Vanderbilt Institute of Chemical Biology



Discovery at the VICB







SAN1 - A New Player in DNA Cross-Link Repair


By: Carol A. Rouzer, VICB Communications
Published:  July 10, 2018



SAN1 is a 5′ exonuclease that helps cells survive upon exposure to agents that cause DNA interstrand cross-links.


Interstrand cross-links (ICLs) are a particularly toxic form of DNA damage that covalently links the two complementary strands of DNA together. ICLs, which block both replication and transcription, are also difficult to repair, requiring a combination of nucleotide excision repair, translesion synthesis, and homologous recombination. The Fanconi anemia (FA) pathway, comprising >20 proteins, is an ICL repair mechanism present only in animals. FA pathway-mediated repair involves the cooperation of multiple different nucleases. Among these are XPF-ERCC1 that cleaves on the 3′ side of the lesion. XPF-ERCC1 may also cleave on the 5′ side of the ICL, but other nucleases appear to participate at this site. Now, VICB member Ian Macara and his laboratory report the discovery of SAN1, a 5′ exonuclease that plays a role in the repair of ICLs [A. M. Andrews, et al. Nat. Commun., (2018) 9, 2592].

Although the protein encoded by the gene Fam120b was originally described as a transcriptional co-activator of PPAR-γ, the Macara lab researchers noted that it also contains a nuclease domain comprising seven acidic residues closely resembling those found in the FEN1 family of structure-specific nucleases. This led the investigators to express the murine protein in E. coli, using an attached C-terminal Strep-tag to facilitate purification. They confirmed that the purified protein possessed 5′ exonuclease activity using two single stranded substrates, each 50 nucleotides long. A D90A mutation, which neutralized one of the acidic residues in the nuclease domain of the protein, eliminated its activity. The researchers named the protein SAN1 (senataxin-associated nuclease 1) to describe both its enzymatic activity and its ability to interact with the protein senataxin, as will be described below.

To further characterize SAN1's activity, the researchers expressed the human protein in HEK 293T cells, adding both a Strep2 and a FLAG tag, to enable a two-step affinity purification scheme. They found that the 5′ exonuclease activity co-purified with SAN1 through both steps. They then tested the ability of the enzyme to act as a nuclease on a variety of different structures (Figure 1). Using single stranded DNA, they discovered that short chains (~25 nucleotides) were not good substrates, whereas longer ones (~50 nucleotides) were readily cleaved. Using this simple single stranded substrate, kinetic studies revealed KM and kcat values similar to those of other FEN family nucleases. SAN1's action generated 3 to 7 nucleotide fragments, indicating that the enzyme was not processive. Although the SAN1 did not work well with short single stranded substrates, it did display some activity with a double stranded substrate containing an ~20 nucleotide 5′ single stranded overhang. Much better activity resulted using a splayed structure, comprising a double stranded segment with unwound single stranded ends. However, if the 3′ end of the splayed structure was double stranded, producing a 5′ flap, activity was substantially reduced, and if the 5′ end of the splayed structure was double stranded (3′ flap), no activity was observed. Similarly, no activity resulted using substrates that did not have a single stranded 5′ end of reasonable length, including gapped and nicked structures. Together, the findings suggested a model in which the enzyme binds to DNA downstream of where it cleaves, possibly by interacting with the other strand of a splayed structure, and then cuts 5′ to the binding site about 3-7 nucleotides away (Figure 2).



FIGURE 1. Examples of substrates used to evaluate the activity of SAN1. In each case, the gold star represents the 5′ end of one of the two DNA strands.





FIGURE 2. Possible mechanisms for the 5′-nucleotidase activity of SAN-1. In the case of a single-stranded or splayed substrate, it is possible that the enzyme can interact with one single strand of DNA while the other strand is cleaved. Such an interaction is not possible for the 5′-flap substrate. This might explain the much higher activity of the enzyme for the single stranded and splayed structures. Figure reproduced under the Creative Commons Attribution 4.0 International License 4.0 from A. M. Andrews, et al. Nat. Commun., (2018) 9, 2592.



In their proposed model of SAN1 activity, the researchers envisioned that the enzyme binds to DNA through an interaction between the N-terminal nuclease domain and the highly conserved C-terminal domain of the protein. These two domains are separated by a region of repeated 12 residue motifs containing QEVPM sequences. The investigators discovered that they could delete the repeat region with no effect on nuclease activity; however, the C-terminal domain was required. Furthermore, when they expressed the N- and C-terminal domains individually, they found that they interacted with each other. These results supported the hypothesis that an interaction between the N- and C-terminal domains is important for SAN1's enzymatic activity.

Using CRISPR/Cas9 technology, the researchers created a HeLa cell line in which the Fam120b gene had been knocked out. These SAN1-/- cells expressed no SAN1 protein, as expected, but they also displayed no obvious change in growth or morphology under routine culture conditions. Similarly, the cells exhibited no difference from wild-type HeLa cells in their sensitivity to ionizing radiation, which causes single strand and double strand breaks in DNA, or to hydroxyurea or camptothecin, which cause replication stress. However, SAN1-/- cells were more sensitive to the toxicity of mitomycin C (MMC) and cisplatin, both of which cause ICL formation. This increased sensitivity was reversed by expression of wild-type SAN1 but not the inactive D90A mutant SAN1 in the SAN1-/- cells. These findings suggested that enzymatically active SAN1 is involved in ICL repair, but not homologous recombination or non-homologous end-joining.

In hematopoietic cells from patients suffering from Fanconi anemia due to a mutation in an FA pathway gene, exposure to DNA cross-linking agents leads to the formation of abnormal chromosomal structures referred to as radial chromosomes. Structures such as these were prevalent in SAN1-/- cells treated with MMC but rare in identically treated wild-type cells (Figure 3). Similarly, SAN1-/- cells treated with MMC exhibited higher levels of γ-H2AX and 53BP1 (both indications of DNA damage) than wild-type cells following exposure to MMC (Figure 4). These observations indicated that, in the absence of SAN1, ICLs that are formed in response to agents such as MMC and cisplatin are not repaired efficiently, leading to further DNA damage and chromosomal aberrations.



FIGURE 3. Metaphase spreads showing chromosomes from untreated (a and b) and MMC-treated (c and d) wild-type (a and c) or SAN1-/- (b and d) HeLa cells. Very few radial chromosomes are visible in the absence of MMC treatment in either cell population. MMC causes some radial chromosome formation in wild-type cells (red arrow), but the number of such aberrant chromosome structures is much higher in SAN1-/- cells. Figure reproduced under the Creative Commons Attribution 4.0 International License 4.0 from A. M. Andrews, et al. Nat. Commun., (2018) 9, 2592.



FIGURE 4. Photomicrographs showing the presence of DNA damage as indicated by γ-H2AX (pink) in wild-type left and SAN1-/- (right) HeLa cells that had (bottom) or had not (top) been treated with MMC. Levels of γ-H2AX are low in both cell populations in the absence of MMC treatment. The increase in γ-H2AX is greater in SAN1-/- cells than in wild-type cells. Nuclei are stained with DRAQ5 (blue). Figure reproduced under the Creative Commons Attribution 4.0 International License 4.0 from A. M. Andrews, et al. Nat. Commun., (2018) 9, 2592.



The apparent involvement of SAN1 in ICL repair led the investigators to hypothesize that it may play a role in the FA pathway. They tested this hypothesis by using siRNA to knockdown expression of FANCD2, a key FA pathway protein, in SAN1-/- and wild-type HeLa cells. They discovered that FANCD2 knockdown in SAN1-/- cells caused a synergistic increase in sensitivity to ICL-inducing agents over that observed with SAN1 or FNACD2 deletion alone. They obtained similar results when they knocked down expression of the FA pathway nuclease XPF. In contrast, the investigators found that knockdown of SNM1a and FAN1, nucleases known to act independently of the FA pathway, did not lead to further sensitivity to ICL-inducing agents in SAN1-/- cells. These results suggested that SAN1 acts independently from the FA pathway, possibly in concert with SNM1a and FAN1.

To better understand how SAN1 is involved in ICL repair, the investigators performed a genome-wide yeast two-hybrid screen in search of proteins with which it interacts. Among the proteins identified by the screen, the researchers were particularly intrigued by senataxin, a protein believed to be a helicase involved with DNA transcription and repair. They confirmed an association between SAN1 and senataxin by co-immunopreciptation studies, and they demonstrated that exposure of cells to MMC increases binding of SAN1 to senataxin. Depletion of senataxin in HeLa cells using siRNA increased their sensitivity to cisplatin and MMC to a similar degree as knockout of SAN1, but depletion of senataxin in SAN1-/- cells did not further increase their sensitivity to ICL-inducing agents. A mutant SAN1 protein lacking the central repeat segments was unable to interact with senataxin. Notably, expression of this protein in SAN1-/- cells did not reduce their sensitivity to MMC or cisplatin. These results confirmed that SAN1 interacts with senataxin, most likely through SAN1's central repeat domain, and this interaction is important to its role in ICL repair.

In most cases, DNA repair pathways are triggered during replication when the replication machinery encounters the damage. However, a second DNA repair pathway occurs during transcription when damage interferes with the process. Stalling of the transcription machinery can lead to the accumulation of R-loops, structures comprising a hybrid of the DNA template strand and nascent RNA along with a loop formed by the coding strand of DNA (Figure 5). Previous studies had demonstrated that senataxin is involved in resolution of R-loops. This led the investigators to hypothesize that SAN1 might similarly be involved in R-loops resolution, and they found that, indeed R-loop levels were higher in SAN1-/- cells treated with MMC than in identically treated wild-type cells. This discovery led the researchers to propose that R-loops, resulting from transcription in the presence of ICLs, recruit senataxin and SAN1 to the site of the ICL, where the two proteins assist in R-loop resolution and ICL repair.


FIGURE 5. Structure of an R-loop. In this case, DNA is shown in red and RNA in blue. There is a section of DNA/RNA hybrid resulting from recent transcription, and the coding strand of DNA corresponding the transcribed region is looped out.


In conclusion, SAN1 is a 5′ exonuclease that plays a role in ICL repair. It does so through an interaction with senataxin, a protein known to be involved in the repair of abnormal RNA-DNA hybrid structures that accumulate during transcription in the presence of DNA damage. The investigators propose that SAN1 is involved in an ICL repair pathway, possibly also involving SNM1a and FAN1. They further hypothesize that SAN1-faciliated repair is most important when ICLs are too abundant to be cleared by the FA pathway. The importance of senataxin to SAN1's function suggests that this repair pathway may also be coupled to transcription. Mutations in the gene encoding senataxin are associated with ataxia with oculomotor apraxia type 2 and juvenile amyotrophic lateral sclerosis, both familial neurodegenerative diseases. It will be interesting to learn if the pathogenesis of these diseases is related to a failure of ICL repair.



View Nature Communications article: A senataxin-associated exonuclease SAN1 is required for resistance to DNA interstrand cross-links







The Vanderbilt Institute of Chemical Biology, 896 Preston Building, Nashville, TN 37232-6304, phone 866.303 VICB (8422), fax 615 936 3884
Vanderbilt University is committed to principles of equal opportunity and affirmative action. Copyright © 2014 by Vanderbilt University Medical Center