Discovery at the VICB
Restoring Replication with HLTF
By: Carol A. Rouzer, VICB Communications
Published: July 6, 2015
Structural and functional studies reveal the role of the mysterious HIRAN domain of the DNA translocase HLTF.
The process of DNA replication may be stalled in a variety of situations, including the presence of unrepaired DNA damage, abnormal secondary structures, aberrant protein-DNA complexes, and lack of sufficient nucleotide substrates. If a replication fork is stalled for a sufficient period of time, it may completely collapse, leading to DNA strand breaks, and potentially even cell death. Consequently, cells have devised multiple ways to prevent or repair stalled forks. Some of these, such as bypass of damaged bases by translesion polymerases, may lead to erroneous base incorporation and ultimately mutations, while others, such as strand switching, can result in error-free replication. The process of fork reversal (also called fork regression) is increasingly being recognized as an important mechanism for repair of stalled forks. In this process, the nascent DNA strands are unwound, separating them from the template strands, and then annealed with each other. The result is a four-armed Holliday junction structure (Figure 1). The advantages of this process are that it minimizes the exposure of single-stranded DNA (ssDNA), it places damaged bases within the context of double-stranded DNA (dsDNA), which promotes repair, and it facilitates the process of strand switching. Yet, despite the importance of fork reversal to genome maintenance, relatively little is known about the proteins that carry it out. Now, Vanderbilt Institute of Chemical Biology (VICB) member Brandt Eichman and his collaborator Karlene Cimprich (Stanford University) report important new findings concerning the structure and function of HLTF, an ATPase-dependent DNA translocase that catalyzes fork reversal [A. C. Kile, et al. (2015) Mol. Cell, published online June 4, DOI:10.1016/j.molcel.2015.05.013].
Figure 1. Diagrammatic representation of the fork reversal process. The newly synthesized nascent strands are first unwound and dissociated from the template parental strands. The parental strands are then annealed. Finally, the free newly synthesized strands are annealed. Note that a short stretch of single-stranded DNA remains because the newly synthesized strands are not exactly the same length. Figure reproduced by permission from Macmillan Publishers, Ltd. from K. J. Neelsen and M. Lopes, (2015) Nat. Rev. Cell. Mol. Biol., 16, 207. Copyright 2015.
In addition to its DNA translocase activity, HLTF also posseses a ubiquitin ligase RING motif, and it is known to polyubiquitinate the DNA replication protein PCNA. HLTF also possesses a HIRAN domain, a very ancient and highly conserved protein structure of unknown function. The Eichman and Cimprich investigators hypothesized that the HIRAN domain in HLTF plays an important role in fork reversal, and they first tested their hypothesis using the iPOND (identification of proteins on nascent DNA) procedure. Developed in the laboratory of VICB member David Cortez, iPOND verified that HLTF is present at the replication fork whether or not the cells were experiencing replication stress. In contrast they did not detect HLTF in mature chromatin.
To see if the HIRAN domain interacts directly with DNA, the investigators expressed and purified the domain from HLTF. They then used the HIRAN domain in an EMSA (electrophoretic mobility shift assay) to show that it possesses a strong affinity for single-stranded DNA (ssDNA) but not double-stranded DNA (dsDNA). They further explored this interaction with DNA in an assay using a single-stranded oligonucleotide attached to biotin. Incubation of the HIRAN domain with the oligonucleotide, followed by streptavadin beads allowed the researchers to capture the oligonucleotide and any bound protein on the beads. They discovered that HIRAN did, indeed, bind to oligonucleotides, but only if the 3’ end of the oligonucleotide was free. Modification of this end by biotin attachment, phosphorylation, or deoxy substitution abrogated HIRAN binding. They obtained the same results using full-length HLTF, but not HLTF lacking the HIRAN domain. These results confirmed that the HIRAN domain is required for the interaction of HLTF with ssDNA and demonstrated a requirement for a free 3′ end in the ssDNA target.
The crystal structure of a complex of HIRAN with a single-stranded oligonucleotide revealed the oligonucleotide with its 3′ end inserted deeply into a binding pocket in the HIRAN protein. The structural data identified a number of amino acids that come into intimate contact with the oligonucleotide, and the importance of these amino acids was confirmed by 15N-HSQC (heteronuclear single quantum coherence) NMR spectroscopy of the interaction of HIRAN with ssDNA. Furthermore, when the investigators created HIRAN proteins in which these residues were mutated, they found that the mutant proteins exhibited markedly reduced affinity for ssDNA.
To further elucidate the role of the HIRAN domain in HLTF function, the researchers created full-length HLTF proteins carrying mutations of key ssDNA binding residues in the HIRAN domain. They found that these mutations had no effect on the proteins’ ATPase activity, confirming that their overall structure remained intact. However, the mutations reduced the ability of the HLTF proteins to reverse a model replication fork as assessed by a gel electrophoresis assay (Figure 2). These experiments also revealed that HLTF does not efficiently reverse a replication fork unless the fork possesses a free 3′ end, consistent with the earlier structural and in vitro data. The requirement that the target DNA carries a free 3′ end had not been previously observed in the case of SMARCAL1, RecG, or UvsW, three other proteins that are known to reverse replication forks, suggesting that the cell possesses multiple proteins that can recognize and repair stalled forks containing a variety of structures.
Figure 2. Schematic diagram of the fork reversal assay. (A) A synthetic model of a stalled fork structure consists of complementary template strands (blue) annealed to two shorter nascent strands (red). (B) The DNA is incubated with proteins to be tested for fork reversal activity. If fork reversal occurs, the two template strands are annealed with one another, as are the two nascent strands. (C) The products are two pieces of double-stranded DNA of different lengths. Separation by gel electrophoresis distinguishes the three species on the basis of different migration on the gel.
Treatment of cells with hydroxyurea leads to inhibition of ribonucleotide reductase, thereby blocking deoxyribonucleotide synthesis. The resulting depletion of deoxyribonucleotide pools interferes with DNA replication in dividing cells, producing stalled forks with free 3′ ends. The investigators used the DNA fiber assay (Figure 3) to monitor fork progression in U2OS cells. They found that hydroxyurea treatment reduced fork progression by 45% in these cells. They were surprised to find that knockout of the gene encoding HLTF restored fork progression in hydroxyurea-treated cells to a rate nearly as high as in wild-type cells incubated in the absence of hydroxyurea (Figure 3). These findings led the investigators to hypothesize that HLTF’s function is to slow replication at the stalled fork, enabling it to undergo remodeling. This replication slowing effect was restored in HLTF knockout cells by overexpression of wild-type HLTF but not HLTF bearing a mutation of a ssDNA-binding residue in the HIRAN domain.
Figure 3. Schematic diagram of the DNA fiber assay. Cells are synchronized in S phase and then incubated for 30 min with IdU. They are then washed and transferred to medium containing CldU with or without hydroxyurea. After the cells are harvested, the DNA is isolated and spread onto a slide as individual fibers. The DNA is stained with antibodies against IdU and CldU, each with a different fluorescent tag. The length of the colored strands under a fluorescence microscope indicates the rate of DNA replication. Knockout of HLTF had no effect on DNA replication rate in the absence of hydroxyurea (HU). Hydroxyurea caused a slowing of DNA replication in wild-type cells, but not in HLTF knockout cells, which synthesized DNA at nearly the same rate as cells incubated in the absence of hydroxyurea.
The Cimprich and Eichman lab investigators used their discoveries to propose a mechanism of fork reversal by HLTF (Figure 4). They suggest that HLTF, always present at the site of DNA synthesis, binds at a stalled replication fork. Its ATPase-dependent DNA translocation activity enables HLTF to begin to anneal the template strands at the fork, moving HLTF in the direction of the nascent strand bearing a free 3′ end. HLTF then captures the free 3′ end using its HIRAN domain and guides it into position for annealing with the other nascent strand. This mechanism provides the foundation for a better understanding of the important process of fork reversal and HLTF’s role in that process. Such information is critical if we are to fully comprehend how cells execute error-free replication in the face of numerous sources of DNA damage.
Figure 4. Proposed mechanism of fork reversal by HLTF. (A) HLTF binds to the stalled replication fork. (B) The ATPase-dependent DNA translocase activity of HLTF begins to anneal the template strands, moving HLTF towards the nascent strand with the free 3′ end. (C) HLTF captures the free 3′ end and then guides it into position for annealing with the other nascent strand.