The Biosynthetic Incorporation of the Intact Leucine Skeleton into Sterol by the Trypanosomatid Leishmania mexicana *

The amino acid leucine is efficiently used by the trypanosomatid Leishmania mexicana for sterol biosynthesis. The incubation of [2-13C]leucine withL. mexicana promastigotes in the presence of ketoconazole gave 14α-methylergosta-8,24(241)-3β-ol as the major sterol, which was shown by mass spectrometry to contain up to six atoms of 13C per molecule. 13C NMR analysis of the 14α-methylergosta-8,24(241)-3β-ol revealed that it was labeled in only six positions: C-2, C-6, C-11, C-12, C-16, and C-23. This established that the leucine skeleton is incorporated intact into the isoprenoid pathway leading to sterol; it is not converted first to acetyl-CoA, as in animals and plants, with utilization of the acetyl-CoA to regenerate 3-hydroxy-3-methylglutaryl-CoA (HMG-CoA). An inhibitor of HMG-CoA synthase (L-659,699) blocked the incorporation of [1-14C]acetate into sterol but had no inhibitory effect on [U-14C]leucine incorporation. The HMG-CoA reductase inhibitor lovastatin inhibited promastigote growth and [U-14C]leucine incorporation into sterol. The addition of unlabeled mevalonic acid (MVA) overcame the lovastatin inhibition of growth and also diluted the incorporation of [1-14C]leucine into sterol. These results are compatible with two routes by which the leucine skeleton may enter intact into the isoprenoid pathway. The catabolism of leucine could generate HMG-CoA that is then directly reduced to MVA for incorporation into sterol. Alternatively, a compound produced as an intermediate in leucine breakdown to HMG-CoA (e.g. dimethylcrotonyl-CoA) could be directly reduced to produce an isoprene alcohol followed by phosphorylation to enter the isoprenoid pathway post-MVA.

Parasitic trypanosomatid protozoa of the genus Leishmania cause diseases in tropical and subtropical regions of the world. The treatment of leishmaniasis still relies upon the drugs introduced many years ago (1), which have toxic side effects, and there is now a great need for more effective new chemotherapeutic drugs (1,2). This has prompted the search for new metabolic targets for drugs and has resulted in the recognition of sterol synthesis inhibitors as a potential candidate (2, 3). The importance of an active sterol biosynthetic pathway in trypanosomatids for growth and viability has been demonstrated using antifungal agents that are inhibitors of sterol biosynthesis. Thus, the imadazole-and triazole-based drugs (e.g. ketoconazole and itraconazole), which inhibit the 14␣-methylsterol 14-demethylase, and the allylamines (e.g. terbinafine), which inhibit squalene epoxidase (2- 11), have been shown to block sterol synthesis in a number of Leishmania and Trypanosoma species with retardation of growth and death of the parasite.
In our studies on sterol biosynthesis in Leishmania species, we have recently demonstrated (12,13) that leucine is the major source of the carbon used for de novo sterol biosynthesis. By contrast, acetate or substrates from which acetyl-CoA is generated by metabolism (e.g. glucose, palmitic acid, alanine, serine, and isoleucine) are very poorly incorporated into sterol, although they are used efficiently for the synthesis of the fatty acid moieties of triacylglycerol and phospholipid. The utilization of leucine for sterol biosynthesis has been shown previously in animal tissues (14 -18), plants (19 -21), and fungi (22,23). In animals and plants, a major route for leucine catabolism has been demonstrated to be located in the mitochondrion (24,25). The pathway proceeds through the production of ␣-ketoisocaproate, isovaleryl-CoA, 3-methylcrotonyl-CoA, and 3-methylglutaconyl-CoA (Scheme 1) to give 3-hydroxy-3-methylglutaryl-CoA (HMG-CoA), 1 which is then cleaved by a lyase to produce acetyl CoA and acetoacetate. The acetyl-CoA generated in this way can be either fed into the citric acid cycle or alternatively transported out of the mitochondrion into the cytosol, where the acetyl-CoA may be utilized for the biosynthesis of a range of compounds including fatty acids and isoprenoids such as sterols. The entry of the acetyl-CoA into the isoprenoid pathway requires the regeneration of HMG-CoA, which is then reduced to mevalonic acid (Scheme 1). Conclusive evidence that leucine enters isoprenoids in plants by this indirect route and involving production of acetyl-CoA has been provided by incubation of 13 C-labeled leucine with a callus culture of Andrographis paniculata (19,20). 13 C NMR analysis of the 13 C-enriched sesquiterpenoid and phytosterols (19,20) produced by the callus showed unequivocally that the leucine was metabolized to acetyl-CoA and acetoacetate prior to incorporation into the isoprenoid pathway. We have demonstrated previously (12,13) that [U-14 C]leucine incubated with Leishmania mexicana and other trypanosomatid species was very efficiently incorporated into sterol and to some limited extent into fatty acids. However, by contrast, [1-14 C]acetate was readily incorporated into fatty acids but poorly utilized for sterol production. These observa-tions are incompatible with a route in Leishmania that requires leucine degradation to proceed to acetyl CoA before reutilization for isoprenoid production (19,20). We have therefore undertaken the studies described here to investigate the metabolic route whereby L. mexicana uses leucine for sterol biosynthesis, since this may produce evidence for a target for new antileishmanial drug development. Cell Culture-The strain of Leishmania used in this study was L. mexicana (MNYC/62/BZ/M379). Promastigotes were cultured at 26°C in HO-minimum essential medium (26), supplemented with 10% (v/v) heat-inactivated fetal calf serum. 14 C-and 13 C-labeled substrates were added to the media to give the final concentrations indicated under "Results." Cell density of cultures was determined by counting using a Neubauer hemocytometer. Cultures were normally established with an initial cell density of 10 6 cells/ml; at the termination of the culture period, the cells were harvested by centrifugation as described previously (12,27). Lovastatin was administered to cultures from a stock solution (2 mg/ml) in Me 2 SO; compound L-659,699 was dissolved (2 mg/ml) in sterile H 2 O saturated with NaHCO 3 for administration to the cultures.

MATERIALS AND METHODS
Amastigotes were obtained after the infection of macrophages isolated from the peritoneal cavities of female CD1 mice with stationary phase promastigotes (72 h of growth) (28,29). Promastigotes were allowed to infect macrophages for 24 h at 32°C before the medium overlying the macrophages was decanted. The cells were rinsed with Locke's solution (containing the following per liter: 9 g of NaCl, 0.42 g of KCl, 0.4 g of CaCl 2 ⅐H 2 O, 0.2 g of NaHCO 3 , and 1 g of glucose) to remove any remaining extracellular promastigotes. RPMI supplemented 15% (v/v) with heat-inactivated fetal calf serum and containing [2-13 C]leucine in place of unlabeled free leucine was then added to the infected macrophages, and the cultures were left at 32°C for 48 h. At this time, amastigote forms were prepared by the method described by Haughan et al. (29).
Lipid Extraction and Analyses-Parasite lipids were isolated after extraction with chloroform/methanol (2:1) as described previously (12,27). Radioactive lipid extracts were analyzed by analytical TLC and radioscanning using silica gel TLC plates and chloroform/ethanol (98:2) as the developing solvent. Sterols were isolated and analyzed by GC or GC-MS as the TMS ether derivatives following previously described protocols (12,27). Steryl acetates were prepared by treatment of the free sterol with pyridine/acetic anhydride (1:1) followed by usual work up of the steryl acetate. The steryl acetates were separated by preparative TLC on silica gel impregnated with 10% AgNO 3 and developed with freshly distilled chloroform. Sterols were quantified by capillary GC analysis using 5␣-cholestane as a standard.
NMR Spectroscopy-The NMR spectra were measured on deuteriochloroform solutions using a Varian INOVA 600 spectrometer operating at 599.9 MHz for protons and 150.9 MHz for 13 C nuclei. 1 H NMR spectra were recorded using the following parameters: 7000 Hz spectrum width, 3-s preacquisition delay, 90°pulse, 5-s acquisition time, 10,000 transients, 64K data points No weighting function was used before Fourier transformation. Broad band proton-decoupled 13 C NMR spectra were obtained using the following parameters: 35,000-Hz spectrum width, 22°pulse, 0.936-s acquisition time, 70,000 transients, 64K data points, line broadening of 1 Hz before Fourier transformation.
N14␣-methylergosta-8,24(24 1 )-dien-3␤-ol was then purified from the lipid by reversed-phase HPLC using an Econosphere C 18 column (250 ϫ 4.6 mm; inner diameter, 5 mm; supplied by Alltech) to separate it from cholesterol and other minor sterols. Compounds were eluted isocratically using acetonitrile/water (9:1), and sterols were detected by UV absorbance at 215 nm. The 14␣-methylergosta-8,24(24 1 )-dien-3␤-ol was eluted at 26 -30 min, and cholesterol was eluted at 32-36 min. The solvent volume of the eluate was carefully reduced by rotary evaporation, and the sterol was extracted into petroleum ether before taking to dryness and storage at Ϫ20°C. The purity of the isolated 13 C-labeled 14␣-methylergosta-8,24(24 1 )-dien-3␤-ol (ϳ3 mg) was checked by GC-MS and then analyzed by 13 13 C]leucine to investigate if the leucine carbon skeleton is incorporated by L. mexicana directly into the sterols by reduction of HMG-CoA to MVA. The major sterols of L. mexicana are cholesterol obtained from the medium, and the biosynthesized ergosta-5,7,24(24 1 )trien-3␤-ol (4,12,13) with smaller variable amounts of ergosta-5,7,22-trien-3␤-ol, stigmasta-5,7,24(24 1 )-trien-3␤-ol, and precursors such as ergosta-7,24(24 1 )-dien-3␤-ol. A sterol mixture of this nature presents certain problems for the type of 13 C NMR study envisaged. First, the complexity of the sterol mixture demands careful purification of one of the biosynthesized ergosta types of sterol so that signal assignments to 13 C-enriched carbons can be made with accuracy and without ambiguity. Second, the major sterols of L. mexicana are ⌬ 5,7 -compounds that are notoriously unstable in small amounts due to oxidation, and this may cause difficulties in the purification, storage, and NMR analysis of these sterols. This problem has been considered in detail by Schroepfer et al. (30), specifically in relation to the NMR analysis of ⌬ 5,7 -sterols. Finally, the success of the study depends upon the extent of enrichment of the biosynthesized sterol with 13 C. A large pool of preexisting sterol from the inoculum will dilute the 13 C-enriched sterol species and could make 13 C-enriched carbons difficult to detect. Accordingly, we looked for a new approach to the problem and decided to use an inhibitor of sterol biosynthesis. An inhibitor was required that would cause the accumulation of a large amount of a relatively stable sterol intermediate that would normally occur in only trace amounts in the parasite, so dilution of the newly synthesized [ 13 C]sterol by preexisting material would not be significant. Sterol biosynthesis inhibitors suited to this purpose are the imadazole and triazole types of antifungal drugs. These compounds block the action of the cytochrome P450-dependent 14␣-methylsterol 14-demethylase with the result that the normal sterols are depleted and one or more 14␣-methylsterols accumulate, often in large amounts (31). It has been demonstrated previously that antifungal imi-dazoles and triazoles also inhibit the 14␣-demethylation step in sterol biosynthesis in several Leishmania species with the resulting appearance of 4␣,14␣-dimethylergosta-8,24(24 1 )-dien-3␤-ol and 14␣-methylergosta-8,24(24 1 )-dien-3␤-ol in appreciable amounts (4,6). These 14␣-methylsterols are considerably more stable during isolation, storage, and NMR analysis than are ⌬ 5,7 -sterols. Consequently, we decided that using the 14demethylase inhibitor ketoconazole offered the best opportunity for the isolation of a pure 13 C-enriched sterol undiluted by preexisting endogenous sterol, which was required for the 13 C NMR analysis to determine the route of incorporation of leucine into the isoprenoid pathway.
Preliminary experiments were first undertaken to determine the optimum conditions for incubation of the L. mexicana promastigotes with ketoconazole to accumulate a 14␣-methylsterol in sufficient amount for isolation and 13 C NMR analysis. The incubation of promastigotes of L. mexicana with ketoconazole (0.1 and 1.0 g/ml) for 72 h followed by isolation and GC-MS examination of the sterols showed that, as anticipated, the ergosta-5,7,24(24 1 )-trien-3␤-ol found in the control was replaced by 14␣-methylergosta-8,24(24 1 )-dien-3␤-ol in the ketoconazole-treated cultures (Table I). Cholesterol taken up from the medium was present in both the control and treated cells. The 14␣-methylergosta-8,24(24 1 )-dien-3␤-ol was identified by the mass spectrum of the TMS ether and by the 1 H NMR spectrum (6,32). Ketoconazole at 0.1 g/ml retarded growth by only about 5% compared with the control, and the total sterol content of the treated cells was ϳ70% of the control value. At the higher ketoconazole concentration (1.0 g/ml), the growth was around 75% of the control, and the total sterol was about 50% of the control. In a further experiment, the L. mexicana promastigotes were incubated with ketoconazole (0.1 g/ml) and [U-14 C]leucine (2.6 Ci) for 72 h to ensure that the accumulating 14␣-methylsterol was being biosynthesized from leucine derived from the medium rather than from some internal source of unlabeled precursor(s). The lipids were extracted and found to contain 4.2% of the radioactivity added to the culture medium, while analytical TLC with radioscanning showed that the 14␣-methylsterol was the major labeled material. Recovery of the labeled sterol from the TLC plate, acetylation, and rechromatography by TLC on silver nitrate-impregnated silica The effects of ketoconazole on the growth and sterol composition of promastigotes of L. mexicana Promastigotes of L. mexicana (1.0 ϫ 10 6 cells/ml) were cultured for 72 h alone (control) or in the presence of 0.1 or 1.0 g/ml ketoconazole. The cell density was determined at the end of the growth period, and the cells were harvested, the lipid was extracted, and the sterols were isolated and analyzed by GC-MS as described under "Materials and Methods"; tr indicates trace amount (Ͻ0.5%). gel showed that the radioactivity accompanied a material with the same R f as 14␣-methylergosta-8,24(24 1 )-dien-3␤-yl acetate.
The above experiments showed that our approach to obtain a pure sterol for the 13 C NMR analysis was feasible. Therefore, multiple cultures (50 ϫ 50 ml) of L. mexicana promastigotes were grown in HO-minimum essential medium (plus 10% (v/v) heatinactivated fetal calf serum) in which the free (i.e. nonprotein) leucine was replaced with [2-13 C]leucine, and 0.1 g/ml ketoconazole was added. The cells were cultured for 72 h and harvested, and the lipid was extracted. The total sterol was isolated from the lipid, and the 14␣-methylergosta-8,24(24 1 )-dien-3␤-ol was then separated from the cholesterol and other minor sterols by HPLC (see "Materials and Methods"). The cholesterol was shown by GC-MS analysis to contain only the natural abundance of 13 C, and there was no detectable labeling from the [2- 13 C]leucine. This observation established unequivocally that the cholesterol in L. mexicana must be taken up from the medium and that it is not the product of de novo synthesis in the parasite. The purity of the isolated 14␣-methylergosta-8,24(24 1 )-dien-3␤-ol (ϳ3 mg) was 95% as judged by GC analysis. The mass spectrum of the TMS ether showed clusters of ions for the molecular and fragment ions arising from several labeled species of the sterol containing from one to six 13 C atoms (Fig. 1). Ions due to unlabeled sterol were very minor, showing that there was excellent incorporation of [2-13 C]leucine into the sterol in accord with our previous studies with [U-14 C]leucine incorporation, which had revealed that at least 80% of the sterol carbon originated from leucine (12,13). The molecular ion region comprised a cluster of ions at m/z 484 (unlabeled sterol), 485, 486, 487, 488, 489, and 490, with the last two predominating (Table II). The fragment ion clusters at m/z 469-475 and 379-385, which arise by loss of a methyl and methyl and TMSOH from the [M] ϩ ion, respectively, showed a similar distribution of molecular species containing 1-6 13 C-enriched positions (Fig. 1). The ions at m/z 304-307 showed fragments containing up to five 13  The 13 C labeling patterns predicted for sterol derived from [2-13 C]leucine by pathways involving either HMG-CoA breakdown to the acetyl-CoA level or directly by reduction of HMG-CoA to MVA are shown in Scheme 2. If the incorporation proceeds indirectly through the intermediacy of acetyl-CoA, the sterol will have 12 13 C-enriched positions (C-2, C-4, C-6, C-8, C-10, C-11, C-12, C-14, C-16, C-20, C-23, and C-25). However, if incorporation results from direct conversion of HMG-CoA to MVA, then only six positions will be enriched (C-2, C-6, C-11, C-12, C-16, and C-23). Clearly, the mass spectral data revealing labeled sterol species containing up to six 13 C atoms pointed to the latter labeling pattern. Accordingly, to determine the exact number of 13 C atoms and their locations in the molecule, the labeled 14␣-methylergosta-8,24(24 1 )-dien-3␤-ol was purified by preparative HPLC and examined by 13 C NMR spectroscopy (Fig. 2). The spectrum showed a very high enrichment of the compound with 13 C and displayed six strong signals indicating the positions specifically labeled from the [ 13 C]leucine. The assignments of these carbon signals were made by comparison with the reported 13 C NMR spectra of other sterols (32).
The singlet signals at ␦ 25.2, 31.2, and 31.5 ppm were readily assigned to C-6, C-2, and C-23, respectively. The three signals centered at ␦ 21.7 (Fig. 2, inset A) comprised a singlet due to C-11 in molecules with no 13 C enrichment at the adjacent positions (C-9 and C-12) and a doublet arising from coupling with C-12 in molecular species that were 13 C-enriched at this position. Similarly, the signals centered at ␦ 30.9 were assigned to C-12 with a singlet in those molecular species lacking 13 Cenrichment at C-11 or C-13 and a doublet due to coupling in molecules with 13 C at position C-11. However, as shown (Fig. 2,  inset B), each of the three signals arising from C-12 was further split by another 13 C-13 C long range coupling. The labeled position responsible for this coupling was assigned to C-16, the signal for which was at ␦ 28.1 (Fig. 2, inset C). Expansion of the signal at ␦ 28.1, which at first sight appeared to be a singlet, showed it was composed of a singlet plus a doublet. The splitting to give the doublet must have resulted from long range coupling in molecules labeled with 13 C at C-12 as well as at C-16. The 13 C labeling pattern determined in the 14␣-methylergosta-8,24(24 1 )-dien-3␤-ol was therefore entirely consistent with the leucine skeleton remaining intact during metabolism and incorporation into the isoprenoid pathway (Schemes 1 and  2). This is in striking contrast to leucine utilization in plants, where it is first degraded to acetyl-CoA before utilization in isoprenoid production (19,20).
Because the inhibitor ketoconazole was used to facilitate the accumulation of the 14␣-methylergosta-8,24(24 1 )-dien-3␤-ol used for the NMR study, there was perhaps a possibility that this drug could have perturbed the metabolic pathways through HMG-CoA. For example, it may have caused an overexpression of HMG-CoA reductase in response to the decline in the normal sterol (i.e. ergosta-5,7,24(24 1 )-trien-3␤-ol). This could have resulted in HMG-CoA being rapidly reduced before it could be cleaved by HMG-CoA lyase to acetyl-CoA and acetoacetate. To check this point, promastigotes were cultured with [2-13 C]leucine in the absence of ketoconazole. GC-MS analysis of the major sterols as their TMS ether derivatives showed that ergosta-5,7,24(24 1 )-trien-3␤-ol had a molecular ion cluster (m/z 469 -475), indicating species of the sterol molecule containing from one up to a maximum of six 13 C atoms (Table  II) with a distribution of molecular species fairly similar to that found previously for 14␣-methylergosta-8,24(24 1 )-dien-3␤-ol. The various fragment ion clusters were also consistent with the presence of labeled species containing 13 C in the positions predicted from the labeling pattern determined for 14␣-methylergosta-8,24(24 1 )-dien-3␤-ol. The mass spectra of the ergosta-5,7,24(24 1 )-trien-3␤-ol TMS ether labeled from either [2-13 C]acetate or [1-13 C]glucose contained the molecular ion at m/z 468 for unlabeled compound, presumably produced from leucine, and a series of further molecular ions of diminishing abundance containing 1-10 13 C atoms (above this, the ions were too weak to determine whether molecular species with the theoretical maximum of 12 13 C atoms were present). The incorporation of acetate and glucose into the sterol was consistent with our previous investigations (12,13), which indicated that acetate can provide up to 20% of the carbon needed for sterol production in L. mexicana but with the major portion (ϳ80%) arising from leucine.
Incorporation of [2-13 C]Leucine into Sterol by Amastigotes of L. mexicana-The above experiments were performed with L. mexicana promastigotes. We have previously reported that the amastigote form of this parasite can also use [U-14 C]leucine for sterol biosynthesis (12). This has now been confirmed by incubation of L. mexicana amastigotes cultured in macrophages with [2-13 C]leucine. The mass spectrum of the isolated ergosta-5,7,24(24 1 )-trien-3␤-ol, analyzed as the TMS ether, had a strong molecular ion at m/z 468 for unlabeled sterol, but this was accompanied by an ion of similar abundance at m/z 473 for sterol with five atoms of 13 C, together with less abundant ions for molecules containing one, two, three, four, and six atoms of 13 C (Table II) in proportions similar to that seen in promastigotes. The unlabeled ergosta-5,7,24(24 1 )-trien-3␤-ol must be  14 C]leucine into the major sterol of (a) L. mexicana promastigotes incubated with ketoconazole (0.1 g/ml), (b) L. mexicana control promastigotes, (c) L. mexicana amastigotes The sterols were analysed by GC-MS of their TMS ether derivative as described in the Methods. The relative abundance of the ions has been corrected to take into account the natural abundance of 13 C-carbon in the compounds. M ϩ , molecular ion; M ϩ ϩ 1, molecular ion with one 13 C atom; M ϩ ϩ 2, molecular ion with two 13 C atoms, etc.  from the preexisting sterol pool in the amastigotes produced prior to exposure to [2-13 C]leucine. In the 48-h incubation used for this experiment, the amastigotes will have undergone about two cell divisions; therefore, the amount of newly synthesized sterol labeled with 13 C must be insufficient to dilute the unlabeled sterol to the extent seen with the promastigotes (Table  II). Thus, it can be concluded that the promastigote and amastigote forms of the parasite both utilize leucine as a main carbon source for sterol biosynthesis that proceeds by the direct route.
Effects of an Inhibitor of HMG-CoA Synthase-An inhibitor of HMG-CoA synthase should block [1-14 C]acetate incorporation into isoprenoids but have no effect on the direct incorporation of [U-14 C]leucine into sterol (Scheme 1). Accordingly, [U-14 C]leucine and [1-14 C]acetate were incubated separately with L. mexicana promastigotes in the presence and absence of the compound L-659,699, a fungal metabolite, which is a competitive inhibitor of HMG-CoA synthase (33,34). An analysis of the labeled lipid by TLC and radioscanning showed that after the incorporation of [1-14 C]acetate, about 2-3% of the radioactivity was in sterol, with the remainder distributed between triacylglycerol (ϳ35%) and phospholipid (ϳ60%) as observed previously (12). However, incubation with [1-14 C]acetate in the presence of L-659,699 (20 g/ml) resulted in the triacylglycerol and phospholipid remaining labeled in about the same proportions as in the control, but there was a complete abolition of label from the sterol that was consistent with the inhibition of HMG-CoA synthase. The results obtained with the [U-14 C]leucine incubations are presented in Table III and show that the presence of L-659,699 had no apparent inhibitory effect on the incorporation of radioactivity into total lipid and sterol or the distribution of radioactivity between the labeled products even at the highest concentration of L-659,699 (50 g/ml). These results are consistent with the view that in L. mexicana the incorporation of leucine does not require breakdown to the acetyl-CoA level as an essential step.
When cells were cultured with [2-13 C]leucine and increasing concentrations of the HMG-CoA synthase inhibitor L-659,699 followed by GC-MS analysis of the ergosta-5,7,24(24 1 )-trien-3␤ol, the results provided evidence for the dual sources of carbon from either leucine metabolism or acetyl-CoA to fuel the isoprenoid pathway. In the absence of the inhibitor, the predominant molecular species of the sterol TMS ether (M ϩ at m/z 473) had five 13 C atoms rather than six, indicating a contribution from unlabeled precursors probably via the acetyl-CoA These results can be explained by the inhibition of the HMG-CoA synthase, resulting in no contribution from the acetyl-CoA pool in the cell and all of the sterol then being derived from 13 C-labeled leucine. Moreover, they are also consistent with unlabeled carbon introduced in the absence of HMG-CoA synthase inhibitor being largely derived from acetyl-CoA obtained from a carbohydrate, fatty acid, or ketogenic amino acid source rather than directly from unlabeled leucine of protein origin.

Effects of Lovastatin, an Inhibitor of HMG-CoA Reductase-
The effects of an HMG-CoA reductase inhibitor on the incorporation of leucine into sterol were tested using lovastatin, and the results are presented in Table IV. Lovastatin had little or no inhibitory effect on growth of the cultures at concentrations of 5 and 10 g/ml, but growth retardation became apparent as the concentration was increased to 25 and 50 g/ml. Lovastatin at 5 g/ml caused an inhibition of incorporation of [U-14 C]leucine into the total lipid and sterol. The extent of inhibition became progressively greater as the lovastatin concentration was increased to 50 g/ml. This provided good evidence for either the intermediacy of HMG-CoA and the action of HMG-CoA reductase in the pathway or for the operation of a similar type of reductive reaction to that catalyzed by HMG-CoA reductase, which is sensitive to lovastatin inhibition. Analysis of the sterols recovered from the incubations with lovastatin showed that there was a decline in the amount of sterol in the cells, which paralleled the decline in [U-14 C]leucine incorporation. The decline in sterol concentration was accompanied by a change in the composition of the sterol mixture with a progressive increase in the proportion of cholesterol (cholest-5-en-3␤-ol) and corresponding drop in the amount of ergosta-5,7,24(24 1 )-trien-3␤-ol. Cholesterol is not synthesized by the parasite but is derived from the fetal calf serum of the culture medium (12,13), whereas the ⌬ 5,7 -sterols with a C-24 substituent in the side chain are produced by de novo biosynthesis in the protozoa (4). It was noticeable that in these experiments the proportion of stigmasta-5,7,24(24 1 )trien-3␤-ol was observed to increase up to a concentration of 25 g/ml lovastatin. This could be accounted for by the utilization of the 24-methylenesterols as substrate by the second C-24 1transmethylase; the 24-methylenesterols were then not being replaced because of the block in sterol production imposed by the lovastatin. Leishmania sp. may have a growth requirement for a sterol with the structural features of the endogenous sterols (i.e. a ⌬ 5,7 -ring system and a C-24 methylene, methyl, or ethylidine side chain) or alternatively perhaps only newly synthesized sterol can play some important role in sustaining cell growth (27). However, the inhibition of HMG-CoA reductase by lovastatin could also inhibit the production of other isoprenoidderived compounds (e.g. dolichols, prenylated proteins, ubiquinone side chain). This could lead to growth inhibition by starvation of the cell of other vital compounds (e.g. dolichols, prenylated proteins) in addition to the sterols.
The addition of excess MVA to the L. mexicana culture was tested to determine the effect on the incorporation of [U-14 C]leucine into sterol and on culture growth. The incorporation of MVA into sterols by Leishmania species has been demonstrated previously (4,5,(11)(12)(13). Cultures (10 ml, 2 ϫ 10 6 cells/ml) were incubated with [U-14 C]leucine (2 Ci) in the presence or absence of MVA (1 mg/ml). The cultures were stopped after 72 h, the sterols were extracted, and the radioactivity was determined. The addition of MVA had a marked effect and significantly reduced (p ϭ 0.005) the radioactivity incorporated from [U- 14 14 C]leucine into the sterols and other lipids of L. mexicana promastigotes and the sterol composition of the protozoa In experiment (Expt.) 1, cultures (10 ml, 10.2 ϫ 10 6 cells/ml) at midlog phase were incubated for 2 h with lovastatin at the concentrations shown.
[U-14 C]leucine (2 Ci) was then added, and the cells were harvested after a further 22-h growth. The lipids were extracted, and the distribution of radioactivity in the component lipids was determined by separation using TLC as described under "Materials and Methods." In experiment 2, cells were cultured as for experiment 1. The sterols were recovered from the lipid by preparative TLC and analyzed as their TMS ether derivatives by GC and GC-MS; tr indicates trace amount (Ͻ0.5%). Expt from leucine metabolism. Moreover, the growth inhibition of the cells by lovastatin was reversed by the exogenous MVA (Fig. 3), showing that it could enter the isoprenoid pathway to provide the requirements of the cell for sterols and/or other essential isoprenoid-derived compounds.

DISCUSSION
The incorporation of leucine into cholesterol by animal tissues has been demonstrated (14 -18), and it was reported that in rats this proceeded with the prior breakdown of the leucine to acetate (14). Likewise, the incorporation of leucine into a plant sterol was also shown to require the catabolism of the leucine to acetyl-CoA before being incorporated into the isoprenoid pathway (19). By contrast, we have now established unequivocally by MS and 13 C NMR methods that the trypanosomatid L. mexicana can incorporate the leucine skeleton intact into the isoprenoid pathway for sterol production without breakdown first to acetyl-CoA. This could occur by a pathway leading to HMG-CoA, and the HMG-CoA could then be directly reduced to MVA by HMG-CoA reductase (Scheme 1). The fact that some label from [U- 14 C]leucine appears in the fatty acids of triacylglycerols and phospholipids (Tables III and IV) (12) is best explained by the formation of intermediary HMG-CoA, which can yield labeled acetyl-CoA by the action of HMG-CoA lyase. A mitochondrial HMG-CoA reductase has recently been characterized in Leishmania and Trypanosoma species (35,36). The existence of this enzyme could provide the opportunity for a portion of any HMG-CoA produced from leucine metabolism to be reduced to MVA and thus channeled directly into the isoprenoid pathway for sterol biosynthesis. The operation of pathways from leucine and acetyl-CoA, which merge at HMG-CoA, followed by reduction to MVA, is consistent with the following observations. (i) Inhibition of HMG-CoA synthase does not lower leucine incorporation into sterol (Table I). (ii) Lovastatin inhibits the incorporation of leucine (Table IV) (Fig. 3). However, there is an alternative route for the direct incorporation of leucine into sterol that would also be compatible with some of the above criteria. This requires the reduction of dimethylcrotonyl-CoA to dimethylallyl alcohol, followed by phosphorylation to yield dimethylallyl diphosphate, which is a constituent of the isoprenoid pathway (Scheme 1) and is interconvertible with isopentenyl diphosphate by an isomerase-catalyzed reaction. This route would effectively be a reversal of the mevalonate shunt that has been demonstrated in some animal tissues (24,37,38). The reduction of dimethylcrotonyl-CoA to dimethylallyl alcohol would be mechanistically similar to the conversion of HMG-CoA to MVA and could perhaps be catalyzed by the HMG-CoA reductase or a very similar enzyme that may also be susceptible to lovastatin inhibition. If leucine carbon is being channeled along this route, the lowered incorporation of [2-14 C]leucine by added unlabeled MVA could be explained in two ways. Either the conversion of MVA into an appreciable unlabeled pool of dimethylallyl diphosphate/isopentenyl diphosphate results in dilution of the leucine-derived radioactive dimethylallyl diphosphate/isopentenyl diphosphate or the MVA is converted into excess sterol that may inhibit leucine utilization by a feedback inhibition mechanism.
The utilization of the intact leucine skeleton for sterol production may make an important contribution to the metabolic economy of the Leishmania cell. The use of leucine could spare the need for acetyl-CoA produced from glucose or fatty acid catabolism, which would therefore remain available for energy production or other biosynthetic reactions. However, amino acids derived from the breakdown of exogenous proteins are recognized as important energy and carbon sources in trypanosomatids (39 -43). Leucine and other amino acids (glutamate, proline) are reported to be taken up readily from the growth medium by Leishmania sp. and T. cruzi and catabolized to provide acetyl CoA or other metabolites that can be oxidized to provide energy (39 -43). Our work has shown that several Leishmania species, Trypanosoma cruzi, and Endotrypanum monterogeii can all utilize leucine as a carbon source not only for sterol production but also for fatty acid biosynthesis (12,13). Part of the HMG-CoA produced from leucine could be channeled into breakdown by HMG-CoA lyase (Scheme 1) to produce the acetyl-CoA needed for the synthesis of fatty acids. Additionally, this route of leucine catabolism could provide some acetyl-CoA for oxidation in the tricarboxylic acid cycle (41,43,44) or to support the mitochondrial acetate-succinate CoA transferase cycle for the generation of ATP and acetate (45). We have demonstrated that the metabolism of [U-14 C]leucine by L. mexicana promastigotes produces 14 CO 2 . 2 Part of this 14 CO 2 will arise from the decarboxylation of the labeled leucine (Scheme 1) and by C-4 demethylation of a labeled sterol intermediate (4,32), but some could arise from the oxidation of acetyl-CoA generated from the leucine (43,44). Clearly, there must be coordinated regulation of the metabolism of amino acids, glucose, and fatty acids to maintain the balance of acetyl-CoA needed for cell metabolism under conditions of varying availability of these substrates to the promastigote or amastigote forms of the Leishmania parasite.
The key position of HMG-CoA in the catabolism of leucine and in the production of isoprenoids poses questions regarding the cellular compartmentation and regulation of the pathways and enzymes involved in HMG-CoA metabolism. In animals and plants, leucine breakdown is a mitochondrial event (24,25). It has been reported that leucine aminotransferase and ␣-ketoisocaproate dehydrogenase are present in cytosolic and mitochondrial preparations from T. cruzi (46). Similarly, we have found that leucine aminotransferase is located in the mitochondrion of L. adleri. 3 These facts suggest that leucine FIG. 3. The effects of lovastatin and lovastatin plus mevalonic acid (MVA) on the growth of L. mexicana promastigotes. Cells were cultured as described under "Materials and Methods." Either lovastatin (30 g/ml) or lovastatin (30 g/ml) plus MVA (1 mg/ml) was added to cultures at the start of the growth period, and samples were withdrawn at intervals to determine the growth by cell counting using a Neuberger hemocytometer. catabolism to produce HMG-CoA could be located in the mitochondrion of trypanosomatids. HMG-CoA lyase is a mitochondrial enzyme in the mammalian liver cell (47), and there is an HMG-CoA synthase also located in the mitochondrion of mammalian cells (47), but these enzymes have not yet been studied in trypanosomatids. The remaining enzyme of HMG-CoA metabolism, HMG-CoA reductase, is associated mainly with the endoplasmic reticulum in mammalian cells. The HMG-CoA reductase of trypanosomatids has been studied in Trypanosoma brucei (48), T. cruzi (49,50), and Leishmania major (35,51) and variously described as either a microsomal, a glycosomal, or a soluble enzyme. However, recent investigations have now revealed that the HMG-CoA reductase of T. cruzi and L. major (51) and T. brucei (36) are predominantly located in the mitochondrion. Thus, the mitochondrion may be perhaps a major cellular site for the first stages in the production of isoprenoids in trypanosomatids. Certainly, a mitochondrial location of HMG-CoA reductase could provide for an efficient integration of the isoprenoid pathway with a mitochondrial leucine degradation sequence of reactions and thus facilitate the efficient utilization of leucine carbon for sterol synthesis.