Faculty Summaries
Hong Yan, PhD
Hong Yan, PhD
Associate Professor
Hong.Yan@fccc.edu
Office Phone: 215-728-2514
Fax: 215-728-3616
Office: R304
  • 1. Analysis of The Mechanism for The 5' -> 3' Strand Processing at DNA Double-Strand Breaks
    A conserved pathway for the 5' strand-specific processing at DSBs
    A conserved pathway for the 5' strand-specific processing at DSBs

    Three major pathways have been identified to repair DSBs in eukaryotes: 1) non-homologous end joining (NHEJ); 2) homologous recombination (HR); and 3) single-strand annealing (SSA). NHEJ is an error-prone pathway that re-ligates DNA ends, usually after the ends are made ligatable by other enzymes like DNA polymerases or nucleases. HR uses a homologous sequence on the sister chromatid or a homologous chromosome as the template to repair the missing information. SSA occurs when the break is between two direct repeat sequences and the final result is effectively the deletion of one of the two repeats. HR and SSA are both dependent on sequence homology and often collectively termed as homology-dependent DSB repair. NHEJ has been well studied by a combination of genetic analysis in vivo and biochemical reconstitution in vitro. In contrast, our understanding of homology-dependent DSB (HR and SSA) repair has been more limited due to the lack of an in vitro system that can reconstitute these two pathways. Using the nucleoplasmic extracts (NPE) derived from nuclei reconstituted in Xenopus egg extracts, we have succeeded in establishing SSA repair between two repeats at the ends of DNA. Moreover, we have found that the Xenopus Werner syndrome protein (xWRN or FFA-1) plays an important role in SSA. WRN, a member of the highly-conserved RecQ helicase family, has been implicated in replication fork restart and DSB repair, but its exact role is poorly understood. Our study provides the first evidence that directly links WRN to a particular DSB repair pathway.

    Further analysis of xWRN’s mechanistic role in SSA has led us to the discovery of a novel pathway for the 5'-> 3' processing of DNA ends. The first step for the homology-dependent DSB repair pathways is the strand-specific processing of a DNA end to generate a 3' single-strand (ss-) tail. The 3' ss-tail then either invades a homologous template to form a D-loop (for HR) or anneal with the complementary 3' ss-tail from the other end (for SSA). Despite the pivotal role of end processing in homology-dependent DSB repair, the underlying mechanism is only beginning to be revealed. We have found that DNA ends are rapidly processed into 3' ss-tails in NPE. The reaction can be divided into two steps: first the unwinding of ds-DNA ends and then the degradation of 5' ss-tails. The unwinding step is dependent on ATP, but the 5' ss-tail degradation step is not. More importantly, we have found that xWRN plays an important role in DNA end processing by promoting end unwinding but not 5' tail degradation. There is no significant 5'-> 3' ss-DNA exonuclease activity in NPE, so the final product after end unwinding and 5' ss-tail degradation is the 3' ss-tail (Figure: A conserved pathway for the 5' strand-specific processing at DSBs). By biochemical fractionation, we have purified the 5’ -> 3’ ss-DNA exonuclease from Xenopus egg extracts. Mass spectrometry analysis identified it as the Xenopus homolog of the budding yeast DNA2 helicase/nuclease. While this highly conserved protein has been reported to possess mostly an endonuclease activity with no obvious polarity bias, we have found that the 5’ -> 3’ ss-DNA exonuclease activity is dramatically stimulated by the eukaryotic ss-DNA binding protein RPA. Immunodepletion of xDNA2 with antibodies from NPE leads to a significant reduction in 5’-strand resection and SSA, confirming its important role in DSB repair. This basic architecture, first unwinding the end with a RecQ-type helicase and then degrading the 5’ strand by a 5’ -> 3’ ss-DNA exonuclease, is now found to be conserved in E. coli, yeast, and, most likely, also in human cells (Figure: A conserved pathway for the 5' strand-specific processing at DSBs).

    A RecQ-type helicase and a 5’ -> 3’ ss-DNA exonuclease are not the only proteins required for end processing. EXO1, a protein that has been reported to possess 5’ -> 3’ ds- and ss-exonuclease activities, appears to form a separate pathway for end processing in yeast and human cells. MRE11 (in complex with RAD50 and NBS1), which possesses both a 3’ -> 5’ ds-DNA exonuclease activity and a ss-DNA endonuclease activity, appears to have some effect on end processing in several (but not all) studies. The human CtIP protein and its putative yeast homolog, Sae2, are also required for end processing. The major questions in the field are exactly what mechanistic roles these proteins play, if at all, in end processing. By combining the Xenopus system with enzymatic analysis, we have made significant inroads into answering these questions. In addition, we have begun to identify additional proteins that bind to DNA ends and might thus play key roles in end processing and DSB repair.

    Top
  • 2. Mechanistic Analysis of DNA Lesion Replication
    Two mechanisms for DNA lesion replication
    Two mechanisms for DNA lesion replication

    While most of the lesions in DNA are rapidly repaired, inevitably some lesions are missed and might consequently stall the progression of replication forks. A major advance in recent years is the recognition that overcoming replication stalling is essential to genome maintenance and requires the concerted action of a myriad of proteins traditionally considered as being involved in replication, repair, recombination, and checkpoint response. Many models, largely based on genetic analyses, have been proposed to explain the extensive genetic data on the interplay among replication, repair, recombination, and checkpoints. However, the intrinsic limitation of classical genetic methods makes it difficult to directly test these models. The mechanism by which cells overcome replication stalling still poorly understood.

    To rigorously analyze the mechanism of how fork stalling DNA lesions are replicated, we have endeavored to reconstitute lesion replication in the Xenopus egg extract. In this extract, DNA, usually sperm chromatin, induces nuclear formation around itself and the DNA within the nucleus is replicated once. A more recent development is to use NPE to induce plasmid DNA replication. The Xenopus system is the only eukaryotic replication system that relies solely on cellular replication proteins such as ORC, MCMs, CDK2, CDC45, etc. This is in sharp contrast to viral replication systems, such as SV40, which depend on viral proteins like T antigen for DNA replication. Compared to normal egg extracts, NPE has an advantage in that it can also support high efficiency replication of plasmid DNA. In collaboration with Y. Matsumoto we have prepared plasmid DNA containing a specific lesion (apurine/apyrimidine (AP) site) as the substrate for studying replication fork stall and restart. AP sites are the common intermediate of base excision repair of many DNA lesions in vivo with a steady state level as high as 50,000–200,000 per genome in human and rodent tissues. In vitro AP sites are strong stalling sites for many DNA polymerases.By using AP-containing DNA as thesubstrate for replication in NPE, we have succeeded in developing the first biochemical system to study how a fork-stalling lesion is replicated in a truly eukaryotic cellular replication system. The use of a site-specific lesion makes it possible for us to rigorously analyze the stalling site in relation to the lesion and the nucleotide(s) inserted opposite the AP lesion after replication. Our data provide convincing evidence that the AP lesion does stall DNA replication. However, this stalling is transient and the lesion is eventually replicated. Furthermore, we have found that two distinct mechanisms are used to replicate the AP lesion: one error-prone and one error-free (Figure: Two mechanisms for DNA lesion replication). We expect these biochemical studies to finally reveal the molecular nature of error-prone mechanism and errorfree mechanism, how the two mechanisms are controlled, and what, if any, role proteins like WRN and BLM might play.

    Top