LAUSR

Guanine-rich regions of DNA can fold into a particular structure called a G-quadruplex (G4) (Figure 1, A). The G4 structure is stabilized by Hoogsteen hydrogen bonds that form between the guanine bases (Moon et al., 2015). Previous studies have proposed a role for G4s in diverse biological processes, such as DNA replication, telomerase maintenance, and gene regulation (Bao et al., 2019; Sparks et al., 2019; Lago et al., 2021). G4s play a controversial role in regulating transcription; they have been associated with both high and low transcription activity depending on transcription factor recruitment, template, or the coding strand that possesses the G4 (Raiber et al., 2012; Lago et al., 2021). Over the last few decades, computational sequence analyses and high-throughput methods have been developed to investigate the genome-wide distribution of G4s in mammals. Although putative G4-forming sequences have been computationally recognized in plant genomes, their experimental validation and functional characterization are lacking (Ge et al., 2019; Cagirici and Sen, 2020; Cagirici et al., 2021). In this issue of Plant Physiology, Feng et al. (2021) used the BG4 engineered antibody, which has a high affinity for G4, to perform BG4-DNA-IP-seq analysis to identify and characterize G4-folded regions in the rice (Oryza sativa) genome. They also investigated the relationship between G4s and epigenomic features to understand the roles of G4s in gene expression. The authors reported that G4s in the promoters of genes are positively corresponded correlated with gene expression but act as repressors when located in the gene body. They also showed that specific chromatin features are enriched at G4 regions. Their experimental G4 maps serve as a valuable dataset for further investigation of G4 function in rice (Feng et al., 2021). Using chromatin immunoprecipitation with BG4 antibodies under G4-favorable conditions followed by high-throughput sequencing, 23,685 G4 peaks, termed dG4s (experimentally detectable G4s), were identified in the rice genome. The accuracy of dG4 localization was confirmed by comparison with putative G4-forming sequences (PQFS) identified in rice by using a previously published script (Fujimoto et al., 2020). Different PGFS patterns were associated with dG4 peaks, demonstrating effective binding of BG4 antibodies to G4-folded regions of the genome. Among these computationally predicted PQFSs, only 5% were detectable in dG4. The 95% predicted but experimentally undetectable G4s (udG4s) were further investigated to understand the difference in their genomic distribution and sequence features compared with dG4. Feng et al. (2021) found that dG4s with PQFS were primarily enriched in promoters and 50-UTRs, while 34.6% of the udG4s were distributed in exons. Also, dG4s exhibited higher GC content and GC/AT skews than udG4s. Finally, the computationally identified sizes of the PQFSs were longer for dG4s than udG4s, and there was less distance between PQFS in dG4s than udG4s. These analyses may explain why the high percentage of predicted G4s is undetectable by BG4 assay; PQFSs in dG4s are more likely to form G4s or could recruit more BG4 compared with udG4s. Another possible explanation could be that the BG4 assay might not efficiently detect all dG4s. To test the effect of G4s on DNA methylation of the rice genome, the authors analyzed publicly available bisulfite sequencing data and found that dG4s were hypomethylated at CG and CHG sites but hypermethylated in CHH sites. Furthermore, the group examined DNA N6-methyladenine (DNA-6mA) levels, known to be associated with gene expression, in both dG4 and udG4, and found higher levels of D-6mA in dG4s compared with udG4. In addition, treating rice seedlings with zebularine, a DNA demethylation agent, N ew s an d V ie w s