|Home | About | Journals | Submit | Contact Us | Français|
S-phase and DNA damage promote increased ribonucleotide reductase (RNR) activity. Translation of RNR1 has been linked to the wobble uridine modifying enzyme tRNA methyltransferase 9 (Trm9). We predicted that changes in tRNA modification would translationally regulate RNR1 after DNA damage to promote cell cycle progression. In support, we demonstrate that the Trm9-dependent tRNA modification 5-methoxycarbonylmethyluridine (mcm⁵U) is increased in hydroxyurea (HU)-induced S-phase cells, relative to G₁ and G₂, and that mcm⁵U is one of 16 tRNA modifications whose levels oscillate during the cell cycle. Codon-reporter data matches the mcm⁵U increase to Trm9 and the efficient translation of AGA codons and RNR1. Further, we show that in trm9Δ cells reduced Rnr1 protein levels cause delayed transition into S-phase after damage. Codon re-engineering of RNR1 increased the number of trm9Δ cells that have transitioned into S-phase 1 h after DNA damage and that have increased Rnr1 protein levels, similar to that of wild-type cells expressing native RNR1. Our data supports a model in which codon usage and tRNA modification are regulatory components of the DNA damage response, with both playing vital roles in cell cycle progression.
Saccharomyces cerevisiae tRNA methyltransferase 9 (Trm9) methylates wobble uridines to facilitate the synthesis of 5-methoxycarbonylmethyluridine (mcm5U) and 5-methoxycarbonylmethyl-2-thiouridine (mcm5s2U) in specific tRNAs encoding arginine, lysine, glutamine and glutamic acid.1-3 Trm9-catalyzed tRNA modifications have been implicated in differentiating between cognate and near cognate codons in mixed codon boxes and optimizing codon-anticodon interactions.1,4 Reporter studies support that Trm9-catalyzed tRNA modifications enhance binding to codons ending in A in arginine and glutamic acid mixed codon boxes to increase the speed of translation in a codon-dependent context.3 In vivo, the influence of Trm9-dependent tRNA modifications on translation has been demonstrated for components of the ribonucleotide reductase (RNR) complex, with Rnr1 and Rnr3 protein deficiencies noted in asynchronous trm9Δ cells.3 High-throughput and targeted studies have also demonstrated that trm9Δ cells are sensitive to killing by ionizing radiation and alkylating agents.3,5-7 When taken together, the reported work on Trm9 and corresponding mutants demonstrates that there is a link between tRNA modifications, translation and the DNA damage response.
The RNR complex plays an important role in regulating cell cycle transitions and the DNA damage response. The RNR complex has four protein subunits: Rnr1, Rnr2, Rnr3 and Rnr4. It catalyzes the rate-limiting step in the biosynthesis of deoxyribonucleotides (dNTPs) from ribonucleotides,8 with the concentration of dNTP pools limiting for DNA replication and DNA repair. During DNA replication, dNTPs need to be replenished;9,10 thus, high RNR activity is required to complete S-phase. During conditions of DNA damage, DNA replication requires a further increase in dNTP levels to promote efficient repair synthesis.11,12 The survival of S. cerevisiae in response to DNA damaging agents has been previously shown to depend on increased dNTP levels.13 Hence, factors that reduce dNTP levels, by reducing Rnr1, Rnr2 and Rnr4 levels or RNR activity, render cells hypersensitive to DNA damaging agents as well as to the RNR inhibitor hydroxyurea (HU).14 DNA synthesis is highest during S-phase and is lowest during G1-phase; hence, RNR activity is tightly coupled with phases of the cell cycle.15-18 RNR activity is restricted in the G1-phase of the cell cycle, and as the cell transitions from G1 to S-phase,19,20 there is a sudden surge in RNR activity triggered by the phosphorylation-mediated degradation of Sml1, subcellular re-localization of subunits and increased transcription of RNR1 and RNR3.8,21-24
In S. cerevisiae, 5-methylcytosine (m5C) tRNA modifications are increased in response to oxidative stress,25,26 suggesting a dynamic role for the tRNA wobble position in regulating protein synthesis. Changes in the levels of many other tRNA modifications occur under conditions of increased reactive oxygen species, which include Cm, m22G, mcm5U, Um and Am modifications,25 suggesting that global tRNA reprogramming is part of the cellular stress response. The observed increase in m5C and mcm5U modifications has provided strong support for the theory that the wobble position of the anticodon plays a regulatory role during stress responses. Cell cycle oscillations are tightly coupled to efficient stress and DNA damage responses; consequently, timely cell cycle progression can promote efficient responses to stress. We reasoned that increased tRNA modification and translation should also have roles in the surge in RNR activity required for cells to transition from G1 to S, which should be especially prominent under conditions of DNA damage.
In the following S. cerevisiae-based study, we demonstrate that the levels of 16 tRNA modifications oscillate during the course of the cell cycle, with a distinct collection occurring in HU-induced S-phase. HU and MMS activate the intra-S checkpoint and increase the transcription of RNR genes to promote the DNA damage response.27-29 We show that mcm5U tRNA modification increases in HU-induced S-phase, relative to G1 and G2, to promote the codon-dependent translation of RNR1 and boost RNR activity. Further, we demonstrate that mutants lacking Trm9 display a DNA damage-induced cell cycle phenotype characterized by delayed transition from G1- to S-phase, relative to wild-type, in response to DNA damage. Notably, we also show that a change in RNR1 codon usage can rescue the damage-induced cell cycle phenotype associated with trm9Δ cells and increase Rnr1 protein levels in trm9Δ cells. Ultimately, our study demonstrates that RNR1 is translationally regulated by the Trm9-dependent tRNA modification mcm5U. Our study also highlights new regulatory roles for gene-specific codon usage patterns and oscillations in tRNA modifications during the cell cycle.
We synchronized cells in G1-phase using α-factor, in S-phase using hydroxyurea and in G2-phase using nocadazole, we also treated cells with MMS for 2 h to align them in S-phase. We then measured Rnr1 protein levels (Fig. 1A) using immunoblots. Relative to wild-type, Rnr1 protein levels were lower in trm9Δ cells synchronized in G1, S, G2 and after treatment with MMS. The difference between Rnr1 levels in wild-type and trm9Δ cells was most extreme in HU-induced S-phase. RNR1 is known to be transcriptionally induced in S-phase,21 which helps account for the modest G1 to S increase in Rnr1 protein levels in trm9Δ cells. Previously, we have used northern blots and polysome profiles to demonstrate that the transcription of RNR1 is similar in wild-type and trm9Δ cells and that translation of RNR1 is perturbed in asynchronous trm9Δ cells.3 Combined, our data support the idea that RNR1 translation and tRNA modifications are regulated differently during the cell cycle.
To analyze tRNA modification levels as a function of the cell cycle, wild-type cells were grown to mid-log phase and synchronized in G1, S and G2, as well as treated with MMS for 2 h (Fig. 1B). Small RNA species were isolated and tRNA modifications were quantified by LC-MS/MS analysis (Fig. 1C; Table S1A‒C). RNA extracted from cells treated with MMS for 2 h was heavily degraded and not suitable for further analysis. The quantitative tRNA modification data specific to G1-, HU-induced S- and G2-phases was analyzed relative to asynchronous cells, clustered and viewed as a heat map to identify modifications that change as a function of cell cycle stage (Fig. 1C). We have identified 12, 2 and 9 significant (p < 0.05) changes in tRNA modification levels in G1-, S-, and G2-phases of the cell cycle, respectively. In both G2 and G1, we observed significant decreases in I, m5U, m5C, Cm, m7G, m1I and m1A modifications relative to asynchronous cells. Modifications to positions 34 (wobble base) and 37 (adjacent to the anticodon) of tRNA are well represented in this list and include m5C, Cm, and m1I. We were very interested in the S-phase data, as efficient DNA synthesis after DNA damage requires optimal Rnr1 protein levels in this phase, and we predicted that tRNA modification levels would be distinct. We observed a significant ~2-fold increase (p < 0.003) in mcm5U modifications during S-phase, relative to asynchronous, G2 or G1 cells (Fig. 1D), with an ~2-fold reduction (p < 0.04) in the mcm5s2U modification. This could suggest that dethiolation of mcm5s2U is driving the increase in mcm5U levels; however, other explanations are also possible. These include increased transcription or selective degradation of specific tRNAs. Nevertheless, our results support the notion that increased mcm5U wobble bases, which are Trm9-dependent, help promote the translation of RNR1 to provide the burst in RNR activity needed to support DNA synthesis in S-phase after DNA damage.
If mcm5U modifications increase the translation of RNR1 in S-phase after DNA damage, then the loss of mcm5U should lead to a DNA damage-induced cell cycle phenotype in trm9Δ cells. We tested our hypothesis using wild-type, trm9Δ and trm9Δ + TRM9 cells treated with MMS, followed by propidium iodide (PI)-based staining of DNA content and fluorescent activated cell sorting (FACS) to characterize cell cycle populations. We observed that 95% of the wild-type cells had progressed into S-phase 2.5 h after treatment with 0.04% MMS, whereas 25% of trm9Δ cells remained in G1-phase (Fig. 2). We were able to alleviate the MMS-induced cell cycle phenotype in trm9Δ cells by re-expressing TRM9. Similar experiments and results were also performed and obtained using ionizing radiation and HU. We observed that 99% of the wild-type cells were outside of G1-phase 3 h after exposure to 200 Gy of IR, whereas 20% of the trm9Δ cells remained in G1-phase (Fig. S1). After HU treatment of trm9Δ cells, we also observed a delayed transition into S-phase (Fig. S2), a phenotype similar to what we have described for MMS.
Our cell cycle results support the conclusion that trm9Δ cells transition slowly into S-phase after DNA damage relative to wild-type. To further test our prediction, we performed cell synchronization experiments. First, wild-type and trm9Δ cells were synchronized in G1-phase with α factor treatment, washed and then cultured in YPD or YPD containing MMS. We observed that ~80% of wild-type cells transition into S-phase 30 min after being released into MMS (Fig. 3). In contrast, only ~50% of the G1 synchronized trm9Δ cells transition into S-phase by 30 min. Notably it took 3 h after release into MMS for ~80% of the trm9Δ cells to have transitioned into S-phase. Cumulatively, our cell cycle results provide evidence that there is a DNA damage-induced cell cycle phenotype (i.e., delayed transition into S-phase) in trm9Δ cells.
The RNR1 transcript is upregulated in S-phase relative to G1 and after exposure to HU or MMS.21 Increased RNR activity is vital for S-phase and an efficient DNA damage response. We have previously demonstrated that poor translation of RNR1 mRNA in asynchronous trm9 mutants leads to lower Rnr1 levels.3 Hence, we reasoned that reduced RNR activity, via an Rnr1 deficiency, is responsible for our observed cell cycle phenotype in trm9Δ cells after DNA damage. We expressed RNR1 from the ADH1 promoter in both wild-type and trm9Δ cells. In theory, this could partially override the translational defect associated with deficiencies in Trm9 and provide higher levels of the RNR1 transcript and protein in both wild-type and trm9Δ cells. Each of the cell types was treated with MMS, and a cell cycle profile was determined over 3 h. We observed that overexpression of RNR1 transcripts partially alleviated the MMS-induced cell cycle phenotype and increased the number of trm9Δ cells that transitioned into S-phase (Fig. 4).
Mec1-Rad53-Dun1-dependent phosphorylation of Sml1 will promote degradation of Sml1 to provide a rapid increase in dNTP levels, with this occurring during S-phase or and after DNA damage.22 Deletion of SML1 has been reported to suppress the DNA damage phenotype associated with the decreased RNR subunit transcription.30 In general, lower Sml1 protein levels will increase RNR activity and lead to increased dNTP levels and resistance to DNA damage-induced cell death.22 We deleted SML1 in trm9Δ cells and observed increased numbers of cells outside of G1-phase in sml1Δ-trm9Δ cells after IR or MMS damage (Fig. 5; Fig. S3), when compared with trm9Δ cells. Results from both our RNR1-overexpression and sml1Δ experiments support the idea that low Rnr1 levels and reduced RNR activity in trm9Δ cells results in a delayed transition into S-phase after DNA damage.
Our data on tRNA modifications, Rnr1 protein levels and cell cycle progression support the idea that increased mcm5U levels promote the translation of RNR1 transcripts in S-phase during the DNA damage response. The RNR1 transcript has a distinct codon usage, characterized by over usage of AGA and GAA codons.3 We suggest that the increase in mcm5U tRNA modification affects translation in a codon- and thus gene-dependent fashion. To test this model, we used a dual reporter system to analyze the influence of Trm9-catalyzed modifications on the translation of 9 different codon runs. The dual reporter consists of Renilla luciferase connected in frame by an intervening 12 base pair sequence (5′-CCCGGGGAGCTC-3′) to Firefly luciferase.31 Both Renilla and Firefly luciferase are transcribed on the same open reading frame, and efficient translation of Firefly luciferase requires efficient translation of the 12 base pair sequence (i.e., 4 codon) linker region. The resulting Renilla-linker-Firefly activities are, in theory, a single protein with Renilla luciferase activity serving as an internal control. The dual luciferase construct is under the control of an ADH1 promoter and CYC1 terminator (cloned in pJD375 plasmid). The Firefly/Renilla ratio is an indicator of the translational efficiency across the intervening 12 base pair (4 codon) sequence. In the test reporter plasmids, the 12 base pair intervening sequence was replaced by either four AGA, AGG, AAA, AAG, GAA, GAG, CAA, CAG or GGA codons in a row (designated a 4X codon run), and the reporters were individually transformed into the wild-type and trm9Δ cells. Each of the cells was grown to log phase, and the Firefly/Renilla ratio was determined (Fig. 6A). We observed that Trm9 significantly enhances the levels of Firefly activity associated with 4XAGA, 4XGAA and 4XCAA codon runs. In theory, the 4XAGA corresponds to mcm5U in tRNA-Arg; 4XGAA corresponds to mcm5s2U in tRNA-Glu; and 4XCAA corresponds to mcm5s2U in tRNA-Gln. The 4XAGA, 4XGAA and 4XCAA results support the idea that Trm9-catalyzed tRNA modifications are required for the optimal translation of specific codons. In trm9Δ cells, we observed significantly higher Firefly activity specific to the 4XGAG and 4XTTG reporters, corresponding to glutamic acid and leucine, respectively. The increased reporter activity suggests that GAG and TTG codons were translated more efficiently in the absence of Trm9-catalyzed modifications.
AGA and TTG codons are specific to amino acids encoded by six codons, with four of these codons belonging to a family codon box. The native RNR1 gene has many of these individual codons (18 AGA and 22 TTG), and we hypothesized that a codon-optimized RNR1 gene that increased Rnr1 protein levels could rescue the DNA damage-induced trm9Δ cell cycle phenotype. To test this hypothesis, we used a codon optimized RNR1 gene, with the goals of providing the most translatable unit and, where possible, removing codons that are regulated by Trm9-catalyzed tRNA modifications. We have previously used this codon optimized RNR1 gene (Opt-RNR1),4 to support the idea that native RNR1 transcript is susceptible to translational infidelity. Opt-RNR1 uses all high-usage codons and has all AGA and TTG codons converted to CGT and CTA, respectively. Opt-Rnr1 replaced the endogenous RNR1 at its native locus in both wild-type and trm9Δ mutants. We studied the cell cycle progression of wild-type and trm9Δ cells having either native or codon-optimized RNR1 genes (Fig. 6B; Fig. S4). We observed that trm9Δ cells with the codon-optimized RNR1 had many more cells in S-phase after DNA damage relative to trm9Δ cells with native RNR1 and had a cell cycle profile similar to wild-type cells containing the native RNR1 gene. In contrast, trm9Δ cells expressing native RNR1 had fewer cells transition to S-phase after exposure to MMS. Our cell cycle results predicted that Rnr1 protein levels would be higher in the trm9Δ cells expressing the codon-optimized RNR1 gene, relative to trm9Δ cells expressing native RNR1. We analyzed Rnr1 protein levels in the above-described cells (-/+ MMS) and show that, indeed, Rnr1 protein levels are higher in the trm9Δ cells expressing the codon optimized RNR1 gene (Fig. 6C) relative to trm9Δ cells expressing native RNR1. These results support the idea that Rnr1 protein levels are regulated by gene-specific codon usage patterns and tRNA modifications during the cell cycle. They also show that tRNA modification defects can be rescued by changes in codon usage patterns. It is interesting to note that MMS-treated wild-type cells expressing the codon optimized RNR1 transitioned more rapidly into S-phase after DNA damage, and their Rnr1 protein levels were higher than wild-type cells expressing the native RNR1 gene (Fig. 6B and C). These observations further support the ideas that codon usage plays a regulatory role during cell cycle progression and that Rnr1 protein levels are rate-limiting in regard to transition into S-phase.
Microarray analysis has clearly shown that transcripts can be regulated in a cell cycle-specific fashion.32 Our data demonstrates that, similar to transcript levels, the levels of specific tRNA modifications can change as a function of cell cycle stage. This implies that translation is regulated as a function of cell cycle, and we have provided the example of increased RNR1 translation being dependent upon increased mcm5U levels. Other transcripts rich in the AGA codons are likely to have increased translation in S-phase after DNA damage, with likely candidates being RNR2 and RNR3 (13 and 14 AGA codons, respectively).3 In total, we have demonstrated that levels of 16 tRNA modifications significantly change when comparing G1, G2 or HU-induced S-phase to asynchronous cells. G2- and G1-specific data account for 14 of these changes in modifications and suggest that translation is distinct outside of DNA replication. Clearly, codon-specific translation will have a more complicated code in G1 and G2, relative to S-phase, as many modifications are decreasing. We propose a generalized model in which particular tRNA modifications regulate the translation of specific transcripts based on codon usage patterns (Fig. 7). There is a dynamic aspect to our model, as individual tRNA modifications increase or decrease as a function of cellular state in order to translationally regulate codon-specific transcripts. At its simplest, translation in S-phase after DNA damage can be used to increase the levels of specific AGA-rich transcripts. This codon-specific translational regulation is used to increase the levels of Rnr1, a vital component that ensures proper cell function during DNA replication under damaging conditions. Studies addressing Trm9 are the prototype for codon-specific translational regulation, because mutants are only deficient in mcm5-based modifications at the wobble position.4 Other trm mutants have pleotropic deficiencies in tRNA modification status,25 so that defining specific rules for these systems will be difficult.
Our study also demonstrates that a reduction of Rnr1 levels in trm9 mutants is a major contributor to the delayed transition into S-phase after DNA damage. Low Rnr1 levels, and presumably low dNTP levels, have previously been associated with a cell’s failure to transition into S-phase. Rothstein and coworkers have demonstrated that when sml1–4SA (mutant form of Sml1 that does not undergo degradation) is overexpressed in rnr1-W688G (a strain with compromised Rnr1 activity) mutants, the cells persist in G1 and delay entry into S-phase.33 Efficient transition into S-phase may require a burst of RNR activity. In support, Chabes and Stillman have demonstrated that constitutively high dNTP levels impede the entry of cells into S-phase by slowing down the initiation of replication.34 They have also demonstrated a large difference in dNTP levels between G1- and S-phases as well as an increase after DNA damage. Our observations with trm9Δ cells are in line with these published results. The fact that we can partially rescue the cell cycle phenotype in trm9Δ mutants by increasing the levels of the RNR1 transcript or Rnr1 protein by genetic means is further proof that RNR activity and dNTP levels are key determinants influencing transition into S-phase. In addition, we have identified trm9Δ mutants as a novel model for studying decreased RNR activity.
Our observation that Trm9-dependent tRNA modifications play an important role in promoting transition into S-phase also demonstrates another layer of regulation for cellular RNR activity. We alluded to translational regulation for Rnr1 and Rnr3 in our previous work,3 but we failed to show increased tRNA modification levels one hour after MMS induced DNA damage. Most likely, this was because many cells are still in G1-phase 1 hour after MMS treatment (Fig. 2). Our current study demonstrates that mcm5U modification levels increase in DNA damage-associated S-phase relative to G1 and G2, with Trm9 activity being required for a translation-based regulatory boost. RNR activity is thus regulated transcriptionally, translationally, post-translationally and by subcellular localization. The correlation of tRNA modification defects with cell cycle phenotypes is not without precedent. In S. cerevisiae, other proteins involved in generating mcm5-based modifications, including Trm112, Sit4 and Sap185/190, are linked to G1 to S cell cycle transition phenotypes.35-38 Additionally, the sit4 mutants have been shown to be sensitive to the DNA damaging agents MMS39 and doxorubicin (SGD database). We predict that these enzymes also play roles in the translational regulation of RNR1 or other important cell cycle genes. It is important to note that other components of the DNA damage response have been linked to tRNA. Mec1 and Rad53 signaling through Los1 is required for the accumulation of unspliced tRNA in the nucleus and subsequent delay in transition for G1 to S-phase after DNA damage.40 Cumulatively, there is much evidence to link tRNA biology to cell cycle progression after DNA damage.
We note that studies of yeast Trm9 should help elucidate the regulatory role for mammalian proteins, as the mouse and human ALKBH8 gene products are homologous to yeast Trm9.41 ALKBH8 is a tRNA methyltransferase that can produce mcm5U modifications, among others, and ALKBH8 has been demonstrated to be transcriptionally upregulated in an ATM-dependent fashion after DNA damage.7 We suggest that dynamic changes in tRNA modifications will also occur in humans after DNA damage, and that increased tRNA modification should be downstream of ATM activation. This implies that there will be translational regulation of specific transcripts after DNA damage, where changes in tRNA modification levels influence the translation of specific codons. This, in turn, influences the levels of specific proteins via a gene-specific translational regulatory mechanism. We have observed both positive (AGA, GAA and CAA) and negative (GAG and TTG) regulation of codon runs by Trm9. While we have primarily concentrated on the increased translation of specific transcripts by wobble base modifications, we cannot ignore the possibility that a Trm9-dependent decrease in the translation of codon specific transcripts could also occur. Reduced or restricted translation of specific codon usage patterns by a given tRNA modification may be a way to decrease the levels of proteins detrimental to a biological process; however, this idea is merely speculative. Our model (Fig. 7) begs the question of why cells would regulate translation using codon usage and changes in tRNA modification status. One possible reason is the need for cells to quickly respond to a given cellular state. A rapid change in the level of tRNA modifications would allow cells to use existing transcripts more or less efficiently, and this could provide a vital cellular response to stress or changes in environmental conditions.
Tables S2 and 3 list the S. cerevisiae strains and the oligonucleotides used in the study. The trm9Δ mutants were made by transforming a URA3 knockout cassette derived from a trm9Δ heterozygous mutant (YSC4427–98798995) from Open Biosystems. Mutants were selected on SD–Ura plates and were confirmed by PCR. Media preparation and other yeast manipulations were performed using standard methods.42 The sml1Δ cells were generated using the corresponding KanMX cassette derived from the S. cerevisiae Gene Deletion library. The RNR1 gene was cloned by PCR based approaches in pHybcI/HK (Invitrogen) using SacI and SalI cloning sites. The generated plasmid, pHybcI/HK-RNR1, has the RNR1 gene under the control of the ADH1 promoter and ADH1 transcription terminator. It was transformed into cells and selected for on SD-His plates.
Wild-type and trm9Δ cells were grown to ~5 × 106 cells/ml in YPD and treated as described in Figure 1 legend. Protein extracts were prepared as described,4 and protein concentrations were measured using a BCA Protein Assay Kit (Pierce). Western blots were performed using anti-Rnr1 goat pAB (Santa Cruz Biotechnology), donkey anti-goat HRP conjugate secondary (Santa Cruz Biotechnology), rabbit pAB to β-tubulin (Abcam) and stabilized goat anti-rabbit HRP conjugate secondary (Thermo Scientific).
Cells were grown to ~1 × 107 cells/ml in YPD and treated as described in the legend. Chemicals included MMS (Sigma-Aldrich), nocadazole (Sigma-Aldrich), α-factor (Zymo Research) and hydroxyurea (MP Biomedical). Samples were spun down immediately and fixed overnight in 75% ethanol. Fixed cells were spun down and washed twice with distilled water. Cells were re-suspended in 1 ml of 50 mM sodium citrate pH 7.0 and 80 µg of RNase A (Fisher Bioreagents) was added. Samples were then incubated for 1 h at 55°C, and then 250 µg of Proteinase K (Amresco) was added to each sample, followed by incubation for 1 h at 55°C. DNA staining was facilitated by the addition of 20 mg/ml propidium iodide (Sigma-Aldrich) to each sample, followed by a 30 min incubation at room temperature. DNA content was analyzed using a BD LSRII flow cytometer. Histograms depicting cell cycle state were generated using WinMDI 2.8 software (Scripps Research Institute), and the percentage of cells in G1-phase was calculated using ModFit LT 3.2 software (Verity Software House).
Cells were arrested in G1, S and G2 or asynchronously grown as described above. Modified ribonucleosides in tRNA were identified and quantified as reported previously.25 Cells were suspended in 1.5 ml of Trizol reagent with deaminase inhibitors (5 mg/mL coformycin, 50 µg/ml tetrahydrouridine) and antioxidants (0.1 mM deferoxamine mesylate, 0.5 mM butylated hydroxytoluene) to prevent nucleoside modification artifacts,43,44 S. cerevisiae cell pellets were lysed by three cycles of bead beating in a Thermo FP120 Bead Beater set at 6.5 m/s for each 20 sec cycle, with 1 min of cooling on ice between cycles. Lysates were extracted with 0.3 ml volume of chloroform at ambient temperature for 5 min, with 3 min of phase separation. The solutions were centrifuged at 12,000 × g for 15 min at 4°C and the aqueous phase was collected. Absolute ethanol was added to the aqueous phase (35% v/v), and small RNA species were then isolated using the PureLink miRNA Isolation Kit according to manufacturer’s instructions. The quality and concentration of the resulting small RNA mixture was assessed with a Bioanalyzer (Agilent Small RNA Kit), with tRNA comprising > 95% of the small RNA species present in the mixture. Following the addition of deaminase inhibitors, antioxidants and [15N]5-2'-deoxyadenosine internal standard (6 pmol), tRNA (6 μg) in 30 mM sodium acetate and 2 mM ZnCl2 (pH 6.8) were hydrolyzed with nuclease P1 (1 U) and RNase A (5 U) for 3 h at 37°C. The resulting products were dephosphorylated with alkaline phosphatase (10 U) and phosphodiesterase I (0.5 U) for 1 h at 37°C following the addition of acetate buffer to 30 mM, pH 7.8. The enzymes were removed by filtration (Microcon YM-10). Ribonucleosides were resolved on a Thermo Scientific Hypersil GOLD aQ reverse-phase column (150 X 2.1 mm, 3 µm particle size) eluted with the following gradient of acetonitrile in 8 mM ammonium acetate at a flow rate of 0.3 ml/min and 36°C: 0–18 min, 0–1%; 18–23 min, 1%; 23–28 min, 1–6%; 28–30 min, 6%; 30–31 min, 6–100%; 31–41 min, 100%. The HPLC column was coupled to an Agilent 6410 Triple Quadrupole LC/MS mass spectrometer with an electrospray ionization source operated in positive ion mode with the following parameters for voltages and source gas: gas temperature, 350°C; gas flow, 10 L/min; nebulizer, 20 psi; and capillary voltage, 3500 V. The first and third quadrupoles (Q1 and Q3) were fixed to unit resolution, and the modifications were quantified by pre-determined molecular transitions. Q1 was set to transmit the parent ribonucleoside ions and Q3 was set to monitor the deglycosylated product ions, except for Y, for which the stable C-C glycosidic bond led to fragmentation of the ribose ring; we used the m/z 125 ion for quantification. The dwell time for each ribonucleoside was 200 ms. The retention time, m/z of the transmitted parent ion, m/z of the monitored product ion, fragmentor voltage, and collision energy of each modified nucleoside, and 15N-labeled internal standards are as follows: D, 1.9 min, m/z 247→115, 80 V, 5 V; Y, 2.5 min, m/z 245→125, 80 V, 10 V; m5C, 3.3 min, m/z 258→126, 80 V, 8 V; Cm, 3.6 min, m/z 258→112, 80 V, 8 V; m5U, 4.2 min, m/z 259→127, 80 V, 7 V; ncm5U, 4.3 min, m/z 302→170, 90 V, 7 V; ac4C, 4.4 min, m/z 286→154, 80 V, 6 V; m3C, 4.4 min, m/z 258→126, 80 V, 8 V; Um, 5.1 min, m/z 259→113, 80 V, 7 V; m7G, 5.1 min, m/z 298→166, 90 V, 10 V; m1A, 5.7 min, m/z 282→150, 100 V, 16 V; mcm5U, 6.4 min, m/z 317→185, 90 V, 7 V; m1I, 7.3 min, m/z 283→151, 80 V, 10 V; Gm, 8.0 min, m/z 298→152, 80 V, 7 V; m1G, 8.3 min, m/z 298→166, 90 V, 10 V; m2G, 9.4 min, m/z 298→166, 90 V, 10 V; I, 10.9 min, m/z 269→137, 80 V, 10 V; mcm5s2U, 14.2 min, m/z 333→201, 90 V, 7 V; [15N]5-dA, 14.4 min, m/z 257→141, 90 V, 10 V; m22G, 15.9 min, m/z 312→180, 100 V, 8 V; t6A, 17.2 min, m/z 413→281, 100 V, 9 V; Am, 19 min, m/z 282→136, 100 V, 15 V; yW, 34.2 min, m/z 509→377, 120 V, 10 V, and i6A, 34.4 min, m/z 336→204, 120 V, 17 V. Signal intensity of each ribonucleoside was normalized with the signal intensity of [15N]5-dA from its sample before being analyzed with hierarchical clustering with the software Cluster 3.0, in which the ratios of modification levels in treated cells relative to untreated cells were transformed to log2 scale and were clustered using the Pearson correlation algorithm. The results were presented as a heat map using Java Treeview.
Wild-type and trm9Δ cells were transformed with reporter plasmids derived from pJD375 (having a 4X codon run in between Renilla and Firefly luciferase).31 Transformed yeast cells were grown overnight in SD-Ura, then diluted into fresh YPD to ~1 × 106 cells/ml and grown to ~107 cells/ml. Cells were pelleted, washed twice with 0.5 ml of cold lysis buffer containing 1 mM PMSF (Pierce), then resuspended in cold lysis buffer and broken by agitation with glass beads (0.5 mm Sigma-Aldrich). Lysates were clarified by centrifugation, and supernatants were transferred to pre-chilled tubes. Luminescence reactions were initiated by the addition of 50 µl of Promega DLR (Promega) to 5 µl of clarified cell lysates and measured using a Victor Plate Reader (PerkinElmer).
A codon optimized RNR1 gene (Opt-RNR1) was developed computationally and was chemically synthesized by GeneScript, as previously described.4 Opt-RNR1 was cloned into pRS413 using a SacI (2178) to XhoI (2091) ligation strategy. The RNR1 constructs along with the HIScx marker from pRS413 were than PCR amplified using primers that containing 40 base pairs of homology to the upstream or downstream regions of the native RNR1 locus. Transformants were selected for their ability to grow on media lacking histidine and the presence of the insert was verified by PCR.
Financial support was provided by the National Institute of Environmental Health Sciences (R01 ES015037, R01 ES017010 and P30 ES002109), the MIT Westaway Fund and the Singapore-MIT Alliance for Research and Technology.
No potential conflicts of interest were disclosed.
Ashish Patil helped design the experiments, constructed strains, performed damage-induced cell cycle studies and some of the western blots. He also analyzed the data and helped write the original manuscript. Madhu Dyavaiah contributed to experimental design, performed western blot experiments and some cell cycle experiments. Fraulin Joseph performed western blot experiments and some cell cycle experiments. John Rooney performed some western blot experiments. Clement Chan measured tRNA modification levels, contributed to experimental design and helped write the manuscript. Peter Dedon contributed to experimental design, data analysis, discussed results and helped write the manuscript. Thomas Begley contributed to experimental design, data analysis, discussed results and helped write the manuscript.
Supplemental material may be found here: www.landesbioscience.com/journals/cc/article/21919
Previously published online: www.landesbioscience.com/journals/cc/article/21919