DNA polymerase ι is acetylated in response to SN2 alkylating agents

DNA polymerase iota (Polι) belongs to the Y-family of DNA polymerases that are involved in DNA damage tolerance through their role in translesion DNA synthesis. Like all other Y-family polymerases, Polι interacts with proliferating cell nuclear antigen (PCNA), Rev1, ubiquitin and ubiquitinated-PCNA and is also ubiquitinated itself. Here, we report that Polι also interacts with the p300 acetyltransferase and is acetylated. The primary acetylation site is K550, located in the Rev1-interacting region. However, K550 amino acid substitutions have no effect on Polι’s ability to interact with Rev1. Interestingly, we find that acetylation of Polι significantly and specifically increases in response to SN2 alkylating agents and to a lower extent to SN1 alkylating and oxidative agents. As we have not observed acetylation of Polι’s closest paralogue, DNA polymerase eta (Polη), with which Polι shares many functional similarities, we believe that this modification might exclusively regulate yet to be determined, and separate function(s) of Polι.


Results
p300 Interacts with Polι. Our previous results suggested a link between Polι polyubiquitination and inhibition of the lysine acetyltransferase activity p300 25 , a large, multi-domain protein (Fig. 1a) influencing multiple processes in the cell [reviewed in 12 ]. To further corroborate the relationship between Polι and p300, we initially investigated whether p300 physically interacts with Polι. We confirmed the interaction by performing FLAG co-immunoprecipitation assays in HEK293T cells transiently transfected with FLAG-Polι and a p300-c-myc expression vector (Fig. 1b). Additionally, we have also demonstrated the interaction between endogenous p300 and Polι proteins by performing a proximity ligation assay (PLA) (Fig. 1b).
Because p300 can both acetylate and ubiquitinate substrates 16 , we next examined whether the RING ubiquitin ligase domain, which is present in the p300 protein, or the HAT domain, which is responsible for the acetyltransferase activity of p300, is required for the interaction with Polι. The catalytic regions of p300 form a compact module in which the RING domain is juxtaposed with the HAT domain. We performed immunoprecipitation experiments with transiently transfected HEK293T cells using either of the two HA-tagged p300 deletion constructs and FLAG-tagged Polι. p300ΔC (amino acids 1-1287) lacking the HAT domain failed to interact with Polι, while p300ΔN (amino acids 1046-2414) interacted strongly with the protein, indicating that C-terminal half of the p300 protein, including the HAT domain, is required for binding to Polι (Fig. 1c).
To further map the interaction between Polι and p300 and identify a region in Polι necessary for an interaction with the p300 protein, we generated a series of Polι deletion constructs lacking domains involved in protein-protein interactions including the ubiquitin binding motifs UBM1 and UBM2, the Rev1-interacting region (RIR) and the PCNA interacting peptide (PIP) (Fig. 2a). We used the HA-immunoprecipitation assay in HEK293T cells transiently transfected with HA-tagged p300ΔN and full-length, or truncated versions of Polι. As shown in Fig. 2b, HA-p300ΔN interacted with all of the Polι constructs lacking the known protein interacting domains, which suggests that the Polι catalytic domain located at the N-terminus of the protein is sufficient for Left panel shows the results of immunoprecipitation assays (IP) performed with anti-FLAG resins incubated with extracts (E) from cells co-expressing p300-c-myc and FLAG-Polι or p300-c-myc and empty FLAG vector as a negative control. Right panel shows the results of the proximity ligation assay (PLA), signal (red) provides evidence that endogenous Polι and p300 are close together in the cell, actin cytoskeleton is shown in green (Alexa488 Phalloidin staining), nucleus in blue (Hoechst staining). Scale bar is 10 µm. (c) Immunoprecipitation assays (IP) were performed with anti-FLAG resins that were incubated with extracts (E) from cells expressing FLAG-Polι or FLAG empty vector as a negative control and either of HA-tagged p300 deletion constructs p300ΔN or p300ΔC. (c) Immunoprecipitation assays were performed with 10μl of anti-HA resins that were incubated with equal amounts of extracts from cells co-expressing HA-p300ΔN and full-length FLAG-Polι or FLAG-Polι lacking the N-terminal catalytic domain (IP-immunoprecipitation). In (b,c) the input controls contain 20μg or 10μg of protein extracts, respectively. www.nature.com/scientificreports www.nature.com/scientificreports/ the interaction with p300. To verify this hypothesis, we tested the interaction of HA-p300ΔN with Polι lacking the catalytic domain, but possessing all known protein-interacting regions (409-715aa). The recombinant Polι protein lacking its N-terminus did not interact with p300, implying that the presence of the catalytic domain of Polι is required for its interaction with the p300 protein (Fig. 2c).

polι is Acetylated in vivo and in vitro.
Our current studies show that Polι directly binds the C-terminal half of p300 protein including the HAT domain. This interaction prompted us to test whether Polι is acetylated by p300 in vivo and in vitro. As demonstrated in Fig. 3a, acetylation of FLAG-tagged Polι was clearly detected in an immunoprecipitated (IP) complex using anti-FLAG resins, when cells co-expressing FLAG-Polι and p300-c-myc were pre-treated for 24 h with a deacetylation inhibitor (DKAc) cocktail, in contrast to cells treated with DMSO. To verify whether p300 is responsible for Polι acetylation, we co-transfected FLAG-Polι with either full-length wild-type p300 protein, or a variant with a D1399Y mutation that reduces acetyltransferase activity 52 . The acetylation of Polι decreased dramatically when co-expressed with the p300 D1399Y mutant with the acetyltransferase-deficiency, suggesting that Polι can be acetylated by p300 (Fig. 3b). In addition, we proved that Polι can be acetylated by p300 by performing in vitro experiments, in which full-length p300 was incubated with a purified His-tagged Polι protein and acetyl coenzyme A (Fig. 3c).

Identification of Polι Acetylation sites.
To further explore acetylation of Polι, we performed a series of mass spectrometry analyses to identify residues of Polι that are acetylated in vivo and in in vitro. We initially used anti-FLAG resins to purify FLAG-tagged Polι from HEK293T transiently transfected with the FLAG-Polι plasmid. These cells were then incubated with, or without, DKAc inhibitors for 24 hours (Fig. 4a) before the purified proteins were run on an SDS-PAGE gel and analyzed by mass spectrometry. Additionally, small amounts of the purified proteins were evaluated by western blot using antibodies against acetylated lysines and Polι. Compared to Polι isolated from untreated cells, we observed a faint acetylation signal from FLAG-Polι purified from cells . In vivo and in vitro acetylation of Polι by p300. (a) Immunoprecipitation assay of extracts from HEK293T cells co-transfected with FLAG-Polι and wild-type p300-c-myc. As indicated, prior to harvesting, cells were treated either with DMSO or with deacetylase inhibitor (DKAc) cocktail. (b) Immunoprecipitation assay used extracts from HEK293T cells co-transfected with FLAG-Polι and wild-type p300-c-myc, or FLAG-Polι and p300 containing the D1399Y mutation that abolishes the acetyltransferase activity of p300. 24hrs prior to harvesting, cells were treated with DKAc inhibitor. In panels (a,b), Polι acetylation was verified by western blot using antibodies against acetylated lysines (KAc). Polι was visualized using polyclonal rabbit antibodies against a C-terminal fragment of Polι. (c) In vitro acetylation of recombinant His-tagged Polι by p300. The reaction products were resolved by SDS-PAGE and analyzed by western blot using antibodies against KAc and against Polι.
www.nature.com/scientificreports www.nature.com/scientificreports/ treated with the DKAc inhibitor cocktail. In agreement with the western blot results, mass spectrometry analysis did not reveal any acetylated residues in untreated cells, whereas K550 was identified as the main acetylation site when cells were treated with DKAc inhibitors. In the Polι protein that was acetylated in vitro, we identified four acetylated lysines: K138, K440, K550 and K715 (Fig. S1). K550 is located in the RIR domain that is responsible for Polι's interaction with Rev1, and was the only residue identified in both approaches. To determine if it is the primary site of acetylation, we co-transfected human HEK293T cells with a plasmid expressing p300 and a recombinant plasmid carrying wild-type FLAG-tagged Polι, or one containing a K550R substitution. Transfected cells were treated with DKAc inhibitor cocktail. Proteins pulled-down with anti-FLAG resin were then probed with antibodies against acetylated lysines. In the K550R mutant, the acetylation signal was significantly reduced (Fig. 4b), but still observable, thereby confirming that K550 is the primary acetylation site, but also suggesting the existence of alternate acetylation site(s) in vivo.
Interaction of Acetylated Polι with Rev1. Our observation that K550, which is located in the RIR domain, is the major acetylation site in Polι protein prompted us to check the role of the K550 residue itself and its acetylated form in the interaction of Polι with Rev1. Using anti-HA resins, we performed immunoprecipitation assays on extracts from HEK293T transiently transfected with plasmids encoding HA-tagged full-length Rev1 and either wild-type FLAG-Polι, or a variant carrying a single point mutation at K550. K549 is adjacent to K550 and it is known that if the primary residue is unavailable for posttranslational modification, the neighboring sites could be modified 46 . To determine if K549 might compensate for K550 if it is unavailable for acetylation, we also tested whether a double point mutation of K549 and K550 influences Polι's interaction with Rev1. Initially, we tested lysine-to-arginine (K → R) mutants. Both, the single and double mutant of Polι maintained the interaction with Rev1 (Fig. 5a). To study the effect of the acetylation on the Polι-Rev1 interaction, we replaced the lysine residues with glutamine, since previous studies have suggested that glutamine substitutions simulate acetylation 53,54 . Immunoprecipitation experiments with K → Q Polι mutants indicate that there is no effect on Polι's ability to interact with Rev1 ( Fig. 5b). As a control for these experiments, we used an F546A/F547A Polι mutant, that has been shown to abolish the interaction between Polι and Rev1 55 . Together, our results suggest that K → R mutations, which simulate an unmodified state, or K → Q mutations, which mimic acetylation, do not significantly change the interaction between Polι with Rev1.
polι Acetylation in Response to a Variety of DNA Damaging Agents. The Polι acetylation observed in the previous experiments was mostly detected in cells overexpressing the p300 protein. To determine the physiological conditions under which Polι is acetylated we tested acetylation of the ectopically expressed FLAG-Polι in cells with native levels of p300, after treatment with a variety of DNA damaging agents. We investigated the influence of UV irradiation, which induces cyclobutane pyrimidine dimers and other photoproducts; MMS, a source of alkylating damage; zeocin stimulating DNA double strand breaks; hydrogen peroxide, causing DNA oxidative lesions and two DNA crosslinking agents, used in chemotherapy, mitomycin C and cisplatin. Additionally, we have also tested whether hydroxyurea, which blocks replication fork progression due to decreased dNTPs production, has any influence on the acetylation status of Polι. As shown in Fig. 6a, exposure of cells to most of the DNA damaging agents and hydroxyurea did not result in significant Polι acetylation. In contrast, MMS treatment substantially induced Polι acetylation. We also observed a weak Polι acetylation signal in hydrogen peroxide treated cells. To further analyze the phenomenon of Polι acetylation in response to MMS and hydrogen peroxide, we tested whether other alkylating and oxidizing agents were also able to induce Polι acetylation. We exposed HEK293T cells to increasing concentrations of hydrogen peroxide and potassium bromate, which is known to cause DNA oxidative lesions (Fig. 6b). Both oxidative agents, at higher doses, caused some acetylation of Polι. In comparison to MMS, we investigated the influence of other alkylating agents, EMS and MNNG, on Polι www.nature.com/scientificreports www.nature.com/scientificreports/ acetylation (Fig. 6c). Both alkylating agents caused some acetylation of Polι that was much weaker than MMS. Since alkylation by MMS proceeds via second-order nucleophilic substitution (S N 2) and MNNG is an S N 1 alkylating agent, we examined acetylation of Polι with DMS, another S N 2 alkylating agent. Interestingly both MMS and DMS induced a much higher acetylation level than EMS and MNNG. It is known that compared to MMS, much lower concentrations of DMS induce the same amount of alkylation damage 56 , which might explain the difference in the strength of the acetylation signal when using equimolar amount of MMS and DMS.

Characterization of MMS-induced Polι acetylation.
To further corroborate MMS-induced Polι acetylation, we used mass spectrometry analysis to identify Polι acetylation sites in cells exposed to MMS. HEK293T cells transiently transfected with the FLAG-Polι plasmid were exposed to 200 μM MMS for 1 h. In a control sample from cells not treated with MMS, K550 was again the only acetylated residue. The analysis of Polι protein isolated from MMS-exposed cells also revealed acetylation of K550 and an additional three other lysine residues; K440, K446 and K530 (Fig. S2). Next, by using the K550R mutant, we examined whether K550 is an important site for MMS-induced Polι acetylation. Indeed, MMS-stimulated Polι acetylation was drastically reduced when K550 was unavailable for modification (Fig. 7a). We conclude that K550 is, in general, the main Polι acetylation site after MMS treatment. To test the in vivo properties of the acetylated and unacetylated protein, we examined if Polι variants with K550R and K550Q substitutions could accumulate in DNA replication foci similar to the wild-type protein. Our results show that mutations that either block posttranslational modification(s) at K550 (K550R) or mimic its acetylation (K550Q), do not alter the accumulation of Polι into "replication factories" (Fig. 7b) suggesting that acetylation functions in fine tuning, rather than in the general regulation of Polι's cellular activities. We also speculate that, since the main acetylation site is located outside the catalytic domain of Polι, it is unlikely to directly affect Polι's enzymatic activities. We do not however, exclude that it might indirectly  www.nature.com/scientificreports www.nature.com/scientificreports/ influences enzymatic activities of the polymerase via interactions with other proteins, and this will be a subject of future investigation.
The acetylation of Polι in response to MMS treatment was observed without ectopic expression of p300. This raised the question of whether p300 itself, or other acetyltransferases modify Polι under alkylating conditions. To establish whether p300 acetyltransferase is responsible for the modification, we disrupted the p300 gene in a HEK293T cell line using CRISPR-Cas9 technology and assayed the acetylation of Polι in response to MMS and DMS in the p300 knock-out cells. To our surprise, in two independently isolated p300 knock-out cell lines that were transiently transfected with the FLAG-Polι plasmid, MMS/DMS-induced acetylation of Polι was still observable (Fig. 8a), suggesting that p300 is not the only acetyltransferase responsible for Polι's modification. In parallel, we examined Polι acetylation following MMS treatment in cells exposed to the p300 acetyltransferase inhibitor, L002. HEK293T cells transfected transiently with FLAG-Polι were pre-incubated for 24 hrs with DKAc inhibitor and subsequently exposed, or not (controls), to L002. After a 1 hour exposure, cells were treated with MMS for an additional one, or two hours. Polι acetylation was assessed in extracts collected at the various time points. As shown in Fig. 8b, Polι acetylation is only noticeable in extracts from cells that were not treated with p300 inhibitor, strongly suggesting that incubation of cells with L002 effectively prevented Polι acetylation. These findings imply that p300 is accountable for Polι acetylation in response to MMS exposure. However, in human cells, p300 has a paralogue, CBP, and together they are the only members of one of several classes of Unless otherwise stated, HEK293T cells were transiently transfected with the FLAG-Polι plasmid. 24 hrs post transfection cells were treated for 24 hrs with DKAc inhibitor cocktail. Proteins were visualized by western blot using the following antibodies; polyclonal antibodies against p300 (p300 protein); monoclonal antibodies against acetylated lysines (acetylated Polι); polyclonal rabbit antibodies against a C-terminal fragment of Polι; or monoclonal antibodies against the FLAG epitope, as indicated.
www.nature.com/scientificreports www.nature.com/scientificreports/ acetyltransferase families. It is known that to some extent L002 inhibits CBP acetyltransferase activity and p300 and CBP often acetylate the same substrates, albeit with different potency 16,57 . To verify that CBP can acetylate Polι, we assayed Polι acetylation in HEK293T cells co-transfected with the plasmid carrying FLAG-tagged Polι and a plasmid carrying murine CBP. In the cells treated for 24 hrs with DKAc inhibitor, Polι acetylation was clearly observable (Fig. 8c). To further confirm the ability of CBP to acetylate Polι, we assayed cells co-transfected with FLAG-tagged Polι and either wild-type CBP, or the inactive L1435A/D1436A mutant 58 . In cells expressing the HAT catalytic dead CBP mutant, Polι was not acetylated, thus confirming that CBP also acetylates Polι (Fig. 8c). Finally, we disrupted the CBP gene in a HEK293T cell line using CRISPR-Cas9 technology and verified acetylation of Polι in response to MMS and DMS in the CBP knock-out cells (Fig. 8d). Similar to the p300 KO cells, acetylation of Polι was still observable. Together, our results suggest that both p300 and CBP acetyltransferases can acetylate Polι.
As both over-expression of p300/CBP from a plasmid, or MMS/DMS treatment resulted in increased Polι acetylation, we decided to check whether treatment with S N 2 alkylating agents influences the levels of the p300/ CBP acetyltransferases. In our experiment, we examined the level of native p300 and CBP protein in cells treated with MMS, or DMS. It appears that the amount of p300 protein is slightly increased in MMS treated cells and a stronger effect is observable in cells treated with DMS. Detection of an augmented amount of acetyltransferase in response to those S N 2 alkylating agents is clearer in the case of CBP (Fig. 8e).
Finally, our observation that Polι undergoes p300/CBP-dependent acetylation in response to S N 2 alkylating agents prompted us to determine whether this phenomenon is shared with other Y-family polymerases, particularly with Polη, which is the closest paralogue of Polι in mammalian cells. We investigated acetylation of Polη in cells transfected with FLAG-tagged Polη over-expressing p300 or treated with alkylating agents. It appears that under the conditions where we observe significant acetylation of Polι there is no acetylation of Polη, suggesting that this regulatory mechanism is specific for Polι (Fig. 8f).

Discussion
Polι is a TLS polymerase present in mammals, some other vertebrates and some fungi. The cellular function of the enzyme is still far from being understood. Despite the fact that it was identified two decades ago 59 and possesses very peculiar features that are not found in other Y-family DNA polymerases, such as a dRP-lyase activity and an extremely high mutagenic potential when copying template T, the cellular function of Polι has yet to be determined 42,60 . In an attempt to decipher the circumstances requiring such an enigmatic enzyme, we investigated the mechanism of Polι's regulation. Polι, like other Y-family polymerases, seems to be primarily controlled by the combined action of posttranslational modifications and protein-protein interactions. Similar to other Y-family DNA polymerases, Polι is monoubiquitinated in vivo, and so far, it is only known that such a modification facilitates its physical and functional interaction with Polη 61 . Mass spectrometry analysis previously identified over 27 unique ubiquitination sites in Polι, localized in different functional domains 25 . Additionally, our earlier studies showed that inhibition of p300 acetyltransferase activity induces Polι polyubiquitination 31 . In the current manuscript, we show that Polι not only directly interacts with the p300 acetyltransferase but is subject to acetylation itself (Fig. 3). p300 has two enzymatic activities, an acetyltransferase activity, associated with its HAT domain, and a ubiquitin ligase activity connected with a non-canonical RING domain 15 . We explored whether either of these domains is located in the region required for an interaction with Polι. Our results revealed that Polι does not interact with the N-terminal half of p300 containing the RING domain, but rather with the C-terminal part of p300, including the HAT domain (Fig. 1). Although, we cannot formally exclude the possibility that Polι can also be ubiquitinated by p300, the lack of an interaction with the N-terminal part of the protein that is required for p300 E3 autoubiquitination and E4 polyubiquitination of p53 suggests that p300 ubiquitination of Polι is unlikely.
The main acetylation site of Polι is K550 located in Polι's Rev1 interacting region (RIR motif 539-ASRGVLSFFSKKQMQD-554). K550 is a conserved residue found in many different species of mammals, but not other vertebrates, which correlates with the conservation of the RIR (Fig. S3). Moreover, K550 is conserved in all human and mouse paralogs of Polι, Polη and Polκ 62 . Nonetheless, we demonstrated that both K550R or K550Q mutations, respectively blocking and mimicking acetylation, do not significantly affect the ability of Polι to interact with Rev1. This result is consistent with previous results defining the RIR motif as x-x-x-F-F-y-y-y-y, consisting of two consecutive phenylalanine residues proceeded by no specific residues (x) and followed by any four residues (y) excluding proline 55 . On the other hand, Liu et al., observed significant reduction in the strength of the interaction between Rev1 and Polκ when the corresponding lysine in Polκ was mutated to glycine (K571G) 63 . Lysine 550 is located near the interface with Rev1 and can also serve as one of many possible ubiquitination sites 25 . In our previous studies, we speculated that attaching ubiquitin to K550 would obstruct the interaction of Polι and Rev1 25 . Following this idea, we can hypothesize that acetylating K550 might prevent its susceptibility to ubiquitination, thereby assuring an opportunity for the two polymerases to interact. In general, the RIRs of TLS DNA polymerases play an essential role in regulating TLS activity, however, the interaction between the C-terminal domain of Rev1 and RIR motifs of Polη, Polι and Polκ are not exceptionally tight and are often competitive with one another. So, it seems that a Rev1-RIR interaction results from a dynamic association, rather than a stable connection. Even though the results of our immunoprecipitation experiments show that in general, blocking, or mimicking, Polι acetylation does not notably alter the interaction with Rev1, we cannot exclude the possibility that it can influence polymerase engagement in the microenvironment of a particular replication block.
To further corroborate the physiological conditions leading to Polι's modification, we investigated the acetylation in response to a variety of DNA damaging agents. The results of our experiments showed that Polι is acetylated in response to alkylating agents and in much lower extent to oxidative agents (Fig. 6).
It is worth noting that similar to Polι acetylation, Lee et al., observed global acetylation of proteins in cells exposed to genotoxic agents, particularly MMS and other alkylating agents and to a lower extent to hydrogen peroxide 64  www.nature.com/scientificreports www.nature.com/scientificreports/ p300/CBP, our results show that these two acetyltransferases are responsible for Polι acetylation (Figs 3 and 8). Additionally, our results suggest that, at least partially, increased Polι acetylation in response to MMS and DMS could be a consequence of higher cellular levels of the two acetyltransferases, as we observed slight induction of both p300 and CBP in MMS or DMS treated cells (Fig. 8d). We are unaware of any information in the literature reporting the inducibility of p300/CBP in response to DNA damaging agents. However, Cohen et al., showed that CBP protein accumulates in the cytoplasm in response to UV radiation 65 .
Although, our results point to MMS/DMS as the main facilitator of Polι acetylation, we noticed that oxidative agents (hydrogen peroxide and potassium bromate) can also induce Polι acetylation (Fig. 6b), but to a much lower extent. Since both alkylation and oxidation of DNA, as well as other DNA damaging treatments used in this research, can activate the ATR/ATM dependent cascade of kinases governing DNA damage response (DDR), one can speculate that Polι acetylation reflects the level of DDR induction. However, based on the literature data on phosphorylation of the landmarks of DDR: Chk1 66 , Chk2 67 or γH2AX 68 , in response to the DNA damaging treatments inducing or not Polι acetylation, we conclude that the Polι acetylation does not correlate with activation of DDR. While the exact mechanism of how Polι is influenced by the cellular responses is not known, it was previously shown that Polι can bypass alkylation or oxidative lesions both in vitro and in vivo [32][33][34]39,[69][70][71][72][73] .
Both alkylation and oxidative DNA lesions are substrates of base excision repair (BER). It is also worth mentioning that Polι displays BER activity in vitro and in vivo and functionally interacts with the BER scaffold protein XRCC1 39,60,74 . Furthermore, Polι possesses deoxyribose phosphate (dRP) lyase activity, a feature characteristic for BER polymerases 58,68 . Strikingly, proteins involved in BER represent a significant group among all proteins involved in DNA damage response that were found to be acetylated 5,[75][76][77][78] . This suggests that protein acetylation can considerably influence the activity of BER enzymes, even though it cannot be unequivocally proven that it activates or represses BER.
Interestingly, the level of Polι acetylation is particularly high in cells treated with two S N 2 alkylating agents: MMS and DMS, in contrast to rather moderate acetylation in response to EMS (S N 1/S N 2) and MNNG (S N 1). One could speculate that the DNA lesions produced by S N 2 rather than S N 1 alkylating agents might serve as a signal for Polι acetylation. The stronger response to S N 2 than S N 1 alkylators is intriguing, since the sites methylated in an S N 2 reaction in duplex DNA are also modified by S N 1 alkylating agents. However, in single-stranded DNA, some sites are more reactive with MMS than with MNNG 79 . Accordingly, DNA repair dioxygenases from the AlkB family cause resistance to cell killing by S N 2 but not the S N 1 alkylating agents [80][81][82] . On the other hand, S N 2 and S N 1 alkylating agents differently modify amino acids residues 83 ; therefore, at the moment we cannot exclude the possibility that Polι acetylation results from S N 2 alkylation of proteins, or other cellular constituents. However, in light of the known DNA damage tolerance/repair activities of TLS polymerases, the functional connection of Polι acetylation with DNA rather than protein modifications by S N 2 alkylating agents seems more credible. Nonetheless, this phenomenon needs to be investigated in detail and will be the subject of future studies.
Despite the fact that Polι, similar to many other cellular proteins, is acetylated in response to MMS, its closest paralogue, Polη, does not follow the same pathway. Additionally, it has been recently shown that Polκ, another member of the Y-family DNA polymerases, is not a substrate for in vitro CBP/p300 acetylation 22 . Polι was thought to be a less efficient and more error prone back-up polymerase for Polη 31,84 . Our results, however, suggest that despite many similarities between these two Y-family enzymes, there are cellular requirements and conditions devoted specifically for Polι and acetylation of the enzyme might play a role in adjusting its performance to these circumstances. Plasmids. Vectors expressing truncated or mutated variants of Polι (pJRM258 -pJRM262) were generated by sub-cloning respective PCR fragments into pJRM46 61 expressing N-terminal FLAG-tagged full-length human Polι. Plasmids pJRM264, pDH77, pGS11, pGS7 and pNWA10 are pJRM46 derivatives with single, or double lysine to arginine, lysine to glutamine or lysine to alanine substitutions that were generated by chemically synthesizing appropriate DNA fragments (Genscript) and subsequently sub-cloned into pJRM46. Vectors pJRM242 and pJRM244, expressing N-or C-terminal fragments of p300 proteins tagged with an HA-epitope, were generated by sub-cloning respective PCR fragments into the pCMV6AN-HA vector (Origene). Plasmid pGS23, expressing full-length wild-type human Rev1 was generated by sub-cloning the codon optimized synthesized REV1L gene (Genscript) into pCMV6AN-HA. Plasmids pGS24 and pGS25 expressing eGFP-tagged full-length Polι with single K550R, or K550Q mutations, respectively, were generated by chemically synthesizing appropriate DNA fragments (Genscript) and their subsequent sub-cloning into pDH24 85 . The full list of all the plasmids used in this study is shown in Table 1. www.nature.com/scientificreports www.nature.com/scientificreports/ Sigma DUO92004-100RXN) was performed using 40 μL of reaction per slide according to the manufacturer's instruction. In brief, slides after overnight incubation with primary antibodies recognizing Polι and p300 were washed with PBS (3x) and incubated with two PLA secondary antibodies with probes, anti-mouse MINUS and anti-rabbit PLUS for 1 hour at 37 °C. Slides were then washed with buffer A, followed by the ligation reaction, for 30 min at 37 °C. Subsequently the amplification reaction was performed for 100 min at 37 °C, followed by washing with buffer B (all ingredients and buffers were provided by manufacturer) and mounting with VECTASHIELD Mounting Medium with DAPI for nucleus visualization. Images of immunofluorescently stained cells were acquired using confocal microscopy. The confocal system consisted of a Zeiss LSM800 Exciter microscope equipped with lasers that produced light at 405, 488, and 594 nm wavelengths was used for the acquisition of pictures of PLA positive cells, 40×/1.30 objective was used to scan samples. A series of continuous optical sections at 0.4 µm intervals along the z-axis of a cell were scanned for all fluorescent signals and stored as a series of 1024 × 1024 pixel images. The laser power and image acquisition settings were kept constant.

Plasmid Transfection and Protein
Protein Purification. Full-length recombinant human Polι was purified as described previously 87 . Briefly, pJM868 plasmid encoding His-tagged Escherichia coli-codon optimized human Polι 87 was expressed in the E. coli strain RW644 88 . The His-Polι protein was purified on HisPur Ni-NTA Superflow Agarose (Thermo Scientific) as recommended by the manufacturer. The eluate containing Polι was dialyzed in buffer including 20 mM sodium phosphate pH 7.3, 10 mM sodium chloride, 10% glycerol, 10 mM 2-mercaptoethanol, and applied to HP Q Sepharose (GE Healthcare). Polι was eluted in a step gradient of sodium chloride. Full-length recombinant p300 was purchased from Active Motif.
Mass Spectrometry. LC-MS analysis of gel slices was conducted in Mass Spectrometry Lab, Institute of Biochemistry and Biophysics, Polish Academy of Science, where they were analyzed by mass spectrometry as a custom contract service.
In vitro Acetylation. Acetylation reactions were performed in acetylation buffer (50 mM Tris pH8, 150 mM sodium chloride, 1 mM DTT, 10 mM sodium butyrate, 5% glycerol) with the addition of 100 μM acetyl coenzyme A, 1 mM PMSF, 100 ng of p300 (Active Motif) and purified recombinant Polι for 1 hour at 30 °C. The reaction was then stopped by the addition of Laemmli buffer and followed by SDS-PAGE.