To assess whether pyrimidinones inhibit P. falciparum
growth, we examined the effects of 157 compounds in this class and related Biginelli and Ugi multicomponent condensation-derived compounds on the uptake of [3
H]hypoxanthine into infected human erythrocytes. The hypoxanthine assay provides a rapid, quantifiable read-out of parasite viability, and the compounds assayed included several recently described agents,18
as well as precursors and structurally related analogs.
The impact of JAB75 (see Section 3) on [3
H]hypoxanthine up-take is shown in . In this and all other assays, we used the chloroquine (CQ)-resistant Dd2 clone and employed CQ as an internal control because CQ is known to inhibit Dd2 with an IC50
value of ~0.2 μM.23,24
In this experiment, the IC50
for JAB75 was calculated to be ~0.3 μM (, second column) and the IC50
for CQ was 0.19 μM (data not shown). From our initial analysis of 157 compounds, we identified nine molecules (; see Section 3) with IC50
values between 30 nM and 1.6 μM (, second column).
Figure 1 JAB75 inhibits [3H]hypoxanthine uptake into red blood cells infected with P. falciparum. The uptake of [3H]hypoxanthine into infected human erythrocytes and the IC50 value were determined as described in Section 3. The raw data from a single assay were (more ...)
Effects of most potent P. falciparum inhibitors on the steady-state ATPase activities of P. falciparum, yeast (Ssa1), and human (HsHsp70)
Figure 2 Select pyrimidinones inhibit hypoxanthine uptake into P. falciparum-infected red blood cells. The depicted compounds inhibited P. falciparum replication with IC50 values of 30 nM–1.6 μM (see ). The structures were drawn to maximize (more ...)
To ensure that the compounds were not generally cytotoxic, we also determined the 50% growth inhibitory concentrations (GI50
) for each of these nine agents in two human cell lines, HepG2 hepatocellular carcinoma cells and WI-38 embryonic diploid lung cells, as previously described.18
Based on this analysis, all GI50
values in these cells were >10 μM, which is well above the concentration needed to inhibit P. falciparum
growth (). As a control for this experiment, we found that the GI50
values for paclitaxel in HepG2 and WI-38 cells were 1.0 ± 0.6 nM and 13.7 ± 0.2 nM, respectively (data not shown).
We previously showed that a subset of pyrimidinones inhibit the activity of Hsp70.18,19
Therefore, we next assessed the effects of the nine compounds on Hsp70 ATPase activity. We first examined the ability of each pyrimidinone to modulate the ATP hydrolytic rate of a purified yeast Hsp70, Ssa1, as previously published.18–20
In addition, we wished to compare the effects of these agents on human and P. falciparum
Hsp70. The three chaperones are 71–74% identical to one another at the amino acid level and, not surprisingly, Hsp70s from different species have been reported to substitute functionally for one another. For example, the growth of bacteria containing a mutation in the gene encoding DnaK, the Hsp70 homolog, can be rescued at high temperature by expression of a P. falciparum
One might also envision, however, that a specific inhibitor could selectively target the P. falciparum
chaperone but have no effect on Hsp70s from different species.
To compare the effects of the lead compounds on the ATPase activities of the Hsp70s, we first modified purification schemes for both the H. sapiens and P. falciparum chaperones (see Section 3). The peak fractions from the P. falciparum Hsp70 purification are shown , and the single-turnover ATPase activities of the enzyme in the absence and presence of a J domain chimera (see below) are shown in .
Figure 3 Purification and analysis of P. falciparum Hsp70. (A) A Coomassie Brilliant blue-stained gel summarizing individual steps during the P. falciparum Hsp70 purification is shown: lane 1: crude E. coli lysate from the P. falciparum Hsp70 over-expressing strain; (more ...)
Next, each of the nine compounds was incubated with the P. falciparum
, yeast and human chaperones, and steady-state ATPase assays were performed. The results presented in , columns 3–5, indicate that the compounds display a range of activities and alter the activities of each chaperone distinctly. For example, JAB75, MAL2-61, and MAL2-215 reduced the rate of ATP hydrolysis by the human enzyme by 30%, but had no effect or stimulated the activities of the parasite and yeast enzymes. In contrast, MAL2-39 compromised the ATPase activities of all three enzymes, but mainly HsHsp70. MAL3-39 inhibited all enzymes to a very similar extent. Perhaps most intriguing, DMT2264 affected the ATPase activity of P. falciparum
Hsp70 significantly more than Ssa1 and HsHsp70. These data indicate that the pyrimidinones can be classified based on their in vitro effects on Hsp70s from different species. Nevertheless, it is important to note that these analyses were performed at a final concentration of 300 μM in order to maximize the effects. We also selected this concentration because the action of pyrimidinones was first studied under these conditions;18
however, when used at concentrations that inhibited P. falciparum
replication (), the steady-state ATPase activity of Hsp70 was unaltered, suggesting the possibility of secondary targets or other issues related to compound metabolism or accumulation (see below).
The cellular activity of Hsp70 most often requires interaction with J domain-containing Hsp40 partners.5,6
In fact, one of our most potent pyrimidinone-based inhibitors of breast cancer cell proliferation has no effect on endogenous ATPase activity but compromises the ability of a J domain-containing co-chaperone to enhance ATP hydrolysis.19,21
Therefore, it was possible that the compounds might have greater effects on the J domain-stimulated ATPase activity of the Hsp70s. Because the purification of a P. falciparum
Hsp40 has not been reported, we instead chose to use a chimeric protein that contains the J domain from Hlj1, a yeast Hsp40, fused to glutathione-S
Recent work indicates that the chimera interacts promiscuously with a variety of Hsp70s and augments their ATPase activities.36
For comparison, we also used a full-length ‘type-I’ Hsp40 co-chaperone from yeast, Ydj1.
We added each of the nine P. falciparum
inhibitors in single-turnover ATPase reactions containing either the parasite (PfHsp70) or yeast Hsp70 (Ssa1) in the presence or absence of the Hlj1 chimera and Ydj1 (see ). Furthermore, the activity of human Hsp70 in the presence or absence of Ydj1 and one of its known Hsp40 partners, Hdj1, was examined in the presence or absence of each pyrimidinone. We chose to use single-turnover conditions for this analysis because the level of J domain-mediated stimulation of Hsp70 ATPase activity is significantly greater for single-turnover assays as compared to steady-state assays (see, e.g., ). Furthermore, the effects of compounds on enzyme kCAT
values can specifically be monitored and the fold change measured. As shown in , these nine compounds exhibited a range of effects. First, relatively subtle changes were observed when the effects of the compounds were examined in the presence of the J domain protein. For example, a <50% decrease (in the presence of MAL2-213 or DMT2264) and a ~40% increase (in the presence of MAL2-61 or MAL3-39) was measured when the J domain was added into the assay in the presence of these chemicals (, ‘fold change PfHsp70 kCAT
+ Hlj1’). In contrast, two of the compounds induced a potent ‘burst’ of ATP hydrolysis in the single-turnover assay in the absence of a J domain (‘fold change PfHsp70 kCAT
’): Based on a fit of the data to a single exponential, MAL2-39 and MAL2-61 enhanced P. falciparum
Hsp70 ATPase activity by 7.2 and 6.5-fold, respectively. Each of the other compounds also enhanced ATP hydrolysis to varying extents, an effect that has been previously noted for certain other pyrimidinones.18,19
Because the cycle of Hsp70 ATP binding/hydrolysis is coupled with substrate binding and release, altered rates of endogenous or J domain-stimulated ATP hydrolysis will correspondingly alter the efficacy of substrate binding. Therefore, enhanced rates of ATP hydrolysis can lead to defects in the ability of Hsp70s to act as a molecular chaperone, especially under stress conditions.27,28
Moreover, recent data confirm that the individual chaperone cycles and conformations have been tailored to match the folding of specific substrates.29
In the future, it will be important to examine whether these agents affect the binding of known Hsp70 substrates. Moreover, it is critical to note that effects on ATPase activity in steady-state assays may result from alterations in any one of a number of steps in the hydrolytic cycle, including the kCAT
, ATP binding, and inorganic phosphate and/or ADP release.
Effects of the P. falciparum inhibitors on the single-turnover ATPase activities of a malarial parasite, yeast, and human Hsp70s
An examination of the endogenous and J domain-stimulated ATPase activities of the yeast enzyme in the presence or absence of the compounds yielded quite different results (, ‘fold change Ssa1 kCAT’ and ‘fold change Ssa1 kCAT + Hlj1’). We found that MAL2-215 exerted significant effects on both the endogenous and J domain-stimulated activity of Ssa1, but the compounds with the greatest impact on the P. falciparum enzyme (e.g., MAL2-39, MAL2-61, and MAL2-213) only modestly altered the activity of the yeast enzyme. Notably, the strongest effects of each pyrimidinone were observed when the activity of the human enzyme was examined in the presence of its partner, Hdj1. At this point, the substructure features that mediate these distinct phenomena remain to be elucidated. Our data nevertheless indicate that the continued examination of pyrimidinones with diverse Hsp70s will prove worthwhile, especially if a specific binding site on the chaperone for this class of modulators can be identified. Efforts toward this goal are underway.
It is important to note that the relative effects of the compounds on the endogenous or J domain-stimulated P. falciparum
ATPase activities do not correlate with the IC50
values in the [3
H]hypoxanthine uptake assay (compare and ). There are several explanations for this fact. First, it is possible that some compounds have secondary cellular targets, which may enhance the antimalarial effects. Second, some of the compounds may be metabolized when added to the infected erythrocytes to produce derivatives that may be more or less potent, depending on the nature of the modification. Third, the less potent agents in the [3
H]hypoxanthine uptake assay might be actively excluded from cells due to the action of multi-drug transporters or other gene products that are known to mediate drug resistance in this parasite. 30
Assays with radiolabeled or fluorescently labeled compounds will help clarify this possibility by enabling measurements of compound accumulation in the parasite; however, these results are not yet available. Fourth, the bonafide
Hsp70 that is a target of the active compounds may be any one of the other five Hsp70s that are encoded by the P. falciparum
genome. It is striking how distinctly some compounds affect the activities of the yeast Ssa1 and human proteins and the P. falciparum
Hsp70 utilized in this study (PfHsp70-1; PF08_0054) even though these proteins are >70% identical. Because Hsp70-1 and the mitochondrial Hsp70 in P. falciparum
(PfHsp70-3; PF11_0351) are only 48% identical, the compounds are also expected to exhibit distinct effects on parasitic Hsp70s. The problem of identifying the target(s) of these compounds could, in principle, be rectified by the purification of each of the six P. falciparum
Hsp70s and 43 P. falciparum
J proteins so that each combination could be tested in ATPase assays in the presence of the modulators. A streamlined approach would be to prepare activated, affinity tagged derivatives of our novel P. falciparum
inhibitors and then identify potential cellular target(s) using an unbiased screen. This experimental regimen would ideally isolate the ‘correct’ Hsp70 and/or Hsp40 chaperone target(s).
We note structural trends that relate to potency in the P. falciparum replication assay. Specifically, there are several common and distinct features amongst the identified inhibitors depicted in . For example, all nine compounds share an ester pyrimidine core, substituted at C-4, and eight of the nine are alkylated at N1. Among the compounds, DMT3024, DMT2264, and MAL3-39 are structurally closely related. They share a benzyl ester pyrimidine core that is substituted at C-4 with an arene moiety, and an N-alkylated amide side chain that is attached via a 3–5 carbon linker. They differ in their side chain lipophilicity and in the presence of the morpholine moiety on DMT3024 and MAL3-39, which is absent in DMT2264. The most potent compounds, MAL2-215 and MAL2-213, have a slightly different ester substitution on the pyrimidine core, as well as a distinct tetrasubstituted pyrrole side chain. This tetrasubstituted pyrrole appears to be an important determinant of potency because modifications in ester identity (Bn vs Et), N1 linker length (C-4 vs C-6), and C-4 aryl substitution (Ph vs NO2) do not appear to affect activity. Finally, MAL2-61, MAL2-39, and MAL2-29 are truncated derivatives that still have the signature pyrimidine heterocycles but a minimal N1 side chain substituent lacking an amide function, i.e. H or Bn (JAB75), butyl (MAL2-39), and hexanoic acid (MAL2-29).
In summary, we describe the discovery of a novel class of anti-malarial agents. In the future, second-generation chemical libraries of these pyrimidinone sub-classes should be analyzed for their effects on P. falciparum
, but already we have identified compounds that have equal or greater potency in the [3
H]hypoxanthine uptake assay as established antimalarial agents and that are synthetically readily modified.31
Based on the success of our initial efforts, reported herein, we are confident that compounds with greater potencies and improved pharmacological properties are within reach.