A tale of two strands: Decoding chromatin replication through strand-specific sequencing

Abstract

DNA replication, a fundamental process in all living organisms, proceeds with continuous synthesis of the leading strand by DNA polymerase ε (Pol ε) and discontinuous synthesis of the lagging strand by polymerase δ (Pol δ). This inherent asymmetry at each replication fork necessitates the development of methods to distinguish between these two nascent strands in vivo. Over the past decade, strand-specific sequencing strategies, such as enrichment and sequencing of protein-associated nascent DNA (eSPAN) and Okazaki fragment sequencing (OK-seq), have become essential tools for studying chromatin replication in eukaryotic cells. In this review, we outline the foundational principles underlying these methodologies and summarize key mechanistic insights into DNA replication, parental histone transfer, epigenetic inheritance, and beyond, gained through their applications. Finally, we discuss the limitations and challenges of current techniques, highlighting the need for further technological innovations to better understand the dynamics and regulation of chromatin replication in eukaryotic cells.

Keywords: DNA damage repair; DNA replication; epigenetic inheritance; genome maintenance; nucleosome assembly; parental histone transfer; strand-specific sequencing.

Figure 1.

General principles of library preparation for strand-specific DNA sequencing. Commonly used methods to preserve strand-specificity during library preparation include i) marking 3’-end of ssDNA, first developed for the sequencing of ancient DNA; ii) splinted adapter ligation to ssDNA in a single reaction, widely used by methods like OK-seq; iii) ligation of hairpin or Y-shaped adapters to dsDNA, adopted by various methods, such as SCAR-seq and END-seq; and iv) tagmentation-based methods that rely on oligo-replacement, widely used by methods including eSPAN and ssMeDIP-seq. After strand separation and PCR amplification, DNA fragments with differentially marked 5’ and 3’ ends are barcoded and subjected to next generation sequencing. The mapped reads are aligned to the reference genome and used for strand-specific analysis.

Figure 2.

Selected strand-specific sequencing methods to probe chromatin replication. (A) Simplified workflow of OK-seq, which purifies and sequences short nascent Okazaki fragments. (B) eSPAN originally devised in yeast, which combines ChIP and BrdU IP for detection of proteins at nascent chromatin, with Pol ε and Pol δ eSPAN as examples. (C) eSPAN in mammalian cells, which combines CUT&Tag and BrdU IP for detection of proteins at nascent chromatin. (D) Replicon-seq, which combines targeted MNase digestion and long-read sequencing to detect replicons at the single-molecule level. See text for more detailed descriptions.

Figure 3.

Parental histone transfer during DNA replication. (A) Replication-coupled distribution of parental histones is mediated by pathways centered around the replisome. Pol ε mediates parental histone transfer to the leading strand, whereas parental histone transfer to the lagging strand involves several chaperones, including MCM2, Pol α, Pol δ, and various co-factors (Ctf4, Tof1, Mrc1, PCNA, FACT and others). Importantly, the FACT complex co-chaperones with MCM2, and binds to histone hexamers, consisting of one H3–H4 tetramer and one H2A-H2B dimer evicted from parental nucleosomes. Mrc1 directly binds H3–H4 tetramer and promotes its deposition onto both the leading and lagging strands. Whether and how Pol α, Pol δ and Pol ε bind to histone H3–H4 tetramers or hexamers remain to be structurally determined. Please note that factors involved in the deposition of newly synthesized histones are not depicted in this diagram for simplification. (B) Asymmetric transfer of parental H3K9me3 to head-on L1 elements at the leading strand. The HUSH complex, comprising TASOR, MPP8 and PPHLN1, works with SETDB1, an H3K9 methyltransferase, to establish H3K9me3 at L1 retrotransposons. During DNA replication, HUSH and Pol ε interact with each other and coordinate parental H3K9me3 deposition to the leading strand, while HUSH travels along the leading strand to propagate H3K9me3 asymmetry. Low level local transcription likely plays a role in H3K9me3 asymmetry, and H3K9me3 on the leading strand likely provides a template for the restoration of H3K9me3 symmetry post replication.

Figure 4.

Insights into genome maintenance revealed by strand-specific sequencing methods. (A) Upon replication stress, replicative helicases and DNA polymerases are uncoupled, leading to fork arrest and accumulation of ssDNA, which is coated by RPA. ATR/Mec1 is subsequently activated by RPA, ATRIP, ETAA1 and other factors. The downstream kinase CHK1/Rad53 is then activated by Mec1 through the adaptor protein Claspin/Mrc1. Activated Rad53 in turn attenuates the replication function of Mrc1, which slows down the CMG helicase, thereby coordinating DNA synthesis on the leading and lagging strands during stress response. (B) Cohesin is a multi-subunit protein complex, including SMC1/3 and various regulatory factors, which form a ring-like structure encircling the sister chromatids. Loading of PCNA to the leading and lagging strands is preferentially mediated by Ctf18-RFC and Rfc1-RFC, respectively. PCNA loading to the leading strand by Ctf18-RFC recruits Eco1, which acetylates SMC3 and promotes sister chromatid cohesion. PCNA unloading from the lagging strand is primarily mediated by the Elg1-RFC complex and regulated by Rad6-Rad18 mediated ubiquitylation. (C) Different end structures form at the nCas9-induced nicks on the leading and lagging template strands, as revealed by END-seq. A nick on the leading template strand generates a single-ended DSB (seDSB), leading to loss of the CMG helicase. A nick at the lagging template strands produces a double-ended DSB (deDSB) as the fork bypasses, where CMG is retained and DNA synthesis continues. Both forms of DSBs are repaired by HR, which depends on BRCA1 for RAD51 loading, but not for end resection, a distinction from the repair of canonical DSBs.