The anti-parasitic agent suramin and several of its analogues are inhibitors of the DNA binding protein Mcm10

Minichromosome maintenance protein 10 (Mcm10) is essential for DNA unwinding by the replisome during S phase. It is emerging as a promising anti-cancer target as MCM10 expression correlates with tumour progression and poor clinical outcomes. Here we used a competition-based fluorescence polarization (FP) high-throughput screening (HTS) strategy to identify compounds that inhibit Mcm10 from binding to DNA. Of the five active compounds identified, only the anti-parasitic agent suramin exhibited a dose-dependent decrease in replication products in an in vitro replication assay. Structure–activity relationship evaluation identified several suramin analogues that inhibited ssDNA binding by the human Mcm10 internal domain and full-length Xenopus Mcm10, including analogues that are selective for Mcm10 over human RPA. Binding of suramin analogues to Mcm10 was confirmed by surface plasmon resonance (SPR). SPR and FP affinity determinations were highly correlated, with a similar rank between affinity and potency for killing colon cancer cells. Suramin analogue NF157 had the highest human Mcm10 binding affinity (FP Ki 170 nM, SPR KD 460 nM) and cell activity (IC50 38 µM). Suramin and its analogues are the first identified inhibitors of Mcm10 and probably block DNA binding by mimicking the DNA sugar phosphate backbone due to their extended, polysulfated anionic structures.


Fluorescence polarization assay controls, hit confirmation, proteins, and data analysis
The LOPAC pilot screen identified the sulfhydryl reagent 4-chloromercuribenzoic acid as an inhibitor in the fluorescence polarization (FP) assay. As there was no known Mcm10 inhibitor, 4-chloromercuribenzoic acid (30 μM) was used as a positive control on initial screening plates. Following the discovery of suramin as a Mcm10 inhibitor, suramin (7 μM) was used as a positive control on subsequent screening plates. To confirm activity, screening hits were tested in dose response experiments (8 concentrations in duplicate) using xMcm10-ID. Similar dose response experiments were conducted with suramin and its analogs to confirm activity for hMcm10-ID (1 μM) and xMcm10-FL (2 μM) and to determine selectivity relative to RPA70AB (0.4 μM). Saturation experiments for each protein were conducted using increasing concentrations of protein with a fixed probe concentration (12.5 nM). FP was measured in mP units, and ∆mP values for each well were calculated as the difference between the measured mP and the average of the "probe only" controls. All raw mP values were corrected for background measurements, using the average of the buffer controls. ∆mP values from compound wells were then expressed as a percent of the average ∆mP values of control wells that did not contain any compaound. The parallel and perpendicular mP values were also used to calculate total fluorescence intensity (TFI) determined by the sum of parallel mP + (2 x perpendicular mP). TFI was used to detect fluorescence interference by compounds [1]. For the HTS, all compounds showing > 25% inhibition of the ∆mP signal and < 3x the standard deviation of the TFI were defined as hits. IC50 values were calculated using the four parameter logistic equation in Prism 6.0 (GraphPad). Ki values were obtained using an online Ki calculator according to the equations of Nikolovska-Coleska [2] (http://sw16.im.med.umich.edu/soft-ware/calc_ki/). KD values were calculated using the one site or two site binding equations in Prism 6.0 (GraphPad).

Screening collections, compounds, and reagents
Suramin sodium salt was obtained from Sigma. NF023, NF110, NF157, NF279, NF340, NF449, NF546, pyridoxalphosphate-6-azophenyl-2', 5'-disulfonic acid tetrasodium salt (iso-PPADS), and pyridoxalphosphate-6-azophenyl-2', 4'-disulfonic acid tetrasodium salt (PPADS) were all purchased from TOCRIS. Protein production and purification xMcm10-ID (amino acid residues 229 to 427) and hMcm10-ID (residues 237 to 436) were expressed in E. coli strain BL21(DE3). The fragments were cloned into a modified pET28a vector with the human rhinovirus (HRV) 3C protease cleavage site following the N-terminal 6xHis-tag. Transformed bacteria were grown at 37 °C in LB medium to an OD600 of ~0.5, upon which isopropyl--D-1-thiogalactopyranoside (IPTG) and ZnCl2 were added to final concentrations of 0.5 mM and 50 M, respectively. Protein expression was allowed to continue overnight at 18 °C. E. coli cells were pelleted at 4,500 x g for 20 min, resuspended in 50 mM Tris-HCl pH 7.4, 0.5 M NaCl, 5 mM imidazole, 10 mM -mercaptoethanol, and lysed by the addition of 0.5 mg/mL (final) hen egg white lysozyme and sonication. Samples were then centrifuged at 64,000 x g for 1 hr. The cleared lysate was passed through a Ni-NTA column, which was subsequently washed with the same buffer containing 20 mM imidazole, and the bound protein was eluted with a linear imidazole concentration gradient up to 300 mM. The eluted protein was pooled, treated overnight with HRV 3C protease at 4 °C to remove the His-tag, and further purified by gelfiltration over a HiLoad 26/60 Superdex 200 column operating with the running buffer 20 mM Tris-HCl pH 7.4, 0.5 M NaCl, 10 mM -mercaptoethanol. The full-length xMcm10 was expressed with MBP fused onto the Nterminus and 6x-His tag attached to the C-terminus of the protein [3]. Protein expression and purification were performed as described above, except that 100 g/mL (final) RNaseA was added during the lysis step, the lysis buffer contained 1.0 M NaCl, and a Superdex 200 (10/300) column was used for gel-filtration. The MBP-tag was kept uncleaved as it helped with protein solubility and did not interfere with DNA-binding. The purified Mcm10 proteins were concentrated by ultrafiltration, flash frozen in small aliquots by liquid nitrogen, and stored at -80 S3 °C. The protein concentrations were determined based on UV absorption at 280 nm and theoretical extinction coefficient calculated from the amino acid sequences. Vector construction and production of recombinant human RPA70AB (181-422) in a pSV281 vector that incorporates an N-terminal 6xHis-tag containing a TEV cleavage site has been reported previously [4]. To summarize briefly, RPA70AB was expressed in the Escherichia coli host BL21(DE3) cells using Terrific Broth medium containing kanamycin at 37 C. After cell lysis, the soluble fraction was run over a Ni affinity column, and then the 6xHis-tag was cleaved using TEV protease of RPA70AB. Further purification was achieved using heparin chromatography followed by Superdex-75 gel filtration. The final yield was 15-20 mg/liter of culture. The purity of the sample was confirmed using SDS-PAGE and electrospray mass spectrometry.

Compound purity analysis by LCMS
All purchased compounds were analyzed for purity by analytical LCMS. Compounds were dissolved in MilliQ water to make aqueous stock solutions ranging from 8-17 mM. The stock solutions were then diluted 500-fold using MilliQ water and filtered (0.2 μm, PTFE) prior to purity analysis using an Acquity UPLC (Waters Corporation) equipped with an Acquity BEH UPLC C18, 1.7 μm (2.1 x 50 mm) column for separation. Compound absorbance was detected at 214 nm using a photo diode array detector. Mass data was acquired using a Micromass ZQ mass spectrometer. LC was completed using a gradient method with each of the following mobile phase systems: Mobile Phase A2: 95% water, 5% acetonitrile, 10 mM ammonium acetate; Mobile Phase B2: 95% acetonitrile, 5% water, 10 mM ammonium acetate; Mobile Phase A1: 95% water, 5% methanol, 10 mM ammonium acetate; Mobile Phase B1: 95% methanol, 5% water, 0.1% 10 mM ammonium acetate. All compounds except iso-PPADS and PPADS were found to have > 95 % purity. PPADS had >91 % purity in mobile phase A1/B1 while iso-PPADS had > 90 % purity in mobile phase A2/B2.

SPR competition and control experiments
For ABA competition SPR experiments, the CM5 chip surface was prepared as described in Materials and Methods. Solution A was prepared with 50 μM NF023 in running buffer while solution B was prepared with the competitor near its KD value (NF023 = 5 μM, suramin = 0.5 μM, and iso-PPADS = 30 μM) ± 50 μM NF023 in running buffer. Data were collected at a rate of 40 Hz and analyzed with the Biacore S200 Evaluation Software. For these experiments, hMcm10-ID was immobilized to the chip surface and NF023 at a concentration 10-fold higher than its KD. A running buffer control was performed to establish the baseline of the experiment. When buffer and NF023 were injected as solution B, there was a small decrease in RU due to slight buffer mismatch effects (Figure 5a). Buffer injected alone as solution B resulted in no response. NF023, suramin, and iso-PPADS were chosen as competitors to be tested at concentrations equal to their respective KD values as determined by SPR.
Although the complex kinetics of NF546 is likely due to conformational selection and induced-fit conformational change, it is possible that the complex binding observed for NF546 by SPR is due to avidity effects that can often plague SPR experiments with higher density ligand surfaces. In an effort to explore whether this was the case, a lower density surface (3000 RU) of hMcm10-ID was prepared and suramin was chosen as a test compound. Similar KD values were obtained for suramin between the high density and low density surfaces (Supplementary Figure S7). As such, it is unlikely that a lower density surface would greatly change the SPR results. Supplementary Figure S5. iso-PPADS and PPADS weakly and incompletely inhibit the binding of the 5'-6FAM probe to Xenopus Mcm10 internal domain (xMcm10-ID), indicating a species difference between Xenopus and human proteins. Compare to the activity of these inhibitors with hMcm10-ID (Fig. 3a) and xMcm10-FL (Fig. 3b). Error bars represent ± SEM where n ≥ 3. The tailing up of the dose-response curves for both compounds at concentrations of 1 mM could be due to lack of solubility, a detergent-like effect of these amphipathic molecules, or fluorescence interference.