There are five SMYD homologs (SMYD1-5) in vertebrates, such as Danio rerio, Xenopus (Silurana) tropicalis, Gallus gallus, Mus musculus, Bos taurus, Bos mutu, Pan troglodytes, Pan paniscus, and Homo sapiens. The SMYD family was divided into two main clades (Fig. 1). One branch consisted of SMYD1, 2 and 3, in which SMYD1 clustered with SMYD2 first (78% ML bootstrap support) forming a sister group to SMYD3 (89% ML). The other branch contained SMYD4 and SMYD5 (29% ML). The monophyly of SMYD1-5 was maximally supported (100% ML). This result was consistent with the domain similarity of the SMYD family (Spellmon et al. 2015): (1) SMYD1-3 shared similar domains, including the N terminal SET domain, MYND domain, and the C-terminal tetratricopeptide repeats (TPR) (Jiang et al. 2011; Sirinupong et al. 2010, 2011); (2) the domain structure of SMYD4 and SMYD5 was more divergent (Spellmon et al. 2015); SMYD4 had an extra TPR domain on its N terminus and an extended C-terminal domain, while SMYD5 lacked the C-terminal domain. The sequence differences between SMYD4 and SMYD5 partially contributed to the low branch supporting value (29% ML).
In contrast to mammals, only one SMYD homologue is present in the genomes of many ciliates and plants (Spellmon et al. 2015). Ciliate SMYD and plant SMYD formed two well-separated clades (70% and 93% ML, respectively), indicating their diverged evolutionary paths from mammalian homologues. The clustering of ciliate SMYD also reflected the phylogenetic relationship of the constituent species which are divided between two fully supported sister branches: the oligohymenophoreans Paramecium tetraurelia and Tetrahymena thermophila and the hypotrichs Stylonychia lemnae and Oxytricha trifallax (Gao et al. 2016).
The expression levels of SMYD1 during the life cycle of Tetrahymena thermophila were evaluated by RT-PCR (Fig. 2a). As previously reported (Miao et al. 2009), SMYD1 is expressed during vegetative, starvation, and conjugation stages; in the present study, SMYD1 expression was highest during conjugation (Fig. 2a); however, we mainly focus on SMYD1's function at the vegetative stage. To detect its cellular localization, we initially generated a somatic SMYD1-CHA strain by introducing a short sequence encoding the hemagglutinin (HA) tag to the C-terminus of the SMYD1 gene, but we failed to detect any immunofluorescence signal, probably due to the low expression level of SMYD1 during the vegetative stage (Fig. 2a). To facilitate tracking SMYD1's distribution, an SMYD1 overexpression mutant was generated (SMYD1-CHA-overexpression) by placing SMYD1 and the C-terminal HA tag coding sequence under the cadmium-inducible MTT1 promoter (Fig. 2b). During the vegetative stage, SMYD1 was localized in both the cytoplasm and the macronucleus (Fig. 2c). As the confocal microscope scanned different layers of the Tetrahymena cell, SMYD1 showed different localization patterns: in layer 1, signals were mainly in the cytoplasm, and those in the macronucleus were weak; in layer 2, signals in the macronucleus became as strong as those in the cytoplasm; in layer 3, SMYD1 mainly appeared in the macronucleus and on the cell membrane. To remove the interference of SMYD1 signals from the cytoplasm and to reveal the presence of nuclear SMYD1, immunofluorescence staining was carried out on macronuclei purified from SMYD1-CHA-overexpression cells. Strong SMYD1 signals were detected throughout the transcriptionally active macronucleus and on the periphery of the transcriptionally inactive micronucleus (Fig. 2d), corroborating the localization of SMYD1 in the nucleus. The localization of SMYD1 in both the cytoplasm and the nucleus suggests that SMYD1 has the potential to methylate both histone and non-histone substrates (Al-Shar'i and Alnabulsi 2016; Calpena et al. 2015; Tracy et al. 2018).
Figure 2. Localization and expression of SMYD1 protein. a Gene expression profile for SMYD1 in wide-type cell. The different time points are shown along the X-axis. Lm stands for growing cells at the density of ~ 3.5 × 105 cells/ml. S3, S9 stands for starved cells (~ 2×105 cells/ml) collected at 3 h and 9 h after starvation, respectively. C3, C9 stands for conjugative cells (equal volumes of CU427 and CU428) collected at 3 h and 9 h after mixing, respectively. The normalized gene expression levels, as previously described (Cheng et al. 2016), are shown along the Y-axis. b Schematic diagram of SMYD1-CHA-overexpression (SMYD1-CHA-OE) plasmid. c SMYD1 signals in different layers of the same SMYD1-CHA-OE cell (vegetative stage) scanned by confocal microscopy. Arrows represent the micronucleus. Layers 1-3 indicate three different layers of the same cell. d SMYD1 signals within the nucleus obtained by nucleus purification. Arrows show the micronucleus. e pulse-labeling of SMYD1-CHA-OE (vegetative stage) cells. 0 h and 1 h stand for SMYD1 signals of cells collected at 0 h and 1 h after the removal of CdCl2, respectively. DAPI, nuclear signals; HA, signals of target protein; merge, combined signals of target protein and nucleus
To trace the dynamic change of SMYD1 protein, we performed a pulse-chase experiment by transiently inducing the expression of SMYD1 with CdCl2 (Fig. 2e). SMYD1 was initially detected in the cytoplasm immediately after the cadmium induction (0 h), suggesting that SMYD1 was present in the cytoplasm. Signals in the cytoplasm decayed, while the macronuclear signals increased 1 h later (1 h), demonstrating that SMYD1 can shuttle between the cytoplasm and the macronucleus.
To reveal SMYD1's functions in Tetrahymena, we generated SMYD1 knockout cells (∆SMYD1) and characterized the phenotype. To investigate the roles of SMYD1 in regulating gene transcription, we carried out RNA-seq analysis in wild type (CU428) and the isogenic ∆SMYD1 cells. In total, 21, 461 well-annotated genes were included, among which 4841 (23%) were up-regulated (> 2-fold) and 1507 (7%) were down-regulated (< 0.5-fold). KEGG pathway analysis (Table 1) revealed that the most affected pathways in ∆SMYD1 cells were a metabolic pathway (ko01100) and biosynthesis of secondary metabolites (ko01110), supporting the assertion that SMYD1 plays important roles in regulating metabolic genes in Tetrahymena. The result of Gene Ontology (GO) term enrichment analysis (Fig. 3) was consistent with that of the KEGG pathway analysis; this revealed the single organism metabolic process as the most enriched pathway for the up-regulated genes in ∆SMYD1 cells. These results suggest that SMYD1 is involved in regulation of metabolism.
Pathways associated with up-regulated genes ko01100 Metabolic pathways (98) ko01110 Biosynthesis of secondary metabolites (52) ko01130 Biosynthesis of antibiotics (30) ko01120 Microbial metabolism in diverse environments (25) ko01200 Carbon metabolism (20) ko01230 Biosynthesis of amino acids (17) ko00970 Aminoacyl-tRNA biosynthesis (16) ko00230 Purine metabolism (13) ko00010 Glycolysis/gluconeogenesis (12) ko04141 Protein processing in endoplasmic reticulum (12) Pathways associated with down-regulated genes ko03008 Ribosome biogenesis in eukaryotes (17) ko01100 Metabolic pathways (16) ko04142 Lysosome (7) ko01110 Biosynthesis of secondary metabolites (5) ko01130 Biosynthesis of antibiotics (4) ko00270 Cysteine and methionine metabolism (4) ko00230 Purine metabolism (3) ko04141 Protein processing in endoplasmic reticulum (3) ko01120 Microbial metabolism in diverse environments (3) ko04113 Meiosis-yeast (3) Numbers within () correspond to genes mapped onto this pathway
Table 1. KEGG pathway analysis for differentially expressed genes in ΔSMYD1 strain
Figure 3. GO (GO term: biological process) analysis of up-regulated genes in ∆SMYD1. Different biological processes are shown along the X-axis. Y-axis is the gene ratio. The higher the values, the closer the relationship between the SMYD1 protein and the target process
To examine if there was DNA replication deficiency in ∆SMYD1 cells, immunofluorescence staining was performed for two indicative markers in the DNA damage response (DDR) system-γH2A.X (phosphorylation form of H2A.X, indicator of double strand DNA breakage) and RPA1 (single strand DNA binding protein, indicator of single strand DNA accumulation) (Gao et al. 2013). There was an increase of γH2A.X levels in ∆SMYD1 cells, similar to the phenotype of the DNA replication deficient strain ∆TXR1 (Gao et al. 2013), indicating accumulation of double strand DNA breakage produced by abnormal DNA replication (Fig. 4a). RPA1 was slightly induced at mRNA (Fig. 4b) and protein (more RPA1 foci) (Fig. 4c, white arrows) levels, weaker than ∆TXR1 cells but stronger than WT cells. Moreover, genes significantly induced in ∆TXR1 cells were mostly up-regulated (though to a more moderate degree) in ∆SMYD1 cells, including many key players in ssDNA sensing/binding, DNA alkylation repair, nucleotide excision repair (NER), mismatch repair (MMR), homologous recombination (HR), and non-homologous end joining (NHEJ) (Fig. 4b). Taken together, these results argue that SMYD1 deletion resulted in mild replication stress.
Figure 4. Phenotype of ∆SMYD1 cells. a Accumulation of double strand DNA breakage (DSBs) in ∆SMYD1 cells. ∆TXR1, positive control; WT, negative control, wide type; γH2A.X., indicator of double strand DNA breakage, the phosphorylation form of the histone variant H2A.X. b Heat map of relative gene expression for DNA damage response-related genes in ∆SMYD1 cells, normalized against WT cells. ∆TXR1/WT, positive control, data from (Gao et al. 2013); NER, Nucleotide excision repair; MMR, mismatch repair; HR, homologous recombination; NHEJ, non-homologous end joining. Up-regulated genes are the reddest (∆SMYD1/WT > 2) down-regulated genes are the greenest (∆SMYD1/WT < 0.5), constant genes are represented by color between them. c Accumulation of single strand DNA in ∆SMYD1 strain. White arrows show RPA1 foci, indicating the accumulation of single strand DNA in the ∆SMYD1 strain. ∆TXR1, positive control; WT, negative control, wide type; RPA1, indicator of single strand DNA, single strand DNA binding protein; DAPI, nuclear signals; merge, combined signals of the target protein and nucleus. d Growth curve of ∆SMYD1 and WT strains. X-axis, different time points; Y-axis, cell number
It should be noted that no obvious growth defect was observed for ∆SMYD1 cells, since doubling time of ∆SMYD1 strain (Fig. 4d) showed little difference from that of WT cells (4.4 vs. 4.7 h). This suggested that SMYD1 was not essential for vegetative growth, or SMYD1 deficiency could be successfully coped with by activating the DNA damage responses (Fig. 4b).
To further study the mechanism underlying SMYD1 functions, immunoprecipitation (IP) of SMYD1-CHA-overexpression cells was carried out to assess the interactome of SMYD1 protein (Fig. 5a). Mass spectrometry analysis (Supplementary Table 2) revealed the heat shock protein 90 (HSP90: TTHERM_00080030) as a potential interacting partner with SMYD1, which is consistent with previous studies on mammalian systems (Donlin et al. 2012; Sima and Richter 2018; Spellmon et al. 2017). This mass spectrometry result was corroborated by the silver staining of the IP sample (Fig. 5a), which revealed a ~ 98 kDa band corresponding to the predicted size of HSP90 (blue arrow in Fig. 5a), as well as the bait protein (SMYD1) band (red arrow in Fig. 5a, ~ 53 kDa).
Figure 5. Interactome analysis of SMYD1 protein. a Silver staining of the SMYD1-CHA-OE IP sample. Red arrow indicates SMYD1 protein (53 kDa); blue arrow indicates the potential HSP90 protein. WT, wild type (negative control). b Purification of proteins interacting with SMYD1 using gel filtration chromatography. IP sample (without TCA precipitation) of SMYD1-CHA-OE strain was used as input. 21-32, different sized protein complex in different collection tubes: 24-25, protein complex of ~ 150 kDa; 29-32, bait protein of ~ 53 kDa (SMYD1). c Western blot of IP samples for HSP90-Nflag/SMYD1-CHA-overexpression strains with or without Cd2+ induction. SMYD1-CHA-overexpression strain with Cd2+ induction was used as negative control. d Co-localization of HSP90 and SMYD1 in Cd2+-induced HSP90-Nflag/SMYD1-CHA-overexpression strain. DAPI, nuclear signals; HSP90, HSP90 protein signals, detected by the α-Flag antibody; SMYD1, SMYD1 protein signals, detected by the α-HA antibody; Merge, combined signals of target proteins and nucleus
To further confirm the interaction between HSP90 and SMYD1, gel filtration chromatography was used to separate proteins in the IP sample of SMYD1-CHA-overexpression cells (Fig. 5b). As a control, the bait protein band (SMYD1) was enriched in fractions 29-32 (calculated size is ~ 53 kDa). It was also enriched in fractions 24 and 25 (calculated size ~ 150 kDa), corresponding to the predicted size of the SMYD1-HSP90 complex. This result raised the possibility of direct interaction between SMYD1 and HSP90. To support this, we introduced a flag tag to the N-terminus of HSP90 in SMYD1-CHA-overexpression cells (HSP90_Nflag/SMYD1-CHA-OE) and carried out the co-precipitation of SMYD1 and HSP90 (Fig. 5c). By performing immunoprecipitation with anti-Flag M2 beads, we detected SMYD1 in the IP sample by the anti-HA antibody at the expected size (~ 53 kDa) (Fig. 5c lane 2). As negative controls, no SMYD1 band was detected without cadmium induction (Fig. 5c, lanes 3 and 4) or in IP samples from cells without flag-tagged HSP90 (Fig. 5c, lanes 5 and 6). These results support the interaction between SMYD1 and HSP90.
To investigate whether SMYD1 can methylate HSP90, SMYD1 in vitro methyltransferase assay was performed. We only detected a ~ 53 kDa radioactive band (Supplementary Fig. 1, red arrow), representing either the self-methylated SMYD1 or SMYD1 binding of H3-labeled S-adenosylmethionine (SAM). Though we failed to detect any radioactive band of HSP90, the ~ 53 kDa radioactive band provides evidence that SMYD1 protein possesses the methyltransferase activity.
Phylogenetic analysis of SMYD family methyltransferases
Localization and expression of SMYD protein in Tetrahymena thermophila
Phenotypic analysis of ∆SMYD1 cells
Interactome of SMYD1 protein
A total of 47 SMYD amino acid sequences (Table 2) of representative eukaryotic species were downloaded from the National Center for Biotechnology Information (NCBI) Database (https://www.ncbi.nlm.nih.gov/). Sequences were aligned by MUSCLE 3.7 (Edgar 2004), provided on the web server "Phylogeny.fr Robust Phylogenetic Analysis For The Non-Specialist" (http://phylogeny.lirmm.fr/phylo_cgi/one_task.cgi?task_type=muscle). The alignment was used for the subsequent phylogenetic tree construction.
Species name and protein name GenBank accession no. Species name and protein name GenBank accession no. Danio rerio SMYD1 NP_001034725 Gallus gallus SMYD4 NP_001025886 Xenopus (Silurana) tropicalis SMYD1 XP_012811600 Mus musculus SMYD4 AAH95952 Gallus gallus SMYD1 NP_989486 Bos taurus SMYD4 XP_005220153 Mus musculus SMYD1 NP_001153599 Pan paniscus SMYD4 XP_003816904 Bos taurus SMYD1 DAA24603 Homo sapiens SMYD4 NP_443160 Pan paniscus SMYD1 XP_003805903 Drosophila virilis SMYD5 XP_002053397 Homo sapiens SMYD1 NP_938015 Branchiostoma floridae SMYD5 XP_002609030 Danio rerio SMYD2 NP_001013568 Xenopus laevis SMYD5 NP_001085635 Xenopus (Silurana) tropicalis SMYD2 XP_00293475NP_9380151 Danio rerio SMYD5 NP_001004614 Gallus gallus SMYD2 NP_001264500 Gallus gallus SMYD5 NP_001012912 Mus musculus SMYD2 EDL13024 Bos taurus SMYD5 NP_001073717 Bos mutus SMYD2 NP_001069832 Mus musculus SMYD5 NP_659167 Pan troglodytes SMYD2 XP_003308794 Pan paniscus SMYD5 XP_003808242 Homo sapiens SMYD2 NP_064582 Homo sapiens SMYD5 NP_006053 Danio rerio SMYD3 NP_001032477 Stylonychia lemnae CDW88943 Xenopus (Silurana) tropicalis SMYD3 XP_004914684 Oxytricha trifallax EJY75465 Gallus gallus SMYD3 XP_015139481 Paramecium tetraurelia XP_001437474 Bos mutus SMYD3 XP_005216902 Tetrahymena thermophila XP_001022867 Mus musculus SMYD3 NP_081464 Volvox carteri f. nagariensis XP_002956692 Pan paniscus SMYD3 XP_514316 Physcomitrella patens XP_001778213 Homo sapiens SMYD3 NP_001161212 Oryza sativa AAS07242 Branchiostoma floridae SMYD4 XP_002589088 Arabidopsis lyrata subsp. lyrata XP_002886136
Table 2. Accession numbers of species used in the phylogenetic tree
A Maximum-Likelihood (ML) tree was constructed with RAxML-HPC2 on XSEDE v 7.2.8 (Stamatakis 2006; Stamatakis et al. 2008), provided by the CIPRES Science Gateway (Miller et al. 2010), using plant species, Volvox carteri f. nagariensis, Physcomitrella patens, Oryza sativa, Arabidopsis lyrata subsp. lyrata and Glycine max, as the outgroup. The MTART model of Protein Substitution Matrix selected by ProtTest 3 (Darriba et al. 2011) and other defaulted parameters were used for the maximum-likelihood (ML) analysis. The robustness of internal branches was estimated by 1000 bootstrap replicates.
The wide-type Tetrahymena thermophila strain CU428 (provided by Tetrahymena Stock Center, Cornell University, Ithaca, NY), from which all mutant strains were derived, was cultured in SPP medium (Orias et al. 1999) at 30 ℃. Cells in mid-exponential growth phase (~ 2×105 cells/ml) were used for subsequent experiments.
In the current study, 6 constructs, including ∆SMYD1, SMYD1-CHA, ∆SMYD1-RPA1-CHA, SMYD1-CHA-overexpression, and HSP90-Nflag were generated, and all primers used here are listed in Table 3. Construct RPA1-CHA was generated as previously described (Gao et al. 2013). Constructs ∆SMYD1 and SMYD1-CHA were generated according to previous studies (Feng et al. 2017; Liu et al. 2007; Noto et al. 2015). For the SMYD1-CHA-overexpression construct, the target fragments were cloned into the newly constructed CHA-overexpression vectors. CHA-overexpression vectors were generated based on the inducible MTT1 and MTT3 promotors (Cd2+ inducible) to investigate proteins of low expression level. Primers used are shown in Supplementary Table 1. The MTT1-MTT3 region, including the 3′ and 5′ untranslated regions (5.4 kb in total), were amplified with Platinum Tag DNA Polymerase (Invitrogen, 11304-011) and cloned into pBlueScript SK (-) vector. The MTT3 locus was replaced with neo4 coding region, providing the paromomycin resistance for cells. The MTT1 locus was replaced by Flag-HA tag with Sbf I cutting site on its N terminus.
Name Sequence SMYD1_5f715_NotI AGTTCTAGAGCGGCCGCGATTATTCGCCTATAGTTGATGG SMYD1_3r5096_Not I ACCGCGGTGGCGGCCGCGTTTCCTACTCAGTCTCTTGC SMYD1_Nf1849_GHA CCCTACGACGTCCCCGACTACGCCTACTAGGTAGAAAATCTTAATAATTAGTATCG SMYD1_Nr1848_GHA GTCGGGGACGTCGTAGGGGTATCCCATATATATCTTTGATTTTCTAAATTAATTG SMYD1_Cf3335_GHA CCCTACGACGTCCCCGACTACGCCTGATATTCTTTAAAATAAAATAAAAAAAAG SMYD1_Cr3334_GHA GTCGGGGACGTCGTAGGGGTATCCATTATATTTCATTTTTATCTCAC SMYD1_3f3956_neo4 CTGACGTCGCACCATCCCGTTGTTGCATAGGATTGTTTTCG SMYD1_3r3928_neo4 GTCAGGTGCCTGGTACCCGTTTATTTAAAAAGCAGTAGC SMYD1_f2734 CTTTTTGTGTGAATGTAAAAGGTG SMYD1_r2862 CCTTAAGTGTTACAGTCAGAGC α-Tubulin_f TCAGTAACCTTCTTCTTCACC α-Tubulin_r CACTGGTTTCAAGGTCGGTAT MTT1_SMYD1_f1724_NHA CGTCCCCGACTACGCCTACTAGGTAGAAAATCTTAATAATTAG MTT1_SMYD1_r3241_NHA CATATTTATTTCACCTATTATATTTCATTTTTATCTCACTTTTTATATC MTT1_SMYD1_f1698_CHA CTTAAAATAATGGATCCTTACTAGGTAGAAAATCTTAATAATTAG MTT1_SMYD1_r3217_CHA CGTCGTAGGGGTATCCATTATATTTCATTTTTATCTCACTTTTTATATC HSP90_5f1223_Not I AGTTCTAGAGCGGCCGCATCAAAGTATGAAGAAGACAGG HSP90_3r6007_Not I ACCGCGGTGGCGGCCGCTAATCAAATAAATCTCTCTGTTCTG HSP90_Nf2158_Flag GGAGACTACAAGGACGACGATGACAAGTCTCAACAAGCTGAACACTTTGC HSP90_Nr2157_Flag GTCATCGTCGTCCTTGTAGTCTCCCATTTCTTATGATATATCTTTTTTTTTAAT HSP90_3f5043_BSR CTGACGTCGCACCATCCCGTGAAGTTTTTTGATATTATCACAC HSP90_3r5004_BSR GTCAGGTGCCTGGTACCCACTTTTATATCAGTGAAAATGGAG Italic characters represent the adaptors
Table 3. Primers used for plasmid construction and RT-PCR
All of the above constructs except RPA1-CHA and HSP90-Nflag were introduced into CU428 by standard biolistic transformations (Cassidy-Hanley et al. 1997). RPA1-CHA was introduced into ∆SMYD1 strain, and HSP90-Nflag into SMYD1-CHA-overexpression strain, respectively. Paromomycin, Cycloheximide or Blasticidin was used for subsequent transformant selection according to the drug cassette. Complete somatic replacement was validated by quantitative-PCR as previously reported (Zhao et al. 2017).
SMYD1-CHA-overexpression stain was cultured overnight (~ 18 h) in 1L 1 × SPP containing 0.5 μg/ml CdCl2. Mid-log-phase (~ 2×105 cells/ml) cells were collected and the macronucleus purification was carried out as described (Chen et al. 2016). The purified macronuclei were washed with nuclear wash buffer (50 mmol/L pH 7.4 Tris, 2 mmol/L MgCl2) once and resuspended in 200 μl nuclear wash buffer for subsequent immunofluorescence staining.
A volume of 15 ml of target cells in mid-exponential growth phase was collected and fixed in 2% paraformaldehyde (diluted with 1 × PBS). Permeabilization was then accomplished with 0.4% Triton X-100 (diluted with 1 × PBS), after which antibodies (details in Table 4) were incubated with cells (Gao et al. 2013; Liu et al. 2007). Digital images were captured using an Olympus BX43 microscope and an Olympus DP73 camera.
Antibody dilution Incubation condition Primary antibody α-HA (Rabbit monoclonal, Cell Signaling, 3724S) 1:2000 4 ℃, overnight α-γH2A.X(Mouse monoclonal, Millipore, 05636) 1:5000 RT, 2 h α-Flag (Mouse monoclonal, Sigma, F1804) 1:5000 4 ℃, overnight Secondary antibody Goat Anti Mouse 555 IgG (Invitrogen, A32727) 1:5000 RT, 1 h Goat Anti Rabbit 555 IgG (Invitrogen, A27017) 1:5000 RT, 1 h Goat Anti Rabbit 488 IgG (Invitrogen, A32731) 1:5000 RT, 1 h
Table 4. Antibodies for immunofluorescence staining
SMYD1-CHA-overexpression cells were cultured overnight in 800 ml 1 × SPP containing 0.5 μg/ml CdCl2. Cells at mid-log phase were collected by centrifugation and immunoprecipitation (IP, details shown in Supplementary Methods) was carried out, after which the IP sample was sent to Proteomics Resource Facility (Department of Pathology, University of Michigan) for LC-MC analysis.
The HSP90-Nflag/SMYD1-CHA-overexpression strain was cultured overnight in 800 ml 1 × SPP with or without 0.5 μg/ml CdCl2. At the same time, the SMYD1-CHA-overexpression strain was cultured overnight in 800 ml 1 × SPP as a negative control. IP was carried out on these strains as described above. IP samples were used for western blotting with primary antibodies α-Flag (Mouse monoclonal, Sigma, F1804, 1:5000) and α-HA (Rabbit monoclonal, Cell Signaling, 3724S, 1:2000).
An IP sample (without TCA precipitation) of SMYD1-CHA-overexpression strain was used for the in vitro methyltransferase activity test. Immunoprecipitation was modified from the protocol outlined in the Supplementary Methods, in which T0 buffer (30 mmol/L Tris HCl, 30 mmol/L Tris Base, 20 mmol/L KCl and 2 mmol/L MgCl2) was used instead of T150 buffer (30 mmol/L Tris HCl, 30 mmol/L Tris Base, 20 mmol/L KCl, 2 mmol/L MgCl2 and 150 mmol/L NaCl) and HA elution (eluted with 250 μg/ml HA peptide at RT for 15 min) was used as a final product for the subsequent methyltransferase activity test.
The in vitro methyltransferase activity test was modified from previous studies (Wu et al. 2013; Zhou et al. 2016). For each test, H3-labeled S-adenosylmethionine (SAM, final concentration 10 mmol/L, from Perkin Elmer, NET155H250UC) was added to 20 μl IP sample (without TCA precipitation, in T0 buffer) or 20 μl T0 buffer (negative control). The reaction was carried out at 25 ℃ overnight, after which the methylation was detected by autoradiography.
A volume of 10 ml of cells was collected at indicated time points. TRIzol™ Reagent (Invitrogen, 15596026) was used to extract the total RNA, after which DNA was removed with the Turbo DNA-free kit (Ambion, AM 1907). Complementary DNA (cDNA) was reverse-transcribed using Superscript Ⅲ Reverse Transcriptase kit (Invitrogen, 18080-051) with parameters as follows: 50 ℃ for 50 min, 85 ℃ for 5 min, and maintained at 4 ℃.
RT-PCR, with cDNA as template, was carried out on a CFX96™ Real-Time System (BIO-RAD, USA) with the Radiant™ Green Lo-Rox qPCR Kit (Alkali Scientific, QS1020). The reaction was carried out as previously described (Gao et al. 2013). Each reaction was performed in duplicate using primers SMYD1_f2734 and SMYD1_r2862 (Table 3). α-Tubulin_f and α-Tubulin_r (Table 3) were used as internal controls (Cheng et al. 2016). The 2-ΔΔCt method was used to analyze the real-time PCR data.
Total RNA of log-phase ∆SMYD1 and wild-type CU428 cells (negative control) were extracted using Qiashredder (Qiagen, 79654) and RNeasy Kit (Qiagen, 74624) according to the protocol provided in TetraFGD (Xiong et al. 2013). The Qubit RNA Assay Kit in Qubit 2.0 Flurometer (Life Technologies, CA, USA, Q32852) and the RNA Nano 6000 Assay Kit of the Agilent Bioanalyzer 2100 system (Agilent Technologies, CA, USA, 5067-1511) were used to measure the RNA concentration and integrity, respectively.
In total, 3 μg RNA was used for sample preparation. The sequencing library was produced by NEBNext Ultra RNA Library Prep Kit for Illumina (New England Biolabs, MA, USA, E7530L) according to manufacturer's recommendations, and details are as previously described (Zhang et al. 2015). The library was sequenced on the Illumina Hiseq 2500 platform provided by Novogene Bioinformatics Institute (Beijing, China), and 125 bp paired-end reads were generated.
FASTX-Toolkit (Gordon and Hannon 2010) was used to remove adapters and reads of low quality from the raw data, after which the remaining reads were mapped onto the Tetrahymena thermophila genome (http://ciliate.org/index.php/home/downloads, June 2014). Gene expression levels were calculated by RSEM v1.2.7 (Li and Dewey 2011). A heat map of genes differentially expressed between ∆SMYD1 and CU428 was generated with MultiExperiment Viewer (v4.9) (Mar et al. 2011). The KEGG (Kyoto Encyclopedia of Genes and Genomes) pathway analysis was carried out using the KAAS server (http://www.genome.jp/kegg/kaas/). Gene Ontology (GO) term enrichment analysis on the up-regulated genes was performed using BiNGO v3.0.3 (p.adjust < 0.05), which was integrated in Cytoscape v3.4.0, and the plot was generated by the R package, ggplot2 (Kohl et al. 2011; Maere et al. 2005; Wickham 2016).
The IP sample (without TCA precipitation) of the SMYD1-CHA-overexpression strain was used as input of the purification. Purification was carried out by the AKTApurifier HPLC system (including Pump P-900 and Superose 6 column, GE Healthcare) according to previous studies (Hubert et al. 2014; Hughes et al. 2013). Firstly, filtered distilled water was pumped into the system to remove any contamination and bubbles, after which filtered T0 buffer was pumped into the system as loading buffer. Secondly, formula between the Retention Volume and Molecular Weight was calculated using two marker proteins, BSA (66 kDa) and WDR5 (36 kDa); that is Rv = (2.9661-logMw)/0.073, where Rv is Retention Volume (ml) and Mw is Molecular Weight (kDa). Thirdly, 300 μl IP samples of the SMYD1-CHA-overexpression strain were added into the system by 1 ml syringe and run at a flow rate of 0.5 ml/min to purify proteins of different sizes. Samples (10 μl) from each fraction (500 μl/tube) were retained for western blot analysis.
The SMYD1-CHA-overexpression strain was cultured in 60 ml 1 × SPP at 30 ℃ overnight. The mid-log-phase cells were pulse-labeled by adding CdCl2 (final concentration 1.5 μg/ml) into the medium. Two hours later, CdCl2 was removed by washing with 1 × SPP twice. Cells were resuspended in 60 ml fresh 1 × SPP and collected at 1 h intervals during the chase.