Small-molecule activation of OGG1 increases oxidative DNA damage repair by gaining a new function

(1)

Karolinska Institutet http://openarchive.ki.se

This is a Peer Reviewed Accepted version of the following article, accepted for publication in Science.

2022-08-19

Small-molecule activation of OGG1 increases oxidative DNA damage repair by gaining a new function

Michel, Maurice; Benitez-Buelga, Carlos; Calvo, Patricia A.; Hanna, Bishoy M.F.;

Mortusewicz, Oliver; Masuyer, Geoffrey; Davies, Jonathan; Wallner, Olov; Sanjiv, Kumar;

Albers, Julian J.; Castaneda-Zegarra, Sergio; Jemth, Ann-Sofie; Visnes, Torkild;

Saste-Perona, Ana; Danda, Akhilesh N.; Homan, Evert J.; Marimuthu, Kartrick; Zhenjun, Zhao; Chi, Celestine N.; Sarno, Antonio; Wiita, Elisee; von Nicolai, Catharina; Komor, Anna J.; Rajagopal, Varshni; Müller, Sarah; Hank, Emily C.; Varga, Marek; Scaletti, Emma R.;

Pandey, Monica; Karsten, Stella; Haslene-Hox, Hanne; Loevenich, Simon; Marttila, Petra;

Rasti, Azita; Mamonov, Kirill; Ortis, Florian; Schömberg, Fritz; Loseva, Olga; Stewart, Josephine; D'Arcy-Evans, Nicholas; Koolmeister, Tobias; Henriksson, Martin; Michel, Dana;

de Ory, Ana; Acero, Lucia; Calvete, Oriol; Scobie, Martin; Hertweck, Christian; Vilotijevic, Ivan; Kalderén, Christina; Osorio, Ana; Perona, Rosario; Stolz, Alexandra; Stenmark, Pal;

Warpman Berglund, Ulrika; de Vega, Miguel; Helleday, Thomas

Science. 2022 Jun 24;376(6600):1471-1476.

American Association for the Advancement of Science (AAAS) http://doi.org/10.1126/science.abf8980

http://hdl.handle.net/10616/48169

If not otherwise stated by the Publisher's Terms and conditions, the manuscript is deposited under the terms of the Creative Commons Attribution-NonCommercial-NoDerivatives License (http://creativecommons.org/licenses/by-nc-nd/4.0/), which permits non-commercial re-use, distribution, and reproduction in any medium, provided the original work is properly cited, and is not altered, transformed, or built upon in any way.

(2)

Submitted Manuscript: Confidential Template revised February 2021

1

Small molecule activation of OGG1 increases repair of oxidative DNA damage by a new function

Authors: Maurice Michel¹*†, Carlos Benítez-Buelga^1,2†, Patricia A. Calvo³‡, Bishoy M. F.

Hanna¹‡, Oliver Mortusewicz¹‡, Geoffrey Masuyer^{4, 5}‡, Jonathan Davies⁵‡, Olov Wallner¹‡,

5

Kumar Sanjiv¹‡, Julian J. Albers¹, Sergio Castañeda-Zegarra^1,6, Ann-Sofie Jemth¹, Torkild Visnes⁷, Ana Sastre-Perona⁸, Akhilesh N. Danda¹, Evert J. Homan¹, Karthick Marimuthu¹, Zhao

Zhenjun¹, Celestine N. Chi⁹, Antonio Sarno¹⁰, Elisée Wiita¹, Catharina von Nicolai¹, Anna J.

Komor¹¹, Varshni Rajagopal¹,Sarah Müller¹, Emily C. Hank¹, Marek Varga¹, Emma R.

Scaletti^5,12, Monica Pandey^{1, 13}, Stella Karsten¹, Hanne Haslene-Hox⁷, Simon Loevenich⁷, Petra

10

Marttila¹, Azita Rasti¹, Kirill Mamonov¹, Florian Ortis¹, Fritz Schömberg¹⁴, Olga Loseva¹, Josephine Stewart¹, Nicholas D’Arcy-Evans¹, Tobias Koolmeister¹, Martin Henriksson¹, Dana Michel¹⁵, Ana de Ory¹⁶, Lucia Acero⁸, Oriol Calvete¹⁷, Martin Scobie¹, Christian Hertweck^{11, 18},

Ivan Vilotijevic¹⁴, Christina Kalderén¹, Ana Osorio^{17, 19}, Rosario Perona^{2, 19}, Alexandra Stolz²⁰, Pål Stenmark^{4, 12}, Ulrika Warpman Berglund¹§, Miguel de Vega³, Thomas Helleday^{1, 13}*

15

Affiliations:

1Science for Life Laboratory, Department of Oncology-Pathology, Karolinska Institutet; 171 76 Stockholm, Sweden.

2Instituto de Investigaciones Biomédicas Alberto Sols (CSIC/UAM); 28029, Madrid, Spain.

3Centro de Biología Molecular 'Severo Ochoa' (CSIC-UAM); 28049 Madrid, Spain.

20

4Department of Pharmacy and Pharmacology, Centre for Therapeutic Innovation, University of Bath; BA2 7AY Bath, United Kingdom.

5Department of Biochemistry and Biophysics, Stockholm University; 106 91 Stockholm, Sweden.

6Department of Clinical and Molecular Medicine (IKOM), Norwegian University of Science and Technology; 1, 7491, Trondheim, Norway.

25

7Department of Biotechnology and Nanomedicine, SINTEF Industry; N-7465 Trondheim, Norway.

8Experimental Therapies and Novel Biomarkers in Cancer, Hospital La Paz Institute for Health Research (IdiPAZ); Madrid, Spain.

9Disciplinary Domain of Medicine and Pharmacy, Faculty of Medicine, Department of Medical

30

Biochemistry and Microbiology, Uppsala University; Uppsala, Sweden.

10Department of Environment and New Resources, SINTEF Ocean; N-7496 Trondheim, Norway.

11Leibniz Institute for Natural Product Research and Infection Biology - Hans Knöll Institute, Department of Biomolecular Chemistry; 07745 Jena, Germany.

12Department of Experimental Medical Science, Lund University; Lund, Sweden.

35

13Sheffield Cancer Centre, Department of Oncology and Metabolism, University of Sheffield; S10 2RX Sheffield, United Kingdom.

(3)

2

14Institute of Organic and Macromolecular Chemistry, Friedrich Schiller University Jena; 07743 Jena, Germany.

15Chemical Processes & Pharmaceutical Development, Unit Process Chemistry I, Research Institutes of Sweden – RISE; 151 36 Södertälje, Sweden.

16Department of Molecular Biosciences, The Wenner-Gren Institute, Stockholm University; 106

5

91 Stockholm, Sweden.

17Familial Cancer Clinical Unit, Human Cancer Genetics Programme, Spanish National Cancer Research Centre (CNIO); 28029 Madrid, Spain.

18Institute of Microbiology, Friedrich-Schiller-University Jena; 07743 Jena, Germany.

19Centro de Investigación Biomédica en Red de Enfermedades Raras (CIBERER); 28029 Madrid,

10

Spain.

20Institute of Biochemistry II and Buchmann Institute for Molecular Life Science, Goethe University Frankfurt; 60590 Frankfurt, Germany.

*Corresponding authors. Email: maurice.grube@scilifelab.se; thomas.helleday@scilifelab.se

†These authors contributed equally to this work

15

‡These authors contributed equally to this work

§Present Address: Oxcia AB; 113 34 Stockholm, Sweden.

Abstract: Oxidative DNA damage is recognised by 8-oxoguanine (8-oxoG) DNA glycosylase 1 (OGG1), which excises 8-oxoG, leaving a substrate for apurinic endonuclease 1 (APE1), initiating repair. Here, we describe a small molecule (TH10785) that interacts with the Phe319 and Gly42

20

amino acids of OGG1, increases the enzyme activity 10-fold and generates a novel β,δ-lyase enzymatic function. TH10785 controls the catalytic activity mediated by a nitrogen base within its molecular structure. In cells, TH10785 increases OGG1 recruitment to and repair of oxidative DNA damage. This alters the repair process, which no longer requires APE1 but instead is dependent on polynucleotide kinase phosphatase (PNKP1) activity. The increased repair of

25

oxidative DNA lesions with a small molecule may have therapeutic applications in various diseases and ageing.

One-Sentence Summary: A small molecule activator of OGG1 generates a new enzymatic function and increases oxidative DNA damage repair.

30

Main Text: The well-established role of OGG1 in DNA repair, cancer and inflammation led to the development of OGG1 inhibitors, among them O8, SU0268 and TH5487 with potential uses as anti-cancer or anti-inflammatory agents (1–3). Around the same time, small molecule activators of OGG1 were reported (4, 5). Applications of these molecules targeting OGG1 in high oxidative stress conditions, such as Alzheimer’s and obesity, have been suggested (6-8). In the literature,

35

prominent examples of enzyme activators exhibit their effect by allosteric control (9-12). In the case of OGG1, the excised 8-oxoG nucleobase has been reported to assist catalysis (13) and we have previously identified potential secondary druggable sites in OGG1 (14). Importantly, while OGG1 has only a weak β-elimination function, other DNA glycosylases are bifunctional and process AP-Sites using a β,δ-lyase activity (Fig. 1A)(15). Thus, to be able to apply small molecule

40

(4)

3

activators of OGG1 in disease models, we sought to investigate the mode of action at a molecular level.

We synthesized previously reported compounds (4) (Details SI) and assessed their ability to thermally stabilize OGG1 using differential scanning fluorimetry (DSF) (Fig. S1). We found that TH10785, Compound C in (4), stabilized OGG1 by 3.8 ± 0.7°C (100 µM). Next, we measured in

5

vitro OGG1 activity using a fluorescence-based biochemical assay(3)in presence of TH10785 or APE1, to increase the reaction rate in a coupled enzyme assay (Fig. 1B). Surprisingly, TH10785 alone, in the absence of APE1, increased the reaction rate in a concentration-dependent fashion and reached a maximal enhancement at 6.25 µM. At higher concentrations the turnover progressively decreased to the levels of the reaction rate of the DMSO control, producing a bell-

10

shaped activity profile (Fig. S2). This increase in activity was specific for OGG1 and its substrate (Fig. S3, Table S1). To test if TH10785 stimulates the OGG1 β-lyase activity on AP-sites, we performed the assay using an uracil-containing substrate in the presence of human uracil-DNA glycosylase 2 (UNG2). The AP-sites generated by UNG2 were resolved with increasing efficiency as well (Fig. S4), suggesting a stimulation of the residual β-lyase activity of OGG1. To confirm

15

this, we analysed the reaction products generated after in vitro incubation of OGG1 with a radioactively labelled double-stranded DNA substrate containing an internal 8-oxoG:C pair or an AP-site. In the presence of TH10785 the β-lyase activity of OGG1 was stimulated for both substrates (Fig. 1C and 1D). To exclude an effect on the DNA glycosylase activity of OGG1 we measured the number of AP-sites generated in the presence of TH10785 at 6.25 µM. No effect was

20

observed (Fig. S5), suggesting that OGG1 activation by TH10785 is confined to its lyase activity.

Unexpectedly, the presence of TH10785 yielded two products as observed using gel electrophoresis. One corresponds to the expected 18mer-3’-phosphate unsaturated aldehyde (PUA), and a second with higher electrophoretic mobility (Fig. 1C and 1D). Investigating the product composition (Fig. 1E)(16), we observed that neither of the products were elongated by

25

adding DNA Polymerase X(17) + dGTP (lane d), indicating the absence of a 3’-OH end. However, treating the degradation products with T4 polynucleotide kinase, which has a robust 3’- phosphatase activity, the electrophoretic fastest product was converted to the more slowly migrating 18mer-3’OH product (lane g). This product was able to be elongated by DNA Polymerase X + dGTP, yielding a 19mer species (lane h). Complete repair to the original 34mer

30

was obtained by adding T4 DNA ligase (lane i). This in vitro pathway reconstitution demonstrates that TH10785 allows OGG1 to efficiently cleave AP-sites through a novel β-δ activity.

To understand how an OGG1 β,δ-lyase activity is controlled by TH10785, we solved the co-crystal structure of TH10785 bound to OGG1, confirming binding to the catalytic site (Fig 1F, Fig S6).

Additionally, we confirmed the need of Lys249 to form a hOGG1-abasic site intermediate adduct

35

(Fig. S7). Complementary, we confirmed active site binding by producing relevant enzyme mutants and assessed their activity in the presence of TH10785 (Fig. S8). Altogether, these results demonstrate that catalytic site binding is critical for OGG1 activation by TH10785. Asking how OGG1 activators can engage the active site and not inhibit the occurring enzymatic reactions, we performed molecular dynamics simulations (Movie S1, S2 and Fig. S9) and KD measurements

40

using fluorescence spectroscopy (Table S2) (18). Both indicated the affinity of TH10785 to OGG1 to be rather low (Table S2). Affinity of TH10785 to OGG1 (KD = 5.5 ± 1.7 µM) increased when adding AP-sites analogue containing dsDNA (KD = 1.3 ± 0.3 µM). To investigate this apparent preference for a ternary complex of DNA, OGG1 and TH10785, we determined the kinetic parameter of the corresponding reactions. Using the AP-site substrate, the enzymatic turnover,

45

vmax, increased dose-dependently to 20 fold compared to the lowest concentration (Fig. 1G and 1H, Fig. S10). For the 8-oxoA substrate, the enzymatic turnover increased as well, but followed the aforementioned bell-shaped curve going through a maximum at 6.25 µM of TH10785. The

(5)

4

results of a DNA binding affinity assay further showed that at the relevant concentration of 6.25 µM, TH10785 does not compete with 8-oxodG for binding to OGG1, unlike the OGG1 inhibitor TH5487 (Fig. S11). Collectively, this indicates that TH10785 acts as an essential cofactor by a de facto increased enzyme activity through stimulation of substrate turnover. As evident by the kinetic analysis, higher concentrations of TH10785 compete with the nucleobase-containing substrate but

5

not with the AP-site containing substrate for binding to OGG1 (Fig. 1G and H).

These results led us to the hypothesis that TH10785 acts in a similar manner as 8-oxoG in the product-assisted catalysis by 8-oxoG postulated by Fromme et al.(13). To test this, we built a ternary complex of OGG1, substrate DNA and TH10785, which was docked in the emptied binding pocket using induced-fit simulation (Fig. S12, S13 and Movie S3). Using this model, we

10

hypothesized the 1-N position to be the center of a proton abstraction at the C’2 position of the open-chain deoxyribose. To investigate this, we systematically evaluated a structure activity relationship based on activation potential (AC50), stabilization in DSF, pKa, KD and product composition in the electro-mobility shift assay used above (Fig. S14-S20). Based on the results obtained, we concluded that TH10785 acts as a catalyst, enhancing β- and inducing a novel δ-

15

elimination by OGG1. By binding into the OGG1 active site with the DNA substrate bound via a transient Schiff-Base intermediate, TH10785 provides an appropriately positioned nitrogen base to mediate proton abstraction. As a consequence of this well-balanced affinity and activity, induction of a β,δ-elimination cascade is established.

To see whether TH10785 also increases OGG1 activity in cells, we first confirmed target

20

engagement to OGG1 using the cellular thermal shift assay (19) (Fig. S21). Next, using live-cell confocal microscopy, we measured OGG1-GFP recruitment kinetics to laser micro-irradiated sites.

While TH5487 reduced, TH10785 increased OGG1-GFP recruitment to laser-induced DNA damage sites (Fig. 2A and Fig. S22). We hypothesized, that TH10785-mediated OGG1 recruitment is controlled by increased turnover of oxidative DNA damage repair through the novel AP-lyase

25

function. Consequentially, a quicker product release would occur. This is strengthened by the observed higher OGG1 mobility using fluorescence recovery after photo bleaching and the absence of an effect of TH10785 towards the DNA binding capacity of OGG1 (Fig. S23 and S11).

To test this, we challenged U2OS cells with KBrO3 for 1 h and measured the accumulation of OGG1 substrates by qPCR, at regions prone to oxidation such as the telomeres(20) or a G4

30

quadruplex forming sequence at the MYC promoter(G4_MYC) (21). As a negative control, a region within the MYC promoter that lies outside the G4 quadruplex forming sequence was included (NO/G4_MYC). At the control region, we found unchanged levels of oxidative DNA damage. In contrast, both telomeric and G4_MYC regions accumulated oxidative DNA damage in response to KBrO3 and the corresponding levels were either reduced or increased by TH10785

35

or TH5487 treatments respectively (Fig. 2B and 2C). In agreement with this data, we performed a modified comet assay and obtained reduced tail moments with TH10785 present (Fig. S24).

Since KBrO3 generates both 8-oxoG and AP-sites, qPCR or Comet assays do not allow discrimination of glycosylase and AP lyase activities. Therefore, we quantified 8-oxodG levels in U2OS cells exposed to KBrO3 for 1 hour by LCMS-MS and immunofluorescence. After treatment,

40

we monitored the resulting levels of 8-oxodG at different time points (4 h and 8 h, Fig. 2D or 3h and 7h, Fig.S25). We found declining 8-oxodG levels over time when using different TH10785 concentrations. In contrast, inhibition of the glycosylase activity was observed with TH5487.

Considering that TH10785 is not a potent OGG1 DNA glycosylase inhibitor (Fig. S5), decreased repair of 8-oxodG in presence of TH10785, compared to DMSO, may reflect the redirection of

45

OGG1 function towards resolving AP-sites or competition for 8-oxoG substrates. To test this hypothesis, we treated U2OS cells with KBrO3 for 1h, followed by incubation with an aldehyde reactive probe (1mM/3h) (22) to label AP-sites and monitor by FACS (Fig. 2E and 2F). While

(6)

5

both OGG1 modulators reduced KBrO3-induced AP-sites (Fig. 2G), we found fewer DNA strand breaks (γH2AX) with TH5487 (Fig. 2H), demonstrating that impairment of OGG1 glycosylase activity leads to reduced number of AP-sites. In contrast, we found more DNA strand breaks (γH2AX) with TH10785 (Fig. 2H), confirming the generation of DNA strand breaks by the catalytic activity of TH10785 in cells. Collectively, these results point towards a new cellular role

5

of TH10785 activated OGG1, favouring AP-sites over 8-oxoG.

Next, we tested to what extent TH10785 can induce β,δ-elimination in cells. We hypothesized that simultaneous stimulation of β,δ-elimination and blocking of PNKP1 activity should overload the system with unrepaired DNA single-strand breaks (Fig. 1A). Thus, DDR was measured by IF using the markers γH2AX as well as 53BP1, in U2OS cells exposed to OGG1 inhibitor or activator (Fig

10

3A, Fig. S26) and analogue compounds (Table S6 and Fig. 3B) alone or in combination with PNKP1i. We found that the PNKP1 inhibitor only induced a strong DDR in combination with OGG1 activators causing β,δ-lyase activity in vitro. To assess this causality, we monitored transcriptional changes using RNA sequencing and found that PNKP1i in combination with TH10785, but not the single treatments, induced a marked transcriptional upregulation of key

15

players of the recognition and repair of DNA double-strand breaks (Fig. 3C). Additionally, cell viability was reduced in combination with PNKP1 inhibition for TH10785, but not for TH5487 (Fig. 3D and 3E). These results demonstrate that activation of the OGG1 β,δ-lyase activity by TH10785 occurs both in vitro and in cells, and that PNKP1 is essential to avoid an accumulation of DNA damage and consequential cell death.

20

In conclusion, we present a novel concept that allows the induction of an OGG1 β,δ-lyase activity by an enzyme directed small molecule catalyst, binding into the active site of the enzyme (Fig. 3F, S27 and S28). The new catalytic function caused by the presence of TH10785 prefers AP-sites over 8-oxoG and creates a dependency of PNKP1 in vitro and in cells. Improving or rerouting of repair pathways dealing with oxidative DNA damage has implications in many diseases such as

25

inflammation, cancer, Alzheimer’s or ageing and the concept outlined here allows for the possibility to control and reroute repair pathways in a new manner (24).

References and Notes

1. N. Donley, P. Jaruga, E. Coskun, M. Dizdaroglu, A. K. McCullough, R. S. Lloyd, Small

30

Molecule Inhibitors of 8-Oxoguanine DNA Glycosylase-1 (OGG1). ACS Chem. Biol. 10, 2334–2343 (2015).

2. Y. Tahara, et al. Potent and Selective Inhibitors of 8-Oxoguanine DNA Glycosylase. J. Am.

Chem. Soc. 140, 2105–2114 (2018).

3. T. Visnes, et al. Small-molecule inhibitor of OGG1 suppresses proinflammatory gene

35

expression and inflammation. Science. 362, 834–839 (2018).

4. B. A. Baptiste, S. R. Katchur, E. M. Fivenson, D. L. Croteau, W. L. Rumsey, V. A. Bohr, Enhanced mitochondrial DNA repair of the common disease-associated variant, Ser326Cys, of hOGG1 through small molecule intervention. Free Radical Biology and Medicine. 124, 149–162 (2018).

40

5. D. Schniertshauer, D. Gebhard, H. van Beek, V. Nöth, J. Schon, J. Bergemann, The activity of the DNA repair enzyme hOGG1 can be directly modulated by ubiquinol. DNA Repair. 87, 102784 (2020).

(7)

6

6. S. Oka, et al. MTH1 and OGG1 maintain a low level of 8-oxoguanine in Alzheimer’s brain, and prevent the progression of Alzheimer’s pathogenesis. Sci Rep. 11, 5819 (2021).

7. S. S. B. Komakula, et al. The DNA Repair Protein OGG1 Protects Against Obesity by Altering Mitochondrial Energetics in White Adipose Tissue. Sci Rep. 8, 14886 (2018).

8. S. S. B. Komakula, B. Blaze, H. Ye, A. Dobrzyn, H. Sampath, A Novel Role for the DNA

5

Repair Enzyme 8-Oxoguanine DNA Glycosylase in Adipogenesis. International Journal of Molecular Sciences. 22, 1152 (2021).

9. W. You, et al. Structural Basis of Sirtuin 6 Activation by Synthetic Small Molecules. Angew Chem Int Ed. 56, 1007–1011 (2017).

10. B. P. Hubbard, et al. Evidence for a Common Mechanism of SIRT1 Regulation by Allosteric

10

Activators. Science. 339, 1216–1219 (2013).

11. S. J. Gardell, et al. Boosting NAD + with a small molecule that activates NAMPT. Nat Comm.

10, 3241 (2019).

12. Z. Huang, et al. Identification of a cellularly active SIRT6 allosteric activator. Nat Chem Biol.

14, 1118–1126 (2018).

15

13. J. C. Fromme, S. D. Bruner, W. Yang, M. Karplus, G. L. Verdine, Product-assisted catalysis in base-excision DNA repair. Nat Struct Mol Biol. 10, 204–211 (2003).

14. M. Michel, et al. Computational and Experimental Druggability Assessment of Human DNA Glycosylases. ACS Omega. 4, 11642–11656 (2019).

15. T. Visnes, M. Grube, B. M. F. Hanna, C. Benitez-Buelga, A. Cázares-Körner, T. Helleday,

20

Targeting BER enzymes in cancer therapy. DNA Repair. 71, 118–126 (2018).

16. T. A. Rosenquist, D. O. Zharkov, A. P. Grollman, Cloning and characterization of a mammalian 8-oxoguanine DNA glycosylase. PNAS. 94, 7429–7434 (1997).

17. B. Baños, J. M. Lázaro, L. Villar, M. Salas, M. de Vega, Editing of misaligned 3’-termini by an intrinsic 3’-5’ exonuclease activity residing in the PHP domain of a family X DNA

25

polymerase. Nucleic Acids Res. 36, 5736–5749 (2008).

18. I. Boldogh, et al. Activation of Ras Signaling Pathway by 8-Oxoguanine DNA Glycosylase Bound to Its Excision Product, 8-Oxoguanine. J Biol Chem. 287, 20769–20773 (2012).

19. R. Jafari, et al. The cellular thermal shift assay for evaluating drug target interactions in cells.

Nat Protoc. 9, 2100–2122 (2014).

30

20. J. M. Baquero, et al. Small molecule inhibitor of OGG1 blocks oxidative DNA damage repair at telomeres and potentiates methotrexate anticancer effects. Sci Rep. 11, 3490 (2021).

21. S. Roychoudhury, et al. Endogenous oxidized DNA bases and APE1 regulate the formation of G-quadruplex structures in the genome. PNAS. 117, 11409–11420 (2020).

(8)

7

22. K. Kubo, H. Ide, S. S. Wallace, Y. W. Kow, A novel sensitive and specific assay for abasic sites, the most commonly produced DNA lesion. Biochemistry. 31, 3703–3708 (1992).

23. P. Srivastava, A. Sarma, C. M. Chaturvedi, Targeting DNA repair with PNKP inhibition sensitizes radioresistant prostate cancer cells to high LET radiation. PLOS ONE. 13, e0190516 (2018).

5

24. N. Owen, I. G. Minko, S. A. Moellmer, S. K. Cammann, R. S. Lloyd, A. K. McCullough, Enhanced cytarabine-induced killing in OGG1-deficient acute myeloid leukemia cells.

PNAS. 118, e2016833118 (2021).

25. D. Zheng, Z. Chen, J. Liu, J. Wu, An efficient route to 1-aminoisoquinolines viaAgOTf- catalyzed reaction of 2-alkynylbenzaldoxime with amine. Org. Biomol. Chem. 9, 4763–4765

10

(2011).

26. S. Gianni, et al. The Kinetics of PDZ Domain-Ligand Interactions and Implications for the Binding Mechanism. J Biol Chem. 280, 34805–34812 (2005).

27. B. M. F. Hanna, T. Helleday, O. Mortusewicz, OGG1 Inhibitor TH5487 Alters OGG1 Chromatin Dynamics and Prevents Incisions. Biomolecules. 10, 1483 (2020).

15

28. T. Visnes, et al. Targeting OGG1 arrests cancer cell proliferation by inducing replication stress. Nucleic Acids Res. 48, 12234-12251 (2020).

29. W. C. Tse, D. L. Boger, A Fluorescent Intercalator Displacement Assay for Establishing DNA Binding Selectivity and Affinity. Acc. Chem. Res. 37, 61–69 (2004).

20

Acknowledgments: We would like to thank Judith E. Unterlass, Mari Kullman Magnusson, Louise Sjöholm, Therese Pham, Sofie Demir, Athina Pliakou, Teresa Sandvall, Kristina Edfeldt, and Michael Sundström for support and discussions. We thank the scientists at stations I04 and I04-1 of the Diamond Light Source, Didcot, Oxfordshire (UK) for their support during data collection (allocations MX15806 and MX21625).

25

Funding:

European Research Council Grant TAROX-695376 (TH) The Swedish Research Council Grant 2015-00162 (TH) The Swedish Research Council Grant 2018-03406 (PS)

MCIN/AEI/10.13039/501100011033 and “ERDF A way of making Europe” grant

30

BFU2017-83900-P (MdV)

Crafoord Foundation grant 20190532 (PS) Alfred Österlund Foundation grant (PS)

The Swedish Pain Relief Foundation grant (TH)

The Swedish Cancer Society grant CAN2018/0658 (TH)

35

The Swedish Cancer Society grant CAN 2017/716 (PS)

(9)

8

The Torsten and Ragnar Söderberg Foundation grant (TH)

The Dr. Åke-Olsson foundation for haematological research grant 2020-00306 (MM) Thomas Helleday Foundation for medical research postdoctoral stipends (MM, CBB) NTNU Enabling Technology Programme on Biotechnology to Valentyn Oksenych (SCZ) EMBO Short-Term Fellowship 9005 (SCZ)

5

FEBS Short-Term Fellowship (SCZ)

SSI Grants for Scandinavian Exchange (SCZ)

German Research Foundation (DFG) 239748522 (AJK, CH) Sonderforschungsbereich (SFB) 1127 (AJK, CH)

The Leibniz Award (CH)

10

The Norwegian Research Council grant 303369 (TV)

Karolinska Institutet Research Foundation Grant 2020-02186 (MM) Lars Hiertas Minne Stiftelse grant (MM)

Asociacion Española Contra Cancer grant Postdoctoral AECC 2020, Nº POSTD20042BENI (CBB)

15

This study was funded by “Instituto de Salud Carlos III CP19/00063 and PI20/00329”, co- funded by European Social Fund “Investigating in your future” and “European Regional Development Fund” (ASP)

CNIO studies are partially funded by Instituto de Salud Carlos III, project reference PI19/00640, cofounded by the European Regional Development Fund (ERDF), “A way to

20

make Europe” (AO)

This project has received funding from the Innovative Medicines Initiative 2 Joint Undertaking (JU) under grant agreement No 875510 (MM, EJH, EW, AS)

This work has received funding from the European Union’s Horizon 2020 research and innovation programme under the Marie Skłodowska-Curie grant agreement No 722729

25

(TH, BMFH).

Author contributions:

Conceptualization: MM, CBB, TH

Data curation: MM, CBB, BMFH, OM, GM, JD, OW, KS, SCZ, ASJ, TV, ASP, JJA, EJH,

30

ZZ, CNC, AS, AJK

Formal analysis: MM, CBB, PAC, BMFH, GM, JD, OW, KS, SCZ, ASJ, TV, ASP, JJA, EJH, ZZ, CNC, AS, AK, DM, AdO, LA, OC

Methodology: MM, CBB, HP, FTGS, CW, JRK, NJB, PRB, JLS, EH

Resources: MM, TV, CNC, AS, MS, CH, IV, CK, AORP, AS, PS, UWP, MdV, TH

35

(10)

9

Investigation: MM, CBB, PAC, BMFH, OM, GM, JD, OW, KS, SCZ, ASJ, TV, ASP, JJA, AND, EJH, KM, ZZ, CNC, AS, EW, CvN, AJK, VR, SM, ECH, MW, ERS, MP, SK, HHH, SL, PM, AR, KM, FO, FS, OL, JS, NDE, TK, MH, AdO, LA, OC

Visualization: MM, CBB, BMFH, OM, GM, MdV

Funding acquisition: MM, CBB, SCZ, TV, CH, IV, CK, AO, RP, AS, PS, UWP, MdV, TH

5

Project administration: MM

Supervision: MM, CBB, OM, OW, KS, TV, MS, CH; IV, CK, AO, RP, AS, PS, UWB, MdV, TH

Writing – original draft: MM, CBB, TH Writing – review & editing: MM, CBB, TH

10

Competing interests: TV, OW, TK and TH are listed as inventors on as US patent no.

WO2019166639A1, covering OGG1 inhibitors. The patent is fully owned by a non-profit public foundation, the Thomas Helleday Foundation for Medical Research, and TH and UWB are members of the foundation board. CK and UWB are employees of Oxcia AB, a company holding a license to WO2019166639A1. OM, OW, KS, ASJ, TV, EJH, EW, OL, TK, MH,

15

MS, CK, PS, UWB, and TH are shareholders of Oxcia AB. All remaining authors declare that they have no competing interests.

Data and materials availability: Transcriptome data have been deposited at the Gene Expression Omnibus (GEO) with the accession number GSE188779. The atomic coordinates and structure factors (codes 7AYY, 7AYZ, 7AY0) have been deposited in the Protein Data

20

Bank (www.wwpdb.org/). The supplementary materials section contains additional data. All other data needed to evaluate the conclusions in this paper are present in either main text or the supplementary materials.

Supplementary Materials Materials and Methods

25

Figs. S1 to S39 Tables S1 to S10 References (25-29) Movies S1 to S3

30

Fig. 1. OGG1 activator TH10785 induces a de-novo β-δ-elimination in vitro allowing for AP- sites as novel substrates: (A) The repair fate for substrates of monofunctional and β,δ-bifunctional DNA glycosylases is either APE1 or PNKP1 dependant. OGG1 as a bifunctional glycosylase with residual product assisted β-elimination activity recognizes and excises 8-oxoG. Due to the inefficient process, APE1 performs incision of the AP-site and excision of 3′-Phosphate

35

Unsaturated Aldehyde (PUA) to generate a 3′-hydroxyl end compatible with a DNA polymerase.

Other DNA glycosylases, such as endonuclease VIII-like proteins 1 and 2 (NEIL1/2), can catalyse an additional β,δ-elimination in addition to the DNA glycosylase function which includes incision of the AP-site and removal of resulting PUA. This generates a 3′-hydroxyl end, which is only compatible with a DNA polymerase after additional elimination of the generated 3′-phosphate by

40

PNKP1. Thus, these enzymes are functioning independently of APE1; (B) TH10785 dose-

(11)

10

dependently stimulates the excision of 8-oxoA:C by OGG1 in the absence of APE1. Data are presented as averages ± SD of 2 technical replicates from 2 independent experiments. nFU, normalized fluorescence units generated after subtracting the DMSO control; (C) A radioactively labelled 8oxoG:C containing DNA substrate confirms an AP-lyase activity of OGG1 but yields two distinct products. 1 nM of the [³²P]5’-labelled 8oxoG-containing substrate was incubated with

5

25 nM OGG1 and either 10% DMSO or 6.25 μM TH10785, as indicated. Samples were analysed by 7M urea-20% PAGE and autoradiography. Position of products as indicated; (D) Using 1 nM U:C containing DNA substrate and 6 nM E. coli UDG in the same assay set-up confirms AP-sites as substrates and yields two distinct products; (E) APE1 independent de-novo β,δ-elimination of AP-sites by OGG1 is confirmed in the presence of TH10785. 1 nM of the 8oxoG containing

10

substrate was incubated with 6.25 µM TH10785 and 50 nM OGG1 either in the presence (+) or absence (-) of 0.2 nM APE1. After incubation for 30 min at 37ºC, reactions were further incubated in the presence of the indicated protein. The position and a scheme of the reaction products are shown; (F) Close-up view of ligand TH10785 (magenta) binding to human OGG1. Important residues in the binding site are marked, hydrogen bond interactions are shown as black dashed

15

line. A central π-stacking of the quinazoline ring with Phe319Ala was observed supporting the inactivity of the Phe319Ala with TH10785. Further stabilisation of the complex is added by a hydrogen bond between the proton of the secondary amine and Gly42. In addition, the cyclopropyl group is located within the shallower lipophilic pocket surrounded by Pro266 and Met271.

Remarkably, in the hOGG1 complex, TH10785 makes a hydrogen bond with the acidic side chain

20

of Asp268, whilst Lys249 is oriented away from the catalytic pocket; (G) and (H) Kinetic parameter vmax determined from saturation curves with 8-oxo (G) and AP-site (H) containing substrates reflect an increased turnover of AP-site substrate in presence of TH10785. Although binding to the active site, no competitive behaviour is observed for AP-sites. In contrast, at saturating concentrations TH10785 competes with the 8-oxo containing substrate. Data is

25

represented as mean values ± SD. Data is the average of 2 independent experiments.

Fig. 2: TH10785 allows OGG1 to increase DNA repair by addressing AP-sites: (A) In the presence of TH10785 more OGG1 is recruited to laser damaged sites. (left): Representative live- cell confocal images of microirradiated U2OS OGG1-GFP cells treated with OGG1 inhibitor (TH5487, 10 µM), OGG1 activator (TH10785, 1 µM) or vehicle (DMSO). (right): Connecting

30

lines ± SEM of OGG1-GFP recruitment kinetics. Scale bar, 5 µm. NFU, normalized fluorescence units (the complete set of images as well as comparative analysis of Maximum OGG1-GFP intensity at laser-irradiated sites, are represented in Fig.S22). (B) and (C) TH10785 mediated repair by OGG1 decreases oxidative damage in guanine rich regions of the genome; OGG1 incision assay on DNA extracted from U2OS cells treated with 20mM KBrO3 for 1h alone or in

35

combination with TH10785 (10µM), TH5487(10µM) or DMSO. Each bar represents the mean ±

SEM of (B) Telomere Delta Ct, (C-left) G4_MYC Delta Ct and (C-right) Control NO/G4_MYC.

Delta Ct values are calculated as Target Ct for the DNA incubated with OGG1-Target Ct for the DNA incubated with the buffer. Data is the average 3 independent experiments with 3 or 4 replicates each. Statistical significance was calculated by Mann-Whitney test (ns, non-significant;

40

*, P<0.05; ***, P<0.001). (D) With TH10785 present, accumulated 8-oxodG in U2OS cells after oxidative stress declines slower than in the DMSO but faster than in the TH5487 control. Cells were treated with 20 mM KBrO3 for 1 h and afterwards, cells were exposed to 1 µM/10 µM TH10785, 10 µM TH5487 or DMSO at indicated doses for 4 h and 8 h. The amount of genomic 8-oxodG was quantified with LC-MS/MS. Each bar represents the mean ± SD of three replicates

45

from two independent experiments. Statistical significance was calculated using two-tailed unpaired t-test test (ns, non-significant; *, P<0.05; **, P<0.01; ***, P<0.001). (E-H) TH10785

(12)

11

increases AP-sites cleavage during BER initiation. (E) Scheme of treatment and incubation with ARP_probe in U2OS cells. (F) Histogram overlay for ARP-STREP_FITC geometric mean of cells at different treatment conditions. (G) Comparative analysis of ARP-STREP_FITC signal induction over DMSO in percentage. Each bar represents the mean± SD. Data is the average of two independent experiments with 3 and 4 biological replicates each. Statistical significance was

5

determined using two-tailed unpaired t-test (ns, nonsignificant; *, P< 0.05; ***, P< 0.001). (H) Comparative analysis of the average number of γH2AX foci per nucleus using quantitative microscopy. Each bar represents the mean foci/nuclei ± SEM of five independent experiments. For each experiment a minimum of 200 cells were analysed. Statistical analysis was performed using Mann-Whitney test (***, p<0.001, ns, non-significant). Mean ± SEM values for Fig.2A and

10

Figures 2B, 2C, 2H are reported in Table S8 and S9, respectively.

Fig. 3: TH10785 induced OGG1 β,δ-lyase activity shifts cells towards PNKP1 dependence.

(A) and (B) OGG1 repair products are causing DNA damage when β,δ-elimination is established.

(A) Quantification of γH2AX foci per nuclei. Each bar represents the mean ± SEM of three independent experiments. Statistical analysis was performed using Mann-Whitney test (ns, non-

15

significant; *, P< 0.05; **, P< 0.001; ***, P< 0.0001). (B-Up) Quantification of 53BP1 nuclear intensity in U2OS cells treated for 24 h with the indicated conditions. PNKP1i, PNKP1 inhibitor;

C_TH# (Combination of TH# and PNKPi). Each bar represents the mean ± SEM from three independent experiments in U2OS cells treated for 24 h with the indicated conditions. For each experiment a minimum of 200 cells was analysed. Statistical analysis was performed using Mann-

20

Whitney test (****, p<0.0001). (B-Down) Duplicated image from Supplementary Figure S20 to illustrate that OGG1 activators causing β,δ-lyase activity in vitro induce DDR activation when combined with PNKP1i in cells. (C) OGG1 activation through β,δ-elimination in combination with PNKP1 inhibition causes an upregulation of members of the DNA damage response. Heat map showing the expression of genes involved in DNA Double Strand Break Procession (Gene

25

Ontology gene set). U2OS cells were treated with 10µM TH10785, 10µM PNKP1i, 10µM TH10785 + 10µM PNKP1i or DMSO for 24 h (biological n = 3). Colour indicates row-wise z- scored DESeq2-normalized counts. Samples and genes are clustered hierarchically (Euclidean distance, complete linkage). (D) and (E) OGG1 activation through β,δ-elimination in combination with PNKP1 inhibition decreases cell survival. Connecting lines showing mean ± SEM normalized

30

viability values of U2OS cells treated for 72h hours with the indicated doses PNKP1i in combination with TH10785 (D) or TH5487 (E). Data are average of 3 independent experiments.

(F) Molecular mechanism of OGG1 enzyme-directed small molecule catalysis: TH10785 establishes active site binding of OGG1 by addressing key amino acid residues Ph319 (green, π- stacking) and Gly42 (blue, H-bonding). An appropriately placed nitrogen base enables abstraction

35

of the 2’C-proton and increases β-elimination. Further, a de-novo δ-elimination is installed by the catalyst, possibly occurring simultaneously to product release. Mean ± SEM values for Fig.3A and Figures 3B as well as 3D, 3E are reported in Table S9 and S10, respectively.

(13)

A damaged nucleobase

monofunctional β,δ-bifunctional

AP sites P

APE1 PNKP1

DNA polymerase

B

C D

E

G H

F

5'³²P-GTACCCGGGGATCCGTAA8GCGCATCAGCTGCAG-3' CATGGGCCCCTAGGCATTCCGCGTAGTCGACGTC

5'³²P-CTGCAGCTGATGCGCUGTACGGATCCCCGGGTAC-3' GACGTCGACTACGCGCCATGCCTAGGGGCCCATG

0.01 0.1 1.0 10 100 1000

0 2 4 6 8 10

log[TH10785]/µM Vmax(ΔFºsec-1)

8-oxo

Vmax(ΔFºsec-1) 0 10 20 30

0.01 0.1 1.0 10 100 1000

log[TH10785]/µM abasic site U E.coli UDG

DMSO TH10785

19 34

16 34

34-mer

18-mer-3′PUA 18-mer-3′P

34-mer

15-mer-3′PUA 15-mer-3′P

Ligation product

18-mer-3′PUA 18-mer-3′P 19-mer 18-mer-3′OH

30min

incubation - + + + + + + + + +

- + - - + - - - - ^hOGG1^hAPE1

10min additional incubation

- + + - - + + + - - - - - - - - + + + - - - +

PolX + dGTP hAPE1 PNK Ligase 34-mer

10 20 30 40 50 60

-1000 0 1000 2000 3000 4000 5000 6000

Time [min]

nFU

TH10785 100 µM TH10785 50 µM TH10785 6,25 µM TH10785 3.125 µM TH10785 1.56 µM TH10785 0.39 µM TH10785 0.19 µM APE1, w/o OGG1 APE1 w/ OGG1 Glycosylase

5'³²P-GTACCCGGGGATCCGTAA8GCGCATCAGCTGCAG-3' CATGGGCCCCTAGGCATTCCGCGTAGTCGACGTC

19 34

a b c d e f g h i

0 2 4 8 16 32 0.25 0.5 1 2 4 time(min) 0 2.5 5 10 20 40 60 0.25 0.5 1 2 4 8 time(min)

(14)

DMSO Legend Legend

before 30 sec 120 sec

1 µM TH1078510 µM TH5487DMSO 0.91.01.11.21.31.4

0 20 40 60 80 100 120

DMSO

1 µM TH10785 10 µM TH5487

0 2 4 6 8 10

DMSO KBrO

3 + DMSO KBrO

3 + TH10785 KBrO

3 + TH5487

DMSO KBrO

3 + DMSO KBrO

3 + TH10785 KBrO

3 + TH5487 DMSO

KBrO 3 + DMSO

KBrO 3 + TH10785

KBrO 3 + TH5487 0

2 4 6 8 10

0 2 4 6 8 10

OGG1-GFP NFU ΔCt Tel (OGG1-Buffer)

ΔCt G4_MYC (OGG1-Buffer) ΔCt NO/G4_MYC (OGG1-Buffer)

DMSO KBrO

3 + DMSO DMSO

1 µM TH1078510 µM TH1078510 µM TH5487 DMSO

1 µM TH1078510 µM TH1078510 µM TH5487

8oxoG per MdN

0 5 10 15

0 20 40 60 80 100

10³ 10⁴

FITC-A

Normalized to Mode (cell number)

0 50 100 150

DMSO KBrO

3 + DMSO KBrO

3 + TH10785 KBrO

3 + TH5487

ARP-STREP FITC signal (%)

DMSO KBrO

3 + DMSO KBrO

3 + TH10785 KBrO

3 + TH5487

ɣH2AX Foci/ Nuclei

0 5 10 15 20

16h 1h 3h

Compound incubation

KBrO3 20 mM

compound incubation ARP 1mM

compound incubation

-IF:ɣH2AX -FACS: ARP ns

* ***

ns

*

ns

4h recovery 8h recovery

ns** *

***

**** ****

A B

C D

E

F G H

G4_MYC NO/G4_MYC TSS MYCT

GGGAGGCGTGGGGGTGGGACGGTGGGGTACAGA

*

***

ns time(s)

DMSOKBrO3 KBrO3 + TH10785 KBrO3 + TH5487

(15)

PNPKi 0 µM

µM PNPKi 1.25 µM

µ M

PNPKi 2.5 µM

µM PNPKi 5 µM

µM PNPKi 10 µM

µM PNPKi 20 µM

µM SLX1A

WDR70 C14orf39 RAD50 ATMSMARCAD1 RBBP8 MRE11 BRCA1 NBNRNF138 HELBDNA2 BARD1 UBE2V2 BLMBRIP1 RAD52 SPO11 EXD2SLX1B KAT5SLX4 SETMAR UBE2N

treatment

Z-score

DMSO PNKP1i TH10785 C_TH10785

4

-4 0

8 6 4 2 0

Non-treated PNKP1i

10µM TH10785C_TH10785 C_TH5487 10µM TH5487

γH2AX Foci/ Nuclei

10µM TH10785C_TH10785 C_TH5487 10µM TH5487

Non-treated 0.5mM HU PNKP1i 8-Br-G C_8-Br-G

TH12161

C_TH12161 TH11738 C_TH1

1738 0.95

1.10

1.05

1.00

Relative 53BP1 signal intensity (RFU)

DMSO TH10785 TH5487 TH12161 TH11738 8-Br-G

0 5 10 15 20

0 20 40 60 80 100

TH10785 (µM)

0 5 10 15 20

TH5487 (µM) 0

20 40 60 80 100

Cell survival (%) Cell survival (%)

16mer-P′UA 16mer-P

34mer

C

A B

time(min) Alk. c 1.5 2.5 5 1.5 2.5 5 1.5 2.5 5 1.5 2.5 5 1.5 2.5 5 1.5 2.5 5

D E

F

*

**

***

ns

****

PNPKi 0 µM PNPKi 1.25 µM PNPKi 2.5 µM PNPKi 5 µM PNPKi 10 µM PNPKi 20 µM

O OH P'5

P'3 O HN

H N O NH

O

N NH 2

Lys249 Phe319

O N

Gly42

1' 2' pka 10.15

+H - 8-oxoG + TH10785

O OH P'5

P'3 O

Lys249 Phe319

O N

Gly42

N N

CyNH

H 2' 1' pka 6.55

H H

OH OH P'5

P'3 O Lys249

N N

CyNH

H 2' 1'

H P'5 OH ^OH

P'3 OH N

N CyNH

1' H2N 2' O

Lys249

~H -H

β,δ-bifunctional β,δ-bifunctional

N

N N

(16)

1

Supplementary Materials for

Small molecule activation of OGG1 increases oxidative DNA damage repair by gaining a new function

Maurice Michel¹*†, Carlos Benítez-Buelga^1,2†, Patricia A. Calvo³‡, Bishoy M. F.

Hanna¹‡, Oliver Mortusewicz¹‡, Geoffrey Masuyer^{4, 5}‡, Jonathan Davies⁵‡, Olov Wallner¹‡, Kumar Sanjiv¹‡, Julian J. Albers¹, Sergio Castañeda-Zegarra^1,6, Ann-Sofie

Jemth¹, Torkild Visnes⁷, Ana Sastre-Perona⁸, Akhilesh N. Danda¹, Evert J. Homan¹, Karthick Marimuthu¹, Zhao Zhenjun¹, Celestine N. Chi⁹, Antonio Sarno¹⁰, Elisée Wiita¹,

Catharina von Nicolai¹, Anna J. Komor¹¹, Varshni Rajagopal¹,Sarah Müller¹, Emily C.

Hank¹, Marek Varga¹, Emma R. Scaletti⁵, Monica Pandey^{1, 13}, Stella Karsten¹, Hanne Haslene-Hox⁷, Simon Loevenich⁷, Petra Marttila¹, Azita Rasti¹, Kirill Mamonov¹, Florian

Ortis¹, Fritz Schömberg¹⁴, Olga Loseva¹, Josephine Stewart¹, Nicholas D’Arcy-Evans¹, Tobias Koolmeister¹, Martin Henriksson¹, Dana Michel¹⁵, Ana de Ory¹⁶, Lucia Acero⁸,

Oriol Calvete¹⁷, Martin Scobie¹, Christian Hertweck^{11, 18}, Ivan Vilotijevic¹⁴, Christina Kalderén¹, Ana Osorio^{17, 19}, Rosario Perona^{2, 19}, Alexandra Stolz²⁰, Pål Stenmark^{4, 12},

Ulrika Warpman Berglund¹§, Miguel de Vega³, Thomas Helleday^{1, 13}* Correspondence to: maurice.grube@scilifelab.se; thomas.helleday@scilifelab.se This PDF file includes:

Materials and Methods Supplementary Text Figs. S1 to S39 Tables S1 to S10

Captions for Movies S1 to S3

Other Supplementary Materials for this manuscript include the following:

Movies S1 to S3

(17)

2 Materials and Methods

Chemical Synthesis

All reagents and solvents were purchased from Sigma-Aldrich, Combi-Blocks, Thermo Fischer Scientific, or VWR and were used without purification. Unless otherwise stated, reactions were performed without care to exclude air or moisture. Analytical thin- layer chromatography was performed on silica gel 60 F-254 plates (E. Merck) and visualized under an UV lamp. Flash column chromatography was performed in a Biotage® SP4 MPLC system using Merck silica gel 60 Å (40–63 μm mesh). ¹H and ¹³C NMR spectra were recorded on Bruker DRX-400 MHz, Bruker 250, Bruker Fourier 300, Bruker 600, Bruker 500, and Bruker Avance 400 spectrometer. Chemical shifts are expressed in parts per million (ppm) and referenced to the residual solvent peak. For ¹H and ¹³C measurements, the chemical shift is referred to an internal standard; the remaining protons or respectively the carbons of the corresponding deuterated solvent were used. High-resolution mass spectra (HRMS) were measured with EI or ESI ionization. A chromatographic purification was performed before each measurement. The Thermo Q-Exactive plus device for ESI-mass spectra was coupled to a binary UHPLC system. For EI-measurement, a GC-system was coupled to the Thermo Q-Exactive GC device. Analytical LC–MS were performed on an Agilent MSD mass spectrometer connected to an Agilent 1100 system with: Method ST1090A3: Column ACE 3 C8 (50 × 3.0 mm); H²O (+ 0.1% TFA) and MeCN were used as mobile phases at a flow rate of 1 ml/min, with a gradient from 10% – 90% in 3 min; or Method B0597X3: Column Xterra MSC18 (50 × 3.0 mm); H2O (containing 10 mM NH4HCO3; pH = 10) and MeCN were used as mobile phases at a flow rate of 1 ml/min, with a gradient of 5% – 97% in 3 min. For LC–MS, detection was made by UV (254 or 214 nm) and MS (ESI+). Preparative LC was performed on a Gilson system using Waters C18 OBD 5 μm column (30 × 75 mm) with water buffer (a) 50 mM NH4HCO3 at pH 10 or b) 0.1% TFA) and acetonitrile as mobile phases using a flow rate of 45 ml/min. All final compounds were assessed to be >95%

pure by LC–MS analysis.

4-[4-(1H-imidazol-1-yl)phenoxy]-5-methyl-2-{octahydropyrrolo[1,2-a]pyrazin-2- yl}pyrimidine (TH10715)

0.2 mmol (32.6 mg) of 2,4-dichloro-5-methyl-pyrimidine was dissolved in acetone and 0.2 mmol (32.0 mg) 4-imidazol-1-ylphenol and 0.2 mmol (57.3 mg) potassium carbonate were added. The reaction mixture was stirred at room temperature for 16 hours, quenched with concentrated NaHCO3 solution and extracted three times with DCM. The combined organic phases were dried with Na2SO4, filtered and concentrated under reduced pressure. The residue was purified by RP-HPLC and the product was isolated in 57.5% (57.3 mg) yield and directly used for the next step.

0.115 mmol (33.0 mg) of 2-chloro-4-(4-imidazol-1-ylphenoxy)-5-methyl-pyrimidine was dissolved in isopropanol and 0.115 mmol (14.5 mg) 1,2,3,4,6,7,8,8a- octahydropyrrolo[1,2-a]pyrazine and 0.115 mmol (14.5 mg) DIPEA were added. The reaction mixture was refluxed for 5 hours and the solvent removed under reduced pressure. The residue was purified by RP-HPLC (b) using a gradient of acetonitrile and water as gradient. The product was isolated as the TFA salt in 21.0% (9.1 mg) yield.

1H–NMR (600 MHz, MeOD): δ = 8.25 (s, 1H), 8.13 (s, 1H), 7.75 (t, J = 1.3 Hz, 1H), 7.70 – 7.67 (m, 2H), 7.33 – 7.30 (m, 2H), 7.08 (bs, 1H), 3.13 (d, J = 5.1 Hz, 1H), 2.47 (dt,

(18)

3 J = 3.7 Hz, 1.8 Hz, 4H), 2.09 (s, 3H), 1.87 – 1.65 (m, 4H), 1.25 – 1.21 (m, 4H) ppm. ¹³C- NMR (150 MHz, MeOD): δ = 176.6, 169.4, 169.0, 168.8, 160.5, 145.3, 145.1, 143.4, 139.6, 139.5, 132.7, 132.6, 130.7, 127.7, 127.7, 63.0, 58.2, 51.3, 30.1, 27.7, 26.4, 22.0, 21.2 ppm. HRMS [C21H24N6O]: Calc. [M+H]⁺ 377.2084, found 377.2083

(2S)-2-{[6-methyl-2-(phenylamino)pyrimidin-4-yl]amino}-N-phenylpropanamide (TH10716)

0.2 mmol (32.6 mg) 2,4-dichloro-6-methyl-pyrimidine was dissolved in ethanol and 0.2 mmol (32.6 mg) (2S)-2-amino-N-phenyl-propanamide and 0.2 mmol (35 µL) DIPEA was added. The reaction was stirred at 50 °C for 16 hours. The solvent was removed under reduced pressure and the residue was purified using RP-HPLC (b). (2S)-2-[(2- chloro-6-methyl-pyrimidin-4-yl)amino]-N-phenyl-propanamide was isolated in 20.6%

(12.0 mg) yield and used for the next step.

0.0413 mmol (12.0 mg) of (2S)-2-[(2-chloro-6-methyl-pyrimidin-4-yl)amino]-N- phenyl-propanamide was dissolved in isopropanol and 0.0825 mmol (8 µL) Aniline and 0.0574 mmol (10 µL) DIPEA were added. The reaction mixture was refluxed for 16 hours, evaporated under reduced pressure and purified by RP-HPLC (b). The product was isolated as TFA salt in 50.9% (7.3 mg) yield.

1H–NMR (600 MHz, MeOD): δ = 9.95 (s, 1H), 8.83 (s, 1H), 7.55 (d, J = 7.3 Hz, 2H), 7.47 (d, J = 7.3 Hz, 2H), 7.15 (t, J = 7.9 Hz, 2H), 6.97 (t, J = 7.5 Hz, 2H), 6.89 (t, J = 7.5 Hz, 1H), 6.65 (t, J = 7.2 Hz, 1H), 5.81 (bs, 1H), 4.50 – 4.40 (m, 1H), 2.38 (d2.01 (s, 3H), 1.31 (d, J = 7 Hz, 3H) ppm. ¹³C-NMR (150 MHz, MeOD): δ = 181.8, 181.7, 172.4, 168.8, 150.8, 148.9, 138.4, 138.3, 138.1, 138.0, 132.9, 130.1, 128.8, 128.8, 128.1, 119.5, 105.7, 85.7, 75.4, 74.7, 60.1, 32.8, 30.8, 30.1, 28.00, 24.9 ppm. HRMS [C20H21N5O]: calc.

[M+H]⁺ 348.1819 found 348.1816

1-cyclohexyl-1-(2,4-dichlorophenyl)-2-(1H-imidazol-1-yl)ethan-1-ol (TH10840)

1 mmol (255 mg) 1-(2,4-Dichlorophenyl)-2-imidazol-1-ylethanone was dissolved in 1 mL anhydrous THF and 0.5 mL of a 1M solution of cyclohexyl magnesium bromide in THF was added at 0 °C. The solution was allowed to warm up to room temperature, stirred for 16 hours. The mixture was quenched with a saturated aqueous solution of NH4Cl, extracted three times with DCM. The combined organic phases were dried with Na2SO4, filtered and evaporated under reduced pressure. The residue was repetitively purified using RP-HPLC (a and b). The product was isolated as a mixture of enantiomers in 0.6%

(2.0 mg) yield. NMR contains traces of solvents acetone and dichloroethane.

1H–NMR (600 MHz, CDCl3): δ = 7.36 (s, 2H), 7.09 (d, J = 6.7 Hz, 1H), 5.11 (s, 1H), 4.31 (d, J = 11.7 Hz, 1H), 2.68-2.63 (m, 1H), 1.96-1.93 (m, 2H), 1.71-1.68 (m, 2H), 1.41- 0.72 (m, 11H) ppm. ¹³C-NMR (150 MHz, CDCl3): δ = 138.2, 134.1, 130.7, 130.6, 127.3, 42.0, 29.7, 27.3, 26.7, 26.4, 26.3, 26.1 ppm. HRMS [C17H20Cl2N2O]: calc. [M+H]⁺ 339.1025 found 339.1018

N-cyclohexyl-2-cyclopropylquinazolin-4-amine (TH10785)

0.1 mmol (20.5 mg) 4-chloro-2-cyclopropyl-quinazoline and 50 µL cyclohexylamine were dissolved in 1.5 mL isopropanol and stirred in a microwave reactor for 15 minutes at 130 °C and then for 16 h at 120 °C on a heating plate. Solvents were removed and the residue purified by RP-HPLC(b). The product was isolated as the TFA salt in 20.2% yield.

(19)

4

1H–NMR (600 MHz, CDCl3): δ = 7.81 - 7.80 (m, 1H), 7.70 (d, J = 8.4 Hz, 1H), 7.62 – 7.59 (m, 1H), 7.34 – 7.31 (m, 1H), 4.11 – 4.06 (m, 1H), 2.21 – 2.18 (m, 1H), 2.09 - 2.07 (m, 2H), 1.85 – 1.81 (m, 2H), 1.72 – 1.69 (m, 1H), 1.48 – 1.36 (m, 4H), 1.31 – 1.25 (m, 2H), 1.21 – 1.18 (m, 2H), 1.08 – 1.06 (m, 2H) ppm. ¹³C-NMR (150 MHz, CDCl3): δ = 167.2, 162.4, 162.1, 158.7, 133.0, 125.1, 121.6, 112.7, 50.5, 32.4, 25.6, 25.0, 17.1, 10.2 ppm. HRMS [C17H21N3]: calc. 267.1730 found 267.1727

N-cyclohexyl-2-cyclopropyl-N-methylquinazolin-4-amine (TH11735)

0.1 mmol (20.5 mg) of 4-chloro-2-cyclopropylquinazoline were dissolved in isopropanol and 2.0 mmol (17 µL) cyclohexylamine and 2.0 mmol (20 µL) DIPEA were added. The reaction mixture was stirred at 70 °C for 16 hours and evaporated under reduced pressure. The residue was purified via silica gel chromatography using a gradient of 0-15% MeOH in DCM. The product was isolated in 16.3% yield (4.6 mg).

1H–NMR (600 MHz, CDCl3): δ = 7.86 – 7.84 (m, 1H), 7.79 (d, J = 7.8 Hz, 1H), 7.64 – 7.61 (m, 1H), 7.30 – 7.28 (m, 1H), 4.34 – 4.29 (m, 1H), 3.16 (s, 3H), 2.21 -2.19 (m, 1H), 1.92- 1.87 (m, 4H), 1.73 – 1.71 (m, 1H), 1.66 – 1.59 (m, 2H), 1.44 – 1.36 (m, 2H), 1.17 – 1.14 (m, 2H), 0.99 – 0.97 (m, 2H) ppm. ¹³C-NMR (150 MHz, CDCl3): δ = 166.3, 163.7, 131.9, 127.0, 125.3 (2C), 123.1, 114.7, 59.5, 34.0, 29.9, 25.8, 25.7, 17.9, 9.4 ppm. HRMS [C18H23N3]: calc. 281.1886 found 281.1886

4-(cyclohexyloxy)-2-cyclopropylquinazoline (TH11738)

1.0 mmol (205 mg) of 4-chloro-2-cyclopropylquinazoline was dissolved in DMF and 2.0 mmol (19 µL) cyclohexanol and 3.0 mmol (115 mg; 60% dispersion in mineral oil) NaH were added at 0 °C. The reaction mixture was allowed to warm up to room temperature and was stirred for 16 hours. The reaction was quenched with saturated aqueous solution of NH4Cl solution and extracted three times with DCM. The combined organic phases were dried with Na2SO4, filtered and evaporated under reduced pressure.

The residue was purified via silica gel chromatography using a gradient of 0-15% MeOH in DCM. The product was isolated in 5.9% yield (4.6 mg).

1H–NMR (600 MHz, CDCl3): δ = 8.09 (dd, J = 8.1 Hz, 0.7 Hz, 1H), 7.85 – 7.80 (m, 1H), 7.76 – 7.74 (m, 1H), 7.43 (t, J = 7.9 Hz, 1H), 5.34 – 5.29 (m, 1H), 2.04 – 2.02 (m, 2H), 1.87 – 1.82 (m, 2H), 1.72 – 1.67 (m, 2H), 1.63 - 1.59 (m, 2H), 1.52 – 1.46 (m, 2H), 1.20 – 1.18 (m, 2H), 1.06 – 1.04 (m, 2H) ppm. ¹³C-NMR (150 MHz, CDCl3): δ = 167.7, 166.3, 133.3, 126.4, 125.3, 123.6 (2C), 115.2, 74.4, 31.4 (2C), 25.6 (2C), 23.6, 18.3, 9.8 (2C) ppm. HRMS [C17H20N2O]: calc. 268.1570 found 268.1572

4-(cyclohexylmethyl)-2-cyclopropylquinazoline (TH11873)

0.1 mmol (205 mg) of 4-chloro-2-cyclopropylquinazoline were dissolved in dioxane and 0.01 mmol (11.5 mg) tetrakis(triphenylphosphine)-palladium(0), 0.3 mmol (41.4 mg) potassium carbonate and 0.15 mmol (14.2 mg) cyclohexylmethyl-boronic acid were added. The reaction mixture was stirred for 16 hours at 90 °C and filtered over Celite. The filtrate was concentrated under reduced pressure and the residue purified via silica gel chromatography using a gradient of 0-15% MeOH in DCM. The product was isolated in 16.5% yield (4.4 mg).

1H–NMR (600 MHz, CDCl3): δ = 8.03 (d, J = 8.5 Hz, 1H), 7.90 – 7.88 (m, 1H), 7.80 – 7.76 (m, 1H), 7.50 – 7.46 (m, 1H), 3.06 (d, J = 7.0 Hz, 2H), 2.37 – 2.31 (m, 1H), 1.99 –