Splice site m⁶A methylation prevents binding of DGCR8 to suppress KRT4 pre-mRNA splicing in oral squamous cell carcinoma

Cell culture

Normal oral keratinocytes (NOK) and OSCC cell line HN6 cells were obtained from Cell Bank at the Chinese Academy of Sciences (Shanghai, China) and cultured as previously described (Li et al., 2022).

Cell transfection

Cells were transfected with METTL3, METTL14 or DGCR8 siRNA and siRNA negative control (NC) by Lipofectamine 2000 (Invitrogen, Carlsbad, CA, USA) as previously described (Li et al., 2022). Then cells were collected for following experiments at 48 h after transfection. Sequences of siRNAs were listed in Table 1.

Bioinformatics analysis

RMBase v2.0 database (https://rna.sysu.edu.cn/rmbase/) and RMVar database (https://rmvar.renlab.org/) were used to explore potential m⁶A modification sites in or nearby KRT4 pre-mRNA splicing sites. Besides, ENCORI database (https://starbase.sysu.edu.cn/index.php) was utilized to mine RNA-protein interactions.

Touchdown polymerase chain reaction (PCR)

Touchdown PCR was utilized to detect KRT4 pre-mRNA splicing. First, total RNA was extracted from cells by TRIzol reagent (Invitrogen) followed by cDNA synthesis using PrimeScript RT Reagent Kit (Takara, Dalian, Liaoning, China). Subsequently, touchdown PCR was performed in a volume of 50 µL using Phanta Max Buffer (Vazyme, Nanjing, Jiangsu, China) as follows: 95 °C for 3 min, 95 °C for 15 s, 74 °C for 90 s for 5 cycles; 95 °C for 15 s, 72 °C for 90 s for 5 cycles; 95 °C for 15 s, 70 °C for 90 s for 5 cycles; 95 °C for 15 s, 68 °C for 90 s for 25 cycles followed by 68 °C for 5min. Sequences of primers used for touchdown PCR were listed in Table 2.

Quantitative real-time polymerase chain reaction (qRT-PCR)

After RNA extraction and cDNA synthesis, qRT-PCR was performed by the ABI7300 Real-Time PCR system (Applied Biosystems, Foster City, CA, USA) using TB Green^® Premix Ex Taq™ II (Tli RNaseH Plus) (Takara) as previously described (Li et al., 2022). Sequences of primers used for qRT-PCR were listed in Table 3.

Methylated RNA immunoprecipitation (MeRIP)

NOK and HN6 cells were collected and lysed. Then nucleic acid fragments were interrupted by ultrasound. Next, cell lysate was incubated with 1 µL m⁶A antibody (1:500, #ab208577, Abcam, Cambridge, MA, USA) overnight at 4 °C. Subsequently, m⁶A antibody and methylated RNA fragments were captured by avidin magnetic beads, and the level of methylated RNA was detected by qRT-PCR.

RNA immunoprecipitation (RIP)

RIP was performed by RNA-Binding Protein Immunoprecipitation Kit (Merck Millipore, Billerica, MA, USA). Briefly, HN6 cells were collected and lysed. Then nucleic acid fragments were interrupted by ultrasound. Subsequently, cell lysate was incubated with 1 µL METTL3 (#15073-1-AP, Proteintech; Rosemont, IL, USA), METTL14 (#26158-1-AP; Proteintech) or DGCR8 antibody (#60084-1-Ig; Proteintech) at 4 °C overnight. Next, protein-binding RNA fragments were captured by avidin magnetic beads, and the level of protein-binding RNA was identified by qRT-PCR.

Statistical analysis

Quantitative data of the current study were present as mean ±  standard deviation (SD) and statistical differences were analyzed by SPSS 20 software (SPSS Inc., Chicago, IL, USA) as previously described (Li et al., 2022). Besides, P < 0.05 was considered as statistically significant.

Intron splicing of KRT4 pre-mRNA is suppressed in OSCC

To explore the mechanism downregulating KRT4 in OSCC, splicing of KRT4 pre-mRNA was detected by touchdown PCR. Compared to NOK cell, splicing of intron 2 between exon 2 and 3 (E2/E3), intron 3 between exon 3 and 4 (E3/E4), and intron 5 between exon 5 and 6 in KRT4 pre-mRNA (E5/E6) was inhibited in HN6 cells (Figs. 1A and 1B). Red arrows indicated fragments of KRT4 pre-mRNA containing exon-intron structures (Figs. 1A and 1B). Moreover, results of qRT-PCR confirmed that junctions of exon 2-exon 3 (E2/E3), exon 3-exon 4 (E3/E4) and exon 5-exon 6 (E5/E6) in KRT4 mRNA were dramatically decreased in HN6 cells compared to those in NOK cells, respectively (Fig. 1C). Therefore, these data suggested that intron splicing of KRT4 pre-mRNA was suppressed in OSCC.

m⁶A levels of exon-intron boundaries in KRT4 pre-mRNA is increased in OSCC

Bioinformatics analysis showed that there were m⁶A sites in exon 3 (KRT4-E3(4)) and exon 5 (KRT4-E5(6)) of KRT4 pre-mRNA nearby exon-intron boundaries (Fig. 2A). Besides, results of MeRIP found that m⁶A levels of m⁶A sites in exon 3 (KRT4-E3(4)) and exon 5 (KRT4-E5(6)) of KRT4 pre-mRNA were significantly elevated in HN6 cell compared to those in NOK cells (Figs. 2B and 2C). Thus, m⁶A methylation might be involved in the regulation of KRT4 pre-mRNA splicing in OSCC.

m⁶A methylation of exon-intron boundaries prevents intron splicing of KRT4 pre-mRNA in OSCC

Results of RIP performed by METTL3 and METTL 14 antibodies showed that m⁶A writers METTL3 and METTL14 associated with m⁶A sites in exon 3 (KRT4 pre-mRNA-E3(4)) and exon 5 (KRT4 pre-mRNA-E5(6)) of KRT4 pre-mRNA nearby exon-intron boundaries in HN6 cells (Fig. 3A), suggesting that METTL3 and METTL14 should modify m⁶A levels of m⁶A sites in exon 3 (KRT4 pre-mRNA -E3(4)) and exon 5 (KRT4 pre-mRNA -E5(6)) of KRT4 pre-mRNA in HN6 cells.

Next, METTL3 and METTL14 were silenced by siRNAs in HN6 cells (Fig. S1) to identify the effect of m⁶A methylation on intron splicing of KRT4 pre-mRNA in OSCC. Silence of METTL3 and METTL14 decreased m⁶A levels of m⁶A sites in exon 3 (KRT4 pre-mRNA -E3(4)) and exon 5 (KRT4 pre-mRNA -E5(6)) of KRT4 pre-mRNA in HN6 cells (Fig. 3B). Then intron splicing of KRT4 pre-mRNA was detected by qRT-PCR in HN6 cells. Results indicated that silence of METTL3 and METTL14 reduced structures of intron 2-exon 3 (KRT4-E(2)3), exon 3-intron 3 (KRT4-E(3)4) and exon 5-intron 5 (KRT4-E(5)6) in KRT4 pre-mRNA (Figs. 1A and 3C), which associated with junctions of exon 2-exon 3, exon 3-exon 4 and exon 5-exon 6 in KRT4 mRNA, respectively. By contrast, silence of METTL3 and METTL14 increased mature KRT4 mRNA level in HN6 cells (Fig. 3D). In addition, siRNA NC had no effect on levels of KRT4 pre-mRNA and mRNA (Figs. 3C and 3D).

Above results suggested that m⁶A methylation of exon-intron boundaries prevented intron splicing of KRT4 pre-mRNA in OSCC. Besides, these results also revealed that the m⁶A modification of m⁶A sites in exon 3 of KRT4 pre-mRNA nearby the exon-intron boundary (KRT4 pre-mRNA-E3(4)) could modify not only intron 2-exon 3 structure (KRT4-E(2)3) but also exon 3-intron 3 structure (KRT4-E(3)4). Therefore, silence of METTL3 and METTL14 by siRNAs could simultaneously reduce intron 2-exon 3 structure (KRT4-E(2)3) but also exon 3-intron 3 structure (KRT4-E(3)4).

m⁶A methylation inhibits the binding of DGCR8 to exon-intron boundaries in KRT4 pre-mRNA in OSCC

Bioinformatics analysis of CLIP-seq data from ENCORI database further found that splicing factor DGCR8 could bind to KRT4 mRNA in cancer cells (Fig. 4A). Furthermore, results of RIP demonstrated that DGCR8 associated with boundaries of intron 2-exon 3 (KRT4-E(2)3), exon 3-intron 3 (KRT4-E(3)4) and exon 5-intron 5 (KRT4-E(5)6) in KRT4 pre-mRNA in HN6 cells (Figs. 4B–4D), suggesting that DGCR8 might regulate KRT4 pre-mRNA splicing in HN6 cells. Moreover, silence of METTL3 and METTL14 facilitated the binding of DGCR8 to boundaries of intron 2-exon 3 (KRT4-E(2)3), exon 3-intron 3 (KRT4-E(3)4) and exon 5-intron 5 (KRT4-E(5)6) in KRT4 pre-mRNA (Figs. 4B–4D). All these data together suggested that m⁶A methylation suppressed the binding of DGCR8 to exon-intron boundaries in KRT4 pre-mRNA in OSCC.

Silence of DGCR8 prohibits intron splicing of KRT4 pre-mRNA in OSCC

Next, the role of DGCR8 in intron splicing of KRT4 pre-mRNA was demonstrated in HN6 cells. Results of qRT-PCR showed that silence of DGCR8 by siRNA (Fig. S1) increased structures of intron 2-exon 3 (KRT4-E(2)3), exon 3-intron 3 (KRT4-E(3)4) and exon 5-intron 5 (KRT4-E(5)6) in KRT4 pre-mRNA (Fig. 5A). In contrast, silence of DGCR8 decreased mature KRT4 mRNA level in HN6 cells (Fig. 5B). Besides, siRNA NC had no effect on levels of KRT4 pre-mRNA and mRNA (Figs. 5A and 5B). Thus, these results indicated that silence of DGCR8 prohibited intron splicing of KRT4 pre-mRNA in OSCC.