YAF2-Mediated YY1-Sirtuin6 Interactions Responsible for Mitochondrial Downregulation in Aging Tunicates

ABSTRACT

In budding tunicates, aging accompanies a decrease in the gene expression of mitochondrial transcription factor A (Tfam), and the in vivo transfection of Tfam mRNA stimulates the mitochondrial respiratory activity of aged animals. The gene expression of both the transcriptional repressor Yin-Yang-1 (YY1) and corepressor Sirtuin6 (Sirt6) increased during aging, and the cotransfection of synthetic mRNA of YY1 and Sirt6 synergistically downregulated Tfam gene expression. Pulldown assays of proteins indicated that YY1-associated factor 2 (YAF2) was associated with both YY1 and SIRT6. Protein cross-linking confirmed that YAF2 bound YY1 and SIRT6 with a molar ratio of 1:1. YY1 was bound to CCAT- or ACAT-containing oligonucleotides in the 5′ flanking region of the Tfam gene. Chromatin immunoprecipitation-quantitative PCR (ChIP-qPCR) showed that SIRT6 specifically induced the histone H3 lysine 9 (H3K9) deacetylation of the Tfam upstream region. YY1 and YAF2 accelerated SIRT6-induced H3K9 deacetylation. YY1 and Sirt6 mRNA transfection attenuated mitochondrial respiratory gene expression and blocked MitoTracker fluorescence. In contrast, the SIRT6 inhibitor and Tfam mRNA antagonized the inhibitory effects of YY1 and Sirt6, indicating that Tfam acts on mitochondria downstream of YY1 and Sirt6. We concluded that in the budding tunicate Polyandrocarpa misakiensis, YY1 recruits SIRT6 via YAF2 to the TFAM gene, resulting in aging-related mitochondrial downregulation.

INTRODUCTION

Mitochondria are involved in not only ATP and lipid biosynthesis but also cell differentiation and death (1). The activity of this multifunctional organelle naturally decreases with age (2–4). The mitochondrial free radical theory of aging predicts that reactive oxygen species (ROS) are generated in mitochondrial respiratory chain complexes, which cause the time-dependent accumulation of DNA damage and mitochondrial dysfunction and thus serve as a determinant of life span (4, 5). This theory explains the senescence of long-lived mammals that have sufficient time to accumulate point mutations; however, a number of issues have yet to be resolved (6). In short-lived animals, a mechanism other than point mutations and replication errors (7) may be important for mitochondrial downregulation during aging. We herein present a third epigenetic aging mechanism involving nuclear gene interactions that enable the nonpathological decline of mitochondrial activity.

In the budding tunicate Polyandrocarpa misakiensis, buds arise from locations around the zooidal margin and develop into juvenile zooids approximately 1 week after being pinched off from the parent zooid (Fig. 1A). Bud-forming activity decreases during zooidal aging for 4 to 5 months (Fig. 1B). Although mitochondrial respiratory and gene activities also decrease during aging, they are reversibly restored when adult zooid tissues participate in bud formation (8). A budding-specific polypeptide, TC14-3, is an intrinsic factor involved in the upregulation of mitochondrial activity (8–11).

FIG 1.

PmTfam gene expression and mitochondrial activity during budding and aging in P. misakiensis. (A) Clonal zooids and buds, dorsal view. Scale bar, 0.5 mm. (B) Schematic illustration of budding, bud development, and aging. (C) ISH of PmTfam during aging. (C1) Growing bud, transverse section. Scale bar, 100 μm. (C2) Dorsal area of growing bud. Scale bar, 20 μm. (C3) Dorsal area of juvenile zooid. Scale bar, 20 μm. (C4) Dorsal area of adult zooid. Scale bar, 20 μm. (D) ISH of PmCox1. Scale bars, 20 μm. (D1) Growing bud. (D2) Adult zooid. (D3) Adult zooid transfected with PmTfam mRNA. (E) MitoTracker staining. (E1) Growing bud. Scale bar, 40 μm. (E2) Adult zooid. Scale bar, 40 μm. (E3) Adult zooid transfected with PmTfam mRNA. Scale bar, 20 μm. ae, atrial epithelium; az, adult zooid; c, coelomic cell; db, developing bud; e, epidermis; gb, growing bud; jz, juvenile zooid.

Yin-Yang-1 (YY1) encodes a zinc finger protein that acts as both a transcriptional repressor and activator (12). Sirtuin6 (Sirt6) belongs to the Sir2 family, which functions as a corepressor of gene expression. In yeast, the overexpression of Sir2 results in a prolonged life span, and the loss of Sir2 function antagonizes the longevity effects of calorie restrictions (13). Sir2 and its mammalian homologs serve as NAD-dependent histone deacetylases (14). TC14-3 has been shown to induce the histone methylation of Polyandrocarpa YY1 (PmYY1) and PmSirt6 in order to regulate their expression (10, 11).

Mitochondrial transcription factor A (TFAM) is a high-mobility group protein that promotes mitochondrial transcription activity (15, 16) by specifically binding to the mitochondrial D-loop (mtD-loop) (17, 18). In P. misakiensis, Tfam gene expression is age-dependently downregulated, and synthetic Tfam mRNA introduced into senescent zooids results in the enhanced expression of mitochondrial cytochrome oxidase 1 (Cox1) (10). PmTfam mRNA transfection induced the expression of mtD-loop-driven reporter genes, while the short interfering RNA (siRNA) of PmTfam suppressed reporter gene expression (11). These findings indicate that Tfam plays an indispensable role in the regulation of mitochondrial gene activity in P. misakiensis.

The present study investigated the nuclear gene interactions involved in mitochondrial activities during physiological aging. The results obtained demonstrated that the YY1 and SIRT6 proteins in P. misakiensis were responsible for the downregulation of the PmTfam gene by histone deacetylation, and mitochondrial activity consequently decreased nonpathologically without DNA damage. We also discussed the implications of the decline in mitochondrial activity during aging.

RESULTS

Tfam, YY1, and Sirt6 gene expression during the asexual zooid life span.

PmTfam was expressed in all bud tissues, including the epidermis, atrial epithelium, and mesenchymal cells (11) (Fig. 1C1 and C2). Signals were sustained in juvenile zooids (Fig. 1C3), whereas apparent signals were not detected in the epidermis of adult zooids (Fig. 1C4). The gene expression of PmCox1 was observed in bud tissues but was attenuated in adult tissues, particularly in the epidermis (8) (Fig. 1D1 and D2). When PmTfam mRNA was transfected into adult zooids, the epidermis resumed the expression of PmCox1 (Fig. 1D3). An indicator of mitochondrial membrane potential, MitoTracker also revealed that the respiratory activity of mitochondria declined during aging and was regained in part by the transfection of PmTfam mRNA (Fig. 1E1 to 1E3).

FIG 2.

Aging-related gene expression in P. misakiensis. (A to C) ISH. (A) PmYY1. (B) PmSirt6. (C) PmYaf2. (A1, B1, and C1) Growing buds. (A2, B2, and C2) Developing buds. (A3, B3, and C3) Juvenile zooids. (A4, B4, and C4) Adult zooids. Scale bars in A1 and B1, 100 μm. Scale bar in B4, 40 μm. Other bars, 20 μm. Insets in panels C1 to C4 show lower magnification of buds and zooids. Tissues encircled by squares are enlarged. (D) RT-PCR of genes in growing buds (GB), juvenile zooids (JZ), and adult zooids (AZ). (D1) PmYY1 and PmSirt6 upregulated at the aging stage and beta-actin as an internal reference sample. (D2) PmSirt2, PmHDAC1, PmHDAC3, and PmYaf2 are not upregulated at the aging stage. (E) RT-qPCR of PmYY1, PmSirt6, and PmYaf2. DB shows 2-day-developing buds. Longitudinal bars show standard deviation. ae, atrial epithelium; e, epidermis.

FIG 3.

Effects of synthetic PmYY1 and PmSirt6 mRNA transfection on PmTfam gene expression. (A) Chimeric cDNA consisting of the 5′ UTR of PmRACK1 gene (19) and the ORFs of GFP, PmYY1, or PmSirt6. (B) In vitro translation of synthetic mRNAs. Lane 1, GFP. Lane 2, YY1. Lane 3, SIRT6. Proteins were stained with anti-Myc-tag antibody. Lane 4, Bacterial recombinant YY1. CBB staining. Lane 5, Bacterial recombinant SIRT6. CBB staining. (C) Transfection efficiency of GFP mRNA. RNA was introduced into zooid pieces, and the next day, coelomic cells were smeared and observed by confocal microscopy. Scale bars, 100 μm. (C1 and C2) Lipofection. (C3 and C4) Electroporation. (C1 and C3) Control without mRNA. (C2 and C4) Experiment with mRNA. (D) Effects of PmYY1 and PmSirt6 mRNA transfection and synergistic effects of YY1-Sirt6 and YY1-Yaf2 on PmTfam gene expression. Results of real-time PCR are shown. Horizontal bars show average values. Broken circle shows an exceptional result. (E) Results of ISH of PmTfam. Scale bars, 20 μm. (E1) Control without mRNA. (E2) Experiment cotransfected with PmYY1 and PmSirt6 mRNAs. ae, atrial epithelium; c, coelomic cell; e, epidermis.

FIG 4.

Pulldown and cross-linking assays of protein-protein interactions. (A) Crude extracts and purification of YY1, SIRT6, YAF2, and GFP (arrowheads). (B) Glutathione affinity chromatography and anti-His-tag immunostaining of proteins interacting with GST-YAF2. Asterisk shows a degradation product. (C) Pulldown assay of EDTA-sensitive protein interactions. Arrowhead shows the decreasing amount of PmSIRT6 in the eluate. (D) Native PAGE of proteins prestained with 0.006% CBB-G250. Arrowhead shows a major band of YY1. The bracket (lane 4) shows a smear band shifted upward by adding YAF2 to YY1. (E) SDS-PAGE and anti-His-tag immunostaining of YY1 and YAF2 in the presence of a cross-linker, BS3. (E1) YY1. No homophilic bindings were found. (E2) YAF2. Arrowhead shows a band(s) appearing temporally. (E3) YY1 mixed with YAF2. Arrowheads show bands shifted upward. (F) SDS-PAGE and anti-His-tag immunostaining of SIRT6 and YAF2 in the presence of BS3. (F1) PmSIRT6. (F2) PmYAF2. (F3) SIRT6 mixed with YAF2. Asterisk in panel F1 shows the degradation product of PmSIRT6. Arrowhead in panel F3 shows a de novo band that appears after the protein mixture and cross-linking.

PmYY1 and PmSirt6 were both expressed at low levels in growing buds, while they were prominent in aged animals (Fig. 2A1 to A4 and Fig. 2B1 to B4). The gene expression of YY1-associated factor 2 (PmYAF2) was observed during bud stages and was attenuated during aging (Fig. 2C1 to C4). The results of reverse transcriptase PCR (RT-PCR) were consistent with those of in situ hybridization (ISH) (Fig. 2D1). Reverse transcriptase quantitative PCR (RT-qPCR) confirmed that PmYY1 and PmSirt6 mRNAs both increased during zooid aging, in contrast to PmYaf2 gene expression (Fig. 2E). Even in aged animals, however, the amount of PmYaf2 was markedly larger than the increasing amounts of PmYY1 and PmSirt6. In contrast to PmSirt6, another Sir2 gene, PmSirt2, reduced transcription during aging (Fig. 2D2). A different corepressor gene, PmHDAC1, was strongly expressed during the bud stage and expressed at a low level during the adult stage, while the expression of PmHDAC3 remained constant (Fig. 2D2).

Effects of PmYY1 and PmSirt6 mRNA transfection on PmTfam gene expression.

In preliminary studies, the cDNAs of GFP, PmYY1, and PmSirt6 were ligated to the 5′ untranslated region of PmRack1 (Fig. 3A), which was efficient for reporter gene expression (19). Synthetic GFP mRNA, PmYY1 mRNA, and PmSirt6 mRNA were synthesized in vitro and translated into Myc-tagged proteins (Fig. 3B) with electrophoretic mobilities identical to approximately 28, 46, and 40 kDa, respectively (Fig. 3B, lanes 1 to 3). Although the relative molecular masses of YY1 and SIRT6 were different from their theoretical molecular weights, they showed the same electrophoretic mobility as bacterial recombinant proteins (Fig. 3B, lanes 4 and 5). To examine gene transfection efficiency, GFP mRNA was in vivo introduced into zooids using lipofection and electroporation methods (Fig. 3C). After 1 day, green fluorescence was similarly emitted from coelomic cells using both methods (Fig. 3C2 and C4), which was in contrast to untransfected cells showing only autofluorescence signals (Fig. 3C1 and C3). In subsequent transfection experiments, we used the electroporation method for PCR and the lipofection method for histology because the former sometimes severely injures the epidermis.

The transfection of PmYY1 mRNA was repeated 20 times (Table 1). PmTfam gene expression in the experiment was compared by RT-qPCR with the control, which was electroporated without mRNA. It was decreased to 83.5 ± 32.6% by the transfection of PmYY1 mRNA (Fig. 3D, horizontal bar); however, this value markedly varied in respective replicates. These variations were not directly related to transfection coefficients (Table 1).

TABLE 1.

Relationship between transfection coefficients of PmYY1, PmSirt6, and PmYaf2 mRNAs and PmTfam gene expression

Sample no.	Transfection coefficient; relative tfam gene expression (tfam/G3PD) of:
	YY1 mRNAa	Sirt6 mRNAb + YY1 mRNA	Yaf2 mRNAc + YY1 mRNA
1	17.54; 0.97
2	4.75; 0.26
3	>50; 1.2
4	6.51; 0.41
5	20.1; 0.83
6	7.4; 0.71		21.56; 0.89
7	3.65; 0.72		25.76: 1.15
8	23.95; 1.47	0.7; 1.45
9	7.21; 0.81	13.07; 0.19
10	10.55; 0.87	18.34; 0.45
11	7.55; 0.68	10.12; 0.34
12	34.89; 1.07	8.56; 0.61
13	>50; 0.42	12.12; 0.09
14	>50; 0.95	3.85; 0.15
15	2.05; 0.55		>50; 0.98
16	23.96; 0.66		38.89; 0.75
17	17.45; 1.53		28.82; 1.01
18	20.11; 0.83	3.74; 0.24
19	16.67; 0.71	3.06; 0.39
20	30.12; 1.06	4.51; 0.65

^aThe transfection coefficients of PmYY1 mRNA and its effect on PmTfam gene expression are shown.

^bThe transfection coefficients of PmSirt6 mRNA and the cotransfection effect of PmSirt6 and PmYY1 on PmTfam gene expression are shown.

^cThe transfection coefficients of PmYaf2 mRNA and the cotransfection effect of PmYaf2 and PmYY1 on PmTfam gene expression are shown.

PmSirt6 mRNA transfection by itself had no apparent effect on PmTfam gene expression in all 10 cases (Fig. 3D). However, when PmSirt6 mRNA was cotransfected with PmYY1 mRNA into animals in 10 cases (Table 1), PmTfam gene expression markedly decreased to 34.4 ± 19.9% with one exception (Fig. 3D, broken circle). In this exceptional case (number 8), the transfection of PmSirt6 mRNA failed (Table 1). The cotransfection of PmYaf2 and PmYY1 mRNAs was performed in five cases, and no apparent effects were observed (Fig. 3D; Table 1). This result indicated that the cotransfection of mRNAs by themselves did not decrease PmTfam gene expression.

The results of ISH were consistent with the inhibitory effects of PmYY1 and PmSirt6 cotransfection on PmTfam gene expression (Fig. 3E1 and E2). The results of RT-qPCR and ISH both suggested that interactions between YY1 and SIRT6 proteins regulated PmTfam gene expression.

Protein interactions among YY1, SIRT6, and YAF2.

The recombinant His-tagged YY1 protein, His-tagged SIRT6 protein, and glutathione S-transferase (GST)-tagged PmYAF2 protein were prepared and purified (Fig. 4A, lanes 1 to 6, arrowheads). SIRT6 contained a degradation product visualized by an anti-His-tag antibody (Fig. 4B, lanes 2, 3, and 6, asterisks). In glutathione affinity chromatography, YY1 and SIRT6 were eluted in the flowthrough fraction (Fig. 4B, lanes 1 and 2); however, when these proteins were mixed with the GST-tagged PmYAF2 protein, they were eluted by glutathione (Fig. 4B, lane 3). In the negative control, recombinant GST (Fig. 4A, lanes 7 and 8) instead of GST-tagged YAF2 was mixed with His-tagged YY1 and His-tagged SIRT6. The eluate by glutathione contained negligible signals (Fig. 4B, lane 4). GST-tagged YAF2 was mixed separately with His-tagged YY1 or His-tagged SIRT6. In both cases, YY1 and SIRT6 were efficiently eluted by glutathione (Fig. 4B, lanes 5 and 6). When His-tagged YY1 was combined with GST-tagged SIRT6 (Fig. 4A, lanes 9 and 10), a weak YY1 signal was detected (Fig. 4B, lane 7). These results strongly suggest that YY1 and SIRT6 solely bound to YAF2 and that YY1 may also be associated with SIRT6 with a low binding affinity.

YY1 and SIRT6 are Zn²⁺-binding proteins. They were mixed with GST-tagged YAF2 in the presence of 2 mM EDTA and eluted by glutathione. YY1 showed a clear band irrespective of EDTA (Fig. 4C, lanes 1 and 2), while the SIRT6 band became faint in the presence of EDTA (Fig. 4C, lane 2, arrowhead).

In modified blue native (BN)-PAGE, YY1 showed a major band (Fig. 4D, lane 1, arrowhead), whereas YAF2 and SIRT6 did not have any bands (Fig. 4D, lanes 2 and 3), which may be related to their high isoelectric points (IP) (PmYAF2, 9.96; PmSIRT6, 8.91). When YY1 was mixed with YAF2, the major band disappeared, and a broad band instead appeared upward (Fig. 4D, lane 4, box bracket). When PmSIRT6 was added to this YY1-YAF2 mixture, the broad band was invisible (Fig. 4D, lane 5).

Protein interactions were directly visualized by means of protein cross-linking followed by SDS-PAGE and anti-His-tag immunostaining (Fig. 4E). After treatment with the cross-linker, BS3, YY1 by itself gradually became faint; however, its position on the gel did not change (Fig. 4E1). In contrast, YAF2 of 16 kDa had an additional band of approximately 30 kDa, showing homophilic dimerization (Fig. 4E2, arrowhead). When YY1 was combined with YAF2, new bands of approximately 65 and 80 kDa appeared after 1 min (Fig. 4E3, arrowheads), and a high-molecular-weight band appeared at approximately 150 kDa after 5 min (Fig. 4E3, arrowhead). Protein interactions between SIRT6 and YAF2 were also examined using BS3, SDS-PAGE, and Western blotting. SIRT6 became a smear by itself after the BS3 treatment (Fig. 4F1). YAF2 behaved similarly to that shown in Fig. 4E2 in the presence of BS3 (Fig. 4F2). When SIRT6 was combined with YAF2, a new band of approximately 55 kDa appeared after 1 min (Fig. 4F3, arrowhead). These results indicated that YY1 and SIRT6 interacted with YAF2.

Binding of YY1 to 5′ flanking DNA of the PmTfam gene.

Sequences similar to the YY1-binding consensus motif (ACAT or CCAT) were scattered in the 5′ upstream region of approximately 1.5 kb of the PmTfam gene (Fig. 5A, red letters). Two hundred nucleotides (nt) named territory 1 (Fig. 5A, bold letters) contained three consensus motifs. In gel shift assays, three probes of 30 nt in length were prepared from territory 1 (Fig. 5A, yellow highlights). [³²P]-labeled probes 1 and 2 (ACAT type) had single major and minor bands in the presence of YY1 (Fig. 5B1, lanes 1 and 5, black arrowhead and arrow). In the negative control, YAF2 and SIRT6 showed no apparent bands (Fig. 5B1, lanes 2 and 6). However, following the addition of YAF2 to the probe-YY1 mixture, a major band was shifted upward (Fig. 5B1, lanes 3 and 7, gray arrowheads). When SIRT6 was added to the reaction mixture, the band no longer shifted upward and became weak (Fig. 5B1, lanes 4 and 8). In contrast to probes 1 and 2, probe 3 (CCAT type) gave rise to three major bands when combined with YY1 (Fig. 5B2, lane 1, black and gray arrowheads and arrow). YAF2 and SIRT6 gave no apparent bands by themselves (Fig. 5B2, lane 2). When PmYAF2 was added to the probe-protein mixture, the band in the middle position became weak (Fig. 5B2, lanes 3 and 4).

FIG 5.

Protein-DNA binding by gel shift assay of ³²P-labeled PmTfam gene oligonucleotides and by chromatin immunoprecipitation. (A) The 5′ upstream sequence (1.5 kb) of PmTfam gene. Possible YY1-binding motifs are shown by red letters. Bold letters (200 nucleotides long) (territory 1) have 3 binding motifs. Probes 1, 2, and 3 (30 nt) containing each motif are highlighted. (B) Gel shift assay. (B1) Probes 1 and 2 mixed with YY1, YAF2, and SIRT6. (B2) Probe 3 mixed with the proteins. Black arrowheads (lanes 1 and 5) show a major band in the presence of YY1, and gray arrowheads show a band shifted upward by adding YAF2 to the YY1-oligonucleotide mixture. Arrows show a minor band probably derived from a degradation product of YY1. (C) Blue native PAGE of YY1-oligonucleotide (probe 3) mixture. Arrowhead in lane 1 shows a major band of YY1. The bracket in lane 2 shows the YY1-oligonucleotide (probe 3) mixture giving rise to smear staining at the upper position. (D) YY1 (black arrowhead) cross-linked with oligonucleotide by SPB. Gray arrowhead shows a transient band shifted upward. (E) ChIP-qPCR of Myc-tagged proteins in territory 1 at the 5′ upstream region of PmTfam gene. Juvenile zooids were transfected with mRNA having a Myc tag sequence. After 2 days, the chromatin was extracted and immunoprecipitated by anti-Myc-tag antibody. Each experiment was repeated twice, and the mean values are shown.

YY1-oligonucleotide (probe 3) binding was confirmed by native and SDS-PAGE. In BN-PAGE, YY1 showed a major band (Fig. 5C, lane 1), and the YY1 and probe 3 mixture gave rise to smear staining that moved upward (Fig. 5C, lane 2, box bracket). YY1 was cross-linked to probe 3 in the presence of the cross-linker SPB under UV exposure for definite time periods. A new band of approximately 100 kDa transiently appeared after 2 min (Fig. 5D, gray arrowhead), and an increasing amount of protein appeared at the top of the gel.

Juvenile zooids were transfected with mRNA having a Myc-tagged sequence, and after 2 days, the chromatin was immunoprecipitated using anti-Myc-tag antibody. Transfection experiments were replicated twice. Chromatin immunoprecipitation-quantitative PCR (ChIP-qPCR) showed that Myc-tagged PmYY1 mRNA was more effective in precipitating territory 1 of the PmTfam gene than the control electroporated without Myc-tagged mRNA (Fig. 5E, left). Myc tag-containing PmSirt6 mRNA by itself had no apparent effects (Fig. 5E, right), whereas it was effective in the immunoprecipitation of territory 1 when cotransfected with PmYY1 mRNA having no tag (Fig. 5E, middle). These results afford in vivo evidence that YY1 binds to and recruits Sirt6 to the 5′ flanking region of PmTfam gene.

Tfam histone deacetylation by PmSIRT6.

We examined histone acetylation at H3K9 in territory 1 of the PmTfam gene by ChIP-qPCR and found that anti-acetylated histone H3 lysine 9 (H3K9ac) peaked in developing buds (Fig. 6A). PmSirt6 mRNA was introduced into growing buds and allowed to develop for 2 days. Histone acetylation in territory 1 was examined in developing buds. ChIP-qPCR indicated that PmSirt6 decreased histone H3K9ac to 64% but had no effects on H3K14ac or H3K27ac (Fig. 6B).

FIG 6.

Histone H3K9 deacetylase activity of PmSIRT6 in collaboration with PmYY1 and PmYAF2. (A) ChIP-qPCR of histone H3K9ac in territory 1 of PmTfam gene in the asexual zooidal life from growing buds (GB) and developing buds (DB) to juvenile zooids (JZ) and adult zooids (AZ). (B) Specificity of PmSIRT6 on histone H3K9 deacetylation in territory 1. PmSirt6 mRNA was introduced in vivo into buds. ChIP data using anti-H3K9ac antibody was compared with those using anti-H3K14ac antibody and anti-H3K27ac antibody. (C) Synergistic effects of SIRT6 protein with YY1 and YAF2 on histone H3K9 deacetylation. Nuclear extracts were treated in vitro with proteins and immunoprecipitated using anti-H3K9ac antibody. The results are expressed as the mean ± standard deviations. n, number of biological replicates. P values were calculated by two-tailed Student's t test.

Polyandrocarpa nuclear extracts were treated in vitro with the SIRT6 protein or in combination with the YY1 and YAF2 proteins. ChIP-qPCR showed that, when mixed with YY1 and YAF2, SIRT6 decreased H3K9ac to less than 20% (Fig. 6C), providing further evidence for the molecular collaboration among YY1, YAF2, and SIRT6.

Effects of YY1 and SIRT6 on mitochondrial downregulation.

Mitochondrial activity is high during bud stages and low during zooidal aging. There is currently no evidence to show that the base substitution in the PmCox1 gene accumulates during zooidal aging; however, the mutation appears to accumulate depending on the age of the asexual strain (Table 2). Accordingly, we examined whether PmYY1 and PmSirt6 affected PmCox1 gene expression using nine samples (sample nos. 9 to 14 and 18 to 20 in Table 1). The results of RT-qPCR showed that the cotransfection of PmYY1 and PmSirt6 mRNAs downregulated PmCox1, similar to their effects on PmTfam (Fig. 7A). The SIRT6 inhibitor OSS-128167 (50% inhibitory concentration [IC₅₀], 89 μM), antagonized the histone deacetylation activity of PmSIRT6 at a concentration of 140 μM (Fig. 7B). This concentration of the inhibitor enhanced the gene expression of both PmTfam and PmCox1 (Fig. 7C and D).

TABLE 2.

Base substitution in mitochondrial Cox1 gene in P. misakiensis

Date of birtha	Date of DNA collection	Age of asexual strains	No. of base substitutionsb/no. of cDNA clones examined (mean ± SD) in:
			Buds	Aged adults
October 2015	January 2016	3 mo	0/30 (0)	0/50 (0)
August 2001	December 2015	14 yrs	3/90 (0.03 ± 0.18)	2/70 (0.03 ± 0.17)
Before 1970c	April 2014	>44 yrs	16/90 (0.18 ± 0.46)	20/90 (0.22 ± 0.57)

^aSexual offspring were obtained by natural spawning of gametes.

^bThe 1,064-nucleotide-long sequence was examined.

^cSeveral asexual animals were collected from the field (40).

FIG 7.

Effects of PmSirt6 and PmYY1 mRNA cotransfection and SIRT6 inhibitor on mitochondrial activity. (A) PmCox1 expression in juvenile zooids treated with mRNAs. (B) Effects of SIRT6 inhibitor on PmTfam gene acetylation in the presence of PmYY1 and PmSirt6 mRNAs. (C) Effect of SIRT6 inhibitor on PmTfam gene expression. (D) Effect of SIRT6 inhibitor on PmCox1 gene expression. The results are expressed as the mean ± standard deviations. n, number of biological replicates. P values were calculated by two-tailed Student's t test.

The effects of PmYY1 and PmSirt6 mRNA transfection on mitochondrial respiratory activity were examined using the mesenchymal cells of juvenile zooids cultured for 4 days after the cotransfection of mRNAs. Cells were stained with the membrane potential-dependent mitochondrial dye, MitoTracker. In comparisons with the control without mRNA (Fig. 8A1 and A2), the total brightness and fluorescent cell number decreased to approximately one-third in the mRNA-treated preparation (Fig. 8B1 and B2; Table 3), whereas fluorescence intensity per single positive cell was similar between the control and the experiment (Table 3). PmTfam mRNA suppressed the inhibitory effects of PmSirt6 and PmYY1 mRNA transfection on mitochondrial activity (Fig. 8C1 and C2; Table 3). The SIRT6 inhibitor was also effective for recovery from mitochondrial downregulation after the PmSirt6 and PmYY1 treatment (Fig. 8D1 and D2; Table 3).

FIG 8.

In vitro MitoTracker staining of coelomic cells harvested from juvenile zooids after mRNA transfection and SIRT6 inhibitor treatment. Scale bars, 20 μm. (A) Control without mRNA transfection. (B) Experiment transfected with both PmYY1 and PmSirt6 mRNAs. (C) Cotransfection of PmYY1, PmSirt6, and PmTfam mRNAs. (D) PmYY1 and PmSirt6 mRNA cotransfection in the presence of SIRT6 inhibitor. (A1, B1, C1, and D1) Bright field microscopy merged with fluorescence confocal microscopy. (A2, B2, C2, and D2) Fluorescence confocal microscopy.

TABLE 3.

Effects of PmSirt6, PmYY1, PmTfam mRNAs, and SIRT6 inhibitor on MitoTracker signals

mRNA or inhibitor	No. of samplesa	Total brightnessb/sample	No. of fluorescent cellsc	Brightness/cell
None	10	4,128.5 ± 520.1	242 ± 75	17.1 ± 2.8
PmSirt6 PmYY1	10	1,320.5 ± 158.6	71 ± 11	18.6 ± 1.7
PmSirt6 PmYY1 PmTfam	9	6,839.3 ± 778.8	236 ± 93	29.8 ± 3.3
PmSirt6 PmYY1 OSS-128167	10	4,534.7 + 751.8	194 ± 85	12.3 ± 5.4

^aA square of 285 × 285 μm² was examined in each sample (image).

^bBrightness beyond a common threshold was counted by means of ImageJ software, shown by the mean ± SD.

^cFluorescent cells were counted manually on the confocal microscopic images.

DISCUSSION

PmTfam is an indispensable transcription factor for mitochondrial activity.

In P. misakiensis, PmTfam gene expression declines with age in accordance with decreased mitochondrial activity (10). The transfection of synthetic PmTfam mRNA enhances mitochondrial Cox1 gene expression and respiration activity (10; this study). Mammalian TFAM promotes mitochondrial transcriptional activity together with transcription factor B (15, 16, 20). It specifically binds to promoters located in the mtD-loop (17, 18, 21, 22). In P. misakiensis, Tfam mRNA transfection induces mtD-loop-driven reporter gene expression, and the RNA interference (RNAi) of PmTfam suppresses reporter gene expression (11). These findings and the present results obtained using mRNA and siRNA indicate that Tfam is indispensable for mitochondrial activity in P. misakiensis (Fig. 9).

FIG 9.

Involvement of YY1, YAF2, and SIRT6 in the mitochondrial function during aging in P. misakiensis. During zooidal aging, YAF2 binds to increasing amount of YY1 and to SIRT6 in a zinc-dependent manner. As YAF2 appears to form the homophilic multimer, it would associate YY1 with SIRT6. Consequently, YY1 recruits PmSIRT6 to the 5′ upstream region of PmTfam. Histone H3K9 is deacetylated there, and PmTfam and mitochondrial gene activities are downregulated during aging. The mitochondrial downregulation does not always bring about disadvantage. In P. misakiensis, aging accompanies the downregulation of ROS scavenger. Therefore, the mitochondrial downregulation may contribute to the protection of cells from nuclear damages.

In mammals, many nuclear gene products have been shown to regulate mitochondrial activity (21). Peroxisome proliferator-activated receptor γ coactivator 1α (PGC1α) plays a crucial role in mitochondrial biogenesis and respiratory gene expression (23). In P. misakiensis, PGC1α gene expression was downregulated during aging (K. Kawamura, unpublished data). However, it currently remains unclear whether the downregulation of PmPGC1 α is responsible for aging-related mitochondrial inactivation. Prohibitin (PHB) is ubiquitously distributed in yeast, plants, and animals (24) and plays an essential role in mitochondrial respiration, biogenesis, and homeostasis (25). In P. misakiensis, prohibitin 2 (PHB2) has been identified as a transdifferentiation-related gene (26). We recently found that PmPHB2 mRNA transfection enhanced MitoTracker staining in adult zooids, similar to PmTfam (Kawamura, unpublished).

YY1 regulates PmTfam gene expression in synergy with Sirt6.

Mitochondrial gene mutations caused by ROS may result in mitochondrial dysfunction and thus serve as a life span determinant (3–5). As demonstrated in the present study, asexual strains of P. misakiensis accumulated the base substitution in the mitochondrial Cox1 gene with increasing time lengths after the birth of sexual offspring. However, the de novo gene mutation has not been found to increase during the aging of individuals from budding to senescence, suggesting that a nonpathological factor(s) in P. misakiensis regulates mitochondrial downregulation during aging. The present study indicated that PmYY1 and its cofactor were potent candidates of nonpathological factors.

YY1 interacts with a number of cofactors to activate and inactivate downstream genes (12). In the present study, PmYY1 mRNA exerted various effects on PmTfam gene expression. In some cases, YY1 appeared to interact with an endogenous coactivator(s), such as histone acetyltransferases, to enhance PmTfam gene expression. YAF2 is a ring finger protein associated with YY1 (27). In the present study, results of RT-qPCR showed that PmYaf2 gene expression decreases during aging, but aged animals still have much larger amount of PmYaf2 than the increasing amounts of PmYY1 and PmSirt6. PmYaf2 was expressed in the epidermis; however, the signals by ISH were weak in contrast with those by RT-qPCR, which may be related to the short length of antisense probe of PmYaf2. The open reading frame (ORF) of PmYaf2 is 420 bp in length.

The inhibitory effects of PmYY1 on PmTfam gene expression were stably reinforced by mixing PmYY1 mRNA with PmSirt6 mRNA; however, the cotransfection of YY1-Yaf2 did not affect the results obtained, suggesting that YY1 specifically functions in synergy with the transcriptional corepressor SIRT6. Pulldown and cross-linking protein assays indicated that YAF2 mediated YY1-SIRT6 binding (Fig. 9). YAF2 cross-linked with YY1 as early as 1 min after the protein mixture in the presence of BS3. The de novo 65-kDa band may be due to binding with a molar ratio of 1:1. The PmYAF2-PmSIRT6 interaction also occurred after 1 min, and the 55-kDa product indicated molecular binding of 1:1. PmYAF2 may form a homophilic dimer. Histone deacetylase (HDAC) class I (HDAC1, HDAC2, and HDAC3) is another corepressor (28). In P. misakiensis, aging did not influence the gene expression of HDAC1, HDAC3, or HDAC class III (Sirt2) (11; this study). Therefore, it is reasonable to assume that PmYY1-PmHDAC interactions, if any, are indifferent to aging or budding.

In the pulldown assay, the binding of SIRT6 to the YY1-YAF2 complex was sensitive to EDTA. SIRT6 is a zinc-binding protein, similar to YY1 and YAF2 (27). SIR2 family proteins contain four cysteines in the conserved zinc finger-like loop within a small domain (29). In archaebacteria, when Cys is replaced with Ala or the protein is deprived of zinc ions, the enzyme abolishes histone deacetylase activity (30); however, the zinc ion is distant from the enzyme active site in bacteria and humans (29). In the gel shift assay performed in the present study, PmSIRT6 did not exert any apparent effects on protein-DNA binding in the Tris-borate-EDTA (TBE) buffer containing EDTA. These results indicated that zinc ions are essential for the binding of SIRT6 to the YY1-YAF2-DNA complex (Fig. 9).

YY1 binds the consensus motif on the PmTfam gene.

YY1 has been shown to recognize the CCAT or ACAT core consensus motif in nt sequences (31). In the 5′ upstream region of PmTfam, territory 1 contained three core motif sequences. The results of the gel shift assay clearly showed that YY1, but not YAF2 or SIRT6, bound to both CCAT- and ACAT-containing probes. When YAF2 was added to YY1, the radioactive probe shifted upward. Based on the theoretical molecular weight of YAF2 (16 kDa), the change in electrophoretic mobility appeared to be too large. Since PmYAF2 had a high IP value (9.96), it may have caused an overall increase in the positive charge and been responsible for the slow electrophoretic mobility of the YY1-YAF2-oligonucleotide probe complex.

We herein describe how an additional major band appeared in the gel shift assay when YY1 was combined with the CCAT probe (probe 3). Probe 3 contains GCCGCCATTATT, which is very similar to the longer DNA-binding motif of YY1 (GCCGCCATTTTG) (32). The YY1-probe 3 mixture showed a broad band on BN-PAGE and had a de novo single band of approximately 100 kDa 2 min after protein-DNA cross-linking, suggesting that YY1 bound to probe 3 with a molar ratio of 2:1. This dimerization of YY1 may be responsible for the additional major band in the gel shift assay. Mammalian YY1 has been shown to form a dimer to facilitate enhancer-promoter interactions (33).

In addition to in vitro evidence for YY1-DNA binding, YY1 was shown to bind in vivo to the 5′ flanking region of the PmTfam gene by using anti-Myc-tag antibody precipitation. The results of ChIP-qPCR also showed that YY1 recruited SIRT6 to DNA. In this experiment, territory 1 of PmTfam was immunoprecipitated at an unexpectedly high level in the absence of Myc-tagged mRNA. It is possible that the anti-Myc-tag antibody reacted with an endogenous protein(s) other than Myc-tagged protein.

YY1 recruits histone H3K9 deacetylase to the PmTfam gene.

ChIP-qPCR showed that in the 5′ upstream region of the PmTfam gene, histone H3K9 acetylation was the strongest in developing buds, which is consistent with the expression pattern of the histone H3K9 acetyltransferase gene PmGCN5 (34). Histone acetylation in adult zooids may come from the pharynx and gonad that heavily undergo nuclear H3K9 acetylation (34). Consistent with H3K9ac, the pharynx continues to express Tfam gene during zooidal stages (11). Human SIRT6 exhibits weak but evident histone H3K9 deacetylase activity (35, 36) and, in contrast to other SIR2 family proteins, contains an extended long zinc-binding loop (36). PmSIRT6 had an eight-residue insertion instead of two residues between the second set of cysteines (data not shown). This zinc-binding loop may be necessary for enzyme activity (36).

PmSirt6 mRNA transfection specifically deacetylated H3K9 in territory 1 of the PmTfam gene without affecting H3K14 or H3K27. H3K9 deacetylation in territory 1 was markedly accelerated in vitro by the addition of YY1 and YAF2 proteins to SIRT6, and the SIRT6 inhibitor consistently stimulated histone H3K9ac of the PmTfam gene, indicating that PmSirt6 encodes histone deacetylase. YY1 may stimulate the enzyme activity of SIRT6; however, it is more reasonable to assume that YY1 recruited histone deacetylase to the proper position of the 5′ upstream region of PmTfam (Fig. 9).

SIRT6 is responsible for mitochondrial downregulation via Tfam gene deacetylation.

The decrease observed in mitochondrial function is generally considered to reflect dysfunctions in mitochondria and cells (4). However, it is important to discriminate nonpathological decreases in the activity of mitochondria from irreversible dysfunction. In P. misakiensis, mitochondrial gene activity decreases during zooidal aging and is reversibly restored during budding (8). The present study showed that SIRT6 caused mitochondrial downregulation in collaboration with YY1 and that TFAM neutralized this deleterious effect of SIRT6 and YY1. The SIRT6 inhibitor stimulated PmTfam gene expression and upregulated mitochondrial function, similar to PmTfam mRNA. These results indicated that TFAM functioned downstream of SIRT6 and YY1 and antagonized SIRT6 (Fig. 9). We recently found that PHB2 was more effective than TFAM (Kawamura, unpublished).

We do not assume that mitochondrial downregulation is always disadvantageous for life. Nonpathological and reversible mitochondrial downregulation may be indispensable for aged animals. In P. misakiensis, many nuclear genes decrease in transcriptional activity during aging, and the gene for Zn/Cu-superoxide dismutase (SOD1) is one such case (37). If mitochondria remain active despite low SOD1 activity, excess amounts of ROS may accumulate and induce severe damage in the nucleus. Therefore, the low activity of mitochondria may have adaptive significance in aged animals (Fig. 9).

In conclusion, a novel protein complex consisting of YY1, YAF2, and SIRT6 negatively controls mitochondria, and this complex appears to have the adaptive significance of suppressing uncontrollable mitochondrial activities (Fig. 9), which is expected to guarantee intracellular quality management during normal aging.

MATERIALS AND METHODS

Polyandrocarpa strains.

Several asexual strains of P. misakiensis were separately cultured on glass plates in culture boxes placed at the Uranouchi Inlet near the Usa Marine Biological Institute, Kochi University. Some glass plates containing zooids and buds were brought back to the laboratory every week for experiments. F1 zooids were obtained by keeping sexually mature zooids in the dark in a seawater tank overnight and then exposing them to light in order to induce spawning.

Mitochondrial staining.

MitoTracker Red CMXRos (Invitrogen, Eugene, OR, USA; catalog no. M7512) was dissolved in dimethyl sulfoxide (DMSO) as a 1 mM stock solution and diluted 10,000-fold in sterile seawater immediately before use. Living animals were incubated in MitoTracker-containing seawater for 20 min, fixed in 4% formaldehyde in phosphate-buffered saline (PBS; pH 7.4), and cryocut at a thickness of 20 μm. Sections were observed with confocal microscopy (Eclipse C1si system; Nikon Co., Ltd., Tokyo, Japan). The strength of fluorescence was quantified by pixel values per micrometer squared using the ImageJ tool (38).

Preparation and transfection of synthetic mRNA.

The open reading frames (ORFs) of PmTfam (GenBank accession no. AB920559), PmYY1 (GenBank accession no. LC348389), and PmSirt6 (GenBank accession no. AB920558) were ligated to the pCMV vector (Stratagene, La Jolla, CA, USA). Green fluorescent protein (GFP) cDNA was prepared for the pilot transfection experiment. cDNAs were ligated to the 5′ untranslated region (UTR) of the PmRACK1 gene (Fig. 3A) to confirm proper and efficient translation (19). 5′-Capped mRNA was synthesized from the T3 promoter using an mRNA synthesis kit (Ambion, Austin, TX, USA). Immediately before electroporation (100 V, 250 μF) into 20 buds or zooids in 2-mm cuvettes, mRNA was diluted with HEPES-buffered saline (HBS; pH 7.2) to a concentration of 0.1 to 0.2 μg/μl. In the control, electroporation was performed in HBS without mRNA. RNAs for lipofection were diluted in 50 μl seawater at a final concentration of 0.2 μg/μl and mixed with the same volume of seawater containing 2 μl Lipofectamine 3000 (Invitrogen). Specimens were dissected and incubated in RNA-liposome solution for 60 min and then returned to natural seawater. Transfected animals were then cultured in natural seawater.

Antibodies.

A rabbit anti-Myc-tag polyclonal antibody (MBL; Nagoya, Aichi, Japan) was diluted 500-fold with PBS immediately before use. An alkaline phosphatase-labeled anti-rabbit secondary antibody (Vector Laboratories, Burlingame, CA, USA) was used after a 200-fold dilution. A mouse anti-His-tag monoclonal antibody (Qiagen, Düsseldorf, Germany) and anti-mouse IgG labeled with horseradish peroxidase (Vector Laboratory) were diluted 200-fold.

Recombinant proteins.

The coding regions of PmYY1, PmSirt6, and PmYaf2 (GenBank accession no. LC348390) having BamHI and SalI adaptors at the respective 5′ and 3′ ends were cloned into the pGEM-T vector (Promega, Madison, WI, USA). The respective cDNAs were digested with BamHI and SalI and ligated to the pQE30Xa vector (Qiagen). PmYaf2 was also ligated to the pGEX vector (Amersham Pharmacia Biotech, Uppsala, Sweden). GST-tagged PmYAF2 recombinant protein was induced in the bacterial BL21 strain with 0.1 mM isopropyl-β-d-thiogalactopyranoside (IPTG) at room temperature for 6 h. Recombinant 6× His-tagged YY1, SIRT6, and YAF2 were produced in the RB791 strain. Soluble proteins were extracted by sonication in PBS and purified by nickel bead or glutathione bead affinity chromatography followed by gel filtration high-performance liquid chromatography. Proteins were also synthesized in vitro by the wheat germ extract translation system according to the instruction manual (Promega).

Pulldown assay and Western blotting.

GST-tagged proteins and His-tagged proteins (2 μg/μl each) were mixed in 100 μl PBS containing glutathione beads and incubated at room temperature for 20 min. After washing three times with PBS, bound proteins were eluted for 60 min with 5 mM glutathione in 0.1 ml Tris-buffered saline (TBS; pH 8.0). Proteins were separated by SDS-containing 12.5% polyacrylamide gel electrophoresis (SDS-PAGE). After electrophoresis, proteins were blotted onto a nitrocellulose membrane and stained with antibodies.

Protein-protein and protein-DNA cross-linking.

Bis(sulfosuccinimidyl)suberate (BS3) (Dojindo Laboratories, Kumamoto, Japan) was dissolved in DMSO as a 100 mM stock solution and diluted 50-fold in 10 mM HEPES buffer (pH 7.5) containing 150 mM NaCl, 3 mM MgCl₂, and 1 mM dithiothreitol (DTT). Proteins (20 μg) were cross-linked in 100 μl of the reaction solution at 25°C for 0, 1, 2, 5, and 10 min. Regarding YY1-DNA cross-linking, 40 μg of protein was preincubated with 1 nmol oligonucleotides (probe 3) in 100 μl of PBS-EDTA containing 50 μg succinimidyl-[4-(psoralen-8-yloxy)]-butyrate (SPB; Pierce Biotechnology) on ice for 20 min, and the mixture was then irradiated with UV rays at 25°C for 0, 2, 5, 10, and 20 min.

Gel shift assay and native PAGE.

Double-stranded DNA probes were prepared by annealing the following complementary oligonucleotides: probe 1, 5′-ACATAGCCTACATTTTCACATCTAACTTGC-3′ and 5′-AAGCAAGTTAGATGTGAAAATGTAGGCTAT-3′; probe 2, 5′-TTCTCTGTTATATGATGTCCTCGGTGCTTC-3′ and 5′-CGGAAGCACCGAGGACATCATATAACAGAG-3′; and probe 3, 5′-TACTTGGGGGAATAATGGCGGCAGTGCAGC-3′ and 5′-TAGCTGCACTGCCGCCATTATTCCCCCAAG-3′.

The 5′ ends of both strands were labeled with Redivue [γ-³²P] ATP (GE Healthcare, Buckinghamshire, UK). Recombinant proteins of approximately 1 μg were incubated with the labeled probes in 20 μl of binding buffer containing 10 mM Tris-HCl (pH 7.5), 50 mM NaCl, 0.5 mM DTT, 4% glycerol, 0.5 mM EDTA, 1 mM MgCl₂, and 0.05 mg/ml of poly(dI-dC). After a 20-minute binding reaction, the reaction mixtures were fractionated on a 5% polyacrylamide gel. An autoradiogram was obtained using a BAS-2500 IP reader (Fujifilm, Tokyo, Japan). Native PAGE was performed with 4% polyacrylamide in Tris-borate-EDTA (TBE) buffer. Before electrophoresis, proteins were weakly prestained with 0.006% Coomassie brilliant blue (CBB) G250, a method modified from the protocol for blue native (BN) PAGE.

ChIP assay.

According to the cross-linking method for ChIP (39), approximately 2 × 10⁸ cells were fixed in 1% paraformaldehyde in sterile seawater for 5 min and neutralized with 200 mM glycine for 5 min. Protein G beads (Novex, Oslo, Norway; catalog no. 10003D) were pretreated with rabbit anti-Myc-tag antibody or anti-acetylated histone H3 lysine 9 (H3K9ac) antibody (Upstate, Millipore Corp., CA, USA). Cross-linked DNA and histones were separated from each other, and DNA was used as the template for PCR as described elsewhere (10, 34).

RT-PCR.

Poly(A)-positive [poly(A)⁺] RNA was extracted and purified using an mRNA isolation kit (Roche, Mannheim, Germany; catalog no. 11741985001). After the DNase I treatment, single-stranded DNA complementary to poly(A)⁺ RNA was synthesized at 55°C for 30 min using a Transcriptor First Strand cDNA synthesis kit (Roche; catalog no. 04379012001) and was then amplified by a PCR. Glycerol-3-phosphate dehydrogenase (G3PD) was used as the reference to normalize the expression of the other genes.

qPCR.

DNA templates for RT-qPCR were generated from approximately 0.2 μg of mRNA. qPCRs were executed using the StepOne Plus real-time PCR system (Applied Biosystems). The PCR solution (10 μl/tube) was composed of 1 volume of 8-fold diluted DNA (100-fold for input DNA of ChIP-qPCR), 5 volumes of the SYBR green PCR master mix (Applied Biosystems, Waltham, MA, USS), and 0.5 mM (each) the forward and reverse primers. Primer sets were designed to amplify short cDNAs (200 to 300 bp) without any additional bands. PCR was performed for 40 thermal cycles, denaturation at 95°C for 30 s, annealing at 60°C for 30 s, and extension at 72°C for 30 s. Data were analyzed by both the relative standard curve method and comparative threshold cycle (C_T) (ΔΔCT) method (32). Experiments were repeated five times (biological replicates; n = 5) unless otherwise stated. In each replicate, results were expressed as the mean ± standard deviations of six measurements (technical replicates; n = 6). P values were calculated by the two-tailed Student's t test.

SIRT6 inhibitor.

OSS-128167 (IC₅₀, 89 μM) was purchased from Selleck (Houston, TX, USA). It was dissolved in DMSO at a concentration of 50 mM. Immediately before use, the stock solution was diluted with sterile seawater.

Histone deacetylation.

Growing buds were transfected with PmSirt6 mRNA and allowed to develop for 2 days. Acetylated histone was collected from 40 developing buds using anti-H3K9ac, anti-H3K14ac, and anti-H3K27ac antibodies. PCR products prepared from untreated chromatin lysates were compared with those from the lysates treated with PmSIRT6. To evaluate in vitro the deacetylation activity of SIRT6 protein in collaboration with YY1 and YAF2, chromatin lysates were prepared from 50 developing buds in 1 ml of NP-40 buffer followed by brief sonication for 10 s. Lysates were treated with SIRT6 protein or SIRT6/YY1/YAF2 complex at a concentration of 0.2 μg/μl each at 37°C for 60 min in the presence of 1 mM DTT and 1 mM β-NAD. ChIP-qPCR was performed using the anti-H3K9ac antibody.