|Home | About | Journals | Submit | Contact Us | Français|
Conceived and designed the experiments: DJ WS. Performed the experiments: DJ WS. Analyzed the data: DJ WS. Wrote the manuscript: DJ WS.
The success of cisplatin (CP) based therapy is often hindered by acquisition of CP resistance. We isolated NSC109268 as a compound altering cellular sensitivity to DNA damaging agents. Previous investigation revealed an enhancement of CP sensitivity by NSC109268 in wild-type Saccharomyces cerevisiae and CP-sensitive and -resistant cancer cell lines that correlated with a slower S phase traversal. Here, we extended these studies to determine the target pathway(s) of NSC109268 in mediating CP sensitization, using yeast as a model. We reasoned that mutants defective in the relevant target of NSC109268 should be hypersensitive to CP and the sensitization effect by NSC109268 should be absent or strongly reduced. A survey of various yeast deletion mutants converged on the Rad5 pathway of DNA damage tolerance by template switching as the likely target pathway of NSC109268 in mediating cellular sensitization to CP. Additionally, cell cycle delays following CP treatment were not synergistically influenced by NSC109268 in the CP hypersensitive rad5Δ mutant. The involvement of the known inhibitory activities of NSC109268 on 20S proteasome and phosphatases 2Cα and 2A was tested. In the CP hypersensitive ptc2Δptc3Δpph3Δ yeast strain, deficient for 2C and 2A-type phosphatases, cellular sensitization to CP by NSC109268 was greatly reduced. It is therefore suggested that NSC109268 affects CP sensitivity by inhibiting the activity of unknown protein(s) whose dephosphorylation is required for the template switch pathway.
Since its discovery three decades ago, Cisplatin (CP) has been widely used as an effective anticancer agent against a wide variety of solid tumors like tumors of ovary, testis, head and neck, cervix and lung [1,2]. However, treatment success by platinum agents is diminished by both intrinsic and acquired resistance, necessitating a dose escalation that is limited by side effects like nephrotoxicity, ototoxicity, peripheral neuropathy and myelosuppression [1,3,4]. Acquired resistance is often multifactorial in nature, with common mechanisms attributed to decreased cellular drug accumulation through reduced influx or increased efflux [5,6], elevated thiol content and increased ability to repair or tolerate platinum DNA adducts [3,4]. A non-toxic compound used in combination with CP that potentiates sensitivity may increase the therapeutic index of CP, especially in the case of otherwise CP resistant cancers [7-10].
A two-hybrid yeast assay was used by us to screen the National Cancer Institute’s Diversity Set for compounds that can modify the checkpoint response elicited by the topoisomerase I inhibitor camptothecin . Initially, NSC109268 was isolated as such an agent that reduces the checkpoint response to camptothecin. On further study by quantitative survival analysis, it was revealed that NSC109268 increased cellular CP sensitivity . This was in marked contrast to its general tendency of decreasing sensitivity to other DNA damaging agents like nitrogen mustard . CP and nitrogen mustard both create common lesions, namely interstrand crosslinks but with different yield.
NSC109268 had been shown to inhibit the chymotrypsin-like activity of the 20S proteasome in both Jurkat T cell extract and rabbit purified 20S proteasomes using an in vitro assay . In addition, inhibition of phosphatases by NSC109268 had been suspected following molecular modeling, using the human PP2Cα structure in a virtual ligand screening of the Diversity Set of compounds . Using biochemical assays of enzyme activity, NSC109268 was indeed found to strongly inhibit PP2Cα and, less severely, the PP2A group of serine-threonine protein phosphatases .
Described initially in budding yeast, we further confirmed cellular sensitization to CP by NSC109268 in the CP-sensitive ovarian carcinoma cell line 2008 and, even more pronounced, in its CP-resistant counterpart, 2008/C13 . Cellular sensitization to CP by NSC109268 was consistently correlated with a slower S to G2/M phase progression in both yeast and the CP-resistant carcinoma cell line . Although NSC109268 enhanced CP-induced p53 levels, its effect on cell death following CP (i.e. apoptosis and necrosis) was not dependent on p53 .
Given the similarities of the effect of NSC109268 on mediating cellular sensitization to CP in yeast and cancer cell lines, given also the high degree of conservation of DNA repair pathways and the availability of a collection of deletion mutants of non-essential yeast genes, yeast must be considered a valuable model to study the targeted pathway(s) of NSC109268 that are relevant for CP sensitivity . The major target of CP is chromosomal DNA, with the majority of CP adducts comprising of DNA intrastrand crosslinks, mainly diguaninyl crosslinks . Albeit much less frequently, CP also induces the relatively much more lethal interstrand crosslinks . Nucleotide Excision Repair (NER) is the major pathway for bulky platinum adduct removal and thus error-free repair of DNA damage by CP . Consequently, defects in the NER pathway result in hypersensitivity to platinum agents and restoration of NER integrity correlates with reversal of CP sensitivity . Increased expression of the NER gene ERCC1 (RAD10 in budding yeast) is frequently associated with CP resistance in ovarian and gastric tumors .
Interestingly, among various predictors of CP sensitivity examined, such as increased platinum accumulation, decreased glutathione levels, decreased adduct removal or decreased tolerance to platinum-DNA adducts, decreased tolerance was the strongest predictor of CP sensitivity in ovarian cancer cell lines . Furthermore, 2008/C13 cells have been described as being more efficient in replicative bypass of CP lesions than their CP-sensitive counterparts . Therefore, inhibitors of specific DNA repair or tolerance pathways might prove especially efficacious when used in combination with CP.
Activities of polymerase ζ, polymerase η and RAD18, all involved in control of various modes of DNA damage tolerance are required for replicative bypass of CP intrastrand crosslinks . In response to DNA damage, yeast Proliferating Cell Nuclear Antigen (PCNA) is monoubiquitinated or polyubiquitinated at K164. Monoubiquitination of PCNA by RAD18, a ubiquitin ligase, promotes translesion synthesis mediated by Polζ or Polη [23,24]. During translesion synthesis, low fidelity DNA polymerases replicate directly past the lesion in either an error-prone or error-free fashion. Human cells expressing no Polη or reduced levels of REV3, an essential component of the translesion polymerase DNA Polζ, are more sensitive to CP [25-27].
K63-linked polyubiquitination, extending monoubiquitinated K164 of PCNA, depends on Ubc13–Mms2 (forming the E2 enzyme) and Rad5 (the E3 enzyme) and is required for an error-free damage tolerance pathway, most likely mediated by template switch (TS) (Figure S1) [24,28-31]. In contrast to translesion synthesis, the lesion is avoided by a “copy choice” mechanism using an alternate, undamaged DNA template. Upon DNA damage, a lesion in the leading strand template can lead to uncoupling of leading and lagging strand synthesis and single-stranded DNA gaps can be found in both leading and lagging strands. Subsequently, strand invasion mediated by certain Homologous Recombination (HR) factors that include Rad51, Rad52, the complex of Rad55-Rad57, promote the formation of TS intermediates that are dissolved by the action of the Sgs1-Top3 complex (Figure S1) [28,30,31].
In this study, using survival and cell cycle analyses, we identified the Rad5 pathway of damage tolerance by TS as the target pathway of NSC109268 relevant for CP sensitization. Inhibition of this pathway appears to result in the observed delayed S phase traversal of CP + NSC109268 treated yeast cells. Furthermore, our data suggest a critical role of NSC109268 as a PP2Cα and PP2A phosphatase inhibitor in influencing cellular sensitization to CP.
Cisplatin (CP) was purchased from Sigma Aldrich. NSC109268 was initially provided as part of the Diversity Set compound library by the Developmental Therapeutics Branch of the National Cancer Institute and later synthesized by Omm Scientific, Dallas, Texas. NSC109268 was dissolved at a stock concentration of 1 mg/ml in dimethylsulfoxide, CP at 10 mg/ml in dimethylformamide.
S. cerevisiae BY4741 (MATa his3Δ1 leu2Δ0 met15Δ0 ura3Δ0) was the parental haploid wild-type strain used throughout the study unless indicated otherwise. Haploid mutants, each deleted for a defined open reading frame of a non-essential gene, were from the systematic gene deletion collection (purchased from Open Biosystems). BY4741 rad5::HIS3 and BY4741 sml1::KanMX4 rad53::URA3 were constructed in our laboratory. Strains YJK17 (MATα hoΔ hmlΔ::ADE1 hmr::ADE1 arg5,6::HPH::MATa-inc ade1-100 leu2,3-112 lys5 trp1::hisG ura3-52 ade3::GAL::HO), YJK24 (ptc2Δ::URA3MX ptc3Δ::NatMX. YJK17 isogenic), YJK26 (pph3Δ::KanMX, YJK17 isogenic) and YJK70 (pph3Δ::KanMX ptc2Δ::URA3MX ptc3Δ::NatMX, YJK17 isogenic) were kindly provided by Dr. James Haber . Diploid strain D7 (MATa/α ade2-119/ade2-40 trp5-12/trp5-27 ilv1-92/ilv1-92 CYH/cyh2) was originally from Dr. Fritz Zimmermann . Note that this strain does not require adenine and forms white colonies due to intragenic complementation.
All strains used in this study were grown to early logarithmic phase in YPD (1% yeast extract/2% peptone/2% dextrose). Subsequently, cells were washed, resuspended in phosphate-buffered saline (PBS) and treated with CP and NSC109268 at 30°C with constant shaking, typically for 2 h unless indicated otherwise. For strains YJK17, 24, 26 and 70 5 mM phosphate buffer/5 mM sodium chloride (pH 7.0) was used instead of PBS. After treatment, cell suspensions were appropriately diluted and plated onto YPD plates and, in case of D7, also synthetic media plates lacking tryptophan where appropriate. Published protocols and recipes were followed . Plates were incubated at 30°C to allow colony formation. Surviving fractions after drug treatment were calculated by dividing the titer of macrocolony-forming cells in the drug-treated sample by the titer of macrocolony-forming cells in the untreated control sample. Standard deviations are indicated if experiments were repeated at least three times. Otherwise, data of representative single experiments are shown, but in all cases data points were reproduced at least once. Dose Enhancement Factors (DEF) were calculated by determining the ratios of CP doses resulting in 50% or 10% survival, without vs. with inclusion of NSC109268. For DEF calculations, survival values were corrected for killing by NSC109268 alone (usually 10-30%).
Wild-type yeast (BY4741) and isogenic mutant strain rad5 were grown in YPD overnight at 30°C to early logarithmic phase. Cells were then synchronized in G1 using the yeast mating pheromone, α factor, at a final concentration of 10 µg/ml as previously described . After resuspending cells in PBS, 80 µM CP was administered along with 5 µg/ml of α factor for 1 h at 30°C with shaking. Next, cells were washed, resuspended with PBS and α factor and treated for 20 min with 0.7 µM NSC109268. Control samples, containing the appropriate solvent instead of the drug, were incubated in parallel. After treatment, cells were washed and resuspended in fresh YPD to allow for synchronous reentry into the cell cycle. At the indicated time points, samples were collected, fixed in ethanol, sonicated and stained with SYBR Green I (gel staining solution, Lumiprobe) for FACS analysis of DNA content as described . Cellular fluorescence was measured in a FC500 Flow Cytometer (Beckman Coulter Corp.). CXP and ModFit software were used to obtain histograms depicting DNA content versus cell number (of a total of 10,000 cells) and to calculate the fractions of G1, S and G2/M cells.
CP uptake is largely mediated by the plasma membrane copper transporter CTR1 in both yeast and mammals [36,37]. Following internalization, both copper and CP were shown to cause downregulation of CTR1 in ovarian cancer cells by the proteasome-mediated pathway  contributing to CP resistance. Since NSC109268 may prevent the degradation of CTR1 through its 20S proteasome inhibitory activity , we tested CP uptake as a target of NSC109268 in CP sensitization. Haploid wild-type yeast cells were treated with CP for 1 h, then washed to remove CP and treated with NSC109268 for another hour. Following this regimen, NSC109268 clearly sensitized yeast cells to CP (Figure 1) synergistically, as with simultaneous administration of NSC109268 and CP  or pretreatment with NSC109268 before adding CP (data not shown). Since NSC109268 sensitized previously CP-treated cells even in the absence of external CP, we conclude that mechanisms other than increased CP uptake, such as inhibition of DNA repair, must be responsible for sensitization to CP.
In order to identify the pathway targeted by NSC109268 in mediating cellular sensitization to CP we reasoned that, if the critical pathway is already inactive, such mutant cells will be CP sensitive and no sensitization that exceeds an additive effect will be achieved by combination treatment. Several functional screens in S. cerevisiae and Schizosaccharomyces pombe had already identified genes that upon deletion confer sensitivity to CP [39-41]. Genes surveyed by us included those operating in NER (RAD10), DNA damage tolerance (RAD5, MMS2, REV3 and RAD18), interstrand crosslink repair (PSO2) and HR (RAD51). In general, the published relative CP sensitivities of the mutants investigated were confirmed (Figure S2, data not shown). Figure 2 summarizes the effect of NSC109268 on CP sensitivity of all tested deletion mutants, expressed as the factor by which the CP dose resulting in 50% or 10% survival can be reduced in the presence of NSC109268 (dose enhancement factor, DEF), see also Table S1 for data and gene product function. In Figure 3, selected dose response curves are shown.
In surveying a large number of isogenic mutants of haploid strain BY4741, we ruled out several pathways as primary targets of NSC109268 in causing cellular sensitization to CP. For example, PSO2 is a gene known to be specifically involved in interstrand crosslink repair in yeast. CP is known to induce interstrand crosslinks eliciting cell killing in yeast  and as expected, we found the pso2Δ mutant strain to be hypersensitive to CP alone. The combination of NSC109268 and CP was highly synergistic in inducing cell death (Figure 2, Figure 3A) which ruled out the affected interstrand crosslink repair pathway as the target (and possibly also interstrand crosslinks as critical lesion). When testing rad10Δ mutant strain, defective in a key endonuclease component of NER, we found the strain to be hypersensitive to CP, as previously reported , and additional sensitization to CP by NSC109268 was still observed (Figure 2, Figure 3A). Similar results with rev3Δ and rad30Δ mutant strains argued against damage tolerance by translesion synthesis as the target of NSC109268 in mediating cellular sensitization to CP (Figure 2). Rtt101, a component of a novel ubiquitin ligase complex promoting replication through damaged DNA at stalled replication forks  could also be excluded (Figure 2). Similarly, by testing the CP-hypersensitive rad50Δ mutant, we concluded that the MRN/X (Mre11-Rad50-Xrs2) complex and its activity in DNA double-strand break repair is an unlikely target (Figure 2).
As representative examples of the cell cycle checkpoint pathway, we selected mutants deleted for the DNA damage recognition protein Rad24 and the Rad53 kinase. A deletion of the latter is viable in an SML1 deletion background. Both sml1 and sml1 rad53 strains can be sensitized towards CP by NSC109268 to a similar extent (Figure 2, Figure 3B). Since rad24 retained synergism as well, we did not obtain evidence for the checkpoint pathway being a critical target.
In revisiting CP uptake, we found combination treatment to remain synergistic in inducing cell killing in CP-resistant ctr1Δ, a CP uptake mutant (Figure 2). Similar results were obtained for ccc2Δ, deleted for the yeast ortholog of ATP7B, a P-type ATPase that mediates cellular CP efflux and counteracts CP lethality in mammals . Together, these results confirm our previous notion that modification of CP transport is unlikely to be a mechanism by which NSC109268 mediates cellular sensitization to CP.
As previously reported , mutant strain rad5Δ, defective in damage tolerance by TS, was found to be CP hypersensitive. Remarkably, we did not detect additional sensitization upon the administration of NSC109268 during CP treatment of rad5Δ cells (Figure 2, Figure 3C). This result was confirmed by detecting a similar absence of synergism in deletion mutants of other components of the Rad5 mediated TS pathway – RAD18, UBC13 and MMS2 (Figure 2, Figure 3C). As compared to wild-type, no or only a low degree of sensitization by NSC109268 beyond an additive effect was also found for deletions of the RAD51, SGS1 and PSY3, encoding HR components that all are likely to play a role in damage tolerance by TS [28,44] (Figure 2, Figure 3D).
We also tested a deletion of Rad59, an HR protein that stimulates single-strand annealing in complex with Rad52 . Rad59 has been proposed to only participate in the canonical HR pathway but not in the TS pathway of damage tolerance . The combination of CP and NSC109268 remained synergistic in inducing cell death in the rad59Δ strain (Figure 2, Figure 3D). This result implies that HR pathway per se may not be a crucial target of NSC109268 in mediating cellular sensitization to CP.
In a diploid strain (D7, unrelated to BY4741), we tested the influence of CP and NSC109268 on survival and genomic instability events related to interchromosomal recombination. Strain D7 contains two detection systems for such events, the ade2-119/ade2-40 and trp5-12/trp5-27 heteroalleles . Scoring red or pink sectored or pure colonies detects a wider range of events than the trp5 gene conversion system, including chromosome loss events.
For this diploid strain, synergistic interaction of CP and NSC109268 was verified for colony survival (Figure 4A). Frequency of CP-induced gene conversion in the trp5 system and aberrant colony formation in the ade2 system were increased in a dose-dependent manner (Figure 4B,C). Whereas NSC109268 had little effect on its own, it synergistically enhanced CP-induced recombination/aberrant colony frequency in either system. Since gene conversion and mitotic recombination events are dependent on HR , an inhibitory effect of NSC109268 on HR per se appears to be unlikely.
G1-synchronized haploid wild-type and isogenic rad5Δ strains were treated with CP alone or sequentially with CP and NSC109268 before release into nutrient medium to allow for cell cycle progression. By itself, the chosen low dose of NSC109268 caused only a brief delay of G1/S phase transition. To facilitate interpretation, we used a low CP dose that induced only a small effect on cell cycle kinetics in wild-type cells on its own. However, confirming previously published data , the combination of NSC109268 and CP resulted in a notable slow-down of S phase progression, as measured by a delayed increase in G2/M phase cells (Figure 5A, see Figure S3 for FACS profiles). Using the same CP dose, synchronized rad5Δ cells were delayed for G2/M entry as compared to untreated cells (Figure 5B). However, NSC109268 did not lead to any enhancement of this CP-dependent delay beyond the small effect attributable to NSC109268 alone, thus being additive at best (Figure 5B). Preliminary studies suggest that the same is true for the later occurring extensive G2/M arrest (data not shown). This is in marked contrast to the synergistic effect on cell cycle kinetics of NSC109268 in CP-treated wild-type cells (Figure 5A). Thus, cell cycle studies supported our survival analysis hinting at the Rad5 pathway of damage tolerance by TS as a target of NSC109268 in mediating cellular sensitization to CP.
We wished to determine if the known activities of NSC109268 as 20S proteasome inhibitor  or as phosphatase 2Cα and 2A inhibitor  critically affect sensitization to CP by NSC109268. Bortezomib, a known proteasome inhibitor possibly mimicking NSC109268’s proteasome inhibitory activity, was not synergistic in inducing cell death (Figure S4A). Furthermore, NSC109268 and CP combination remained synergistic in ump1Δ (Figure 2, Table S1), deleted for a DNA-damage inducible chaperone involved in 20S proteasome maturation and required for UV resistance . These results argue against direct proteasome inhibition by NSC109268 as critical for mediating sensitization to CP.
Polyubiquitination of PCNA is an essential signal for activation of the Rad5 pathway. As an indirect consequence of proteasome inhibition by NSC109268 due to reduced ubiquitin recycling, we considered that free ubiquitin levels in NSC109268 + CP treated cells may be too low for PCNA polyubiquitination to occur efficiently . This possibility, however, was unlikely since NSC109268 and CP combination remained synergistic in inducing cell death even if cells were overexpressing plasmid-encoded ubiquitin  (Figure S4B). Therefore, we found no evidence that proteasome inhibition by NSC109268 leads to cellular sensitization to CP, directly or indirectly through impairment of free ubiquitin levels.
Next, we considered the phosphatase inhibitory activity of NSC109268. To test for the role of PP2C or PP2A inhibition by NSC109268 in CP sensitization, we first treated single mutant strains defective in Ptc2 or Ptc3, both classified as Ser/Thr protein phosphatase 2C family members, or Pph3, a type 2A-like protein phosphatase A. The combination of NSC109268 and CP was found to remain synergistic in inducing cell death in all single gene deletion mutants of haploid strain BY4741 (Figure 2).
Next, combinations of phosphatase mutations were studied in a different genetic background (YJK). As compared to BY4741, it should be noted that we were able to increase the dose enhancement effect in this strain background while reducing the lethality of NSC109268 alone. At 50%/10% CP survival, DEF between 3 and 5 were determined for WT, pph3 and ptc2ptc3 (Figure 2, Figure 6). However, whereas CP sensitivity without NSC109268 was notably enhanced in the triple mutant ptc2Δptc3Δpph3Δ synergism was greatly diminished, with DEF reduced to less than 2 (Figure 2, Figure 6). While confirming the known redundancy between the studied phosphatases  these results indicate the involvement of both phosphatase 2C or 2A inhibitory activities of NSC109268 in mediating CP sensitization.
Among the platinum family of anti-cancer compounds, CP has been a very effective agent against a host of tumors such as those of the ovary and testis. However, its utility is severely restricted by dose limiting side effects and cellular resistance, either intrinsic or acquired during CP administration. We identified NSC109268 as a compound increasing cellular sensitivity to CP in Saccharomyces cerevisiae . Enhancement of cellular sensitization to CP by NSC109268 was confirmed for CP-sensitive ovarian carcinoma 2008 cells and the derived CP-resistant line 2008/C13 . Shown in both yeast and 2008/C13 cells, the increase in cellular sensitivity to CP by NSC109268 correlated with inhibition of S phase progression . In in vivo studies with xenografted human head and neck carcinomas treated with CP alone, others had shown that a hindrance of S phase traversal correlated with CP responsiveness better than G2/M arrest .
In this study, we explored the possible targets of NSC109268 responsible for enhancing CP sensitivity using the systematic yeast gene deletion collection. We were guided by the concept that absence of the target should cause CP sensitivity but preclude additional sensitization by NSC109268. A similar strategy resulted in the identification of HR as the target of synergism between CP and camptothecin in yeast and human cancer cells . Alterations in uptake or efflux of CP appeared unlikely to be the cause of sensitization since synergism was not diminished if CP was administered first, prior to treating with NSC109268 in CP-free buffer. This notion was further confirmed by mutant analysis.
After investigating various DNA repair and tolerance pathways such as NER, DNA translesion synthesis or interstrand crosslink repair, we identified a DNA damage tolerance pathway, the Rad5 pathway, as critical determinant of the synergism between CP and NSC109268 in budding yeast. This pathway is believed to bypass DNA damage at replication forks, that have been stalled in S phase due to DNA adducts, by using the newly synthesized strand of the sister chromatid as a template (Figure S1) [28-31,53]. Fork regression or recombinative sister strand junction formation have been discussed as the underlying mechanisms in this process . CP primarily causes bulky DNA adducts which lead to a slow down of the replication fork, eventually leaving behind single-stranded gaps to be filled by Rad5 mediated TS pathway, involving certain HR components [28,31]. Inhibition of Rad5 mediated lesion bypass activity by NSC109268 may also result in fork collapse leading to genetic instability – as shown here with reporter strain D7 – and finally cell death, thus mediating cellular sensitization to CP. It has been debated if the Rad5 mechanism is critical for S-phase progression in the presence of DNA damage, as suggested here, or acts predominantly in G2-phase to fill remaining single-stranded gaps [30,54-56]. The discrepancies in timing found in the literature may very well depend on level and kind of DNA damage.
Our findings seem to fit particularly well with the model of filling of single-stranded gaps by template switch using a subset of recombination functions (Figure S1) [28,31]. We found proteins involved in initial PCNA monoubiquitination, such a Rad6 or Rad18, and in subsequent polyubiquitination, such as Rad5, Ubc13 or Mms2, to be required for CP+NSC109268 synergism. But we also identified HR proteins such as Rad51 or the junction-resolving enzyme Sgs1 as participants of the targeted mechanism(s). This includes Psy3, a member of the Shu complex proposed to participate in TS following strand invasion . All of these proteins that are required for synergism of CP and NSC109268 are exclusively or non-exclusively involved in the TS pathway.
Interestingly, the mutant group showing no or very reduced synergism did not include Rad59 which has been shown by others to play a role in canonical HR but not in the TS pathway . We also demonstrated that CP-mediated and HR-dependent gene conversion events  are increased by NSC109268 and not decreased. (Current models of the Rad5 pathway do not involve the homologous chromosome in a template switch mechanism, so this result does not contradict the inhibition of the Rad5 pathway.) Based on these observations, we do not favor that NSC109268 acts as a general HR inhibitor to exert its function on CP sensitivity, however, a more rigorous demonstration is still required.
A screen for single gene deletion mutants of budding yeast conferring exquisite CP sensitivity had previously revealed RAD5 and other key players of the same pathway as top hits . In humans, this pathway was initially regarded as a tumor suppressor pathway since the essentially error-free bypass should counteract genetic instability [29,57]. On the other hand, the human ortholog of the yeast Rad5, HLTF, was reported to be overexpressed in radiation resistant recurrent human cervical carcinoma and a knockdown of HLTF in HeLa cells lead to a decrease in cellular proliferation . Conceivably, targeting HLTF in chemotherapy-resistant cancers may lead to increase in cellular CP sensitivity.
Upon investigating the known activities of NSC109268 as a 20S proteasome and phosphatase 2C and 2A inhibitor [13,14], we found no evidence for the importance of proteasome inhibition in mediating CP sensitivity in yeast. In contrast, our survival studies indicated 2C and 2A phosphatase inhibition by NSC109268 to be responsible for mediating CP sensitization. Analysis of the yeast single mutants and double mutants ptc2Δ, ptc3Δ (each one defective for functionally redundant PP2C phosphatases), pph3Δ (defective for PP2A-like phosphatase) compared to the triple phosphatase mutant ptc2Δptc3Δpph3Δ revealed a marked defect in the viability of ptc2Δptc3Δpph3Δ cells in the presence of CP, while neither of the single mutants were similarly sensitive to CP. These results suggested redundancy of phosphatase PP2C and phosphatase PP2A-like mediated pathways in CP damage responses, as was already reported for other agents such as hydroxyurea or methyl methanesulfonate . If phosphatase redundancy is removed in the triple mutant, sensitization of CP-treated cells by NSC109268 is greatly reduced suggesting phosphatases as the relevant target.
Phosphatases Psy2 or Ptc2 and, in a redundant fashion, Ptc3 have been shown to be required for turning off the DNA damage-induced cell cycle checkpoint by dephosphorylating kinase Rad53 [59,60]. However, upon treatment of rad53Δ mutant (sml1Δ background), Rad53 was excluded as a major target of NSC109268 in mediating CP sensitization. Thus, a reduced ability to dephosphorylate Rad53 checkpoint kinase and to release checkpoint arrest is not responsible for the cellular sensitization to CP by NSC109268.
Although the actual protein target(s) remain to be determined, this study underlines the critical importance of protein phosphatases in DNA damage responses that goes beyond recovery from checkpoint arrest. Even in a model organism such as yeast, significant gaps in our knowledge will need to be filled. For the first time, this study shows an influence of phosphatases on the Rad5 tolerance pathway. Additionally, the notion of intertwined HR and TS pathways is supported [28,30,31]. Interestingly, the triple phosphatase mutant pph3Δptc2Δptc3Δ used by us was reported to be defective in early steps of HR  that may also be necessary for sister chromatid junction formation and processing within the Rad5 TS pathway [28,30,31]. Interestingly, Rpa2, a subunit of the single-strand DNA binding complex Replication Protein A, represents another target candidate since its phosphorylation status, which may be enhanced by NSC109268-mediated phosphatase inhibition, is inversely correlated with RAD51 foci formation .
Taken together, our studies with NSC109268 in budding yeast demonstrate its potential as a cellular chemotherapy sensitizer, most likely by acting through the inhibition of dephosphorylation of critical protein(s) of the Rad5 pathway mediating CP damage tolerance during S phase. NSC109268 and related compounds may thus be especially valuable in CP combination therapy of the subset of cancers that developed CP resistance due to alterations within the Rad5 pathway. It remains to be seen if this is a mechanism of resistance that is widespread among cancer patients.
Model for DNA damage tolerance by template switch. Provoked by a bulky lesion (star) in the leading strand template, the proposed interplay of recombination proteins, PCNA ubiquitination and DNA polymerases in filling a single-stranded gap is depicted. Adapted from .
Relative cisplatin sensitivity of haploid yeast mutants in agar diffusion assays. For these semiquantitative streak tests, early-logarithmic phase BY4741 wild-type or mutant cell cultures were concentrated to 4x107 cells/ml. Of these suspensions, 10 µl samples were streaked on YPD plates in a radial fashion, and 250 µl of 1.5 mM CP were pipetted in the circular center hole (0.5 inch diameter). Plates were kept at 4°C for 3 hours to allow for CP diffusion before incubating at 30°C. Streaks were photographed after 30 hours.
Flow cytometric DNA profiles of G1-synchronized cells released into fresh medium after treatment with cisplatin, NSC109268 or both. Wild-type (A) or RAD5-deleted cells (B) of BY4741 were synchronized with α-factor and then sequentially incubated with CP and NSC10268 before release into fresh YPD medium, as described in Material and Methods. Samples were taken for a period of 100 or 120 min past treatment. Proper control regimens (mock treatment, CP alone, NSC10268 alone) were applied for comparison and are indicated in the figure.
Absence of evidence for direct of indirect influence of inhibition of protein degradation on sensitization to cisplatin by NSC109268. (A) The absence of a non-additive effect of the combination of bortezomib and CP is shown by determining the survival of wild type cells (BY4741) after simultaneous treatment with CP and bortezomib (75, 338 or 375 μM) for 2 h. Surviving fractions of colony forming cells were plotted as a function of CP dose. (B) The absence of an influence of ubiquitin overexpression on the synergism of. CP and NSC109268 in inducing cell killing is shown. The dose responses of wild-type cells overexpressing ubiquitin after treatment with CP and with or without NSC109268 (3 μM) for 2 h are shown. Wild-type cells (SX46A MATa RAD ade2 (ochre) his3-532 trp1-289 ura3-52) had been transformed with a plasmid carrying wild-type ubiquitin gene under the control of the copper inducible CUP1 promoter (YEp96-CUP1-UB) , kindly provided by Dr. Mark Hochstrasser. Strains transformed with YEp96-CUP1-UB or vector plasmid were grown to early logarithmic phase in Trp-dropout medium. To overexpress ubiquitin, CuSO4 was added at 100 μM to the medium for 3 h before treatment and plating onto –Trp dropout plates.
Mutants studied for influence of NSC109268 on cisplatin sensitivity, ranked by dose enhancement factor (DEF) at 50% survival. See also Figure 2.
We would like to thank Dr. James Haber for providing yeast strains and Dr. Mark Hochstrasser for providing plasmids. We also thank Xiangle Sun for assistance with flow cytometry and data analysis.