Final Report Summary - BIOID IN TLS (Identification of novel regulators of translesion DNA synthesis in human cells)
Work progress and achievements during the project
Objective 1: establish BioID screening method
I have set up a newly developed screening method to identify proteins that act during translesion DNA synthesis (TLS) at blocked replication forks in an unbiased manner. I have fused the known TLS proteins RAD18, PCNA and DNA polymerase η with a mutated version of BirA*, a biotin protein ligase that promiscuously biotinylates proteins in its close vicinity. As a negative control I created fusion proteins mutated in such a way that they will not accumulate at sites of DNA damage. Furthermore, I have established stable inducible cell lines that express these BirA fusion proteins at near-physiological levels.
I optimized the protocol described in the paper of Roux et al. (JCB, 2012) to make it suitable for a more dynamic experimental setup. I combined this experimental approach with SILAC based mass spectrometry in collaboration with excellent colleagues in the Proteomics Department at our research center. We performed several mass spectrometry runs with the different fusion proteins under different experimental conditions.
As far as I could conclude from the available data, we have picked up many of the known key players in TLS in the dataset of proteins biotinylated by the BirA* fusion proteins of RAD18, PCNA and polymerase η. This indicates that this particular experimental set up works well for the identification of proteins recruited to stalled replication forks that are potentially involved in DNA damage bypass by TLS. However, unfortunately we have not been able to identify any novel proteins among the strong hits that could be subjected to validation and follow-up experiments. The reason for this might be that the sensitivity of the experimental set up might not be high enough for mass spectrometry.
After many attempts to change the experimental details, in order to increase the sensitivity and specificity of the output in response to treatment with DNA damaging agents, I had to conclude that - even though the method worked up to a certain level - probably the method, in the way I designed it is not sensitive enough to use for mass spectrometry. In other words, even though the fusion proteins are recruited to the DNA damage sites and biotinylate other proteins residing there, there are probably too little individual molecules being biotinylated to identify novel proteins by mass spectrometry.
Deviation from the grant proposal:
We engaged in a very fruitful collaboration with the lab of Professor Matthias Mann (Max Planck Institute of Biochemistry, Munich), a world-leading expert in quantitative proteomics. In this collaboration we developed a new time resolved proteomics approach called CHROMASS (chromatin mass spectrometry), to identify proteins that accumulate at replication forks stalled at DNA interstrand crosslinks. For this approach, crosslinked Xenopus laevis DNA is combined with cell free Xenopus oocyte extract, in which the DNA is allowed to replicate. At different time points all chromatin bound proteins are isolated and can be identified by mass spectometry, which enables us to not only get a full reading of all proteins that are involved in DNA replication, but also proteins that accumulate at stalled replication forks and proteins that are involved in resolving DNA crosslinks. The repair of DNA interstrand crosslinks is an intricate process that requires the actions of several different repair mechanisms such as TLS, the Fanconi Anemia pathway (FA), homologous recombination (HR) and nucleotide excision repair (NER). This new proteomics approach provides a unique and thorough way to identify novel proteins involved not only in TLS, but also in the other pathways involved in faithful repair and bypass of DNA crosslinks that block replication fork progression. The datasets we collected indeed identified most of the known players involved in DNA replication and the different pathways involved in repair of interstrand crosslinks, but also a number of novel proteins that have not previously been linked to any of these processes.
Rather than a deviation from the grant proposal, I believe this project is more of an enrichment to it. With the CHROMASS method we have very successfully identified a number of novel proteins involved in all aspects of replication fork stalling, replication fork bypass and DNA crosslink repair, a goal originally intended with the BioID approach as proposed in my research plan. Moreover, the CHROMASS method has successfully been used in combination with other types of DNA damaging agents to identify novel proteins in other DNA damage response pathways. Furthermore, other labs can now use the CHROMASS method to their own accord in combination with their own experimental setup. This method can potentially be used in any kind of setting for which its desirable to monitor changes in the chromatin bound proteome in time.
Even though I decided to change my focus from BioID to CHROMASS, my work on the BioID project inspired other people to apply BioID to the use of their own research. Researchers in our own lab, collaborating labs at the Center for Protein Research and labs at the University of Copenhagen are converting the BioID method to look at proteins recruited to DNA double strand breaks, DNA ultrafine bridges during mitosis and DNA replication sites. These projects are still ongoing and are yielding promising results. Furthermore, currently some of my colleagues are using the optimized BioID protocol that I established to identify interactors of their favorite protein.
Objective 2: Identify TLS candidates from established proteomics datasets.
I have been mining three proteomics datasets from mass spectrometry screens performed within our laboratory. I selected a number of potentially interesting candidates that were among the top hits in the datasets. I validated 22 candidate proteins for their potential to be involved in the DNA damage response. During the validation, many of these proteins did not seem to be suitable candidates to follow up on, either because of the inability to establish a possible role for them in the DNA damage response or because of the absence of proper reagents (e.g. antibodies, cDNAs, etc.). Three of these proteins had potential for being bona fide DNA damage response proteins. Unfortunately, recently another research group published a paper on one of these proteins, so it may not be worth investigating this candidate anymore.
So far our lab has identified five novel proteins from the aforementioned CHROMASS screen that were recruited to sites of DNA damage, both in Xenopus and in human cells. In addition, a spin-off project from the CHROMASS screen, in which the CHROMASS method was used in combination with an endonuclease that cuts DNA to create DNA double-strand breaks, revealed a number of other interesting candidates that are currently being characterized in the lab.
Objective 3: Follow up on successful candidates from objective 1 and 2 to dissect their function during in the DNA damage response.
As I mentioned under objective 1, unfortunately the BioID method did not prove successful for identification of potential novel TLS-regulating proteins.
However, with the CHROMASS method, we identified at least 5 novel proteins that are involved in the DNA damage response. I have followed up on two of these proteins, SLF1 and SLF2. I discovered that these two proteins form a complex and ‘bridge’ the E3 ubiquitin ligase RAD18 and the SMC5/6 cohesion complex. Both RAD18 and the SMC5/6 complex are known to be required for faithful repair of damaged DNA, but never before have they been associated with each other. Furthermore, the role of both RAD18 and the SMC5/6 complex in the DNA double-strand break response has been poorly understood so far. We have increased the knowledge on these two proteins and shown that they are uniquely linked by the novel proteins SLF1 and SLF2. This data accompanies the CHROMASS screen and the full dataset from the screen, in a resource paper that was published last year in Science.
With this paper, we have provided important new insight into the regulatory mechanisms that protect genome stability in human cells. Furthermore, it provides an enormous resource of knowledge for the scientific community to add to the better understanding of DNA replication and the response to DNA crosslinks.
Currently, I am further investigating the roles of SLF1 and SLF2 in the DNA damage response, as there are still many unanswered questions to address. Hopefully this will lead to another scientific publication, which will increase the understanding on how these proteins collaborate together within the DNA damage response.
Project management
In addition to the scientific progression, I have recruited and supervised an undergraduate student, who worked on a part of my project. She did very well and recently graduated with top grades.
Furthermore, together with some colleagues, at the beginning of 2014, I set up an association for PhD students and post-docs within the Center for Protein Research, my host institution, and I am currently a member of the board. We felt a great need for students and post-docs to unite and organize ourselves. We have been organizing many extracurricular activities that help our development as a researcher and strengthen social and scientific bonds between us. We organized career events, a scientific writing workshop, a scientific retreat; we invited high profile guest speakers from the scientific community and organized a number of social activities. Importantly, we have been given some influence in the center management, since members of the association board have been invited to sit in at (part of) the center management meetings and propose issues on behalf of the student and post-doc community.
During my stay at the Center for Protein Research (CPR), I have been able to manage my research projects with a high degree of independence. I have designed and optimized the methodology for the BioID approach, which will be very beneficial to colleagues in the near future, due to its numerous possible applications. In addition, I have collaborated with scientists from different backgrounds within the CPR for the BioID project. Furthermore, we have set up an international collaboration which was very fruitful, not only for me personally, but also for a number of colleagues who are currently working on some quite developed follow up stories (one manuscript has been published in the Journal of Cell Biology, one manuscript is under revision and a third manuscript is ready to be submitted). Importantly, all of these accomplishments have helped me develop further into becoming an independent scientist.
The main deviation from the planned milestones is that I have so far been unsuccessful in identifying any novel TLS factors from previously existing data sets. Thus far, of the 22 proteins I validated, none have been identified as TLS factors.
In addition, I successfully set up the BioID method, but have been unable to identify novel candidate proteins with it.
However, in addition I collaborated on a very similar project, for which we set up a proteomics based screening method to identify novel factors involved in all aspects of DNA crosslink repair, including TLS. From this screen we identified several novel proteins, on two of which I performed a follow up study. This paper has been published in Science on May 1st 2015.
I would argue that within this project, we used a different mass spectrometry based technique to reach a similar end point. Moreover, the paper contains several datasets of CHROMASS screens in which other DNA repair pathways are triggered, creating a large data resource for other labs to use. Overall, I therefore consider this project to have been highly successful.
Objective 1: establish BioID screening method
I have set up a newly developed screening method to identify proteins that act during translesion DNA synthesis (TLS) at blocked replication forks in an unbiased manner. I have fused the known TLS proteins RAD18, PCNA and DNA polymerase η with a mutated version of BirA*, a biotin protein ligase that promiscuously biotinylates proteins in its close vicinity. As a negative control I created fusion proteins mutated in such a way that they will not accumulate at sites of DNA damage. Furthermore, I have established stable inducible cell lines that express these BirA fusion proteins at near-physiological levels.
I optimized the protocol described in the paper of Roux et al. (JCB, 2012) to make it suitable for a more dynamic experimental setup. I combined this experimental approach with SILAC based mass spectrometry in collaboration with excellent colleagues in the Proteomics Department at our research center. We performed several mass spectrometry runs with the different fusion proteins under different experimental conditions.
As far as I could conclude from the available data, we have picked up many of the known key players in TLS in the dataset of proteins biotinylated by the BirA* fusion proteins of RAD18, PCNA and polymerase η. This indicates that this particular experimental set up works well for the identification of proteins recruited to stalled replication forks that are potentially involved in DNA damage bypass by TLS. However, unfortunately we have not been able to identify any novel proteins among the strong hits that could be subjected to validation and follow-up experiments. The reason for this might be that the sensitivity of the experimental set up might not be high enough for mass spectrometry.
After many attempts to change the experimental details, in order to increase the sensitivity and specificity of the output in response to treatment with DNA damaging agents, I had to conclude that - even though the method worked up to a certain level - probably the method, in the way I designed it is not sensitive enough to use for mass spectrometry. In other words, even though the fusion proteins are recruited to the DNA damage sites and biotinylate other proteins residing there, there are probably too little individual molecules being biotinylated to identify novel proteins by mass spectrometry.
Deviation from the grant proposal:
We engaged in a very fruitful collaboration with the lab of Professor Matthias Mann (Max Planck Institute of Biochemistry, Munich), a world-leading expert in quantitative proteomics. In this collaboration we developed a new time resolved proteomics approach called CHROMASS (chromatin mass spectrometry), to identify proteins that accumulate at replication forks stalled at DNA interstrand crosslinks. For this approach, crosslinked Xenopus laevis DNA is combined with cell free Xenopus oocyte extract, in which the DNA is allowed to replicate. At different time points all chromatin bound proteins are isolated and can be identified by mass spectometry, which enables us to not only get a full reading of all proteins that are involved in DNA replication, but also proteins that accumulate at stalled replication forks and proteins that are involved in resolving DNA crosslinks. The repair of DNA interstrand crosslinks is an intricate process that requires the actions of several different repair mechanisms such as TLS, the Fanconi Anemia pathway (FA), homologous recombination (HR) and nucleotide excision repair (NER). This new proteomics approach provides a unique and thorough way to identify novel proteins involved not only in TLS, but also in the other pathways involved in faithful repair and bypass of DNA crosslinks that block replication fork progression. The datasets we collected indeed identified most of the known players involved in DNA replication and the different pathways involved in repair of interstrand crosslinks, but also a number of novel proteins that have not previously been linked to any of these processes.
Rather than a deviation from the grant proposal, I believe this project is more of an enrichment to it. With the CHROMASS method we have very successfully identified a number of novel proteins involved in all aspects of replication fork stalling, replication fork bypass and DNA crosslink repair, a goal originally intended with the BioID approach as proposed in my research plan. Moreover, the CHROMASS method has successfully been used in combination with other types of DNA damaging agents to identify novel proteins in other DNA damage response pathways. Furthermore, other labs can now use the CHROMASS method to their own accord in combination with their own experimental setup. This method can potentially be used in any kind of setting for which its desirable to monitor changes in the chromatin bound proteome in time.
Even though I decided to change my focus from BioID to CHROMASS, my work on the BioID project inspired other people to apply BioID to the use of their own research. Researchers in our own lab, collaborating labs at the Center for Protein Research and labs at the University of Copenhagen are converting the BioID method to look at proteins recruited to DNA double strand breaks, DNA ultrafine bridges during mitosis and DNA replication sites. These projects are still ongoing and are yielding promising results. Furthermore, currently some of my colleagues are using the optimized BioID protocol that I established to identify interactors of their favorite protein.
Objective 2: Identify TLS candidates from established proteomics datasets.
I have been mining three proteomics datasets from mass spectrometry screens performed within our laboratory. I selected a number of potentially interesting candidates that were among the top hits in the datasets. I validated 22 candidate proteins for their potential to be involved in the DNA damage response. During the validation, many of these proteins did not seem to be suitable candidates to follow up on, either because of the inability to establish a possible role for them in the DNA damage response or because of the absence of proper reagents (e.g. antibodies, cDNAs, etc.). Three of these proteins had potential for being bona fide DNA damage response proteins. Unfortunately, recently another research group published a paper on one of these proteins, so it may not be worth investigating this candidate anymore.
So far our lab has identified five novel proteins from the aforementioned CHROMASS screen that were recruited to sites of DNA damage, both in Xenopus and in human cells. In addition, a spin-off project from the CHROMASS screen, in which the CHROMASS method was used in combination with an endonuclease that cuts DNA to create DNA double-strand breaks, revealed a number of other interesting candidates that are currently being characterized in the lab.
Objective 3: Follow up on successful candidates from objective 1 and 2 to dissect their function during in the DNA damage response.
As I mentioned under objective 1, unfortunately the BioID method did not prove successful for identification of potential novel TLS-regulating proteins.
However, with the CHROMASS method, we identified at least 5 novel proteins that are involved in the DNA damage response. I have followed up on two of these proteins, SLF1 and SLF2. I discovered that these two proteins form a complex and ‘bridge’ the E3 ubiquitin ligase RAD18 and the SMC5/6 cohesion complex. Both RAD18 and the SMC5/6 complex are known to be required for faithful repair of damaged DNA, but never before have they been associated with each other. Furthermore, the role of both RAD18 and the SMC5/6 complex in the DNA double-strand break response has been poorly understood so far. We have increased the knowledge on these two proteins and shown that they are uniquely linked by the novel proteins SLF1 and SLF2. This data accompanies the CHROMASS screen and the full dataset from the screen, in a resource paper that was published last year in Science.
With this paper, we have provided important new insight into the regulatory mechanisms that protect genome stability in human cells. Furthermore, it provides an enormous resource of knowledge for the scientific community to add to the better understanding of DNA replication and the response to DNA crosslinks.
Currently, I am further investigating the roles of SLF1 and SLF2 in the DNA damage response, as there are still many unanswered questions to address. Hopefully this will lead to another scientific publication, which will increase the understanding on how these proteins collaborate together within the DNA damage response.
Project management
In addition to the scientific progression, I have recruited and supervised an undergraduate student, who worked on a part of my project. She did very well and recently graduated with top grades.
Furthermore, together with some colleagues, at the beginning of 2014, I set up an association for PhD students and post-docs within the Center for Protein Research, my host institution, and I am currently a member of the board. We felt a great need for students and post-docs to unite and organize ourselves. We have been organizing many extracurricular activities that help our development as a researcher and strengthen social and scientific bonds between us. We organized career events, a scientific writing workshop, a scientific retreat; we invited high profile guest speakers from the scientific community and organized a number of social activities. Importantly, we have been given some influence in the center management, since members of the association board have been invited to sit in at (part of) the center management meetings and propose issues on behalf of the student and post-doc community.
During my stay at the Center for Protein Research (CPR), I have been able to manage my research projects with a high degree of independence. I have designed and optimized the methodology for the BioID approach, which will be very beneficial to colleagues in the near future, due to its numerous possible applications. In addition, I have collaborated with scientists from different backgrounds within the CPR for the BioID project. Furthermore, we have set up an international collaboration which was very fruitful, not only for me personally, but also for a number of colleagues who are currently working on some quite developed follow up stories (one manuscript has been published in the Journal of Cell Biology, one manuscript is under revision and a third manuscript is ready to be submitted). Importantly, all of these accomplishments have helped me develop further into becoming an independent scientist.
The main deviation from the planned milestones is that I have so far been unsuccessful in identifying any novel TLS factors from previously existing data sets. Thus far, of the 22 proteins I validated, none have been identified as TLS factors.
In addition, I successfully set up the BioID method, but have been unable to identify novel candidate proteins with it.
However, in addition I collaborated on a very similar project, for which we set up a proteomics based screening method to identify novel factors involved in all aspects of DNA crosslink repair, including TLS. From this screen we identified several novel proteins, on two of which I performed a follow up study. This paper has been published in Science on May 1st 2015.
I would argue that within this project, we used a different mass spectrometry based technique to reach a similar end point. Moreover, the paper contains several datasets of CHROMASS screens in which other DNA repair pathways are triggered, creating a large data resource for other labs to use. Overall, I therefore consider this project to have been highly successful.