rise to gametic pronuclei [5,16,17], we explored Snf2 family proteins that localize to the selected hMIC but not to the other three unselected hMICs. Here, we excluded Brg1 Tt , Chd3 Tt , and Chd7 Tt , because previous studies elucidated that these proteins exclusively appear in the MAC and the progeny MAC Anlagen [23,24]. We also excluded Rad16B Tt and Smarcal1B Tt from subcellular localization analysis for the following reasons. The expression profiles in the TetraFGD (http://tfgd.ihb.ac.cn/) indicate that RAD16B Tt transcription is weak during both the vegetative phase and conjugation. SMAR-CAL1B Tt is exclusively and actively transcribed after 8 h from the initiation of conjugation. Since the selected hMIC appears about 6 h after the initiation of conjugation, we postulated that Smarcal1B Tt does not participate in the appearance of the selected hMIC. Figure 2 summarizes the result for Snf2 family proteins involved in hMIC selection, and detailed data are presented in Figures S4-S7. Briefly, 10 Snf2 family proteins showed

Snf2 Family Proteins Appearing in the Selected hMIC
To identify Snf2 family proteins that are involved in hMIC selection, the EGFP-or mCherry-tagged constructs were introduced into WT cells to analyze their subcellular localization. Each fusion protein with EGFP or mCherry was expressed from its endogenous MAC locus or from the cadmium inducible MTT1 promoter [46]. Based on DNA repair and euchromatin formation, the indexes for hMIC selection exclusively appear in one meiotic product that gives rise to gametic pronuclei [5,16,17], we explored Snf2 family proteins that localize to the selected hMIC but not to the other three unselected hMICs. Here, we excluded Brg1 Tt , Chd3 Tt , and Chd7 Tt , because previous studies elucidated that these proteins exclusively appear in the MAC and the progeny MAC Anlagen [23,24]. We also excluded Rad16B Tt and Smarcal1B Tt from subcellular localization analysis for the following reasons. The expression profiles in the TetraFGD (http://tfgd.ihb.ac.cn/) indicate that RAD16B Tt transcription is weak during both the vegetative phase and conjugation. SMARCAL1B Tt is exclusively and actively transcribed after 8 h from the initiation of conjugation. Since the selected hMIC appears about 6 h after the initiation of conjugation, we postulated that Smarcal1B Tt does not participate in the appearance of the selected hMIC. Figure 2 summarizes the result for Snf2 family proteins involved in hMIC selection, and detailed data are presented in Figures S4-S7. Briefly, 10 Snf2 family proteins showed three patterns of nuclear localization when the hMIC selection took place: (1) they appeared in either the MAC (IswiS Tt , Ino80 Tt , Swr1 Tt , Mot1 Tt , and Rad16A Tt ; Supplemental Figure  S5); (2) they appeared in the selected hMIC (IswiG Tt , Lodestar Tt , and Rad5 Tt ; Supplemental Figure S6); or (3) they appeared in both nuclei (Rad54 Tt ; Supplemental Figure S7). Smarcal1A Tt did not localize to any nuclei at the post-meiotic stage but appeared in the MIC undergoing meiotic prophase and in the zygotic nuclei during the 2nd post-zygotic nuclear division (Supplemental Figure S8). These results suggest that Tetrahymena may have acquired multiple Snf2 family proteins that are specialized to manage its nuclear dualism and indicate the possible involvement of four Snf2 family proteins (IswiG Tt , Lodestar Tt , Rad5 Tt , and Rad54 Tt ) in the hMIC selection.

Loss of ISWIG Tt and RAD5 Tt Causes a Defect in hMIC Selection
To elucidate whether any or all of IswiG Tt , Lodestar Tt , Rad5 Tt , and Rad54 Tt are involved in the hMIC selection, the phenotype was observed in conjugating cells in which the respective somatic genes were knocked out. In this observation, we used acetic orcein staining, which allows visualizing Tetrahymena nuclei with light microscopy [47]. For ISWIG Tt , somatic knockout cells of mating types II and VII were obtained from the stock center (https://tetrahymena.vet.cornell.edu/). For LODESTAR Tt , RAD5 Tt , and RAD54 Tt , we created somatic gene knockout mutants for cells of mating types VI and VII. Gene replacement with antibiotic selection with an exogenous NEO cassette [34] can lead to completing the replacement of all 50 target gene copies in the polyploid MAC if the target genes are not essential for vegetative growth, whereas incomplete replacement occurs if they are essential. Phenotypic assortment [33] is attributed to the random distribution of allelic copies in a compound MAC by an amitotic nuclear division. PCR analysis following phenotypic assortment (Supplemental Figure S9) indicated that fragments corresponding to the wild-type loci of ISWIG Tt , LODESTAR Tt , and RAD5 Tt were almost replaced with a fragment for the exogenous NEO cassette in both of the mating types. This result indicates that these three genes are dispensable in vegetative growth for both mating types. In the case of the RAD54 Tt , however, the PCR analysis following phenotypic assortment showed incomplete replacement of the endogenous loci by the NEO4 cassette (data not shown), indicating its indispensable role in vegetative cells. We therefore excluded RAD54 Tt from further analysis. To investigate the impact of depletion of IswiG Tt , Lodestar Tt , and Rad5 Tt for the hMIC selection, we fixed pairing mutant cells every hour after the initiation of conjugation until 16 h and stained fixed cells with acetic orcein to observe nuclear behaviors. The representative images are shown in Figure 3A-D. In LODESTAR Tt ∆ pairing cells, hMIC selection took place at the post-meiosis stage, and the selected hMIC underwent gametogenic mitosis followed by karyogamy and progeny nuclear development ( Figure 3B). At 10 h after the initiation of conjugation, the percentage of cells that reached the progeny nuclear development stage did not differ between the wild-type and LODESTAR Tt ∆ pairs ( Figure 3E), and the aberrant conjugation phenotype was not found in the LODESTAR Tt ∆ pairs ( Figure 3A,B,E). In the ISWIG Tt ∆ pairs and RAD5 Tt ∆ pairs, their knockout did not affect meiosis: four normal hMICs were generated, as in the wild-type pairs ( Figure 3A,C,D). However, none of the hMIC underwent gametogenic mitosis in both ISWIG Tt ∆ and RAD5 Tt ∆ pairing cells at the post-meiosis stage: all hMICs migrated to the posterior region of the cytoplasm, similar to unselected hMIC in the wild-type pair ( Figure 3C,D). These hMICs had disappeared by 10 h after the initiation of conjugation, likely via autophagy ( Figure 3C,D) [7,8], rendering about 43% of ISWIG Tt ∆ pairs amicronucleate and 66% of the RAD5 Tt ∆ pairs amicronucleate ( Figure 3E). These results suggest that IswiG Tt and Rad5 Tt are involved in the hMIC selection.

Depletion of IswiG Tt or Rad5 Tt Does Not Affect PM-DSB Formation but Does Affect DNA Repair and Euchromatin Formation Which Concomitantly Occur in the Selected hMIC
We previously demonstrated that attenuated PM-DSB formation causes a failure in the hMIC selection, resulting in the amicronucleate phenotype [5]. To assess whether PM-DSBs were induced in hMICs after meiosis in ISWIG Tt ∆ pairing cells and RAD5 Tt ∆ pairing cells, immunostaining of γ-H2AX was performed. γ-H2AX foci appeared in four hMICs in both ISWIG Tt ∆ and RAD5 Tt ∆ pairing cells, as wild-type pairing cells ( Figure 4A). This result indicates that PM-DSB formation is independent of either IswiG Tt or Rad5 Tt expression.

Depletion of IswiG Tt or Rad5 Tt Does Not Affect PM-DSB Formation but Does Affect DNA Repair and Euchromatin Formation Which Concomitantly Occur in the Selected hMIC
We previously demonstrated that attenuated PM-DSB formation causes a failure in the hMIC selection, resulting in the amicronucleate phenotype [5]. To assess whether PM-DSBs were induced in hMICs after meiosis in ISWIG Tt Δ pairing cells and RAD5 Tt Δ pairing cells, immunostaining of γ-H2AX was performed. γ-H2AX foci appeared in four hMICs in both ISWIG Tt Δ and RAD5 Tt Δ pairing cells, as wild-type pairing cells ( Figure 4A). This result indicates that PM-DSB formation is independent of either IswiG Tt or Rad5 Tt expression. After the hMIC selection, H3K56 acetylation occurs concomitantly with the disappearance of γ-H2AX foci in the selected hMIC [5]. In addition, H3K18ac, an epigenetic marker representing an enrichment of the euchromatin state appears in the nucleus [5]. We also found that H3K9ac, another marker of euchromatin formation [48], appears in the selected hMIC ( Figure 5). These results indicate that DNA repair involving switching from a heterochromatic to euchromatic chromatin structure occurs in the selected hMIC [5]. At 8 h after the initiation of conjugation, four hMICs in both ISWIG Tt Δ pairing cells and RAD5 Tt Δ pairing cells were undergoing degradation ( Figure 3C,D). γ-H2AX foci persisted in their degrading hMICs ( Figure 4B), suggesting that none of the hMICs in both mutants underwent DNA repair. As additional evidence for this, we performed immunostaining for H3K9ac, H3K18ac, and H3K56ac. As expected, histone H3 acetylation at K9, K18, and K56 was not detected from four hMICs which were undergoing degradation in both ISWIG Tt Δ pairing cells and RAD5 Tt Δ pairing cells ( Figure 5). The persistence of γ-H2AX foci and the absence of H3K9ac, H3K18ac, and H3K56ac strongly suggests that both After the hMIC selection, H3K56 acetylation occurs concomitantly with the disappearance of γ-H2AX foci in the selected hMIC [5]. In addition, H3K18ac, an epigenetic marker representing an enrichment of the euchromatin state appears in the nucleus [5]. We also found that H3K9ac, another marker of euchromatin formation [48], appears in the selected hMIC ( Figure 5). These results indicate that DNA repair involving switching from a heterochromatic to euchromatic chromatin structure occurs in the selected hMIC [5]. At 8 h after the initiation of conjugation, four hMICs in both ISWIG Tt ∆ pairing cells and RAD5 Tt ∆ pairing cells were undergoing degradation ( Figure 3C,D). γ-H2AX foci persisted in their degrading hMICs ( Figure 4B), suggesting that none of the hMICs in both mutants underwent DNA repair. As additional evidence for this, we performed immunostaining for H3K9ac, H3K18ac, and H3K56ac. As expected, histone H3 acetylation at K9, K18, and K56 was not detected from four hMICs which were undergoing degradation in both ISWIG Tt ∆ pairing cells and RAD5 Tt ∆ pairing cells ( Figure 5). The persistence of γ-H2AX foci and the absence of H3K9ac, H3K18ac, and H3K56ac strongly suggests that both IswiG Tt and Rad5 Tt play important roles in DNA repair involving euchromatin formation in the selected hMIC.

Snf2 Family Proteins Localizing to the MAC
To identify the Snf2 family proteins involved in hMIC selection, we generated cells expressing fluorescent-tagged proteins for 10 Snf2 family proteins whose genes are actively transcribed in the early to mid-phase of conjugation and whose subcellular localization has not previously been investigated. Our observations revealed that Snf2 family proteins of Tetrahymena, with the exception of Rad54, exclusively localized to either the MAC or the MIC. In ciliates, active gene transcription occurs only in the MAC. This is responsible for maintaining cellular homeostasis. We found that IswiS Tt , Ino80 Tt , Swr1 Tt , Mot1 Tt , and Rad16A Tt localized exclusively to the MAC. According to previous studies in mammalian cells and yeast, Snf2 family proteins belonging to the Snf2 subfamily, Mi-2 subfamily, Chd7 subfamily, Iswi subfamily, Ino80 subfamily, Swr1 subfamily, and Mot1 subfamily are involved in transcription activation and repression [49][50][51][52][53]. In Tetrahymena, Brg1 Tt (Snf2 subfamily), Chd3 Tt (Mi-2 subfamily), and Chd7 Tt (Chd7 subfamily) exclusively appear in the MAC and play roles in transcription regulation [23,24]. It is reasonable to consider that IswiS Tt , Ino80 Tt , Swr1 Tt , and Mot1 Tt , whose MAC localization was uncovered in this study, also contribute to transcription regulation. Fission yeast Rad16 forms a complex with Rad7 and functions in DNA repair in the genome-wide nucleotide-excision repair system [54]. Rad16A Tt localizing to the MAC may participate in DNA repair and may contribute to the stability of the MAC genome.

Rad54
Rad54 plays an important role in the homologous recombination pathway, one of the major pathways of DNA repair [55]. Rad54 is required for genome stability, and mutations in Rad54 cause chromosome loss [56]. TTHERM_00237490 is a sole gene encoding Rad54 in the Tetrahymena genome, and its protein Rad54 Tt localized to both the MAC and MIC. We attempted to establish RAD54 Tt knockout cells, but the NEO4 cassette could not

Snf2 Family Proteins Localizing to the MAC
To identify the Snf2 family proteins involved in hMIC selection, we generated cells expressing fluorescent-tagged proteins for 10 Snf2 family proteins whose genes are actively transcribed in the early to mid-phase of conjugation and whose subcellular localization has not previously been investigated. Our observations revealed that Snf2 family proteins of Tetrahymena, with the exception of Rad54, exclusively localized to either the MAC or the MIC. In ciliates, active gene transcription occurs only in the MAC. This is responsible for maintaining cellular homeostasis. We found that IswiS Tt , Ino80 Tt , Swr1 Tt , Mot1 Tt , and Rad16A Tt localized exclusively to the MAC. According to previous studies in mammalian cells and yeast, Snf2 family proteins belonging to the Snf2 subfamily, Mi-2 subfamily, Chd7 subfamily, Iswi subfamily, Ino80 subfamily, Swr1 subfamily, and Mot1 subfamily are involved in transcription activation and repression [49][50][51][52][53]. In Tetrahymena, Brg1 Tt (Snf2 subfamily), Chd3 Tt (Mi-2 subfamily), and Chd7 Tt (Chd7 subfamily) exclusively appear in the MAC and play roles in transcription regulation [23,24]. It is reasonable to consider that IswiS Tt , Ino80 Tt , Swr1 Tt , and Mot1 Tt , whose MAC localization was uncovered in this study, also contribute to transcription regulation. Fission yeast Rad16 forms a complex with Rad7 and functions in DNA repair in the genome-wide nucleotide-excision repair system [54]. Rad16A Tt localizing to the MAC may participate in DNA repair and may contribute to the stability of the MAC genome.

Rad54
Rad54 plays an important role in the homologous recombination pathway, one of the major pathways of DNA repair [55]. Rad54 is required for genome stability, and mutations in Rad54 cause chromosome loss [56]. TTHERM_00237490 is a sole gene encoding Rad54 in the Tetrahymena genome, and its protein Rad54 Tt localized to both the MAC and MIC. We attempted to establish RAD54 Tt knockout cells, but the NEO4 cassette could not replace all somatic loci, indicating Rad54 Tt is indispensable for cell viability. We suggest that Tetrahymena Rad54 contributes to genome stability in both nuclei.
When the selected hMIC appeared, a strong fluorescence signal from Rad54 Tt -EGFP was found in the nucleus (Figure 2 and Supplemental Figure S7). Rad51 Tt localizes to the selected hMIC with the disappearance of γ-H2AX, indicating that a homologous recombination pathway contributes to repairing DNA lesions introduced by PM-DSB formation [5]. When Rad51-bound single-stranded DNA invades the complementary strand of donor DNA, Rad54 acts as a motor to relax the donor helix [57]. In the selected hMIC, Rad54 Tt , together with Rad51 Tt , may function in the repair of DNA lesions induced by PM-DSB formation.

Lodestar
Lodestar Tt exclusively localized to the MIC in vegetative cells. During conjugation, Lodestar Tt also localized to the nuclei generated from the MIC, such as hMICs, selected hMIC, pronuclei, and zygotic nuclei (Figure 2 and Supplemental Figure S6). In Drosophila, LODESTAR mutations result in chromatin bridging at the anaphase of mitosis. Lodestar has therefore been considered to facilitate proper chromosome segregation during mitosis [58,59]. In Tetrahymena, the MAC divides by amitosis, whereas the MIC divides by mitosis [60]. During conjugation, the MIC undergoes multiple nuclear divisions, all of which are also mitotic [4]. Therefore, Lodestar Tt , which appeared exclusively in the MIC and nuclei generated from the MIC, may be functionally equivalent to Drosophila Lodestar and participate in the proper segregation of MIC chromosomes during mitotic nuclear divisions. Unlike the Drosophila LODESTAR mutant, however, Tetrahymena LODESTAR Tt ∆ grew as well as the wild-type cells, and no abnormal phenotype was observed in conjugating knockout cells ( Figure 3B,E), suggesting that the function of Lodestar Tt is not essential for mitosis, or that there are other proteins that can compensate for the function of Lodestar Tt in Tetrahymena.

Possible Functions of IswiG Tt and Rad5 Tt for the hMIC Selection
In either ISWIG Tt ∆ pairs or RAD5 Tt ∆ pairs, all hMICs were degraded, indicating that both IswiG Tt and Rad5 Tt play important roles in the hMIC selection. Pair formation occurred in both gene knockout cells as in the wild-type cells, and their MIC began meiosis. No morphological abnormalities were found in the meiotic MIC chromosomes by orcein stain. These indicate that meiosis proceeded normally in both ISWIG Tt ∆ and RAD5 Tt ∆ pairs. γ-H2AX foci were found in the hMICs of both knockout pairs, indicating that depletion of either IswiG Tt or Rad5 Tt did not affect PM-DSB formation which is essential for the hMIC selection [5]. This indicated that the involvement of IswiG Tt and Rad5 Tt in hMIC selection occurs after PM-DSB formation.
After PM-DSB formation, DNA repair factors exclusively localize to the selected hMIC in which γ-H2AX foci disappear [5]. At the same time, H3K9ac and H3K18ac, epigenetic markers for euchromatin formation, appear in the nucleus ( Figure 5) [5]. By contrast, γ-H2AX foci persisted and neither H3K9ac nor H3K18ac appeared in all four hMICs in both the ISWIG Tt ∆ and RAD5 Tt ∆ pairs (Figures 4 and 5). Based on these results, we propose the following two hypotheses as the cause of the amicronucleate phenotype in the ISWIG Tt ∆ pairing cells and Rad5 Tt ∆ pairing cells. First, depletion of either the IswiG Tt or Rad5 Tt causes a failure of the hMIC selection, resulting in the degradation of all hMICs as unselected hMICs. Second is that hMIC selection itself occurs, but that DNA repair with euchromatin formation fails, resulting in the degradation of the selected hMIC as well as the unselected hMICs.
In the selected hMIC, Rad51/Rad54 and DNAPKcs appear simultaneously with the disappearance of γ-H2AX. This indicates that both non-homologous end-joining and homologous recombination repair pathways facilitate repair of DNA lesions caused by PM-DSBs. IswiG Tt localized to the selected hMIC in which γ-H2AX disappeared (Supplemental Figure S10). The mammalian Iswi containing complex is one of the major chromatin remodelers that function in three major DNA repair pathways: homologous recombination, non-homologous end-joining, and nucleotide excision repair [61]. Considering these, IswiG Tt may contribute to DNA repair in the selected hMIC, which is carried out by Rad51/Rad54 and DNAPKcs rather than direct involvement in the hMIC selection. The amicronucleate phenotype observed in the ISWIG Tt ∆ pairs may be due to defects in DNA repair and euchromatin formation in the selected hMIC.
Rad5 Tt localized to the selected hMIC in which γ-H2AX foci disappeared (Supplemental Figure S10). This might suggest that Rad5 Tt also contributes to DNA repair in the selected hMIC. Indeed, yeast Rad5 and its human orthologs, HLTF and SHPRH, are required for tolerance to DNA damage [62]. However, the Rad5-dependent DNA repair pathway only works through DNA replication [63]. During the early-mid phase of Tetrahymena conjugation, DNA replication occurs at meiosis anaphase II or after the gametogenic mitosis, but DNA replication during the post-meiosis stage in which PM-DSB formation and its repair occurs has not been found [4]. If Rad5 Tt is involved in DNA repair and euchromatin formation in the selected hMIC, it may be an independent mechanism from DNA replication. A previous report indicates that Rad5 plays various biological functions, not limited to DNA repair through DNA replication [45]. The possibility that Rad5 Tt participates in DNA repair involving euchromatin formation via an unknown mechanism in the selected hMIC cannot be ruled out.
It should be noted that a part of ISWIG Tt ∆ pairs and RAD5 Tt ∆ pairs could reach the progeny nuclear development stage ( Figure 3E). This may result from incomplete replacement in KO strains, while it was not detected by PCR (Supplemental Figure S9). Alternatively, this may indicate the existence of a pathway that can partially complement the loss of IswiG Tt or Rad5 Tt .

Tetrahymena Monitors the Appearance of the Selected hMIC in Which DNA Repair and Euchromatin Formation Are Completed
A recent study showed that Nrp1, which interacts with histone chaperone Asf1 and core histone proteins, localizes to the selected hMIC [17,18]. Nrp1 is required for DNA repair involving euchromatin formation in the selected hMIC. When Nrp1 is suppressed, γ-H2AX foci persist in all hMICs, and eventually, all hMICs are degraded [17]. Here, we showed that both IswiG Tt and Rad5 Tt participate in DNA repair involving euchromatin formation in the selected hMIC. In ISWIG Tt ∆ pairs and RAD5 Tt ∆ pairs, all hMICs are degraded. As a result, those knockout mutants showed an amicronucleate phenotype. Based on these observations, we suggest that there is a mechanism in the post-meiosis stage that monitors the emergence of mature pronuclei in which PM-DSBs are repaired, euchromatin formation is completed, and that protects mature pronuclei from nuclear degradation. When PM-DSB formation becomes aberrant, or localization of Nrp1, IswiG Tt , or Rad5 Tt to the selected nuclei fails, the DNA repair involving euchromatin formation in the selected nuclei that develop into pronuclei is not completed. Consequently, the selected hMIC degraded in the same manner as the unselected hMIC.

The Involvement of the Iswi Subfamily in Pronuclear Generation with Euchromatin Formation Is Evolutionarily Conserved
In some systems, the sperm paternal genome is packaged as a non-nucleosomal and highly compacted structure [64]. After the fertilizing sperm nucleus has entered the egg cytoplasm, the paternal genome is repackaged into a typical nucleosome structure by maternally providing canonical histone proteins [65]. During this process in Drosophila, maternal Iswi selectively accumulates into the male pronucleus and facilitates nucleosome repacking. Following maternal histone deposition, the chromatin structure in the male pronucleus becomes decondensed [66]. Here, we demonstrated that the Iswi subfamily protein was crucial for chromatin decondensation/euchromatin formation during gamete pronuclei generation in Tetrahymena. This implies that the functional association between Iswi subfamily proteins and gametogenesis is evolutionarily conserved.
Supplementary Materials: The following supporting information can be downloaded at: https:// www.mdpi.com/article/10.3390/microorganisms10122426/s1, Supplemental Figure S1: A schematic diagram showing representative nuclear events during Tetrahymena conjugation. Supplemental Figure S2: The fluorescent-tag knock-in strategy. Supplemental Figure S3: The detailed procedure for DNA transfection into the MAC with electroporation. Supplemental Figure S4: An unrooted ML phylogenetic tree with detailed information such as species, accession IDs, bootstrap value, etc. Supplemental Figure S5: Subcellular localization of Tetrahymena Snf2 family proteins localizing to the MAC. Supplemental Figure S6: Subcellular localization of Tetrahymena Snf2 family proteins localizing to the MIC. Supplemental Figure S7: Subcellular localization of Tetrahymena Rad54 localizing to both the nuclei. Supplemental Figure S8: Subcellular localization of Tetrahymena Smarcal1A. Supplemental Figure S9: Generation of ISWIG Tt ∆, Lodestar Tt ∆, and RAD5 Tt ∆ cells and PCR confirmation. Supplemental Figure S10: Indirect immunofluorescence images indicating the appearance of IswiG Tt and Rad5 Tt in the selected hMIC in which γ-H2AX foci vanished. Supplemental Table S1: List of Snf2 genes found in Tetrahymena genome. Supplemental Table S2: Primer sets for C-terminal EGFPtagging and N-terminal mCherry-tagging. Supplemental Table S3: Antibodies used in the indirect immunofluorescence observations. Supplemental Table S4: Primer sets for somatic gene knockout plasmids. Supplemental Table S5: Primer sets for PCR evaluating somatic gene replacement.
Author Contributions: Y.F. and T.A. contributed to conceptualization; Y.F. and T.A. designed experiments and performed the experiments; Y.F. and T.A. analyzed the data; Y.F., H.B. and K.K. contributed reagents/materials/analysis tools; Y.F. and T.A. wrote the paper; Y.F., T.A., H.B. and K.K. reviewed and edited the paper. All authors have read and agreed to the published version of the manuscript.