Aberrant DNA N⁶‐methyladenine incorporation via adenylate kinase 1 is suppressed by ADAL deaminase‐dependent 2′‐deoxynucleotide pool sanitation

Abstract

Intracellular decay of N⁶‐methyladenine (m6A)‐containing RNA potentially induces aberrant N⁶‐methyl‐2′‐adenine (6mdA) misincorporation into DNA. Biophysically, misincorporated 6mdA may destabilize the DNA duplex in a manner similar to bona fide methylated 6mdA DNA, thereby affecting DNA replication and transcription. Utilizing heavy stable isotope labeling and ultrasensitive UHPLC–MS/MS assay, we demonstrate that intracellular m6A‐RNA decay does not generate free 6mdA species, nor lead to any misincorporated DNA 6mdA in most mammalian cell lines tested, unveiling the existence of a sanitation mechanism that prevents 6mdA misincorporation. Depletion of deaminase ADAL increases the levels of free 6mdA species, concomitant with the presence of DNA‐misincorporated 6mdA resulting from intracellular RNA m6A decay, suggesting that ADAL catabolizes 6mdAMP in vivo. Furthermore, we show that the overexpression of adenylate kinase 1 (AK1) promotes 6mdA misincorporation, while AK1 knockdown diminishes 6mdA incorporation, in ADAL‐deficient cells. We conclude that ADAL together with other factors (such as MTH1) contributes to 2′‐deoxynucleotide pool sanitation in most cells but compromised sanitation (e.g., in NIH3T3 cells) and increased AK1 expression may facilitate aberrant 6mdA incorporation. This sanitation mechanism may provide a framework for the maintenance of the epigenetic 6mdA landscape.

An active sanitation mechanism prevents toxic DNA incorporation of 6mdA derived from the decay of m6A‐containing RNA in mammalian cells.

Introduction

As a prevalent epigenetic DNA modification in prokaryotes (Vanyushin et al, 1968; Fang et al, 2012) and some unicellular eukaryotes (Fu et al, 2015; Beh et al, 2019), DNA N⁶‐methyl‐2′‐deoxyadenosine (6mdA) is also found or proposed to exist in multicellular eukaryotes such as fungi (Mondo et al, 2017), C. elegans (Greer et al, 2015), Drosophila (Zhang et al, 2015), plants (Liang et al, 2018; Zhou et al, 2018), vertebrates (Liu et al, 2016), and mammals (Liu et al, 2016, 2021; Yao et al, 2017; Xiao et al, 2018). Despite its low abundance, eukaryotic DNA 6mdA plays a number of potential roles in the regulation of gene transcription (Fu et al, 2015; Greer et al, 2015), transposon activity (Zhang et al, 2015; Wu et al, 2016; Liang et al, 2018), nucleosome occupancy (Wang et al, 2017a; Beh et al, 2019), and transgenerational progression (Luo & He, 2017; Ma et al, 2019). Biochemically, the presence of the N⁶‐methyl group on adenine of the DNA template induces RNA pol II pausing (Wang et al, 2017b) and partially inhibits DNA replication (Li et al, 2019).

Moreover, epigenetic mRNA N⁶‐methyladenosine (m6A), which shares the same N⁶‐methyladenine base with DNA 6mdA, is an abundant posttranscriptional modification in mammals. mRNA m6A is highly enriched in 3′ untranslated regions, stop codon flanking regions, and long internal exons of mRNA (Dominissini et al, 2012; Meyer et al, 2012) and regulates transcription and processing in the nucleus and translation and decay in the cytoplasm (Yang et al, 2015; He & He, 2021). Notably, mRNA decay is prevalent and exhibits a short half‐life ranging from minutes to hours (Hambraeus et al, 2003; Sharova et al, 2009). Introns and spacer sequences in mRNA during processing; defective mRNA fragments during transcription, processing, and functioning; and scrapped mature mRNAs are degraded as signals by a surveillance system (Sachs, 1993; Valencia‐Sanchez et al, 2006; Houseley & Tollervey, 2009). Along with mRNA degradation, substantial m6A‐related species (e.g., N⁶‐methyladenosine monophosphate (m6rAMP)) are released to the nucleotide pools, or even into the extracellular space (Nyhan, 2005; Jiang et al, 2018). Following purine salvage, the released m6A‐related species might be ultimately converted into premethylated form N⁶‐methyl‐2′‐deoxyadenosine triphosphate (6mdATP), which can be misincorporated into genomic DNA by DNA polymerases (Musheev et al, 2020; Liu et al, 2021) and result in the generation of DNA 6mdA in a replication‐dependent but non‐methyltransferase‐dependent generation of DNA 6mdA.

Chemically, misincorporated 6mdA (i6mdA) shares an identical chemical structure with methylase‐deposited 6mdA, rendering them indistinguishable. Like the bona fide methylated product 6mdA, misincorporated i6mdA should destabilize the DNA duplex and thus affect DNA replication and transcription, and potentially falsify the epigenetic DNA 6mdA landscape (Bochtler & Fernandes, 2021). Collectively, i6mdA, which is unintentionally set on DNA, should be considered as a form of harmful DNA damage.

In this work, we utilized unique heavy stable isotope tracing and ultrasensitive DNA 6mdA assays to investigate the misincorporation of DNA i6mdA in response to intracellular RNA m6A degradation. We found that intracellular RNA m6A degradation could not induce any i6mdA in most of the tested cells. Based on these results, we proposed a 2′‐deoxynucleotide pool sanitation mechanism that blocks the reformulation of RNA decay‐derived m6rA nucleotides into 6mdATP and thus eliminates misincorporated i6mdA. Furthermore, we showed that a deaminase ADAL and adenylate kinase 1 (AK1) play a pivotal role in the proposed sanitation. These data revealed a new framework for maintaining the epigenetic DNA 6mdA landscape.

Results

Tracing of RNA m6A catabolic pathway by heavy stable isotope labeling

To trace the catabolic pathway of RNA m6A, we first exploited heavy stable isotope‐labeled nucleoside [¹⁵N₅]‐adenosine ([¹⁵N₅]‐rA) to treat mouse embryonic stem (mES) cells, human embryonic kidney 293T cells, mouse myoblast C2C12 cells, and mouse embryonic fibroblast NIH3T3 cells. We applied highly sensitive ultrahigh‐performance liquid chromatography–tandem mass spectrometry (UHPLC–MS/MS) to measure RNA labeling in these treated cells (Fig 1A). As revealed by UHPLC–MS/MS analysis, mRNA adenosine (rA) was substantially labeled by nitrogen‐15 (¹⁵N) atoms (over 60% of the total rA bases, Appendix Fig S1A). Notably, the labeled rA was present in two forms, [¹⁵N₄]‐rA and [¹⁵N₅]‐rA (Appendix Fig S1A). The observation of a major form of [¹⁵N₄]‐rA suggested that the labeling reagent [¹⁵N₅]‐rA underwent deamination by adenine deaminase (Ada) and subsequent re‐amination via the purine salvage pathway (Appendix Fig S2A), similar to the previously reported [¹⁵N₅]‐2′‐deoxyadenosine (dA) labeling process (Liu et al, 2017). mRNA m6A was also efficiently labeled (over 56.5% of the total m6A bases) and consistently presented in two forms: major [¹⁵N₄]‐m6A and minor [¹⁵N₅]‐m6A (Appendix Fig S1B).

Figure 1. Stringent control of DNA 6mdA misincorporation from intracellular m6A‐containing RNA decay.

• A
  Schematic diagram of heavy stable isotope labeling and subsequent UHPLC–MS/MS analysis.
• B
  UHPLC–MS/MS chromatograms for detection of the labeled‐m6A ([¹⁵N₅]‐m6A and [¹⁵N₄]‐m6A) in mRNA and the labeled 6mdA ([¹⁵N₅]‐6mdA and [¹⁵N₄]‐6mdA) in genomic DNA of mES, HEK293T, C2C12, and NIH3T3 cells. The cells were treated with [¹⁵N₅]‐rA for 7 days.
• C
  The quantification of labeled‐m6A in mRNA and labeled‐6mdA in genomic DNA.
• D
  UHPLC–MS/MS chromatograms for detection of [¹⁵N₄]‐6mdA in genomic DNA of mES and HEK293T cells. The cells were treated with [¹⁵N₅]‐rA for 0–50 days.
• E, F
  UHPLC–MS/MS chromatograms (E) and the quantitative results (F) of mRNA [D₃]‐m6A and genomic [D₃]‐6mdA of mES, HEK293T, C2C12, and NIH3T3 cells. The cells were treated with [D₃]‐L‐methionine for 7 days, and [¹⁵N₅]‐6mdA nucleoside (50 amol) was spiked in digested DNA samples as external isotopic standard. Error bars = s.d., three independent biological replicates. *The standard was not labeled with a heavy stable isotope. CPS: counts per second. ND—not detected.

Source data are available online for this figure.

The second heavy stable isotope, [D₃]‐L‐methionine, was also applied to trace RNA m6A (Fig 1A). In vivo, the tracer [D₃]‐L‐methionine is transformed into the methyl donor [D₃]‐S‐adenosyl‐L‐methionine ([D₃]‐SAM) by SAM synthetase, and the latter is utilized by diverse methylases as a methyl donor to modify nucleic acid bases (Appendix Fig S2B). To characterize the transformation efficiency of [D₃]‐L‐methionine into the methyl donor [D₃]‐SAM, we measured [D₃]‐5‐methylcytidine ([D₃]‐m5C) in mRNA. The labeling ratios of [D₃]‐m5C to total m5C were over 70% in these four cells (Appendix Fig S1C). Consistently, we detected abundant mRNA [D₃]‐m6A with a labeling efficiency of over 72.5% (Appendix Fig S1D). These data support that intracellular RNA m6A is efficiently labeled with [¹⁵N] or [D] in all tested cells.

The misincorporation of DNA 6mdA from intracellular degraded RNA m6A is tightly regulated

It was reported that intracellular RNA m6A degradation could induce misincorporated DNA i6mdA in a number of cells, including mES cells, HEK293T cells, C2C12 cells, and NIH3T3 cells (Musheev et al, 2020). As noted, the misincorporation of i6mdA exhibited a delayed phase (~ 5 days delay) compared with de novo RNA synthesis in mES and HEK293T cells (Musheev et al, 2020). Following this clue, we treated these four cells with the isotopic [¹⁵N₅]‐rA over 7 days. With a high labeling efficiency (56.5%), the labeled mRNA m6A reached levels of 1.85 per 10³ rC in mES cells, 1.75 per 10³ rC in HEK293T cells, 2.12 per 10³ rC in C2C12 cells, and 2.23 per 10³ rC in NIH3T3 cells (Fig 1B and C). Meanwhile, the levels of labeled 6mdA ([¹⁵N₄]‐6mdA and [¹⁵N₅]‐6mdA) in C2C12 and NIH3T3 cells were 0.23 and 0.83 per 10⁶ dC, respectively (Fig 1C). However, no labeled 6mdA was detected in mES and HEK293T cells (Fig 1B and C). Even after 50 days of labeling to trace RNA m6A degradation, we failed to observe any labeled 6mdA in mES and HEK293T cells (Fig 1D). By contrast, the labeled 6mdA in NIH3T3 cells appeared on day 1 after [¹⁵N₅]‐rA treatment and its level reached a maximum on day 7 (Appendix Fig S3A and B). Notably, our UHPLC–MS/MS method could sensitively detect 6mdA with a limit of detection (LOD) of 3.5 × 10⁻¹⁸ mol (or 3.5 amol). In other words, using 1.0 μg of DNA, we detected seven 6mdA modifications per mouse genome (Appendix Fig S4A and B). To further validate the applicability of our UHPLC–MS/MS method and LOD for 6mdA analysis in real gDNA samples, we spiked [¹⁵N₅]‐6mdA standard into gDNA of [¹⁵N₅]‐rA‐labeled cells. Following enzymatic digestion and ultrafiltration, UHPLC–MS/MS analysis did not detect any MS signals of [¹⁵N₅]‐6mdA in gDNA samples from mES and HEK293T cells, whereas there were significant signals of [¹⁵N₅]‐6mdA in gDNA spiked with 5.0 amol [¹⁵N₅]‐6mdA nucleosides, and the peak heights of these spiked signals were comparable to that of 5.0 amol [¹⁵N₅]‐6mdA standard (Appendix Fig S4C and D). Additionally, we observed that the [¹⁵N₅]‐6mdA level in gDNA of NIH3T3 cells spiked with 100 amol [¹⁵N₅]‐6mdA nucleosides was approximately 365 amol, which was almost the sum of the [¹⁵N₅]‐6mdA level in NIH3T3 gDNA (268 amol) and 100 amol [¹⁵N₅]‐6mdA standard (Appendix Fig S4E). These results demonstrated that our UHPLC–MS/MS method is capable of accurately detecting extremely rare 6mdA in gDNA samples. Moreover, we did not observe labeled 6mdA in other cell lines, including A549, MRC5, T47D, U2OS, HepG2, and T24 cells (Appendix Fig S5A and B). These results consistently support that intracellular RNA m6A degradation cannot induce any misincorporated DNA 6mdA in most of the tested cells.

We also detected abundant mRNA [D₃]‐m6A with a labeling efficiency of over 72.5% in [D₃]‐L‐methionine‐treated cells. In these assays, we spiked the digested DNA with external isotopic [¹⁵N₅]‐6mdA for calibration. The levels of [D₃]‐m6A in mES cells, HEK293T cells, C2C12 cells, and NIH3T3 cells were 2.17, 2.45, 2.16, and 2.12 per 10³ rC, respectively, and the [D₃]‐6mdA in C2C12 and NIH3T3 cells reached levels of 0.55 and 1.08 per 10⁶ dC (Fig 1E and F). However, we still did not detect [D₃]‐6mdA in genomic DNA of mES and HEK293T cells (Fig 1E and F).

The above data supported the absence of the labeled DNA 6mdA (misincorporated 6mdA) regardless of whether intracellular RNA m6A was labeled with [¹⁵N] or [D] in most of the tested cells. These results are consistent with our previous work (Liu et al, 2021), in which we measured only around 0.2–0.4 unlabeled 6mdA per 10⁶ dA and could not detect any isotope‐labeled 6mdA in heavy‐isotope‐traced mES cells. These unlabeled 6mdA are not related to the intracellular m6A metabolic flux. Collectively, we proposed the existence of a 2′‐deoxynucleotide pool sanitation mechanism that blocks the reformulation of intracellular RNA m6A degradation products to form premethylated 6mdATP and thus eliminates DNA 6mdA misincorporation.

The deaminase ADAL preferentially catabolizes 6mdAMP in vitro

Previous studies have demonstrated that the adenosine deaminase‐like (ADAL) protein can catabolize m6rAMP and N⁶‐methyl‐2'‐deoxyadenosine monophosphate (6mdAMP) in vitro (Schinkmanová et al, 2006; Murakami et al, 2011), and based on its activity reducing m6rAMP in plant and human cells (Chen et al, 2018), ADAL was shown to protect RNA from the misincorporation of N⁶‐methyladenine (Chen et al, 2018) and proposed to partly reduce DNA 6mdA misincorporation (Musheev et al, 2020). Here, we speculated that the dual activity of ADAL toward m6rAMP and 6mdAMP prompts it a critical role in the sanitation blocking i6mdA misincorporation. We first verified the catalytic activity of mouse recombinant ADAL protein on m6rAMP and 6mdAMP in vitro (Fig 2A). As detected by UHPLC–MS/MS assays, along with a decrease of m6rAMP (detected in the form of m6A by UHPLC–MS/MS) or 6mdAMP (detected in the form of 6mdA by UHPLC–MS/MS), the nontoxic product inosine monophosphate (rIMP, detected in the form of rI) or deoxyinosine monophosphate (dIMP, detected in the form of dI) was observed (Fig 2B and C). ADAL displays an ability to catalyze the deamination of m6rAMP and 6mdAMP nucleotides in vitro. To further explore the relative catalytic activity, we next measured the catalytic efficiency (K_cat/K_m) of ADAL for both 6mdAMP and m6rAMP. As shown in Fig 2D, in our in vitro reaction system, the catalytic efficiency of ADAL for 6mdAMP was 33.4 ± 3.12 mM⁻¹ s⁻¹. This value was 1.14‐fold higher than that of ADAL for m6rAMP (15.6 ± 2.03 mM⁻¹ s⁻¹) (Fig 2E). In addition, the ADAL protein showed negligible catalyzing activity toward 6mdA deoxynucleoside and 6mdATP (Appendix Fig S6A and B). These data support the dual activity of ADAL toward 6mdAMP and m6rAMP, and its preference for 6mdAMP in vitro.

Figure 2. Preferential hydrolysis of 6mdAMP by recombinant ADAL protein in vitro .

• A
  Diagram of the conversion of m6rAMP and 6mdAMP into rIMP and dIMP by ADAL, respectively.
• B, C
  UHPLC–MS/MS chromatograms of m6rAMP and rI from ADAL catabolyzing m6rAMP (B), and 6mdAMP and dI from ADAL catabolyzing 6mdAMP (C).
• D
  The determination of kinetic constants for recombinant ADAL catabolizing m6rAMP (red) and 6mdAMP (blue). The kinetic data were fitted with the Michaelis–Menten equation.
• E
  The catalytic efficiency (K_cat/K_m) of ADAL to 6mdAMP and m6rAMP. Error bars = s.d., three technical replicates.

Source data are available online for this figure.

The deaminase ADAL functions as a key sanitation enzyme in vivo

To determine the roles of ADAL in the sanitation in vivo, we knocked down Adal mRNA in mES cells via small interfering (si) RNA transfection. As Adal was efficiently knocked down (Appendix Fig S7A and B, Appendix Table S2), both [¹⁵N₄]‐6mdA and [¹⁵N₅]‐6mdA were indeed detected in [¹⁵N₅]‐rA‐traced cells (Fig 3A and B). To further verify this effect, we used clustered regularly interspaced short palindromic repeat/CRISPR‐associated protein 9 (CRISPR/Cas9)‐mediated nonhomologous end joining (NHEJ) to generate Adal‐deficient mES cells (Adal ⁻⁻ ; Appendix Fig S8A). The gene sequencing (Appendix Fig S8B) and western blot analysis (Appendix Fig S8C) confirmed the successful knockout (KO) of the Adal gene in Adal ⁻⁻ mES cells. After treatment with [¹⁵N₅]‐rA for 7 days, the Adal ^−/−‐1 and Adal ^−/−‐2 mES cells also exhibited detectable [¹⁵N₄]‐6mdA and [¹⁵N₅]‐6mdA (Fig 3C and D). The levels of [¹⁵N₄]‐6mdA and [¹⁵N₅]‐6mdA averaged approximately 0.064 per 10⁶ dC and 0.035 per 10⁶ dC (Fig 3D), respectively. Consistently, without the depletion of Adal, labeled 6mdA was not detected in negative control siRNA (siCTRL)‐transfected mES cells (Fig 3A and B) or knockout‐absent wild‐type mES cells (Fig 3C and D). These results indicated that the deaminase ADAL is critically involved in the sanitation for blocking 6mdA misincorporation.

Figure 3. The depletion of ADAL induces misincorporated DNA 6mdA in vivo .

• A, B
  UHPLC–MS/MS chromatograms (A) and quantitative results (B) of DNA 6mdA in [¹⁵N₅]‐rA‐treated mES cells after Adal knockdown for 7 days. Two independent biological replicates.
• C, D
  UHPLC–MS/MS chromatograms (C) and quantification (D) of genomic 6mdA in Adal ^−/− mES cells treated with [¹⁵N₅]‐rA for 3 days. Error bars = s.d., three independent biological replicates.
• E
  The relative levels of Adal in mES, C2C12, and NIH3T3 cells against reference gene TUBB.
• F, G
  The western blot analysis (F) and its quantitative results (G) of ADAL expression in mES, C2C12, and NIH3T3 cells. Unpaired t‐test, error bars = s.d., three independent biological replicates. *The standard was not labeled with a heavy stable isotope. ND—not detected.

Source data are available online for this figure.

Moreover, we also tested the role of ADAL in regulating the misincorporation of genomic 6mdA in NIH3T3 cells. As shown in Appendix Fig S9, ADAL overexpression in NIH3T3 cells reduced the level of labeled 6mdA by 42.7% (Appendix Fig S9A and B), and Adal knockdown significantly increased the level of labeled 6mdA (Appendix Fig S9C and D). In addition, the expression of Adal in mES, C2C12, and NIH3T3 cells was measured. The mRNA levels of Adal in C2C12 and NIH3T3 cells were one‐fold lower than that in mES cells (Fig 3E), and the protein level of ADAL in mES cells was, respectively, 1.7‐fold and 5.25‐fold higher than those in C2C12 and NIH3T3 cells (Fig 3F and G). Combined with the i6mdA levels in the three tested cell lines (Fig 1C and F), ADAL expression levels are inversely proportional to i6mdA levels. In the highest level of ADAL in mES cells, no i6mdA was detected; in the lowest level of ADAL in NIH3T3 cells, the detected 6mdA is the highest. These results strongly support that the downregulation of ADAL in NIH3T3 and C2C12 cells compromises the sanitation and induces the misincorporation of DNA 6mdA.

ADAL blocks the generation of free 6mdA species in intracellular nucleotide pool

If ADAL actively catabolizes cellular 6mdAMP to prevent misincorporation into genomic DNA, the free 6mdA species (6mdA, 6mdAMP, 6mdADP, and 6mdATP) in cellular nucleotide pool should not be detectable in ADAL‐proficient cells. The observation of misincorporated DNA 6mdA essentially suggests that DNA polymerase uses 6mdATP as a substrate for replication, implying that cellular concentrations of free 6mdA species in nucleotide pool should be sufficiently high. Given that incorporated DNA 6mdA is detectable when ADAL is knocked down or KO in mES cells (Fig 3C and D), free 6mdA species should be also detectable. To explore the precise role of ADAL catabolizing 6mdAMP in vivo, we measured both the levels of free 6mdA and m6rA species in cellular nucleotide pool (Fig 4A). By using adequate amounts of mES cells (up to 2 × 10⁷ cells) and improving extraction of free nucleotide pool (see Methods), we showed that the ADAL depletion increases the levels of total free m6rA species (m6rA, m6rAMP, m6rADP, and m6rATP) by over ~4.5 folds (Fig 4B and C). This confirmed the role of ADAL in controlling free m6rA species (Chen et al, 2018). By contrast, we showed that any of free 6mdA species (including 6mdA, 6mdAMP, 6mdADP, and 6mdATP, see Fig 4A) can be detected only in the ADAL‐KO mES cells (Fig 4D and E). These data further support the activity of ADAL catabolizing 6mdAMP in vivo. Noteworthily, free m6rA species can be detected regardless of ADAL knockout, but any of free 6mdA species can be detected only in the ADAL‐KO mES cells. Therefore, if ADAL should have catabolized only m6rAMP but not 6mdAMP in vivo, the un‐eliminated free m6rA species would be converted to free 6mdA species, thus free 6mdA species must be shown up in ADAL‐proficient cells. Taken together, ADAL is one of the key regulation factors in the 2′‐deoxynucleotide pool sanitation.

Figure 4. Quantification of free 6mdA‐ and m6rA‐related species in Adal‐WT and ‐KO cells.

• A
  Schematic illustration of the protocol for extracting free 6mdA, m6rA, dC, and rC.
• B, C
  The UHPLC–MS/MS chromatograms (B) and the quantitation (C) of intracellular free [D₃]‐m6rA‐related nucleotides in WT and Adal ^−/− mES cells.
• D, E
  The UHPLC–MS/MS chromatograms (D) and the quantitation (E) of intracellular free [D₃]‐6mdA‐related nucleotides in WT and Adal ^−/− mES cells. The cells were treated with [D₃]‐methionine for 7 days. Total free [D₃]‐6mdA = Free [D₃]‐6mdA + [D₃]‐6mdAMP + [D₃]‐6mdADP + [D₃]‐6mdATP, Total free [D₃]‐m6rA = Free [D₃]‐m6rA + [D₃]‐m6rAMP + [D₃]‐m6rADP + [D₃]‐m6rATP. CIP: Calf intestinal alkaline phosphatase for dephosphorylation. Error bars = s.d., three independent biological replicates. *The standard was not labeled with a heavy stable isotope. ND—not detected.

Source data are available online for this figure.

Adenylated kinase 1 is an important factor maintaining the 2′‐deoxynucleotide pool sanitation

In addition to Adal, we also explored the impacts of two other deamination reactions on DNA 6mdA misincorporation. AMP deaminase and adenosine deaminase (Ada) were inhibited using the inhibitors cpd3 ((S)‐6‐(4‐(((1‐(isoquinolin‐8‐yl) ethyl) amino) methyl) phenyl) nicotinic acid; Admyre et al, 2014) and DCF (2′‐deoxycoformycin) (Agarwal et al, 1977), respectively. The AMP deaminase inhibitor cpd3 could not reduce the level of RNA [¹⁵N₄]‐m6A (Appendix Fig S10A) and failed to induce any misincorporated 6mdA (indicated by the labeled 6mdA; Appendix Fig S10B). Treatment with DCF (0.1~100 μM) significantly increased the level of RNA [¹⁵N₅]‐m6A and concomitantly reduced the levels of [¹⁵N₄]‐m6A (Appendix Fig S10C). However, labeled 6mdA was not detected in the genomic DNA of DCF‐treated mES cells (Appendix Fig S10D). We performed additional experiments using misincorporated DNA i6mdA‐containing NIH3T3 cells. However, our results showed that both the adenosine deaminase inhibitor DCF and the AMP deaminase inhibitor cpd3 reduced the misincorporation of DNA i6mdA (Appendix Fig S10E and F). If it had an effect on the deamination of the nucleoside m6A/6mdA, the inhibition of adenine deaminase should increase the misincorporated DNA 6mdA. Therefore, we exclude the possibility that adenine deaminase acts as a deaminase for m6rA/6mdA nucleosides. This was also true for the AMP deaminase, as its deamination for m6rAMP and 6mdAMP was excluded.

To further identify the factors involved in the sanitation, we constructed a series of overexpression plasmids for nucleotide kinases functioning in the purine salvage synthesis pathway, including adenosine kinase (ADK), adenine phosphoribosyltransferase (APRT), adenylate kinase 1 (AK1), and nucleoside diphosphate kinase (NDPK). These proteins are hypothesized to facilitate the formation of 6mdATP by catalyzing phosphorylation at a certain step in the purine salvage pathway. Western blot analysis validated the overexpression of the targeted proteins (Fig 5A and B, and Appendix Fig S11A). However, labeled 6mdA was not detected in any of the gene overexpression groups of wild‐type mES cells (Appendix Fig S11B). These results may suggest that such a simple increase in the expression of nucleotide kinases cannot break the misincorporation‐suppressing sanitation process. Then, we replaced wild‐type cells with Adal ^−/− cells. Notably, the level of the labeled 6mdA in the AK1‐overexpressing group (0.78 per 10⁶ dC) was approximately 2.25‐fold higher than that in the control group (EV, 0.24 per 10⁶ dC) (Fig 5C). By contrast, the labeled 6mdA levels in the ADK‐, APRT‐, and NDPK‐overexpressing groups (0.19–0.23 per 10⁶ dC) were similar to that in the control group. Moreover, we also observed a similar enhancing effect in NIH3T3 cells (Fig 5D). The labeled 6mdA content in AK1‐overexpressing cells was approximately 1.69‐fold higher than that in the control group and the ADK‐, APRT‐, and NDPK‐overexpressing groups (Fig 5D). To further explore the role of AK1 in controlling 6mdA incorporation, we knocked down Ak1 gene in NIH3T3 cells with siRNA. As shown by RT–qPCR and western blot analysis (Fig 5E and F), the knockdown of AK1 resulted in a significant decrease in AK1 expression. Concomitantly, the incorporated 6mdA was hardly detected in Ak1‐knockdown cells (Fig 5G), suggesting an irreplaceable role of AK1 in promoting 6 mA incorporation. Taken together, these results consistently suggest that AK1 participates in the sanitation of RNA m6A.

Figure 5. The overexpression of adenylate kinase 1 (AK1) elevates the misincorporated DNA 6mdA.

• A–C
  The western blot analysis of flag‐tagged gene overexpression in Adal ^−/− mES cells (A) and NIH3T3 cells (B) overexpressing ADK, APRT, AK1, or NDPK gene. (C) The quantification of genomic 6mdA in [¹⁵N₅]‐rA‐traced Adal ^−/− mES cells overexpressing flag‐tagged ADK, APRT, AK1, or NDPK gene for 7 days.
• D
  The quantification of genomic 6mdA in NIH3T3 cells overexpressing ADK, APRT, AK1, or NDPK gene. [¹⁵N₅]‐rA treatment began at 24 h after plasmid transfection and lasted for 72 h. Two technical replicates.
• E–G
  The knockdown of Ak1 leads to a significantly decreased i6mdA in NIH3T3 cells. The RT–qPCR (E) and western blot (F) were applied to determine the knockdown of Ak1 in NIH3T3 cells. Two independent biological replicates. (G) The quantitation of misincorporated 6mdA in Ak1‐knockdown cells. [¹⁵N₅]‐rA treatment began at 24 h after siRNA transfection and lasted for 48 h. Error bars = s.d., three technical replicates.

Source data are available online for this figure.

Discussion

Here, we demonstrate that intracellular RNA m6A decay cannot induce free 6mdA species and therefore does not generate misincorporated DNA 6mdA in most of the tested mammalian cells, revealing a 2′‐deoxynucleotide pool sanitation mechanism for suppressing DNA misincorporation resulting from intracellular RNA m6A decay. Furthermore, the depletion of ADAL leads to the presence of the free 6mdA species in cellular nucleotide pool and the misincorporated DNA i6mdA in the genome. Overexpression of AK1 significantly promotes the generation of misincorporated i6mdA in the absence of ADAL, and the knockdown of AK1 almost completely eliminates 6mdA incorporation. The ADAL‐pivoted 2′‐deoxynucleotide pool sanitation suppresses aberrant DNA N⁶‐methyladenine incorporation and maintains epigenetic 6mdA integrity.

By using [¹⁵N₅]‐rA, intracellular RNA m6A can be efficiently labeled and traced. If the degraded RNA m6A could be reformulated to generate 6mdATP via the purine salvage pathway, we should observe the labeled DNA 6mdA (Appendix Fig S12). However, we did not observe any labeled DNA 6mdA in mES cells and HEK293T cells over 7–50 days of [¹⁵N₅]‐rA tracing. Although it was reported that DNA misincorporation occurred at a delayed phase (approximately 5‐day delay) in comparison to RNA labeling (Musheev et al, 2020), our observation of the absence of the labeled DNA 6mdA (misincorporated i6mdA) over 50 days of labeling proves that the m6rA nucleotides generated from RNA m6A decay are hard to be reutilized in DNA replication in most tested cells (with an exception of NIH3T3 and C2C12 cells). By the use of the second strategy involving [D₃]‐L‐methionine tracing, we also observed the labeled RNA m6A ([D₃]‐m6A) but consistently did not observe any labeled DNA 6mdA after 7 or more days of tracing. Notably, both heavy stable isotope tracing strategies showed high labeling efficiency (~60%). Collectively, the above data strongly indicated that intracellular RNA m6A decay is hard to induce misincorporated DNA i6mdA. These results drove us to propose the existence of a 2′‐deoxynucleotide pool sanitation mechanism, which blocks the reformulation of intracellular RNA m6A degradation products to form premethylated 6mdATP and thus eliminates DNA 6mdA misincorporation.

Essentially, the depletion of the deaminase Adal by either knockdown or knockout resulted in the presence of the labeled DNA 6mdA in mES cells, and we could detect free labeled 6mdA species (6mdA, 6mdAMP, 6mdADP, and 6mdATP) in cellular nucleotide pool of ADAL‐KO mES cells. Consistently, both the transcriptional and protein expressions of ADAL in misincorporated i6mA‐detectable NIH3T3 and C2C12 cells were downregulated. The overexpression and knockdown of ADAL remarkably reduced and increased the misincorporated DNA 6mdA in NIH3T3 cells, respectively. Mechanistically, ADAL simultaneously catabolizes the methylated nucleotides m6rAMP and 6mdAMP to the nontoxic nucleotides IMP and dIMP, respectively, eliminating free 6mdA species and blocking the misincorporation of N⁶‐methyladenine via DNA polymerase; thus, the hydrolysis of both m6rAMP and 6mdAMP by ADAL contributes critically to the blockade of replication‐dependent 6mdA incorporation. Moreover, the catabolizing activity of ADAL toward 6mdAMP was 1.14‐fold as high as that toward m6rAMP. Of note, by detecting free 6mdA species in cellular nucleotide pool, we for the first time showed the activity of ADAL in catabolizing 6mdAMP in vivo. Collectively, our data suggested that ADAL functions as a key sanitation enzyme to block the catabolic conversion of intracellular RNA m6A to genomic DNA 6mdA by catabolizing both m6rAMP and 6mdAMP, particularly the latter.

Contrary to the deaminase ADAL, AK1 was identified as a protein involved in the 6mdA incorporation. Overexpression of AK1 did not induce any detectable labeling of DNA i6mdA in the presence of sufficient ADAL but increased the labeled DNA i6mdA in the absence of ADAL. This finding suggests that the phosphorylation of 6mdAMP into 6mdADP is a much slower process than the hydrolysis of 6mdAMP mediated by ADAL. In other words, the preferential substrate for AK1 is not 6mdAMP but dAMP, which is required for DNA synthesis. Overexpression of AK1 can partially compensate for the slow phosphorylation of 6mdAMP by AK1. The slow phosphorylation process would provide a time window long enough for ADAL to catabolize the generated 6mdAMP completely. Moreover, the knockdown of AK1 nearly blocks 6mdA incorporation in NIH3T3 cells, suggesting that AK1 plays an irreplaceable role in i6mdA incorporation. Apart from ADAL, there should be other regulation factors participating in establishing the sanitation. For example, MutT homolog 1 (MTH1), which could efficiently catalyze the hydrolysis of 6mdATP to 6mdAMP in vitro and in vivo, was proved to prevent the incorporation of 6mdATP into DNA and protect the epigenetic state (Scaletti et al, 2020).

It has been reported that NIH3T3 cells have exceptionally high levels of misincorporated 6mdA, approximately 5.0 6mdAs per 10⁴ dA (Musheev et al, 2020), which is comparable to that of mRNA m6A (~30.0 m6A per 10⁴ rA). However, we only observed 6mdA at a level 100–400‐fold lower in NIH3T3 cells (Fig 1C and F). Moreover, we also repeatedly quantified the i6mdA levels of C2C12 and NIH3T3 cells following the methods of (Musheev et al, 2020). The results (0.14 6mdA per 10⁶ dC in C2C12 and 0.42 6mdA per 10⁶ dC in NIH3T3 cells) were 90–1,000‐fold lower than those in the previous report (Appendix Fig S13A and B). The observation of rare misincorporated DNA 6mdA is rational and supports that the misincorporation of DNA 6mdA must be stringently controlled.

The findings on epigenetic mark DNA 6mdA remain elusive due to the analytical bottlenecks in 6mdA detection, such as contamination of coexisting bacteria carrying abundant 6mdA and the absence of reliable 6mdA sequencing technology (Schiffers et al, 2017; O'Brown et al, 2019; Douvlataniotis et al, 2020). However, it is surprising that misincorporated i6mdA must be stringently controlled, while epigenetic 6mdA is rarely found in human cells (Kong et al, 2022). These conflicting observations may hint that true epigenetic 6mdA (methylase‐deposited 6mdA) possess extremely important functions and appear in every cell in certain scenarios, for example, responses to certain stimuli. Otherwise, the mammalian cells should not set such a stringent bar for misincorporation. Interestingly, as the most important and abundant epigenetic mark, the misincorporation of 5mC must also be strictly controlled (Jekunen et al, 1983; Vilpo & Vilpo, 1993). By contrast, DNA 5‐hydroxymethylcytosine (5hmC) is found only in limited cells at moderate abundance and can be misincorporated into genomic DNA (Zauri et al, 2015). In addition, TCGA datasets show that patients with low ADAL expression in tumor tissues exhibit worse survival rates than those with high ADAL expression in some cancers (Appendix Fig S14A–D), suggesting a possible relevance between Adal expression, 6mdA misincorporation, and the survival of cancer patients. These data and underlying logics allow us to infer that the ADAL‐pivoted sanitation and the suppression of 6mdA misincorporation have biological importance, for example, in tumor development and therapy.

In summary, by exploiting a unique heavy stable isotope tracing technology coupled with UHPLC–MS/MS analysis, we found a 2′‐deoxynucleotide pool sanitation mechanism that functions to block potential DNA 6mdA misincorporation. This mechanism involves ADAL‐mediated hydrolysis of both m6rAMP and 6mdAMP, AK1‐mediated restriction of the phosphorylation of 6mdAMP, and other regulatory steps (Appendix Fig S12). The findings will be helpful to the understanding of the mechanism of epigenetic DNA landscape maintenance.

Materials and Methods

Reagents and materials

[¹⁵N₅]‐adenosine ([¹⁵N₅]‐rA), [¹⁵N₃]‐deoxycytidine ([¹⁵N₃]‐dC), [¹⁵N₃]‐cytidine ([¹⁵N₃]‐rC), and [D₃]‐L‐methionine were ordered from Cambridge Isotope Laboratories Inc. (Andover, MA, USA). Adenosine (rA), 2′‐deoxycytidine (dC), cytidine (rC), N⁶‐methyl‐2′‐deoxyadenosine (6mdA), N⁶‐methyladenosine (m6rA), 5‐methylcytidine (m5C), phosphodiesterase I (SVP), nuclease P1 (NP1), cpd3 ((S)‐6‐(4‐(((1‐(isoquinolin‐8‐yl)ethyl)amino)methyl)phenyl)nicotinic acid), and DCF (2′‐deoxycoformycin) were purchased from Sigma‐Aldrich Inc. (St. Louis, MO, USA). [¹⁵N₅]‐6mdA, [¹⁵N₄]‐deoxyinosine ([¹⁵N₄]‐dI), and [¹⁵N₄]‐inosine ([¹⁵N₄]‐rI) were synthesized in our own lab. Benzonase was purchased from Sino Biological Inc. (Beijing, China). Calf intestinal alkaline phosphatase (CIP) was ordered from New England Biolabs Inc. (Ipswich, MA, USA). TRIzol reagent was purchased from Thermo Fisher Scientific Inc. (Waltham, MA, USA). The primary antibodies for ADAL (27369‐1‐AP) and AK1 (14978‐1‐AP) were ordered from Proteintech Group Inc. (Wuhan, China). The anti‐rabbit IgG (HRP‐linked, 7074S) secondary antibody was purchased from Cell Signaling Technology Inc. (Danvers, MA, USA).

Cell culture and stable isotope tracer labeling

Mouse embryonic stem cells (129SvEv) were seeded in the dishes coated with 0.1% sterilized gelatin aqueous solution and cultured in high‐glucose DMEM (Thermo) supplemented with 20% FBS (Thermo), 0.35% penicillin–streptomycin solution (Thermo), 1,000 U/ml mouse leukemia inhibitory factor (mLIF) (Millipore, Germany), 100 μM β‐mercaptoethanol, 3.0 μM Stemolecule CHIR99021 (Stemgent, USA), 1 μM Stemolecule PD0325901 (Stemgent), 100 μM nonessential amino acids (Thermo), 2 mM L‐glutamine (Thermo), and 1 mM sodium pyruvate (Thermo). HEK293T, NIH3T3, and C2C12 cells were cultured in high‐glucose DMEM (Thermo) supplemented with 10% FBS (Thermo) and 1% penicillin–streptomycin solution (Thermo). All of the cell lines were confirmed to be free of mycoplasma contamination and were incubated at 37°C in a bacteria‐free incubator with 5% CO₂.

To be labeled with stable isotope tracers [¹⁵N₅]‐rA or [D₃]‐L‐methionine, cells were seeded into gelatin‐coated 6‐well cell culture clusters at a density of 10⁵ per well. Then, 20 μM [¹⁵N₅]‐rA or 30 μg/ml [D₃]‐L‐methionine was added to the culture medium, and the medium was replaced with fresh medium containing [¹⁵N₅]‐rA or [D₃]‐L‐methionine every 24 h. This treatment was performed over 7 days.

DNA extraction and enzymolysis

Cells to be analyzed were harvested and collected into 1.5 ml aseptic centrifuge tubes. Then, genomic DNA was extracted using a Promega Wizard® Genomic DNA Purification Kit (Promega, USA) according to the manufacturer's instructions on a dedicated sterile operation platform. Notably, to avoid possible contamination, a series of cleaning processes were performed to remove 6mdA nucleosides in the extraction reagents and buffers. The newly extracted DNA was quantitated by the absorbance at UV 260 nm using a NanoDrop 2000 (Thermo).

Before UHPLC–MS/MS analysis, 5.0 μg of DNA was dissolved in 50 μl of enzymatic digestion buffer (10 mM Tris–HCl (pH 8.0) and 1 mM Mg²⁺) and then digested to 2′‐deoxynucleosides with an enzyme mixture of 1 U of benzonase, 0.02 U of SVP, and 1 U of CIP at 37°C for 12 h. Next, the enzymes were removed by ultrafiltration (molecular weight cutoff: 3 kDa; Pall Corporation, USA). The ultrafiltered solutions containing 2′‐deoxynucleosides were subjected to UHPLC–MS/MS analysis.

RNA (and mRNA) extraction and enzymolysis

Total RNA was extracted with TRIzol Reagent (Thermo) according to the manufacturer's instructions. Cells were harvested and collected into 1.5 ml aseptic centrifuge tubes on ice. Then, the cell pellets were lysed and homogenized in TRIzol reagent and immediately subjected to chloroform extraction and ethanol precipitation. The extracted total RNA was dissolved in DEPC‐treated water and quantitated using a NanoDrop 2000 (Thermo). Then, mRNA was extracted from the total RNA using a Dynabeads™ mRNA Purification Kit (Thermo) according to the manufacturer's instructions.

Enzymatic digestion of RNA (2.0 μg in 50 μl enzymatic digestion buffer) was conducted with a mixture of 1 U of benzonase, 0.25 U of NP1, 0.02 U of SVP, and 1 U of CIP in Tris–HCl buffer (10 mM Tris–HCl, pH 8.0, plus 1 mM Mg²⁺) at 37°C for 6 h. Next, the enzymes were removed by ultrafiltration (molecular weight cutoff: 3 kDa; Pall Corporation, USA). The filtered solution containing nucleosides was subjected to UHPLC–MS/MS analysis.

siRNA transfection

Cells were seeded into 6‐well cell culture clusters at a density of 10⁵ per well the night before transfection with 30 pmol of siRNA using Lipofectamine RNAiMAX Transfection Reagent (Thermo) according to the manufacturer's protocols. Control cells were transfected with 30 pmol of negative control sequence. The transfected cells were cultured for another 36 h before RT–qPCR analysis or 48 h before western blot analysis. The siRNA sequences were listed in Appendix Table S1.

Gene overexpression

Gene overexpression transfection was performed with Lipofectamine^® 2000 Reagent (Thermo) for mES cells or Lipofectamine™ LTX Reagent with PLUS™ Reagent (Thermo) for NIH3T3 cells according to the manufacturer's protocols. Gene‐overexpressing plasmids were constructed on the PcDNA3‐flag vector and transfected into cells at a density of 2 × 10⁵ per well in 6‐well cell culture clusters. The negative control group was transfected with the PcDNA3‐flag vector. The transfected cells were cultured for 48 h before western blot analysis. The primers for the construction of expression plasmids were listed in Appendix Table S3.

Gene knockout

Guide RNA (gRNA) fragments for Adal‐KO (forward sequence: 5′‐caccGAGGTGGCCCAACAATAGCCA‐3′; reverse sequence: 5′‐aaacTGGCTATTGTTGGGCCACCTc‐3′) were designed and synthesized with an online tool available at https://www.synthego.com/products/bioinformatics/crispr-design-tool. Then, the gRNAs were cloned into PX458 vectors with a GFP label. mES cells were seeded into 6‐well cell culture clusters at a density of 10⁵ per well before transfection with 2.5 μg of constructed gRNA‐PX458 plasmids using Lipofectamine 2000 Transfection Reagent (Thermo). After culture for 48 h, transfected mES cells were harvested and sorted with a BD FACSAria II flow cytometer (Becton Dickinson, USA). The cells with green fluorescence were collected, diluted into single cells, and cultured for another 7 days. Then, gene sequencing identification and western blot analysis were performed to identify gene‐deficient cell lines.

Western blot analysis

Total protein was extracted with RIPA lysis buffer (Beyotime, China) following the manufacturer's instructions and quantified using a BCA Protein Assay Kit (Beyotime). Then, 30 μg of protein from each sample was separated on 4–12% NUPAGE Bis–Tris gels (Thermo), and the separated proteins were subsequently transferred to polyvinylidene difluoride membranes (Millipore). After they were blocked with 5.0% skimmed milk powder agent for 1 h, the membranes loaded with protein were incubated with a diluted (1:1,000) primary antibody solution overnight at 4°C and subsequently incubated with appropriately diluted (1:1,500) secondary antibodies (HRP‐linked) for another 2 h at room temperature. Finally, the target protein bands were detected with a Multifunctional Chemiluminescence Imaging System (Alliance Q9 ATOM, UVITEC, France). GAPDH was used as a reference control in quantitative analysis.

The extraction of cellular nucleotide pool for detection of free m6rA and 6mdA species

[D₃]‐methionine‐traced cells were collected and washed with PBS twice. Then about 2 × 10⁷ cells were resuspended in 400 μl deionized water, and the cell suspensions were subjected to thermal shock with three cycles of freezing in liquid nitrogen and thawing at 95°C. After thermal shock, the suspensions were centrifuged at 16,000 × g for 15 min at 4°C, and the supernatants were ultrafiltered with ultrafiltration tubes (molecule weight cutoff: 3 kDa; Pall Corporation). The ultrafiltered solution was divided into two aliquots. One aliquot was further treated with calf intestinal alkaline phosphatase (CIP, NEB) for the dephosphorylation of 6mdANPs, transforming total free 6mdA species into 6mdA nucleosides. The analytical sensitivity of MS toward nucleosides is greater than toward corresponding nucleotides. Finally, the CIP‐untreated (CIP‐) and CIP‐treated (CIP+) extracts were subjected to UHPLC–MS/MS assay for the quantification of 6mdA, m6rA, dC, and rC with corresponding external isotopic standards. [¹⁵N₅]‐6mdA standard was used as an external standard spiked into samples for 6mdA analysis.

UHPLC–MS/MS analysis

UHPLC–MS/MS analysis was performed as described previously (Liu et al, 2021). For 6mdA analysis, 1 μg of DNA enzymolysis product was injected into an Agilent 1,290 II UHPLC system coupled with an electrospray ionization (ESI)‐triple quadrupole mass spectrometer (6,495, Agilent Technologies, Santa Clara, CA). A Zorbax SB‐Aq column (2.1 × 50 mm, 1.8 μm particle size, Agilent, USA) was used for separation. The mobile phase consisted of solvent A (2 mM ammonium bicarbonate aqueous solution) and solvent B (pure methanol) at a flow rate of 0.35 ml/min, and the gradient elution was as follows: 0.0–1.0 min, 10% B; 1.0–5.0 min, 10–40% B; 5.0–6.0 min, 40–80% B; 6.0–8.0 min, 80% B; 8.0–8.2 min, 80–10% B; and 8.2–12 min, 10% B. For m6rA analysis, 0.2 μg of RNA enzymolysis product was injected into UHPLC–MS/MS. The separation column and mobile phase were the same as those used in the 6mdA analysis, and the gradient elution was as follows: 0.0–2.0 min, 5% B; 2.0–4.0 min, 5–20% B; 4.0–5.0 min, 20% B; 5.0–5.1 min, 20–5% B; 5.1–6.5 min, 5% B. The quantification of dC, rC, rA, and m5C was performed using a 1,000 times diluted sample. [¹⁵N₃]‐dC and [¹⁵N₃]‐rC were used as internal standard spiked into samples.

The selective MRM transitions for DNA 2′‐deoxynucleoside analysis were as follows: dC: m/z 228 → 112, [¹⁵N₃]‐dC: m/z 231 → 115, 6mdA: m/z 266 → 150, [D₃]‐6mdA: m/z 269 → 153, [¹⁵N₄]‐6mdA: m/z 270 → 154, [¹⁵N₅]‐6mdA: m/z 271 → 151, dI: m/z 253 → 137, and [¹⁵N₄]‐dI: m/z 257 → 141.

Those for RNA nucleoside analysis were as follows: rA: m/z 268 → 136, [¹⁵N₄]‐rA: m/z 272 → 140, [¹⁵N₅]‐rA: m/z 273 → 141, rC: m/z 244 → 112, [¹⁵N₃]‐rC: m/z 247 → 115, m5C: m/z 258 → 126, [D₃]‐m5C: m/z 261 → 136, m6rA: m/z 282 → 150, [D₃]‐m6rA: m/z 285 → 153, [¹⁵N₄]‐m6rA: m/z 286 → 154, [¹⁵N₅]‐m6rA: m/z 287 → 154, rI: m/z 269 → 137, and [¹⁵N₄]‐rI: m/z 273 → 141. The fragmentation voltage for all the MRM transitions was set at 90 V, and the nebulization gas pressure was set at 40 psi. The other conditions were as described previously (Liu et al, 2021).

Activity analysis of ADAL in vitro

An ADAL fusion protein whose N‐terminus was tagged with maltose‐binding protein (MBP) was expressed in E. coli and then purified via a prepacked hydrophobic interaction chromatography column and a prepacked MBP‐Trap HP 5 ml column (GE Healthcare, Uppsala, Sweden). Then, the purity of the isolated ADAL protein was analyzed via SDS–polyacrylamide gel electrophoresis (PAGE), and the corresponding quantification was achieved with a BCA Protein Assay Kit (Beyotime). Next, we prepared 6mdAMP by digesting plasmid DNA that carries abundant DNA 6mdA. The plasmid DNA was extracted from E. coli. Ten microgram of plasmid DNA was digested in a 100 μl reaction mixture containing 10 mM Tris–HCl (pH 8.0), 2 mM Mg²⁺, 1 U of benzonase, and 0.02 U of SVP. After 6 h of incubation, the mixture was filtered by centrifugation through an ultrafiltration tube (MW cutoff: 3 kDa; Pall Corporation, USA). Then, the filtrated solution was mixed with 400 nM m6rAMP. By this protocol, solutions containing two ADAL substrates, 6mdAMP and m6rAMP, were prepared. For characterization of the preparations, 1.5 μg of the product that had been treated with 1.0 U CIP for 1 h and an equivalent amount of untreated product were analyzed by UHPLC–MS/MS for the quantification of 6mdA or m6rA. The difference in 6mdA or m6rA between the CIP‐treated and untreated products represented the content of monophosphorylated N⁶‐methyl‐2′‐deoxyadenine (6mdAMP) or N⁶‐methyladenine (m6rAMP).

The ADAL activity assay was conducted in 50 μl reaction buffer (20 mM Tris–HCl (pH 7.0), 0.1 μg/μl BSA, 2 mM DTT) supplemented with 1.5 μg of the above enzymatic products containing m6rAMP and 6mdAMP and 0 or 0.5 μg of ADAL protein at 37°C for 1 h. After the reactions, the solutions were treated with proteinase K at 55°C for 30 min and heated for enzyme inactivation at 95°C for 10 min and then incubated with 5.0 U CIP for another 1 h. The reaction substrates m6rAMP and 6mdAMP and the products rIMP and dIMP were converted into m6rA, 6mdA, rI, and dI, respectively, for UHPLC–MS/MS analysis. After passing through ultrafiltration tubes (MW cutoff: 3 kDa), the nucleoside solution was subjected to UHPLC–MS/MS analysis.

The catalytic efficiency (K_cat/K_m) of ADAL for m6rAMP and 6mdAMP was analyzed in a 50 μl reaction mixture of 20 mM Tris–HCl (pH 7.0), 0.1 μg/μl BSA, 2 mM DTT, 0.2 μg of ADAL protein, and varying concentrations of the test compound. After incubation at 37°C for 10 min, the reaction mixture was incubated with 5.0 U of CIP for another 1 h to convert all of the nucleotides in the mixture into nucleosides. UHPLC–MS/MS was performed for quantitative analysis of products dI and rI.

Author contributions

Shaokun Chen: Resources; data curation; formal analysis; validation; investigation; visualization; methodology; writing – original draft. Weiyi Lai: Software; supervision; validation; methodology. Yanan Li: Methodology. Yan Liu: Methodology. Jie Jiang: Methodology. Xiangjun Li: Resources; supervision; validation; project administration; writing – review and editing. Guibin Jiang: Resources; supervision. Hailin Wang: Conceptualization; resources; supervision; funding acquisition; validation; writing – original draft; project administration; writing – review and editing.

Disclosure and competing interests statement

The authors declare that they have no conflict of interest.

Supporting information

Appendix

Source Data for Figure 1

Source Data for Figure 2

Source Data for Figure 3

Source Data for Figure 4

Source Data for Figure 5

The EMBO Journal (2023) 42: e113684

Data availability

All data needed to evaluate the conclusions in the paper are present in the paper and/or the supplementary materials. This study includes no data deposited in external repositories. Additional data related to this paper may be requested from the authors.