Polyamines in protozoan pathogens

Abstract

Polyamines are polycationic organic amines that are required for all eukaryotic life, exemplified by the polyamine spermidine, which plays an essential role in translation. They also play more specialized roles that differ across species, and their chemical versatility has been fully exploited during the evolution of protozoan pathogens. These eukaryotic pathogens, which cause some of the most globally widespread infectious diseases, have acquired species-specific polyamine-derived metabolites with essential cellular functions and have evolved unique mechanisms that regulate their core polyamine biosynthetic pathways. Many of these parasitic species have lost enzymes and or transporters from the polyamine metabolic pathway that are found in the human host. These pathway differences have prompted drug discovery efforts to target the parasite polyamine pathways, and indeed, the only clinically approved drug targeting the polyamine biosynthetic pathway is used to manage human African trypanosomiasis. This Minireview will primarily focus on polyamine metabolism and function in Trypanosoma, Leishmania, and Plasmodium species, which are the causative agents of human African trypanosomiasis (HAT) and Chagas disease, Leishmaniasis, and malaria, respectively. Aspects of polyamine metabolism across a diverse group of protozoan pathogens will also be explored.

Introduction

Polyamines are low-molecular-weight, organic polycations that are synthesized from amino acid precursors (1, 2). In eukaryotes, the diamine putrescine is produced by decarboxylation of l-ornithine, and spermidine is synthesized from putrescine by the addition of an aminopropyl group donated by decarboxylated SAM² (dcAdoMet) (Fig. 1A). The core biosynthetic pathway that is common to both higher eukaryotes (e.g. mammals) and many single-cell organisms typically employs three key enzymes: 1) pyridoxal 5′-phosphate (PLP)-dependent ornithine decarboxylase (ODC); 2) pyruvoyl (Pvl)-dependent SAM decarboxylase (AdoMetDC); and 3) spermidine synthase, which conjugates the aminopropyl group from dcAdoMet to putrescine to generate spermidine. In the parasitic protozoa, the complete core polyamine biosynthetic pathway (ODC, AdoMetDC, and spermidine synthase) is present only in Trypanosoma brucei, Leishmania, and Plasmodium species (3). Mammals and other higher eukaryotes also produce the tetraamine spermine, which is synthesized from spermidine by addition of a second aminopropyl group. Additionally, a catabolic pathway that degrades spermine and spermidine back to putrescine via polyamine oxidase (PAO) and spermidine/spermine-N¹-acetyltransferase (SSAT) is also present (2, 4). These catabolic pathways are largely missing from the protozoa (Fig. 1A).

Figure 1.

Polyamine biosynthetic pathway in single-celled parasitic eukaryotes. A, core polyamine biosynthetic pathway. The enzymes found in the various species are indicated by italicized species names as follows: Tb, T. brucei; Ld, L. donovani; Tc, T. cruzi; Pf, P. falciparum; Tg, T. gondii; Cp, C. parvum; Tv, T. vaginalis; Gl, G. lamblia; Eh, E. histolytica. The inset identifies enzymes that are formed as either bifunctional enzymes or require oligomerization with a pseudoenzyme for activity. Abbreviations are defined in the text with the exception of DAP, diaminopropane. B, trypanothione biosynthetic pathway in the trypanosomatids. T(SH)₂, reduced trypanothione; TS₂, oxidized trypanothione. Gene resource sites are as follows: http://tritrypdb.org/tritrypdb/ (101), http://plasmodb.org/plasmo/ (102), http://toxodb.org/toxo/ (103), http://giardiadb.org/giardiadb/, and http://trichdb.org/trichdb/ (104), http://cryptodb.org/cryptodb/ (105), and http://amoebadb.org/amoeba/ (106) (see also Ref. 3). (Please note that the JBC is not responsible for the long-term archiving and maintenance of this site or any other third party hosted site.) Available X-ray structures for parasite enzymes in these pathways are summarized in Table 1.

The polyamine biosynthetic pathway in higher eukaryotes is highly regulated at multiple levels, including transcription, translation, and protein degradation, highlighting the role of polyamines in regulating cell growth (2, 4). Similar mechanisms are absent in the protozoa, although several unique regulatory strategies have been uncovered and will be discussed.

Spermidine serves as a substrate for deoxyhypusine synthase (DHS) in all eukaryotes, which together with deoxyhypusine hydroxylase (DOHH) covalently modifies the translation elongation factor eIF5A with the amino acid hypusine (5–7). DHS transfers the aminobutyl portion of spermidine to a conserved lysine residue on eIF5A, which is the sole protein that carries this modification. Both eIF5A and its hypusine modification are essential in eukaryotes, functioning to relieve ribosomal stalling on mRNAs encoding polyproline tracts and to promote translation termination (8). DHS is present in the full range of parasitic protozoa (Fig. 1A).

Because polyamines are essential for cell growth, their biosynthesis has been extensively studied for its potential to be targeted by drugs that could be used against proliferative diseases. Significant effort has been focused on cancer chemotherapy, but it has proven difficult to identify compounds with sufficient efficacy for this application (2, 9). The availability of dietary polyamines and polyamine transporters reduces the efficacy of biosynthetic inhibitors. Recently, combinations of biosynthetic inhibitors and transport inhibitors are being explored to determine whether more complete cellular polyamine reductions can be obtained for the treatment of cancer, but it remains to be seen whether this strategy will be effective (2). Efforts to study polyamine synthesis and function in protozoan parasites grew from the early work in the cancer field as researchers sought to translate the mammalian discoveries to single-celled eukaryotic pathogens.

Two high-impact discoveries paved the way for both therapeutic applications and for the uncovering of unique biology in these parasitic pathogens. The first of these was the discovery in 1980 by Bacchi et al. (10), who showed the irreversible ODC inhibitor 2-(difluoromethyl)-dl-ornithine (DFMO) (Figs. 2 and 3, A and B) cured Trypanosoma brucei infections in mice. This finding led to the eventual registration of DFMO for treatment of late-stage human African trypanosomiasis (HAT) as a combination therapy with nifurtimox (11, 12), and to its inclusion on the World Health Organization essential medicines list (World Health Organization (WHO) model list of essential medicines accessed August, 2018 (http://www.who.int/medicines/publications/essentialmedicines/20th_EML2017_FINAL_amendedAug2017.pdf?ua=1)).³ The second key discovery was the observation by Fairlamb et al., in 1985 (13), that trypanosomatids conjugate two molecules of GSH to spermidine to generate a unique cofactor called trypanothione (N1,N8-bis(glutathionyl)spermidine), which is used in place of GSH (l-γ-Glu-Cys-Gly) to maintain redox balance in these cells (Fig. 1B). More recent work has uncovered a number of examples of atypical enzyme arrangements and novel regulatory strategies, including bifunctional enzymes and enzyme complexes, that require inactive pseudoenzymes to activate their cognate paralogous enzyme (14–16).

Figure 2.

Structures of ODC and AdoMetDC inhibitors with activity against T. brucei. A, DFMO (PDB code 2TOD) shown with nifurtimox, which is not an ODC inhibitor but is used together with DFMO for combination therapy to treat HAT. B, AdoMetDC inhibitors CGP 40215A (100) (PDB code 5TVF); MDL 73811 (92), and Genz644131 (91); pyrimidineamine UTSAM568 (N⁴-(3,5-dibromophenyl)-6-methylpyrimidine-2,4-diamine) (PDB code 6BM7) compound 44 from Ref. 96. PDB numbers for available co-crystal structures bound to the T. brucei enzymes are in parentheses.

Figure 3.

X-ray structure of T. brucei ODC. ODC K69A bound to DFMO (PDB code 2TOD) (86). A, ribbon diagram of the dimer with monomers colored in teal and tan. DFMO and PLP are shown as spheres. B, ODC active site showing select residues within the 4 Å shell of PLP and DFMO. DFMO is covalently bound to Cys-360. Catalytic residues with known function in catalysis or substrate binding are displayed.

Parasites and their diseases

Protozoan parasites associated with human disease include members of the trypanosomatids (T. brucei, Trypanosoma cruzi, and Leishmania species) (17–19), the apicomplexa (e.g. Plasmodium species (20), Toxoplasma gondii (21), and Cryptosporidium parvum (22)), the anaerobic parasites, including the enteric parasites (e.g. Giardia lamblia and Entamoeba histolytica) (23), and the sexually transmitted Trichomonas vaginalis (24). The trypanosomatids are insect-borne pathogens that together infect 18 million people globally. They are the causative agents of human African trypanosomiasis (T. brucei), Chagas disease (T. brucei), and Leishmaniasis (Leishmania species). Plasmodium species are responsible for malaria, with the most important species Plasmodium falciparum responsible for nearly half a million deaths per year. For both the trypanosomatid and Plasmodium parasites, drug therapy and disease control are challenged by drug toxicity, drug resistance, and/or difficult treatment regiments. Thus, the polyamine pathway has been explored in these parasites for the potential to identify new enzymatic targets for drug discovery. The metabolic pathways, both typical and novel, the essentiality of genes in these pathways, and their utility for being exploited for drug discovery, have all been investigated and will be discussed (reviewed previously in Refs. 14, 19, 25, 26).

Polyamine biosynthetic pathways in trypanosomatids and Plasmodium species

The trypanosomatids have significant biology in common, including their ability to synthesize novel polyamine-containing metabolites (17, 18). All three pathogenic trypanosomatids have also evolved two paralogous gene products for both AdoMetDC and DHS that are required to form the active enzymes providing a unique example of how pseudoenzymes evolve to be enzyme regulators (16, 27–30) (described below). However, there are also many differences between these species that impact their requirement for polyamine biosynthetic machinery (26).

T. brucei is an extracellular parasite that replicates in the bloodstream, where the levels of polyamines putrescine and spermidine in plasma (<1 μm (31)) are not sufficient to support T. brucei growth (32). As a consequence of the extracellular niche, all enzymes that are required for spermidine synthesis are encoded in the T. brucei genome (Fig. 1A and Table 1) (14), and ODC (33, 34), AdoMetDC (both paralogs) (30), and SpdSyn (32, 34) have all been shown to be essential by genetic studies. Intracellular polyamines were first detected in T. brucei by chromatography, and through the use of isotopically-labeled precursors, the ability of T. brucei to make spermidine from l-ornithine and l-methionine was demonstrated (35). Subsequently, quantitative measurements by untargeted LC-ESI/MS found high levels of both intracellular putrescine (1–2 mm) and spermidine (3–4 mm) (36). Spermine has not been detected nor is spermine synthase encoded in the genome. N-Acetylputrescine and N-acetylornithine, but not N-acetylspermidine, have also been detected (37, 38), but their function and biosynthetic origin are both unknown. T. brucei encodes an arginase-like enzyme, but it was found to lack key catalytic residues and to be inactive (39, 40). Labeling studies suggest T. brucei can convert l-arginine to l-ornithine, but the enzymatic machinery used to catalyze this reaction is unknown, and it has been postulated that the primary source of l-ornithine is salvage (40).

Table 1.

X-ray structures of polyamine biosynthetic enzymes from protozoan pathogens

Enzyme	Species	PDB no.	Bound inhibitors, mutations, notes
ODC	T. brucei	1QU4	None
ODC	T. brucei	2TOD	DFMO; K69A
ODC	T. brucei	1NJJ	d-Ornithine + G418
ODC	T. brucei	1SZR	d-Ornithine; K294A
ODC	T. brucei	1F3T	Putrescine
ODC	E. histolytica	4AIB	None
AdoMetDC	T. brucei	5TVO	Inactive monomer
AdoMetDC/prozyme	T. brucei	5TVM	Active heterodimer with prozyme
AdoMetDC/prozyme	T. brucei	5TVF	Active heterodimer with prozyme; CGP 40215A
AdoMetDC/prozyme	T. brucei	6BM7	Active heterodimer with prozyme; UTSAM568 (compound 44)
SpdSyn	P falciparum	2PWP, 4CXM, 2HTE, 1I7C, 2PSS, 2PT6, 3B7P, 4BP1, 4CWA, 2PT9, 3RIE, 4BP3, 4UOE	Various ligands
SpdSyn	T. cruzi	4YUV, 4YUW, 4YUY, 3BWB, 3BWC, 4YUX, 4YUZ, 4YV1, 4YV2, 5B1S, 5Y4Q, 4YV0, 5Y4P	Various ligands
DHSc/DHSp	T. brucei	6DFT	Active heterodimer complex; NAD+
Arginase	Leishmania	4ITY, 4IU0, 4IU1, 4IU4, 4IU5, 5HJ9, 5HJA	Various ligands
TryS	Leishmania	2VOB, 2VPM, 2VPS	Various ligands
TryR	Leishmania	2JK6, 2W0H, 2X50, 2YAU, 4ADW, 4APN, 5EBK	Various ligands
TryR	T. brucei	2WBA, 2WOI, 2WOV, 2WOW, 2WP5, 2WP6, 2WPC, 2WPE, 2WPF, 4NEV, 6BTL, 6BU7	Various ligands
TryR	T. cruzi	1AOG, 1BZL, 1GXF, 1NDA, 4NEW	Various ligands

In contrast, both Leishmania and T. cruzi are intracellular parasites with access to polyamines within the cell of their mammalian host. Thus, as for mammalian cells polyamine transport plays a role in meeting the cellular needs for these nutrients, which has implications for drug discovery. T. cruzi has lost its ODC and relies on salvage to supply putrescine (41), although it still retains an active AdoMetDC (42) and SpdSyn (43). It is auxotrophic for putrescine and spermidine and encodes both saturable and unsaturable transport mechanisms for these amines and for the product of lysine decarboxylation, cadaverine (44, 45). T. cruzi lacks PAO/SSAT activities required for spermidine catabolism consistent with the genome sequence (http://tritrypdb.org/tritrypdb/)³ (101), and no evidence that spermidine can be converted to spermine has been found (45).

Leishmania species encode all three core polyamine biosynthetic enzymes despite occupying an intracellular niche, and they additionally express an active arginase (Fig. 1A). The PAO/SSAT back conversion pathway enzymes are absent from the genome (101),³ and biochemical studies have confirmed the lack of these activities (46). ODC null mutants of Leishmania donovani are unable to grown unless supplemented with putrescine, establishing the requirement for polyamines in these parasites (46). Subsequent studies showed that ODC null mutants were unable to establish an infection in mice, although virulence could be restored by feeding with oral putrescine (47, 48). Likewise, L. donovani null mutants of AdoMetDC (49) and SpdSyn required spermidine for growth in the promastigote stage, and SpdSyn null cells had severely reduced ability to establish an infection in mice (50). Arginase was essential for growth of promastigotes but not for intracellular amastigotes nor was it required to establish an infection in mice (51, 52). The cumulative data suggest that L. donovani amastigotes have access to sufficient pools of l-ornithine to support growth, but putrescine, and to some extent spermidine, sources are not sufficient to fully rescue gene knockouts of ODC or SpdSyn in vivo. Data suggest that, as for mammalian cells (2), levels of dietary polyamines are likely to impact the efficacy of pathway inhibitors on Leishmania intracellular cell growth.

Plasmodium species are the only parasitic apicomplexa to encode a full set of polyamine biosynthetic enzymes (ODC/AdoMetDC and SpdSyn), and they are also unique in being the only organism to encode a bifunctional ODC/AdoMetDC fusion protein (53) rather than the individual enzymes. All three biosynthetic enzymes are absent from T. gondii and C. parvum, however, all apicomplexa encode DHS (Fig. 1A) (3, 15, 25). Intracellular polyamine levels in P. falciparum are nearly 2 orders of magnitude higher than levels in uninfected erythrocytes suggesting that the parasite requires the biosynthetic enzymes to provide sufficient polyamines for growth (54). Plasmodium species synthesize both putrescine and spermidine de novo, and while spermine synthase is not encoded in the genome (http://plasmodb.org/plasmo/),³ (102) characterization of recombinant P. falciparum SpdSyn showed it also converted spermidine to spermine, although at an ∼10-fold lower rate, explaining the presence of spermine in the organism (55). Whereas individual knockouts have not been generated for pathway enzymes in Plasmodium, two genome-wide genetic screens have demonstrated that ODC/AdoMetDC, SpdSyn, and DHS are all essential enzymes for survival of P. falciparum based on saturation transposon mutagenesis (56) and for Plasmodium berghei infection in mice based on analysis of the ability to recover bar-coded knockout mutants (57). DOHH was also essential in the P. falciparum study (56), although it was not analyzed in the P. berghei study.

Polyamine biosynthesis in anaerobic parasites

The anaerobic protozoan parasites have only remnant polyamine biosynthetic pathways. Most encode ODC, but not AdoMetDC or SpdSyn, whereas all encode DHS (Fig. 1A) (3, 58, 59). The pathway in T. vaginalis is the best characterized from this group of organisms (58–60). T. vaginalis synthesizes large amounts of putrescine for excretion, which leads to a characteristically malodorous discharge in patients infected with the parasite (59). T. vaginalis is unable to synthesize spermidine or spermine directly, but the polyamine back conversion pathway (SSAT and PAO) has been reported to be present and was shown to function to convert salvaged spermine to spermidine (59). It is not yet clear whether T. vaginalis relies completely on back conversion to generate spermidine or whether the parasite also has a route to salvage spermidine directly from the host.

Polyamine transporters

As noted above, polyamine transport plays a role in supplementing mammalian cells with polyamines, although not all transporters have been characterized at the molecular level (2). In contrast, polyamine transport in T. brucei does not provide sufficient putrescine or spermidine sources to overcome the need for polyamines to be synthesized by the parasite to support infectivity in vivo. Uptake studies using radiolabeled spermidine have confirmed the absence of high-affinity spermidine transporters, as no spermidine uptake could be detected over an hour time course (32). The lack of high-affinity transporters in T. brucei is consistent with their extracellular environment where levels of polyamines in blood and spinal fluid are low (<1 μm as noted above). Despite the lack of high-affinity transporters, both putrescine and spermidine can be taken up to a limited extent if provided at sufficiently high nonphysiological concentrations, in a process that is not likely to be transporter-mediated. For example, putrescine (0.5 mm) rescues the growth deficit caused by RNAi-induced ODC knockdown (34), whereas spermidine (0.1 mm) restores polyamine pools and rescues growth of SpdSyn (34) and AdoMetDC (30) RNAi-induced knockdown. In contrast, two high-affinity ornithine transporters (K_m ∼5 μm, which is similar to levels in blood and spinal fluid) have been identified in T. brucei, one selective for l-ornithine and the other functioning on both l-ornithine and l-histidine (61). The identification of these transporters supports the hypothesis that T. brucei's primary source of l-ornithine is salvage. Knockout of the l-ornithine transporter affected sensitivity to DFMO (61), consistent with prior studies on DFMO-resistant lines that showed both reduced DFMO and increased ornithine uptake (62). During infection T. brucei releases kinesin heavy chain, which has been linked with increased host arginase activity (63). This pathway may provide increased availability of l-ornithine for salvage to help promote growth.

Both T. cruzi and Leishmania species express high-affinity polyamine transporters (POT1) consistent with their intracellular lifestyle where they have access to host polyamine pools. POT1 from Leishmania transports both putrescine and spermidine (64), whereas POT1 in T. cruzi showed specificity for putrescence and cadaverine but not spermidine or spermine (44). In T. cruzi, POT1 was not essential for the epimastigote insect stage, which also has a nonsaturable mechanism for taking up polyamines; however, the null cell line showed reduced replication at the intracellular amastigote stage (45).

P. falciparum is also able to import putrescine and spermidine (65). Competition experiments suggested that the responsible transporters likely have cross-specificity for putrescine and spermidine, in addition to spermine, ornithine, and to some extent other basic amino acids.

Conserved role of eIF5A

The requirement for spermidine as a substrate to modify translation elongation factor eIF5A with deoxyhypusine and/or hypusine is universally conserved in eukaryotes. In T. brucei, DHS (16) and eIF5A (66) have been shown to be essential using genetic knockouts, and loss of eIF5A was shown to correlate with reduced levels of proteins that were rich in polyproline tracks (66), consistent with the demonstrated role of eIF5A in relieving ribosome stalling in mammalian cells (5, 6, 8). T. brucei DOHH has not been characterized nor is it known whether deoxyhypusine or hypusine represent the dominant modification of eIF5A in the cell, but DOHH is encoded in the genome. Both DHS and DOHH from Leishmania are enzymatically active (67, 68). It seems likely that all three trypanosomatids will be capable of converting deoxyhypusine (dh)-eIF5A to hypusine (h)-eIF5A, but it is not known whether DOHH is essential. In contrast to mammalian cells, in yeast the deoxyhypusine modification is sufficient for full eIF5A function, and DOHH is not essential (7).

Studies have also been conducted in several other parasites. Both DHS and DOHH are essential in P. falciparum (56, 57). T. vaginalis encodes two eIF5A genes and a dhs gene, but dohh has not been identified in the genome (58). TvDHS was reported to be a bifunctional enzyme that also catalyzes the conversion of dh-eIF5A to h-eIF5A (69); however, the TvDHS protein amino acid sequence contains no additional domain capable of catalyzing this reaction, and thus the presence of DOHH activity in T. vaginalis is unlikely. C. parvum also has been shown to have a functional hypusine biosynthetic pathway (70).

Unique metabolism and metabolic products

The trypanosomatids synthesize a polyamine conjugate between spermidine and two GSH molecules, named trypanothione (Fig. 1B) (13). Trypanothione is synthesized in two steps: first through the conjugation of one GSH to generate glutathionylspermidine, and then through the subsequent addition of a second GSH to yield trypanothione. In Crithidia, two separate enzymes catalyze these consecutive reactions (71), but in T. brucei, T. cruzi, and Leishmania, a single enzyme trypanothione synthetase (TryS) is responsible for both steps (72). Additionally, TryS contains an amidase domain that catalyzes the back reaction (73).

Glutathionyl spermidine was also isolated from Escherichia coli by Tabor et al. (74), and the biosynthetic enzyme was shown to have two domains, one catalyzing synthesis and the second an amidase domain that regenerates the substrates (75). Subsequently, the gene encoding glutathionylspermidine synthetase was identified across a range of Enterobacteria species (76). Synthesis of spermidine and GSH conjugates appears restricted to the trypanosomatids and Enterobacteria species, but although the role of trypanothione in the trypanosomatids has been clearly established, the function of glutathionyl spermidine in bacteria remains elusive.

The requirement for trypanothione links polyamine metabolism to the cellular redox state in trypanosomatids, and it represents an example of the versatility of the chemistry of polyamines that facilitates their use as building blocks for specialized molecules in diverse species. Trypanothine has replaced GSH as the key thiol involved in maintaining reduced thiol pools. Trypanosomatids encode a trypanothione reductase (TryR) in place of GSH reductase (77). Depletion of the trypanothione biosynthetic enzymes γ-glutamylcysteine synthetase (78), GSH synthetase (79), TryS (80), and TryR (77) by genetic disruption were lethal in T. brucei in all cases, demonstrating that trypanothione is essential. Additionally, trypanothione-dependent glyoxylase, tryparedoxins, tryparedoxin peroxidase systems, and trypanothione S-transferase take the place of their GSH-based counterparts (81, 82). These systems are essential as highlighted by the finding that gene disruption of tryparedoxin peroxidases leads to ferroptosis-like death (83). These pathways have also been implicated in antimonial resistance in Leishmania species (84).

Pseudoenzymes as activators of polyamine biosynthetic enzymes

The trypanosomatids encode two paralogs for both AdoMetDC and DHS. In the case of AdoMetDC, one paralog retains the active-site residues and the ability to undergo autocatalyic processing to yield the essential pyruvoyl cofactor, whereas the second paralog named AdoMetDC prozyme is a pseudoenzyme that is missing catalytic residues and is inactive (29). The active AdoMetDC paralog (AdoMetDC) is highly impaired when expressed on its own and requires heterodimer formation with prozyme (studied for both T. brucei and T. cruzi AdoMetDC) for activity (fold activation of k_cat upon heterodimerization is ∼1000) (29, 42). Both AdoMetDC paralogs are essential for T. brucei survival, and both are required for spermidine synthesis (30). The two DHS paralogs are each lacking catalytic residues, and both are nearly inactive (16). Formation of a heterotetramer between them is required for enzyme activity, and both paralogs are also essential for growth of T. brucei. Although the requirement for a pseudoenzyme for AdoMetDC activity is only found in the trypanosomatids, the DHS heterotetramer configuration is present in the trypanosomatids and in Entamoeba, but no other species.

T. brucei AdoMetDC prozyme appears to play a regulatory role in the cell. Knockdown or chemical inhibition of AdoMetDC leads to up-regulation of prozyme protein levels, likely through a translational control mechanism that correlates with intracellular dcAdoMet levels (30, 85). Polyamine biosynthesis is highly regulated in other eukaryotic cells (2, 4), but the mechanisms that have been defined in mammalian cells and yeast are not found in the protozoan parasites (14). The regulation of AdoMetDC prozyme expression is the first and only regulatory mechanism that has been described in these parasites, suggesting that the use of pseudoenzymes may have evolved to compensate for lack of other regulatory mechanisms.

The structural basis for activation of both AdoMetDC and DHS by oligomerization with their paralogous pseudoenzymes has recently been elucidated. The X-ray structures of both the active T. brucei AdoMetDC/prozyme heterodimer and of the inactive AdoMetDC monomer were solved (Fig. 4, A and B) (28). These data demonstrated that AdoMetDC is inactive due to autoinhibition by residues in the N terminus and that this autoinhibition is relieved upon heterodimerization with prozyme. Heterodimerization leads to a conformational change in AdoMetDC such that the N terminus inserts into a newly formed pocket generated by rearrangement of the AdoMetDC β-sheet and loop structures at the dimer interface.

Figure 4.

X-ray structure of T. brucei AdoMetDC. A comparison of the inactive monomer (PDB code 5TVO) and the active heterodimer in complex with CGP 40215A (PDB code 5TVF) (28) is shown. A, inactive monomeric structure. B, active heterodimer in complex with prozyme and the inhibitor CGP 40215a. Pyruvoyl (Pvl) is shown as spheres, and the autoinhibitory sequence is shown as sticks. cP31, cis-proline 31; tP31, trans-proline 31.

In contrast to the allosteric mechanism identified for AdoMetDC activation, T. brucei DHS is a heterotetramer of two distinct pseudoenzymes (Fig. 5, A and B) (27). The catalytic unit is a heterodimer between the two DHS paralogs, one named DHSc because it maintains the catalytic lysine and the other named DHSp (for prozyme) because the catalytic lysine is missing. A tetramer is then formed from two such dimers. Although DHSc retains the residues necessary to bind spermidine, the residues in the NAD⁺-binding site have diverged, whereas the opposite is true for DHSp (Fig. 5B). Heterodimer formation restores a single active site across the dimer interface through complementation of deficiencies in the individual paralogs. Despite the altered NAD⁺-binding site, NAD⁺ binds to DHSc in a remnant of the second active site. This “dead-site” NAD⁺ is in close communication with the active-site NAD⁺, and mutagenesis studies suggest that perturbation of the NAD⁺ pocket in the dead site leads to loss of function in the active site (27).

Figure 5.

X-ray structure of T. brucei DHS. A, DHS tetramer structure showing the active-site complementation that leads to formation of one catalytically active site and one dead site across the dimer interface (PDB code 6DFT) (27). B, DHS dimer interface showing the NAD⁺-binding sites with select amino acid residues in the catalytically active site and the remnant dead site. The catalytic lysine (Lys-418) in DHSc and the equivalent inactive residue (Leu-303) in DHSp found in the dead site are marked by #.

Drug discovery

HAT remains the only approved clinical application of a polyamine biosynthetic inhibitor for treatment of a human disease (14). The early validation of ODC as a drug target in T. brucei led to interest in determining whether any other polyamine biosynthetic enzymes might be a drug target in T. brucei or in other protozoan pathogens. Despite significant effort, no strong case has been made that these enzymes would be good targets in species other than T. brucei, and although genetic studies have shown essentiality in some cases, the presence of salvage pathways and transporters seems to limit the ability of enzyme inhibitors to be disease-modifying for the intracellular parasites (e.g. T. cruzi, Leishmania, and Plasmodium species). This may in part be due to the difference between a gene knockout and an enzyme inhibitor, which is unlikely to provide 100% inhibition.

The best-validated biosynthetic enzymes remain T. brucei ODC and AdoMetDC. As noted in the Introduction, DFMO is approved for treatment of late-stage T. brucei gambiense (14) and is used in combination with nifurtimox (NECT) for this application (11). The approval of NECT greatly impacted treatment of HAT, as it is both safer and easier to administer than the toxic alternative, the arsenic-based compound melarsoprol. However, DFMO is not without its limitations; the required dose is high (200 mg/kg every 12 h for 7 days); it is rapidly excreted; it requires i.v. administration; and it is less effective against T. brucei rhodesiense, leaving only melarsoprol for the treatment of this disease (12). DFMO is a mechanism-based irreversible inhibitor of ODC that covalently attaches to an essential active-site cysteine residue (Cys-360). The covalent Michael addition complex has been observed by X-ray crystallography (Fig. 3, A and B) (86). A high-throughput screen (HTS) of 400,000 small molecules was performed in an attempt to find reversible ODC inhibitors that could have better pharmacological properties than DFMO (87). Very few inhibitors were identified suggesting that ODC is not a highly druggable target and that irreversible-based mechanisms are the only path to exploit this target. DFMO does not show selective toxicity toward T. brucei ODC versus the human enzyme as it is an effective inhibitor of both; the selective toxicity has been attributed to differences in intracellular turnover rates (88). DFMO-resistant parasites have been generated in vitro showing a loss of DFMO uptake (62, 89). Characterization of such a mutant led to the identification of the DFMO transporter, which is an amino acid transporter TbAAT6 (90).

A number of studies have suggested that T. brucei AdoMetDC is a druggable target. In addition to the essentiality of the target, MDL 73811 (Fig. 2) was found to be a potent irreversible inhibitor of T. brucei AdoMetDC and of parasite growth in vitro (91, 92). Both MDL 73811 and analogs (e.g. Genz-644131) were shown to cure early-stage T. brucei infections in mice, but they have poor brain penetration and are ineffective against the CNS stage of the disease (92, 93). Attempts to identify MDL 73811 analogs with better brain penetration have been made, but the challenging synthetic route has limited the chemistry effort, and the problem of poor brain penetration has not been overcome within this series (94). To identify better drug-like molecules, an HTS-compatible assay was developed using MS to follow the reaction on a Rapid Fire platform (95). This effort identified 13 chemical scaffolds from a screen of 400,000 compounds that are reversible inhibitors of T. brucei AdoMetDC (IC₅₀ values ∼1–10 μm range) and that showed similar activity on T. brucei in vitro. These chemical series also demonstrated species selectivity versus the human enzyme, and several showed evidence for at least partial on-target parasite killing activity and for good CNS permeability. The hit rate of 0.05% in the screen was lower than ideal, suggesting less than optimal druggability of the target, and it remains to be determined whether any of the identified series can be optimized to yield a drug candidate. Follow-up medicinal chemistry has been reported for a pyrimidineamine series (Fig. 2) (96). Although this effort did not identify analogs with potency greater than the 2–5 μm range, the best compounds from this series showed similar activity on T. brucei in whole-cell assays and good selectivity against the human enzyme. The availability of a co-crystal structure with one analog (UTSam568) should provide a path forward for optimization of the series (Fig. 2 and Table 1).

In Leishmania, despite the initial promise of the findings that ODC and SpdSyn are essential for robust infection in mice, the available polyamine biosynthetic inhibitors have failed to show curative in vivo efficacy, which is most likely related to the parasite's ability to salvage polyamines from the host. Overexpression of ODC in Leishmania led to DFMO resistance, consistent with an on-target activity (97), and studies have suggested that DFMO is more effective against visceral than cutaneous leishmaniasis (reviewed in Ref. 26). However, although DFMO reduces parasite burden in vivo, it is unable to cure L. donovani-infected mice, and the demonstration that oral putrescine can rescue toxic effects highlights the difficulty of targeting ODC in this species, as efficacy would be highly dependent on diet (48). Whether better results could be obtained with different ODC inhibitors, whether SpdSyn might be a better target in Leishmania, or whether a combination of biosynthetic inhibitors and transporter inhibitors would provide improved efficacy remain open questions.

Initial enthusiasm for the potential of TryR or TyrS as drug targets against the trypanosomatids was high based on their essentiality and for their obvious potential to provide selectivity versus the human host. Although well over a thousand TryR inhibitor papers have been published (reviewed in Ref. 98), the conclusion that TryR does not appear to be a druggable target has been evident for some time. Drug-like molecules that inhibit TryR and that are capable of progressing to late-stage lead optimization have evaded discovery. Many of the reported TryR inhibitors have shown off-target activity when tested against the trypanosomatid parasites. Although less effort has focused on TyrS, several screening programs have been undertaken, but overall, the identified compounds have lacked suitable activity on the parasites despite potentially promising results on the enzyme (98).

Efforts to target Plasmodium polyamine biosynthesis have met with similar results to Leishmania (reviewed in Refs. 25, 26). Despite essentiality of the biosynthetic enzymes, and the relatively small pools of polyamines available for salvage in blood cells, DFMO does not show good activity against the erythrocytic stage and was unable to prolong survival of infected mice. AdoMetDC inhibitors such as MDL 73811 arrest growth of the erythrocytic stage P. falciparum. However, potency against the parasite (micromolar range) was too low to translate to in vivo efficacy. Transcriptome analysis after co-administration of both DFMO and MDL 73811 to P. falciparum found lysine decarboxylase and ornithine aminotransferase levels increased 2–3-fold suggesting synthesis of cadaverine and more abundant ornithine might also contribute to lower efficacy (99). Despite the lack of progress toward identifying a suitable lead series, it remains possible that drug-like inhibitors with nanomolar potency against ODC/AdoMetDC or SpdSyn might have sufficient in vivo activity to be progressed for treatment. Given the results of the T. brucei-based ODC screen, AdoMetDC or SpdSyn seem more likely than ODC to bind drug-like inhibitors.

Finally, the potential to exploit the essentiality of DHS in all species has met with significant interest. The key issue in targeting this enzyme is the need for very strong selectivity, because clearly, inhibition of the human enzyme will cause toxicity. The strong similarity of DHS across eukaryotes suggests selectivity will be difficult to achieve in most cases. The exceptions may be the trypanosomatid and Entamoeba species. The requirement for two enzyme paralogs to form the enzyme active site presents an opportunity for the identification of species-selective inhibitors that will not inhibit the human enzyme, but might have the property of being pan-active against all trypanosomatids and possibly Entamoeba (16, 27). The crystal structure of T. brucei DHS combined with the mutagenesis data strongly suggests that species-selective inhibitors of trypanosomatid DHS could be identified based on binding to the dead site NAD⁺ pocket (Fig. 5B), providing a path forward to exploit this essential target in both the trypanasomatids and in Entamoeba species.