Toward Chemical Validation of Leishmania infantum Ribose 5-Phosphate Isomerase as a Drug Target

ABSTRACT

Neglected tropical diseases caused by kinetoplastid parasites (Trypanosoma brucei, Trypanosoma cruzi, and Leishmania spp.) place a significant health and economic burden on developing nations worldwide. Current therapies are largely outdated, inadequate, and face mounting drug resistance from the causative parasites. Thus, there is an urgent need for drug discovery and development. Target-led drug discovery approaches have focused on the identification of parasite enzymes catalyzing essential biochemical processes, which significantly differ from equivalent proteins found in humans, thereby providing potentially exploitable therapeutic windows. One such target is ribose 5-phosphate isomerase B (RpiB), an enzyme involved in the nonoxidative branch of the pentose phosphate pathway, which catalyzes the interconversion of d-ribose 5-phosphate and d-ribulose 5-phosphate. Although protozoan RpiB has been the focus of numerous targeted studies, compounds capable of selectively inhibiting this parasite enzyme have not been identified. Here, we present the results of a fragment library screening against Leishmania infantum RpiB (LiRpiB), performed using thermal shift analysis. Hit fragments were shown to be effective inhibitors of LiRpiB in activity assays, and several fragments were capable of selectively inhibiting parasite growth in vitro. These results support the identification of LiRpiB as a validated therapeutic target. The X-ray crystal structure of apo LiRpiB was also solved, permitting docking studies to assess how hit fragments might interact with LiRpiB to inhibit its activity. Overall, this work will guide structure-based development of LiRpiB inhibitors as antileishmanial agents.

INTRODUCTION

The kinetoplastid protozoan parasites Trypanosoma brucei, Trypanosoma cruzi, and Leishmania spp. are causative agents of the neglected tropical diseases (NTDs) human African trypanosomiasis (HAT), Chagas disease, and leishmaniasis, respectively. These NTDs continue to have a hugely damaging impact on global health and economies; yet, currently available chemotherapeutic options are widely inadequate, and no effective human vaccines are available. Although significant progress has been made in recent years (1), particularly in the treatment of HAT (2), it has remained challenging to develop new drugs for leishmaniasis. Cutaneous leishmaniasis (CL) and visceral leishmaniasis (VL) pose a major health threat to an estimated 1 billion people, with over 1 million cases occurring annually and VL causing 20,000 to 30,000 deaths (3, 4). Although a number of therapeutic candidates are progressing into clinical trials (5), most are at early stages and further leads are required to off-set the often high attrition rates of clinical development. This is particularly true in the case of leishmaniasis where treatment complexity is heightened by the requirements of different regions, which experience infections driven by different leishmanial species that also exhibit differing drug susceptibilities. Thus, sustained efforts to identify new avenues for antileishmanial drug discovery are vital. This need has led to extensive research into the metabolism of Leishmania sp. parasites, with a view to establish areas of biochemical divergence from their hosts that can be exploited to combat them while minimizing potential side effects.

The enzyme ribose 5-phosphate isomerase (Rpi) catalyzes the isomerization of ribose 5-phosphate (R5P) to ribulose 5-phosphate (Ru5P) (see Fig. S1 in the supplemental material) in the nonoxidative branch of the pentose phosphate pathway (PPP) (Fig. 1) (6, 7). Alongside glycolysis and the Kreb’s cycle, the PPP was one of the first identified metabolic pathways and is highly conserved in both prokaryotes and eukaryotes (8). The PPP supports key cellular functions, with the nonoxidative branch providing precursors for nucleotide, amino acid, and vitamin biosynthesis and the oxidative branch contributing to redox regulation (8). Although the oxidative branch of the pathway is confined to eukaryotes, the nonoxidative pathway is common to all organisms. Two physically and genetically distinct forms of Rpi are known to exist and were first characterized in Escherichia coli K-12 (9), which produces both types. RpiA is present in all taxonomic groups but RpiB has only been found in bacteria and lower eukaryotes, including protozoa (10).

FIG 1.

Overview of the pentose phosphate pathway (PPP). Both oxidative and nonoxidative branches of the pathway are depicted. Ribose 5-phosphate isomerase (Rpi) is highlighted in bold. G6PDH, glucose-6-phosphate dehydrogenase; 6-PGL, 6-phosphogluconolactonase; 6-PGDH, 6-phosphogluconate dehydrogenase; RuPE, Ribulose 5-phosphate epimerase; TKT, transketolase; TAL, transaldolase.

The fact that RpiBs are found exclusively in lower eukaryotes, including human-pathogenic species, has led to them becoming the focus of numerous studies aiming to establish their essentiality, solve their protein structure, and identify specific inhibitors. Crystal structures for the RpiBs from E. coli and Mycobacterium tuberculosis have been determined to facilitate targeted drug design (11–13). Inhibition studies have centered on mimicking the structure of the high-energy cis-enediolate isomerization reaction intermediate, such that inhibitors including 4-phospho-d-erythronohydroxamic acid (4-PEH) have been identified (14). However, these inhibitors lack selectivity for RpiBs, given that RpiAs also catalyze isomerization via this high-energy intermediate, meaning they are unlikely to provide routes for the development of novel, selective therapeutics.

The RpiBs from kinetoplastid parasites have also been examined. RpiB downregulation through RNA interference (RNAi) in T. brucei decreased parasite in vitro growth and the infectivity of bloodstream forms toward mice (15). Mice infected with induced RNAi clones exhibited lower parasitemia and a prolonged survival compared with control mice. Phenotypic reversion was achieved by complementing induced RNAi clones with an ectopic copy of the T. cruzi gene (15). A crystal structure for T. cruzi RpiB (TcRpiB) has been determined (16), and research into the enzyme’s substrate specificity and potential inhibition has been conducted (17). Leishmania sp. RpiB has been studied in Leishmania donovani, Leishmania major, and Leishmania infantum (10, 18, 19). In L. infantum, null mutant generation was only possible when an episomal copy of the RpiB gene was provided, and the latter was preserved both in vitro and in vivo in the absence of drug pressure. This finding indicates that the gene is essential for parasite survival (19). Although kinetoplastid RpiBs have also been shown to be susceptible to 4-PEH inhibition, no other specific inhibitors have yet been identified, potentially as studies of these enzymes have been largely dominated by structural modeling and in silico inhibition prediction (10, 20).

Here, we present the findings from a thermal shift (differential scanning fluorimetry) fragment library screening against recombinantly expressed L. infantum RpiB (LiRpiB). Hits obtained were analyzed for their ability to inhibit LiRpiB via the in vitro activity assay and for their antiparasitic potency in cell viability assays. Results indicate that the ability of fragments to interact with and effectively inhibit LiRpiB can be linked to antiparasitic efficacy, strengthening the case for LiRpiB as a validated drug target. We have also determined the first X-ray crystal structure for a Leishmania sp. RpiB, which permitted in silico docking analysis to speculate how the hit inhibitory fragments identified during this research might bind and inhibit LiRpiB activity. Overall, this work provides novel insights that will inform the design of kinetoplastid RpiB-specific leads for drug development.

RESULTS

The LiRpiB recombinant protein was expressed, purified, and subjected to thermal shift analysis. Thermal shift (differential scanning fluorimetry) is reliant on fluorescent dyes that signify when a protein has unfolded (21). One of the most commonly used fluorescent dyes is SYPRO orange (also used here), which forms nonspecific interactions with hydrophobic protein residues (22). This means its signal will be strongest when the protein being analyzed is unfolded and internal hydrophobic residues are exposed. By running an assay in which the reaction temperature is increased in degree increments per minute, it is possible to define the temperature at which a protein unfolds by monitoring its fluorescence. The temperature at which this occurs is designated the protein melting temperature (T_m) (21). Compounds that significantly alter T_m may be potential inhibitors. Initial testing showed LiRpiB was amenable to SYPRO orange thermal shift analysis, with an average T_m of 59.6°C being obtained for the protein.

To permit screening of a fragment library, a positive control for ligand binding had to be established. The obvious choice was the enzyme’s substrate R5P. However, reported K_m values for R5P against other RpiB enzymes were relatively high, with Stern et al. reporting a R5P K_m of 4 mM against TcRpiB (17). Indeed, no significant shifts in LiRpiB T_m in the presence of 5 to 50 mM R5P could be obtained, supporting the suggestion that R5P has a relatively weak interaction with LiRpiB and other parasitic RpiBs.

It was hypothesized that 2-deoxyribose 5-phosphate (dR5P) may serve as an alternative control ligand, as it could be sufficiently similar to R5P to bind LiRpiB but also be retained in the enzyme active site for longer, prolonging the interaction. Testing a gradient of dR5P against LiRpiB (5 to 50 mM) resulted in significant shifts in LiRpiB T_m. To provide the consistency across replicates that would provide Z-factor values for each plate above the confidence threshold (Z-factor, >0.5), 30 mM was selected as the dR5P concentration for positive-control reactions, reproducibly inducing a 6°C shift in LiRpiB T_m (Fig. 2).

FIG 2.

Confirmation LiRpiB is amenable to thermal shift library screening. Fluorescence profiles for 8 replicate LiRpiB negative-control reactions (protein in the absence of any potential ligand) and 8 replicate positive-control reactions (LiRpiB in the presence of 30 mM dR5P) are shown. LiRpiB T_m = 59.3 ± 0.08°C. The presence of 30 mM dR5P produced a 6°C T_m shift (LiRpiB T_m = 65.6 ± 0.11°C).

Approximately 800 fragments (see Fig. S2 in the supplemental material) across 11 fragment plates (Z-factors ranging from 0.5 to 0.8) (see Fig. S3 in the supplemental material) were screened at a 1 mM effective concentration against LiRpiB. Graphs to summarize the data output for each fragment plate were compiled, with a change in T_m relative to the negative control represented for each individual fragment (see Fig. S4 in the supplemental material).

Upon review of the thermal shift data for the fragments, a threshold T_m shift of ±5°C was set for fragment hit selection. This led to the selection of 15 fragment hits (Table 1), which were followed up with enzymatic assays and antiparasitic activity studies.

TABLE 1.

Fragment hits from thermal shift screening of LiRpiB

Initially, fragment inhibition was tested in LiRpiB in vitro activity assays. Inhibition was compared with that achieved with the well-established Rpi inhibitor 4-PEH, which was tested at a 10 mM concentration against both the forward (R5P → Ru5P) and reverse (Ru5P → R5P) isomerization reactions (Table 2). Fragments 328 and 458 were excluded from these assays due to solubility problems. To assess whether fragments could be tested in the forward reaction assay, their absorbances at 290 nm (0.5 mM and 1 mM concentration) were measured. Almost all fragments displayed high absorbance values at 290 nm (optical density [OD], ≥1) (see Fig. S5 in the supplemental material). Therefore, they could not be tested in the forward Rpi activity assay.

TABLE 2.

Inhibitory capacity of compounds against LiRpiB^a

Inhibitor	LiRpiB inhibition (%)		Antiparasitic activity (%)			THP1 viability (%)
	Forward	Reverse	Promastigote WT	Promastigote sKO RpiB	Amastigote
2	10 ± 14	21 ± 2	N.A.	N.A.	30 ± 8	92 ± 1
3	12 ± 9	10 ± 4	14 ± 5	9 ± 0	79 ± 26	99 ± 3
25	6 ± 1	6 ± 7	N.A.	N.A.	40 ± 31	115 ± 26
68	N.T.	19 ± 1	N.A.	N.A.	N.A.	120 ± 18
152	N.T.	13 ± 8	31 ± 11	52 ± 24	N.A.	114 ± 15
278	N.T.	4 ± 4	13 ± 1	28 ± 5	N.A.	19 ± 1
338	N.T.	32 ± 10	100 ± 0	102 ± 2	100 ± 17	77 ± 14
372	5 ± 15	26 ± 3	N.A.	N.A.	N.A.	96 ± 5
383	N.T.	−2 ± 1	N.A.	N.A.	N.A.	121 ± 16
540	N.T.	23 ± 5	26 ± 15	64 ± 5	87 ± 15	71 ± 2
565	N.T.	17 ± 1	N.A.	N.A.	N.A.	96 ± 0
576	39 ± 22	14 ± 5	N.A.	69 ± 28	N.A.	21 ± 20
626	N.T.	−7 ± 8	N.A.	N.A.	N.A.	101 ± 4
4PEH	51 ± 19	24 ± 4	N.A.	N.A.	N.A.	82 ± 1

^aThe enzymatic inhibition of all the compounds was determined at 1 mM, except the 4-PEH that was tested at 10 mM. The values correspond to the mean ± SD of the inhibitory effect (%) relative to control (drug absence) from 2–3 independent assays performed in duplicate. The activity of 100 μM of each fragment against L. infantum promastigotes was determined using the resazurin assay (72 h). For the intramacrophagic parasites, THP1 cells infected with parasites expressing luciferase were used. The cytotoxicity and activity determinations were performed with 100 μM of the fragments in an MTT assay involving phorbol myristate acetate (PMA)-differentiated THP-1 cells. Antiparasite and viability data represented are the average ± SD of at least two independent assays performed in at least triplicate. N.A., not active; N.T., not tested.

Only fragments 2, 3, 25, 372, 540, and 576 were tested in this assay. Results for fragment 540 were also excluded during posterior analysis because, in the presence of R5P, the measured signal became saturated, which hindered a clear interpretation of the measurements obtained. Percentage inhibition values for LiRpiB that could be determined from the forward Rpi assay are shown (Table 2). No significant inhibition was found through this assay system except in the case of fragment 576; 1 mM 576 inhibited the forward reaction to a similar extent as 10 mM 4-PEH (Fig. 3). This result suggests that 576 is (at least) as potent an inhibitor of LiRpiB as 4-PEH.

FIG 3.

Measurements of LiRpiB activity catalyzing the forward reaction in the presence of the fragment 576. Values represent the enzyme activity with (squares) and without (circles) the fragment and with 4-PEH (triangles). The reaction occurred in the presence of 12.5 mM R5P and 0.0025 mg/ml LiRpiB. Fragment 576 and 4-PEH were tested at 1 mM and 10 mM, respectively. The absorbance at 290 nm (OD) was measured every 30 s during 20 min at 37°C. The values obtained with the compound 576 and in the control without the fragment correspond to the mean of duplicates. The values concerning the 4-PEH were obtained from a single enzymatic kinetic reading.

The experimental limitations of the forward reaction assay do not apply to the reverse reaction assay. Therefore, all compounds could be tested in the reverse reaction assay system. Fragments 2, 338, 372, and 540 (1 mM) were capable of inhibiting the enzyme to a similar level as that of 10 mM 4-PEH (Table 2). Surprisingly, the inhibitory capacity of fragment 576 and 4-PEH was not as high in this assay system. This may be due to kinetoplastid RpiB activity favoring the production of R5P, meaning higher levels of these inhibitors may be required in order to inhibit the enzyme in this direction.

The 15 fragment hits and 4-PEH were also tested for their ability to inhibit L. infantum parasite growth at a 100 μM concentration. Compounds were assayed against both L. infantum promastigotes (wild type and LiRpiB single knockout [sKO]) and intramacrophagic amastigotes (Table 2). As 4-PEH is only an inhibitor of LiRpiB at a high millimolar range in protein activity assays (10 mM was routinely used to achieve inhibition during this study), it had been anticipated to show little to no antiparasitic activity in vitro/in vivo (7), which was shown to be the case in treating with 100 μM. However, several of the fragment hits were shown to inhibit parasite growth at this concentration. Interestingly, fragments 152, 278, and 540 were seemingly more effective against the LiRpiB sKO promastigotes than the wild type. However, the same was not evident for fragment 3. Although sKO modifications against essential genes often do not produce significant phenotypic distinctions from wild-type parasites, it is possible that compensatory upregulation mechanisms accounting for the loss of a single LiRpiB allele are acting to protect the parasites from treatment with these compounds (to some extent). Fragment 576 was active against LiRpiB sKO promastigotes but was inactive against the wild type. As the in vitro enzyme assay results point to fragment 576 being a more effective inhibitor of the LiRpiB forward reaction than the reverse reaction, this could indicate that sKO parasites are rendered more susceptible to forward reaction inhibitors. The most active fragments against both L. infantum life cycle forms were 338 (Fig. 4) and 540, which were also among the most potent inhibitors of the LiRpiB reverse reaction. Collectively, these data suggest that a possible mode of action for the observed antileishmanial activity of potent fragments, such as 338, is modulation of the activity of LiRpiB.

FIG 4.

Average dose response curves. The 50% effective concentration (EC₅₀) and 95% confidence interval for miltefosine (A) and fragment 338 (B) antiparasitic activity against wild-type (WT) and single knockout (sKO) RpiB promastigotes. The curves represent the merged output from the data of three independent curves.

To facilitate further development of inhibitor leads, a crystal structure for apo LiRpiB was determined at a 1.6-Å resolution (Table 3). An LiRpiB functional dimer was established, and a LiRpiB tetramer could then be assembled from the monomer of the asymmetric unit via crystallographic symmetry (Fig. 5). This is similar to the TcRpiB tetramer described by Stern et al. (16).

TABLE 3.

Data collection and refinement statistics

Parametera	LiRpiB/SO4
Data collection
Resolution (last shell) (Å)	19.75–1.57 (1.66–1.57)
Space group	F222
Unit-cell parameters
a, b, c (Å)	a = 80.87, b = 83.55, c = 89.35
α, β, γ (°)	α = β = γ = 90
Completeness (last shell) (%)	99.6 (97.8)
Redundancy (%)	6.5
I/σ(I) (last shell)	23.20 (5.09)
Rsym(I) (last shell) (%)	5.27 (34.3)
Refinement
Protein molecule/A.U.	1
Rwork (%)	14.5
Rfree (%)	17.8
RMSD in bond lengths (Å)	0.024
RMSD in bond angles (°)	2.217
Mean B factors (Å2)	17.35
PDB entry code	6FXW

^aRMSD, root mean square deviation; A.U., asymmetric unit.

FIG 5.

Reconstitution of LiRpiB structure. The green monomer represents the asymmetric unit, and additional copies participating in tetramer formation through crystallographic symmetry are depicted in gray.

Each monomer of the dimer (hence also of the tetramer) is based on a Rossmann fold with a five-stranded parallel β-sheet flanked by three α-helices on one side and two on the other. A sixth α-helix (C terminus) extends from the core domain to interact with the second subunit of the dimer. Each monomer subunit of the dimer forms one side of the active site cleft. Within the active site (Fig. 6), key residues involved in substrate interaction that have been identified in RpiB homologues are conserved, namely, Asp13, His14, Cys72, Thr74, and Arg116 from one subunit of the dimer and His105, Asn106, Arg140, and Arg144 from the other subunit of the dimer. This finding indicates that the catalytic mechanism operated by the active site is consistent with other RpiBs.

FIG 6.

LiRpiB active site occupied by a sulfate ion. Key residues for LiRpiB protein activity are labeled and depicted as sticks, and water molecules are depicted as red spheres. Conformation of TcRpiB Arg113, as observed in the TcRpiB apo structure, is highlighted in brown transparent stick.

In comparison to the apo TcRpiB structure, a different orientation of Arg116 (Arg113 in TcRpiB) is observed (Fig. 6). Also, Thr74 is present in LiRpiB in place of Ser71 that occurs in TcRpiB, which is more in line with E. coli RpiB that also carries a Thr residue at this position (11, 16).

Given the orientation of the sulfate ion in the active site and the consequent suggestion that the R5P/Ru5P substrate will be oriented in a similar way (Fig. 7A), these results also indicate that the 4-PEH interaction with LiRpiB will be consistent with that established for other RpiBs (Fig. 7B). Docking analysis predictions for antiparasitic fragment hits 338 and 540 indicates that the interaction of these fragments with the active site are likely to center around residues His14, Arg140, and Arg144 (Fig. 7C and D). These residues are conserved in RpiB active sites but not RpiAs, which could account for the selectivity of these fragments toward combating parasites (Table 2).

FIG 7.

LiRpiB active site illustrating inhibitor substrate and inhibitor binding predictions. R5P/Ru5P (A) and 4-PEH (B), predicted with reference to TcRpiB; 338 (C) and 540 (D) fragment binding predictions from in silico docking analysis. The docking conformations depicted displayed optimal binding energy and highest level of intermolecular interactions.

DISCUSSION

Type B ribose-5-phosphate isomerase has been flagged as an attractive protein target for drug development to combat pathogens, given the critical role of this enzyme combined with its evolutionary divergence from mammalian RpiA. It has long been assumed it may be possible to design RpiB-specific inhibitors; yet, thus far, no such inhibitors have been identified.

In this work, a fragment library screening was conducted for L. infantum RpiB, utilizing thermal shift as the screening technique. Thermal shift assays are commonly used to assess the potential of adding ligands for improving protein stability, which aids crystallization (21, 23, 24). Although not capable of identifying inhibitors directly, they have been widely used to facilitate this process (25–28), assisting in drug discovery efforts. The ability to apply this screening method to any amenable protein, regardless of function, makes thermal shift particularly valuable. This is perhaps most relevant when attempting to establish inhibitors for proteins that lack high-throughput activity-based screening methods, as was the case for LiRpiB. Using this method, LiRpiB was screened against 851 different fragments. A total of 15 hit fragments that produced ±5°C shifts in LiRpiB T_m were selected, progressing to LiRpiB activity assays and cell viability assays against L. infantum parasites. The activity assay results indicate that thermal shift screening was capable of identifying inhibitors, as fragments 2, 338, 372, 540, and 576 showed LiRpiB inhibitory activity. Significantly, fragment hits 338 and 540 that inhibited LiRpiB in vitro also displayed antiparasitic activity toward both L. infantum promastigotes (wild type and LiRpiB sKO) and intramacrophagic amastigotes in an infection model, while also being well tolerated by mammalian THP1 cells in cytotoxicity assays. Overall, fragment 338 was considered the best hit against LiRpiB given its superior potency. The fact that the ability to interact with and inhibit LiRpiB was utilized to identify antiparasitic hits strengthens the case for describing LiRpiB as a potential chemically validated drug target in L. infantum. This is an important observation given that the potential value of RpiB as a drug target in parasitic protozoa has formerly been deemed equivocal (6). Although not specifically examined in this study, the fact that RpiB is highly conserved among Leishmania spp. makes it possible that these findings could also apply to species beyond L. infantum.

Previous research into RpiB has largely focused on understanding the enzyme’s reaction mechanism, particularly in comparison to RpiA; crystal structure elucidation; and/or in silico docking studies. Notably, in the context of protozoan parasites, although the crystal structure established for T. cruzi RpiB has permitted computational ligand docking studies, potential hits have yet to have their antiparasitic properties confirmed (20). For many years, inhibition of Rpi has centered around the well-established inhibitor 4-PEH. However, low potency and lack of RpiB specificity render this compound inappropriate for chemotherapeutic applications. The inhibitory fragments reported here, several of which were markedly more potent than 4-PEH (1 mM fragment concentration was contrasted with 10 mM 4-PEH inhibitory activity), may be able to fill this void and provide initial scaffolds for structure-led rational optimization of RpiB inhibitors. In the case of Leishmania, the first crystal structures for a leishmanial RpiB, solved during this study, can greatly facilitate this process for L. infantum and other Leishmania spp. The structural resolution of LiRpiB indicates that, as might be anticipated from high levels of protein sequence homology, the RpiB structure is highly conserved among protozoan parasites. Thus, there is potential for the design of inhibitors capable of interkinetoplastid impact. Docking results for hit antiparasitic fragments 540 and 338 predict that their interaction with the active site of LiRpiB hinges on residues His14, Arg140, and Arg144. These residues typically coordinate the phosphate moiety of substrate R5P, orienting and stabilizing the substrate in the correct conformation for isomerization. Thus, it is possible fragments 540 and 338 compete with R5P in the binding of these key residues. Future cocrystallization and/or structural modeling of the fragments (and, potentially, further analogues) with LiRpiB may shed further light on the RpiB catalytic mechanism but also, crucially, how to gain inhibitor specificity over RpiA. It will also be important to establish how the antiparasitic fragment hits are turned over by the parasites, as current speculations can be made only as to how intact fragment compounds could interact with LiRpiB.

Furthermore, it is currently unknown whether any off-target effects are also contributing to fragment efficacy, which will be important to determine going forward, as well as antiparasitic efficacy toward other Leishmania spp. Overall, however, this work provides new avenues with which to pursue RpiB as a druggable target in Leishmania spp. and other pathogenic organisms.

MATERIALS AND METHODS

Cloning of the LiRpiB gene.

The LiRpiB gene was PCR amplified from L. infantum genomic DNA (MHOM/MA/67/ITMAP-263), using the following primers: 5′-CAATTTCCATATGCCGAAGCGTGTTGC-3′ and 5′-CCCAAGCGAATTCTCTACTTTCCTTCC-3′ (enzyme restriction sites are underlined). The purified LiRpiB PCR product was NdeI/EcoRI digested and cloned into a pGEM-T Easy vector (Promega). The presence of the LiRpiB open reading frame (ORF) was confirmed via sequencing and was subsequently subcloned into a pET28a(+) expression vector (Novagen).

Expression and purification of recombinant LiRpiB.

The pET28a(+) LiRpiB expression vector was transformed into E. coli BL21(DE3) cells. The recombinant protein was expressed by induction of log-phase cultures in Luria-Bertani media (optical density at 600 nm [OD₆₀₀], 0.6) with 0.5 mM isopropyl-β-d-thiogalactopyranoside (IPTG) for 3 h at 37°C with shaking at 250 rpm/min. Bacteria were harvested through centrifugation (3,077 × g for 40 min at 4°C) and suspended in 20 ml of buffer A (0.5 M NaCl and 20 mM Tris-HCl [pH 7.6]). Samples were then sonicated using a Branson sonifier 250 under the following conditions: output, 4; duty cycle, 50%; and 10 cycles with 15 sec each. Samples were centrifuged (3,077 × g for 60 min at 4°C), and the product supernatant was retained for further processing. Recombinant LiRpiB was purified in one step using Ni²⁺ resin (ProBond) and pre-equilibrated in buffer A. The column was washed sequentially with 2 to 3 ml of buffer A, 20 ml of the bacterial crude extract, 2 ml of buffer A + 25 mM imidazole, 2 ml of buffer A + 30 mM imidazole, 2 ml of buffer A + 40 mM imidazole twice, 2 ml of buffer A + 50 mM imidazole, 10 ml of buffer A + 100 mM imidazole, 5 ml of buffer A + 500 mM imidazole, and 8 ml of buffer B (1 M imidazole, 0.5 M NaCl, and 200 mM Tris [pH 7.6]). The LiRpiB enzyme was eluted in the fractions of buffer A containing 100 or 500 mM imidazole. Desalting was performed against 100 mM Tris-HCl (pH 7.6; storage buffer, reaction buffer for direct reaction), using PD-10 desalting columns (code no 17-0851-01; GE Healthcare).

Differential scanning fluorimetry with LiRpiB.

Differential scanning fluorimetry was set up in 96-well PCR plates employing a total reaction volume of 100 μl. Reactions consisted of 13 μM protein (LiRpiB) in 50 mM morpholinepropanesulfonic acid (MOPS) (pH 8.0) reaction buffer with 5× SYPRO orange dye (Invitrogen) as the fluorescent indicator of protein unfolding (excitation, 492 nm; emission, 610 nm). Fragments from an in-house library at the University of St. Andrews, expanded from the Maybridge Rule of 3 (Ro3) library (25), were screened against RpiB at a 1 mM concentration (0.5% dimethyl sulfoxide [DMSO] final concentration per well). This library incorporates fragments that adhere to the following chemical parameters: ≤300 molecular weight (MW), ≤3.0 cLogP, ≤3 H-bond acceptors, ≤3 H-bond donors, ≤3 rotatable bonds and ≤60 Ų polar surface area. These properties are predicted to increase the probability of viable fragment lead discovery (29), a more stringent application of the “Rule of 5” criteria developed by Lipinski et al. to curate drug-like compound libraries (30). DMSO was used in negative-control reactions and 30 mM dR5P was added to positive-control reactions. Each plate screened included 8-replicate negative and 8-replicate positive-control reactions to permit calculation of a Z-factor (31) (threshold, 0.5). Thermal shift scans were performed in a real-time PCR machine (Stratagene Mx3005P with software MxPro v 4.01) over a temperature range of 25°C to 95°C, ramping at 0.5°C min⁻¹. Data were then exported to Excel for analysis using “DSF analysis.” modified from the template provided by Niesen et al. (21). Melting temperature (T_m) values were calculated through nonlinear regression analysis, fitting the Boltzmann equation to denaturation curves using GraphPad Prism as previously described (25).

Compound preparation for assays.

Compounds were dissolved in 100% DMSO at 100 mM, aliquoted, and stored at −20°C. The following procedure was applied for the solubilization of compounds in reaction buffer (100 mM Tris-HCl [pH 7.6]): 1 hour of vortex mixing, 15 minutes of ultrasound treatment, and 1 h of incubation at 37°C under strong agitation. The optical density (OD) at 290 nm of soluble compounds at a 1 mM or 0.5 mM concentration was measured, and only compounds with low absorbance values (OD, <1) were selected for analysis. The control compound 4-PEH was used as a 100 mM stock in water and stored at −20°C.

LiRpiB activity assays.

Compounds were tested in the forward and/or reverse reaction at a final concentration of 1 mM. Concerning the forward reaction, a direct spectrophotometric method at 290 nm was used to quantify Ru5P formation in the presence of 12.5 mM R5P and 0.0025 mg/ml LiRpiB in a total volume of 300 μl (32). The reaction buffer was 100 mM Tris-HCl (pH 7.6). The absorbance was monitored at 37°C for 20 minutes. The blank for each compound was assumed to be the absorbance at t = 0. Compound inhibitory effect was quantified via measuring OD values in both the presence and absence of compound within the first 4 minutes of the assay and at the endpoint (t = 20). The ratio between these values was used to determine compound percent (%) inhibition. In the inverse reaction, a modification of Dische’s cysteine-carbazole method was used to quantify R5P formation (17). In a total volume of 15 μl, 5 mM Ru5P, 0.0025 mg/ml LiRpiB, and 5 μl of compound were incubated for 10 minutes at room temperature. Reaction buffer was 100 mM Tris-HCl, 1 mM EDTA, and 0.5 mM 2-mercaptoethanol (pH 8.4). For color revelation, 15 μl of 0.5% cysteinium chloride, 125 μl of 75% (vol/vol) sulfuric acid, and 5 μl of a 0.1% solution of carbazole in ethanol were added to 10 μl of the previous mixture. The absorbance at 546 nm was determined following incubation for 30 minutes at room temperature in the dark. A blank without enzyme was always run in tandem with and without compound. The enzyme activity in the presence of compounds was measured by subtracting the OD values obtained in the presence of the enzyme to the blank values. Percent inhibition was assessed via normalization of the activity values against those obtained in the compound negative-control reactions.

Parasite culture.

L. infantum (MHOM/MA/67/ITMAP-263) wild-type promastigotes and single knockout (sKO) LiRpiB promastigotes (19) were maintained at 27°C in complete RPMI 1640 (33). Axenic amastigotes of the same strain expressing firefly luciferase (34) were grown at 37°C with 5% CO₂ in a cell-free medium (35).

Growth inhibition assays.

The percent growth inhibition in promastigotes was determined by incubating the parasites in a 96-well plate with a starting inoculum of 1 × 10⁶ cells/ml promastigotes during 72 h with defined concentrations of the selected fragments. After incubation, 50 μM resazurin was added and incubated for 4 h. Fluorescence was measured at 540 nm and 620 nm excitation and emission wavelength, respectively, using a Synergy 2 multimode reader (Biotek). Activity against intracellular amastigotes was measured using THP1 cells infected with luciferase-expressing amastigotes as previously described (36). Fragments were screened at 100 μM in these assays. THP1 cytotoxicity was determined using a 3-(4,5-dimethyl-2-thiazolyl)-2,5-diphenyl-2H-tetrazolium bromide (MTT) assay following exposure of the THP1 cells to 100 μM of the fragments, as described elsewhere (37).

Production and purification of LiRpiB for crystallization.

The purification protocol was based on that of TcRpiB (17). Briefly, full-length LiRpiB was produced as described above except that expression was induced overnight at 18°C with 0.7 mM IPTG. Cells were harvested via centrifugation (2,500 × g for 30 min at 4°C) and stored at −20°C until use. After thawing, bacterial pellets were suspended in lysis buffer (50 mM Tris-HCl [pH 7.5], 300 mM NaCl, 1 mM DTT, and 20 mM imidazole). Following sonication, the cell homogenate was centrifuged (56,000 × g for 30 min at 4°C), and the soluble tagged protein was purified by affinity chromatography on His60 Ni superflow resin (Clontech). Lysis buffer was used to wash the resin, and protein was eluted with buffer A (50 mM Tris-HCl [pH 7.5], 300 mM NaCl, 1 mM DTT, and 500 mM imidazole). After removal of the His tag (thrombin cleavage, overnight, 4°C), additional size exclusion chromatography (SEC) was carried out on a HiLoad 16/60 Superdex 200 column equilibrated with buffer B (20 mM Tris-HCl [pH 7.5], 150 mM NaCl, 2 mM DTT, and 1 mM EDTA). Fractions containing purified protein were pooled, and protein was concentrated to ∼12 mg/ml by ultrafiltration.

Crystallization and data collection.

Crystallization experiments were carried out using the sitting drop vapor diffusion method in 96-well plates using an Innovadyne nanodrop robot (300-nl protein solution + 300-nl crystallization condition). LiRpiB crystals grew at 277 K in 100 mM HEPES (pH 7) and 2 M ammonium sulfate. Crystals were flash-frozen in liquid nitrogen in crystallization condition supplemented with 22% glycerol. Diffraction data were collected on beamline Proxima 1 (SOLEIL, Saclay, France) on a Pilatus 6M detector.

Structure determination and refinement.

X-ray data were processed with XDS (38) and the software package CCP4 (39). The LiRpiB structure was solved by molecular replacement with Phaser (40, 41) using TcRpiB (Protein Data Bank entry code 3K7O) (16) as the search model. Model building and improvement were conducted by iterative cycles of manual building with Coot (42) and refinement with REFMAC (43). Structural data have been deposited in the Protein Data Bank under entry code 6FXW.

In silico docking analysis.

Docking analysis was conducted using PyRx. Ligands (PEH and fragments 338 and 540) were prepared using Chem3D, using the package’s MM2 structure optimization tool. The LiRpiB functional dimer was prepared as a macromolecule for docking using AutoDock tools (44, 45). PyRx docking analysis (46) was performed using a grid box with dimensions x = 9.217, y = 9.292, and z = 11.459 to encompass the enzyme’s active site, determining the ligand conformations that would provide optimal binding energies (exhaustiveness, 16), which were then studied in relation to the LiRpiB active site structure using PyMOL.