Bis(glutathionyl)spermine and other novel trypanothione analogues in Trypanosoma cruzi.

Trypanosomatids differ from other cells in their ability to conjugate glutathione with the polyamine spermidine to form the antioxidant metabolite trypanothione (N1,N8-bis(glutathionyl)spermidine). In Trypanosoma cruzi, trypanothione is synthesized by an unusual trypanothione synthetase/amidase (TcTryS) that forms both glutathionylspermidine and trypanothione. Because T. cruzi is unable to synthesize putrescine and is dependent on uptake of exogenous polyamines by high affinity transporters, synthesis of trypanothione may be circumstantially limited by lack of spermidine. Here, we show that the parasite is able to circumvent the potential shortage of spermidine by conjugating glutathione with other physiological polyamine substrates from exogenous sources (spermine, N8-acetylspermidine, and N-acetylspermine). Novel thiols were purified from epimastigotes, and structures were determined by matrix-assisted laser desorption ionization time-of-flight analysis to be N1,N12-bis(glutathionyl)spermine, N1-glutathionyl-N8-acetylspermidine, and N1-glutathionyl-N12-acetylspermine, respectively. Structures were confirmed by enzymatic synthesis with recombinant TcTryS, which catalyzes formation of these compounds with kinetic parameters equivalent to or better than those of spermidine. Despite containing similar amounts of spermine and spermidine, the epimastigotes, trypomastigotes, and amastigotes of T. cruzi preferentially synthesized trypanothione. Bis(glutathionyl)spermine disulfide is a physiological substrate of recombinant trypanothione reductase, comparable to trypanothione and homotrypanothione disulfides. The broad substrate specificity of TcTryS could be exploited in the design of polyamine-based inhibitors of trypanothione metabolism.

All living cells require biochemical systems to protect against oxidative and chemical stresses and to maintain redox balance (reviewed in Ref. 1). In most prokaryotes and eukaryotes, glutathione/glutathione reductase and thioredoxin/thioredoxin reductase systems along with associated peroxiredoxins provide these essential biochemical functions (reviewed in Refs. [2][3][4][5]. Although kinetoplastid parasites such as the African trypanosome (Trypanosoma brucei), the American trypanosome (Trypanosoma cruzi), and Leishmania spp. contain glutathione and thioredoxin, they appear to lack the respective oxidoreductases and rely on the unique trypanothione/trypanothione reductase system to maintain these intermediates in the reduced state (6,7). Trypanothione (N 1 ,N 8 -bis(glutathionyl)spermidine) (8) also provides reducing equivalents for two distinct trypanothionedependent peroxidase systems (9 -12), the ascorbate peroxidase (13) and ribonucleotide reductase (14) systems. Trypanothione reductase is essential for parasite survival and virulence (15)(16)(17) and shows pronounced differences in substrate specificity compared with glutathione reductase (18). These characteristics make trypanothione reductase an attractive target for drug design (19,20) and suggest that the biosynthesis of the substrate trypanothione would be an additional attractive target for intervention (21,22).
The biosynthesis of trypanothione represents the convergence of two distinct metabolic pathways (glutathione and polyamine biosynthesis) (23). Although all trypanosomatids studied so far are able to synthesize glutathione de novo (24,25), there are marked differences in polyamine metabolism between T. cruzi and other species (19,26). Notably, T. brucei and Leishmania spp. can synthesize polyamines de novo, whereas T. cruzi epimastigotes cannot (24). In contrast, T. cruzi relies on inducible high affinity diamine transporters to salvage exogenous putrescine and cadaverine (27), which can be converted to spermidine and aminopropylcadaverine, respectively (24). T. cruzi can further convert these metabolites into spermine and bis(aminopropyl)cadaverine, respectively (24), which is unusual for a trypanosomatid.
Therefore, unlike C. fasciculata and E. coli glutathionylspermidine synthetases, TcTryS catalyzes all biosynthetic steps from glutathione and spermidine to trypanothione and all the subsequent reverse hydrolytic steps. An enzyme with wider polyamine substrate specificity could be more favorable to an organism lacking de novo diamine synthesis. Our previous HPLC studies have identified other unknown thiol species in epimastigotes (24,32). In this study, we have identified and characterized several novel trypanothione analogues derived from spermine or other physiological polyamines. We show that TcTryS is able to catalyze both forward and reverse reactions to and from these metabolites. The physiological relevance of these novel thiols in all three life cycle stages of this parasite is discussed.
Metabolite Determination-Intracellular thiols were derivatized with the fluorescent label monobromobimane (34) and analyzed by HPLC as described (35,36). Authentic thiol standards were prepared by reduction of disulfides either enzymatically with trypanothione reductase and required cofactors (24) or chemically with tris(carboxyethyl)phosphine, followed by derivatization with monobromobimane and purification by preparative HPLC. Total intracellular polyamines were extracted, derivatized with dansyl chloride, and analyzed by HPLC (37). Polyamines were quantified with authentic standards using 1,7-diaminoheptane as the internal control.
Purification of Thiol-Bimane Adducts and Mass Spectrometry Analysis-T. cruzi epimastigotes grown for 5 days in RTH/CS were supplemented with polyamines and harvested 1 day later. Approximately 8 ϫ 10 8 cells were derivatized with excess monobromobimane, assuming that 10 8 cells contained 15 nmol of total low molecular mass thiols (38). Thiols were separated on an UltraTechsphere reverse-phase C 18 column (250 ϫ 10 mm; HPLC Technology Ltd.) using the gradient specified (34), but at a flow rate of 4 ml/min. Required fractions (1 min) were collected, dried down by rotary evaporation, and resuspended in 100 l of 40 mM Li-HEPPS and 4 mM DTPA (pH 8.0). Thiols were desalted using a methanol/water/trifluoroacetic acid gradient (39) on an Ultra-Techsphere reverse-phase ion-paired C 18 column (250 ϫ 4.6 mm; HPLC Technology Ltd). Required fractions (0.5 or 1.0 min) were collected, dried down by rotary evaporation, and redissolved in acetonitrile/water to give a concentration of 50 -100 pmol/l thiol. MALDI-TOF mass spectrometry analysis was carried out on a Voyager DE-STR instrument (PerSeptive Biosystems) equipped with a nitrogen laser using ␣-cyano-4-hydroxycinnaminic acid as the matrix. Spectra were obtained from 1-5 l of sample, each averaging 40 -100 laser shots.
Kinetic Studies-Apparent Michaelis constants (K m ) for polyamine analogues were determined spectrophotometrically using purified recombinant TcTryS (29). TcTryS amidase was determined by HPLC (29). The kinetic parameters of bis(glutathionyl)spermine disulfide with recombinant trypanothione reductase from T. cruzi were measured in 40 mM HEPES (pH 7.5), 1 mM EDTA, and 100 M NADPH (40). Kinetic data were fitted to the Michaelis-Menten equation by nonlinear regression with the program GraFit.

Identification of Novel Glutathionyl-Polyamine Conjugates
in Epimastigotes-In our previous study (32), we noted trace amounts of an unknown thiol in cultures supplemented with spermine. To obtain sufficient amounts for analysis, cells were grown in RTH/CS (which lacks amine oxidase activity) and supplemented with 20 M spermine at late exponential phase. Thiol analysis revealed two unknown peaks (peaks U1 and U2) (Fig. 1A), which could be abolished by N-ethylmaleimide treatment prior to derivatization with monobromobimane (data not shown). Peak U1, which corresponds to the unidentified thiolcontaining peak noted in our previous study (32), was purified by preparative HPLC, desalted, and concentrated for analysis by MALDI-TOF mass spectrometry (Fig. 1B). Peak U1 gave three protonated parent ion peaks (m/z 1161.44, 971.38, and 779.29), which correspond to the predicted monoisotopic masses of bis(glutathionyl)spermine-(bimane) 2 (M ϩ H ϩ ) (Fig.  1C), bis(glutathionyl)spermine-bimane (M-bimane ϩ H ϩ ), and bis(glutathionyl)spermine (M-(bimane) 2 ϩ H ϩ ), respectively. In addition to the parent ion peaks, clusters of daughter peaks corresponding to the NH 4 ϩ , Na ϩ , and K ϩ ions of each parent ion were present (Fig. 1B). A variation of Ͻ0.2% was observed between predicted and measured monoisotopic masses. To confirm these assignments, an authentic sample of bis(glutathionyl)spermine was reduced and derivatized with monobromobimane and purified by HPLC. The authentic monobromobimane-derivatized bis(glutathionyl)spermine gave a single fluorescent product with an HPLC retention time and MALDI-TOF spectrum identical to those of peak U1 (data not shown). Thus, the three parent ions observed in peak U1 can be attributed to chemical decomposition of the bisbimane adduct during laser desorption, rather than impurities. Insufficient amounts of purified peak U2 could be obtained from derivatized cell extracts. However, because peak U2 (Fig. 1A) was present only in epimastigotes supplemented with spermine, it was tentatively assigned as monoglutathionylspermine. This assignment was subsequently confirmed by enzymatic synthesis as described below. Radiolabeling studies with [ 14 C]spermine (10 M final concentration, 4.96 kBq/ml) revealed that a small proportion (1.5%) of the total radioactivity was incorporated into a peak that coeluted with authentic bis(glutathionyl)spermine disulfide and that was sensitive to performic acid oxidation or acid hydrolysis (data not shown). The remainder of the radiolabel coeluted with spermine (96%), spermidine (0.8%), and a third unknown compound (0.4%) that was stable against performic acid oxidation or acid hydrolysis.
Because epimastigotes are natural auxotrophs for poly-amines, we investigated whether T. cruzi would also conjugate glutathione with naturally occurring acetylated polyamines (N-acetylputrescine, N 1 -acetylspermidine, N 8 -acetylspermidine, and N-acetylspermine). Thiol analysis of epimastigote cultures supplemented with the above-mentioned polyamines showed that only the latter two gave prominent novel fluorescent peaks (peak U3 in Fig. 2A and peak U4 in Fig. 3A).
Cultures supplemented with N 1 -acetylspermidine also contained a minor peak that coeluted with peak U3 (2% of peak U3 in N 8 -acetylspermidine cultures). To isolate sufficient material for structural analysis, epimastigote cultures were supplemented with N 8 -acetylspermidine or N-acetylspermine and the bimane derivatives purified as described above.
Purified peak U3 gave peaks corresponding to the predicted masses of protonated glutathionyl-N 8 -acetylspermidine-bimane and glutathionyl-N 8 -acetylspermidine (m/z 667.43 and 477.33, respectively; peak U3) (Fig. 2B). The two peaks at m/z 689.43 and 705.40 correspond to the Na ϩ and K ϩ ions of M ϩ H ϩ , respectively. The loss of a bimane and sulfur from M ϩ H ϩ may give a peak at m/z 445.14 (predicted m/z 444.27). Loss of the acetyl moiety would be consistent with the fragment at m/z 624.21. Two other minor peaks (m/z 640.20 and 656.17) could not be unambiguously assigned, although they could represent the sulfoxide and sulfone derivatives of m/z 624.21; the equivalent peaks for the parent ion at m/z 667.43 are not visible.
The retention time of peak U4 is similar to that of trypanothione (Fig. 3A). However, polyamine analysis of epimastigotes grown with N-acetylspermine as supplement in RTH/FCS or RTH/CS showed no significant increase in total spermidine content (data not shown); and therefore, peak U4 could not be trypanothione. MALDI-TOF analysis of purified peak U4 gave a peak at m/z 724.38, which corresponds exactly with the predicted mass of protonated glutathionyl-N-acetylspermine-bimane (peak U4) (Fig. 3, B and C). Peaks at m/z 746.35 and 762.33 correspond to the Na ϩ and K ϩ ions of M ϩ H ϩ , respectively. A minor unassigned peak appeared at m/z 768. 33. Peaks corresponding to the predicted masses of the protonated and sodium ions of glutathionyl-N-acetylspermine (M-bimane) appeared at m/z 534.29 and 556.29, respectively. Another peak at m/z 549.29 (with an associated Na ϩ ion at m/z 571.26) may represent the sulfoxide of m/z 534. 29. Some minor peaks in the region of m/z 600 -700 could not be assigned (Ͻ8% of the intensity of M ϩ H ϩ ) and may be low level impurities. No peaks corresponding to the protonated ions of trypanothione-(bi- Thiol and Polyamine Content of the Different Life Cycle Stages of T. cruzi-The above experiments indicate that the insect (epimastigote) form is able to take up spermidine and spermine and their mono-N-acetyl derivatives and to conjugate them with glutathione. Because amastigotes reside in the cytosol of mammalian cells and are presumed to be incapable of de novo polyamine biosynthesis, we decided to compare the thiol and polyamine content of each of the life cycle stages (Table I).
Epimastigotes grown to late exponential phase in RTH/FCS containing limiting amounts of polyamines (0.22 M putrescine and 0.63 M spermidine) (32) have a markedly lower polyamine content than amastigotes cultured in the polyamine-rich environment of human foreskin fibroblasts (5 M putrescine, 11 M diaminopropane, 700 M spermidine, and 1050 M spermine; calculated assuming that the fibroblast cell volume is 1.78 l/10 6 cells) (41). The total spermidine and spermine content was 12-and 3-fold lower in epimastigotes compared with amastigotes, respectively (Table I). Non-dividing trypomastigotes that transformed from amastigotes and escaped into the culture medium contained similar amounts of spermidine, but 4-fold less spermine. Total polyamine content was highest in amastigotes (5.9 nmol/10 8 cells), followed by trypomastigotes (3.6 nmol/10 8 cells) and epimastigotes (1.1 nmol/10 8 cells). The most abundant thiols in all three life cycle stages were cysteine, glutathione, and trypanothione. Ovothiol A (N 1 -methyl-4-mercaptohistidine) was reduced 5-fold in amastigotes compared with amastigotes and trypomastigotes. This is similar to the situation in Leishmania donovani, where ovothiol A is reduced from 35% of total SH groups in late exponential phase promastigotes (insect stage) to Ͻ3% in amastigotes (intracellular mammalian stage) (38). Both mammalian stages contained trace amounts of monoglutathionylacetylspermidine, presumably the N 1 -glutathionyl-N 8 -acetylspermidine isomer. However, despite containing levels of spermine equivalent to those in epimastigotes, amastigotes and trypomastigotes had levels of glutathionylspermine conjugates that were at or below the limits of detection. Total thiol content (SH groups) was highest in epimastigotes (3.0 nmol/10 8 cells), with similar amounts in amastigotes and trypomastigotes (1.4 and 1.5 nmol/10 8 cells, respectively).
The proportion of free GSH in epimastigotes reflects the availability of diamines and polyamines in the medium (24,32,38). To determine whether spermine exerts a similar effect, epimastigotes were cultured in RTH/FCS supplemented with 10 M spermine or with a mixture of 5 M spermine plus spermidine (Fig. 4). Without supplements, the majority of the total glutathione-containing peptide was present as free GSH (80%), with the remainder conjugated to spermidine and spermine (15 and 5%, respectively). Addition of spermine to the medium significantly increased the total spermine content 17fold from 0.8 Ϯ 0.2 to 14 Ϯ 2 nmol/10 8 cells, with a smaller 3-fold increase in total spermidine content from 0.29 Ϯ 0.08 to 0.87 Ϯ 0.17 nmol/10 8 cells (Fig. 4A). Spermine supplementation reduced the free glutathione content, similar to the effects of spermidine supplementation reported previously (32). However, despite the considerable excess of spermine in the parasites, spermidine-containing conjugate levels (33%, mainly as trypanothione) were higher than spermine-containing conjugate levels (23%, mainly as bis(glutathionyl)spermine) (Fig.  4B). When spermidine and spermine were present in equal amounts, the intracellular concentrations of both polyamines were similar (5.3 Ϯ 0.7 and 5.7 Ϯ 0.7 nmol/10 8 cells, respectively). Despite these equivalent intracellular polyamine concentrations, epimastigotes displayed a pronounced preference to synthesize trypanothione (2.2 Ϯ 0.5 nmol/10 8 cells, 88% of total GSH pool) over bis(glutathionyl)spermine (0.04 Ϯ 0.01 nmol/10 8 cells). This is in contrast to epimastigotes grown in the presence of equal amounts of putrescine (or spermidine) and cadaverine (or aminopropylcadaverine), which synthesize trypanothione and homotrypanothione in equal amounts (24,32). Likewise, trypomastigotes and amastigotes purified from fibroblasts principally conjugated glutathione with spermidine (66 and 43%, respectively) rather than with spermine (Ͻ1%) (Fig. 4). Addition of 50 M spermine and 0.5 mM aminoguanidine (an inhibitor of amine oxidases) to cultures of infected fibroblasts increased glutathionylspermine conjugates in amastigotes only to 4% of the total. Silvio X10-7 epimastigotes also synthesized bis(glutathionyl)spermine and other glutathionyl-polyamine conjugates in equivalent amounts to Y 0 epimastigotes (data not shown), indicating that representatives of both major lineages of T. cruzi (42) have a similar metabolic capability to take up polyamines and to conjugate them with glutathione.
Substrate Specificity of Trypanothione Synthetase-In T. cruzi, a single enzyme (TcTryS) catalyzes ATP-dependent synthesis of trypanothione and homotrypanothione from glutathione and the appropriate polyamine (29), suggesting that spermine and other analogues may also be substrates. Because the enzyme shows pronounced substrate inhibition with GSH, initial rates were determined with 1 mM GSH, saturating MgATP, and varying concentrations of polyamine. Spermine and all of the analogues tested displayed hyperbolic kinetics and were fitted to the Michaelis-Menten equation to obtain kinetic parameters for apparent K m and k cat (Table II). Comparison of the specificity constants (k cat /K m ) indicates that spermine, N-acetylspermine, and N 8 -acetylspermidine are all efficient substrates for the enzyme compared with spermidine, whereas N 1 -acetylspermidine, putrescine, and diaminopropane are markedly less so.
The reaction products of these enzyme assays were derivatized with monobromobimane and analyzed by HPLC. With N 8or N 1 -acetylspermidine as substrate, a single peak with an elution time identical to that of peak U3 was obtained (see Fig.  2). Similarly, incubation with N-acetylspermine yielded peak U4 (see Fig. 3). However, two products were obtained with spermine as substrate corresponding to peaks U1 and U2 (see Fig. 1). Peak U1 corresponds to bis(glutathionyl)spermine; and thus, peak U2 must be monoglutathionylspermine. When a time course of the reaction was analyzed, the rate of formation of monoglutathionylspermine exceeded that of the bis adduct for the first 20 min until a steady-state concentration (ϳ160

FIG. 4. Free and conjugated glutathione and polyamine content in each life cycle stage of T. cruzi.
Parasites (comprising Ͼ90% of each form) were grown in RTH/FCS, harvested, and analyzed for thiol and polyamine content as described under "Experimental Procedures." A, total glutathione and polyamine content. Total glutathione was calculated as free glutathione ϩ glutathionylspermidine ϩ glutathionylspermine ϩ glutathionyl-N 8 -acetylspermidine ϩ 2ϫ trypanothione ϩ 2ϫ bis(glutathionyl)spermine. Total spermidine was calculated as free spermidine ϩ glutathionylspermidine ϩ glutathionyl-N 8 -acetylspermidine ϩ trypanothione. Total spermine was calculated as free spermine ϩ glutathionylspermine ϩ bis(glutathionyl)spermine. B, free glutathione and glutathionyl-polyamine conjugates. Conjugated spermidine was calculated as glutathionylspermidine ϩ glutathionyl-N 8 -acetylspermidine ϩ trypanothione. Conjugated spermine was calculated as glutathionylspermine ϩ bis(glutathionyl)spermine. All values are means Ϯ S.E. of at least three separate cultures. spn, spermine; spd, spermidine. M) was achieved (Fig. 5A). Thereafter, formation of bis(glutathionyl)spermine continued until GSH was exhausted after 2 h. Essentially identical results were obtained with glutathione and spermidine, where the intermediate product, glutathionylspermidine, rose to a similar steady-state level and preceded formation of trypanothione over the first 20 min (Fig. 5B). When presented with an equimolar mixture of spermidine and spermine (5 mM each), formation of spermine conjugates was 3.4 Ϯ 0.3-fold faster than that of spermidine conjugates, consistent with the kinetic data presented in Table II. A second experiment with TcTryS pretreated with iodoacetamide to inactivate the amidase activity gave a similar ratio (3.2 Ϯ 0.2fold), indicating that the amidase activity does not affect partitioning into the end products.
Amidase Activity of Trypanothione Synthetase-The finding that TcTryS preferred spermine to spermidine as substrate was unexpected because the formation of bis(glutathionyl)spermine was not favored compared with trypanothione in cells cultured with spermine (Fig. 4). However, in addition to its biosynthetic activity, TcTryS also displays weak amidase activity, hydrolyzing trypanothione and homotrypanothione to glutathione and the respective free polyamine via the intermediate mono adducts (29). To investigate whether bis(glutathionyl)spermine is a better substrate than trypanothione, 0.5 mM bis(glutathionyl)spermine was incubated with TcTryS in the absence of other substrates, and the products were analyzed by HPLC. Formation of glutathionylspermine and glutathione was linear over the first 10 min (data not shown), with a specific activity of 13.3 Ϯ 2.1 nmol of amide bond hydrolyzed per min/mg, which is four times faster than with the equivalent concentration of trypanothione (3.1 Ϯ 0.8 nmol of amide bond hydrolyzed/min/mg) (29). Because of the low activity of the reverse reaction (Ͻ1% of the forward synthetic reaction), it was not possible to perform a full kinetic analysis.
Bis(glutathionyl)spermine Disulfide as Substrate for Trypanothione Reductase-Bis(glutathionyl)spermine was present as the free thiol in cells supplemented with spermine ( Fig. 1), indicating that it must be maintained as the dithiol form either through thiol disulfide exchange with trypanothione or by direct reduction by trypanothione reductase. Kinetic analysis with recombinant T. cruzi trypanothione reductase yielded an apparent K m of 34 Ϯ 10 M and a k cat of 403 Ϯ 47 s Ϫ1 for bis(glutathionyl)spermine disulfide in the presence of saturating NADPH. The specificity constant for bis(glutathionyl)spermine disulfide (k cat /K m ϭ 12.0 ϫ 10 6 M Ϫ1 s Ϫ1 ) compares favorably with previously determined values for trypanothione disulfide (5.5 ϫ 10 6 M Ϫ1 s Ϫ1 ), homotrypanothione disulfide (4.1 ϫ 10 6 M Ϫ1 s Ϫ1 ), and glutathionylspermidine disulfide (1.9 ϫ 10 6 M Ϫ1 s Ϫ1 ) (24,40), indicating that this metabolite is a physiological substrate in vivo.

DISCUSSION
Our current findings demonstrate that T. cruzi is capable of taking up any of the physiological polyamines found in mammalian cells (spermidine, spermine, N-acetylspermine, and N 1and N 8 -acetylspermidine) and conjugating them with glutathione to form the mono and/or bis adducts. From these and previous studies (24), the physiological order of preference in vivo is spermidine ϭ aminopropylcadaverine Ͼ spermine/Nacetylspermine/N 8 -acetylspermidine Ͼ N 1 -acetylspermidine/ diamines. Although TcTryS can catalyze these biosynthetic reactions in vitro, it remains to be established whether this is the sole metabolic route to glutathionyl-polyamine conjugates in vivo.
Earlier studies on the polyamine substrate specificities of E. coli and C. fasciculata glutathionylspermidine synthetases have shown that the 3-aminopropyl moiety is the preferred site for acylation by glutathione and that the N-4 secondary amine of spermidine is a crucial recognition determinant (21,30). TcTryS behaves like these previously characterized enzymes in that spermine, N-acetylspermine, and N 8 -acetylspermidine are all efficient substrates compared with spermidine, whereas N 1 -acetylspermidine, putrescine, and diaminopropane are markedly less so. Thus, a spacing of three methylene groups between the free amino group and the secondary amine (in polyamines) or the other primary amine (in diamines) is preferred over four (cf. N 8 -with N 1 -acetylspermidine or diaminopropane with putrescine in Table II). Spermine, which has two functionally equivalent aminopropyl moieties, has a 5-fold higher specificity constant than spermidine and N 8 -acetylspermidine, which have only one.
The finding that TcTryS preferred spermine rather than spermidine as substrate was somewhat unexpected. Epimastigotes contained 50-fold less bis(glutathionyl)spermine than trypanothione when cultured with spermine and spermidine, even though the total intracellular concentrations of spermine and spermidine were equivalent. Likewise, amastigotes, which have similar levels of these polyamines, contained Ͼ60-fold less bis(glutathionyl)spermine than trypanothione. Yet, based on the ratios of specificity constants for TcTryS, the concentration of bis(glutathionyl)spermine would be predicted to be 5-fold greater than that of trypanothione.
Several factors could account for this discrepancy. First, the free cytosolic concentration of spermine could be considerably less than that of spermidine (this would be predicted to be ϳ250-fold to account for the above observations). Unfortunately, there are no accurate measurements of free polyamine concentrations in any organism. However, the best available indirect estimate is that 95-99% of spermine and 85-92% of spermidine are bound to macromolecules (DNA, RNA, and phospholipids) and ATP in bovine lymphocytes and rat liver cells (43). Assuming that 10 8 amastigotes have an intracellular volume of 2 l, intracellular spermine and spermidine concentrations are ϳ1.1 and 1.8 mM, respectively. These values are comparable to spermine and spermidine concentrations in the above-mentioned mammalian cell models (43). However, the ratio of free spermidine to spermine in mammalian cells is only 2.6 for bovine lymphocytes and 4.6 for rat liver, which, if applicable to T. cruzi, would result in formation of approximately equimolar amounts of trypanothione and bis(glutathionyl)spermidine. Because this was not observed, either the model does not apply to trypanosomes, or additional selective compartmentalization of spermine into other unique organelles such as the mitochondrion (binding to kinetoplast DNA) or the acidocalcisome (binding to polyphosphate) would have to occur.
A second possibility is that the amidase function of TcTryS could selectively hydrolyze bis(glutathionyl)spermine. However, this was only 4-fold faster than with trypanothione. In addition, the forward synthetic reaction conjugating glutathione with spermine was 2 orders of magnitude greater than the rate of hydrolysis of bis(glutathionyl)spermine. Therefore, the rate of hydrolysis is negligible, in agreement with the time course of product formation shown in Fig. 5. Significantly, inhibition of the amidase activity did not affect the relative incorporation of spermine and spermidine into products, effectively excluding this possibility.
A third explanation could be that rapid oxidation of bis(glutathionyl)spermine occurs in vivo, so it is not detected by the monobromobimane HPLC assay. This possibility is highly unlikely because our kinetic analyses show that bis(glutathionyl-)spermine disulfide is a physiological substrate for the enzyme. Moreover, quantification of amastigote thiols derivatized in the presence or absence of the reducing agent tris(carboxyethyl)phosphine gave identical values (data not shown), indicating that the majority is present as the free dithiol in the cell.
Other possibilities include selective feedback inhibition of the native enzyme in vivo, selective degradation of glutathionylspermine conjugates by an unknown pathway, and selective export of bis(glutathionyl)spermine from the cell. Further experiments are required to examine these possibilities.
Despite the failure to detect bis(glutathionyl)spermine in amastigotes, measurable amounts of a thiol eluting with the retention time of glutathionylacetylspermidine were present. The resolution of our analytical method is insufficient to resolve which isomer was formed, although the specificity of TcTryS would favor glutathionylation of N 8 -acetylspermidine. The concentration of acetylspermidine in uninfected fibroblasts was below the limits of detection (Ͻ3 pmol/10 6 cells or Ͻ1.5 M), i.e. ϳ1000-fold less than spermine, which raises the interesting question of how the amastigote obtains acetylspermidine. Under certain stress conditions, mammalian cells can activate polyamine catabolism through the induction of spermine/spermidine N 1 -acetyltransferase, forming physiologically inert acetylated polyamines that are either excreted or oxidized by polyamine oxidase isoenzymes, producing H 2 O 2 , 3-acetamidopropanal, and putrescine or spermidine, depending on the initial substrate (44,45). Activation of this pathway in response to infection does not appear to have been studied. Mammalian cells can also reversibly form N 8 -acetylspermidine, and inhibition of N 8 -acetylspermidine deacetylase has been reported to promote either growth or differentiation, depending on the cell type (46,47). Whether T. cruzi manipulates host cell polyamine levels or responds to changes in polyamine metabolism induced by the host (48) is an area worthy of further investigation.
Even though C. fasciculata glutathionylspermidine synthetase is able to utilize spermine as substrate (21), C. fasciculata does not synthesize spermine, and bis(glutathionyl)spermine is absent in wild-type cells (34). L. donovani promastigotes grown in RTH/FCS (under conditions identical to those used for T. cruzi epimastigotes) supplemented with 10 M spermine did not contain any detectable spermine or bis(glutathionyl-)spermine (Ͻ0.02 and Ͻ0.002 nmol/10 8 cells, respectively). Presumably, the lack of synthesis/uptake of spermine in other trypanosomatids (35,49,50) prevents the biosynthesis of glutathionylspermine conjugates, although additional controls on the biosynthetic enzymes may exist. T. cruzi therefore appears to be unique among trypanosomatids in its ability to take up/synthesize spermine (24,32) and to conjugate it with glutathione. The wide polyamine substrate specificity of Tc-TryS presents an attractive target that can be exploited in the design of polyamine-based compounds that can inhibit polyamine and trypanothione metabolism (51-53). •, bis(glutathionyl)spermine; Ⅺ, total thiol groups contained in end products; f, total glutathione recovered from assay (free glutathione ϩ monoglutathionylspermine ϩ 2ϫ bis(glutathionyl)spermine). Values were calculated using known amounts of synthetic standards apart from monoglutathionylspermine, which was calculated assuming that the fluorescence response was the same as that of N 1 -glutathionylspermidine. B, time course of formation of monoglutathionylspermidine and trypanothione from glutathione and spermidine. E, monoglutathionylspermidine; •, trypanothione; Ⅺ, total thiol groups contained in end products; f, total glutathione recovered from assay (free glutathione ϩ monoglutathionylspermidine ϩ 2ϫ trypanothione).