CYP51: A Major Drug Target in the Cytochrome P450 Superfamily

Abstract

The cytochrome P540 (CYP) superfamily currently includes about 9,000 proteins forming more than 800 families. The enzymes catalyze monooxygenation of a vast array of compounds and play essentially two roles. They provide biodefense (detoxification of xenobiotics, antibiotic production) and participate in biosynthesis of important endogenous molecules, particularly steroids. Based on these two roles, sterol 14|*alpha*|-demethylases (CYP51) belong to the second group of P450s. The CYP51 family, however, is very special as its members preserve strict functional conservation in enzyme activity in all biological kingdoms. At amino acid identity across the kingdoms as low as 25–30%, they all catalyze essentially the same three-step reaction of oxidative removal of the 14|*alpha*|-methyl group from the lanostane frame. This reaction is the required step in sterol biosynthesis of pathogenic microbes. We have shown that specific inhibition of protozoan CYP51 can potentially provide treatment for human trypanosomiases. Three sets of CYP51 inhibitors tested in vitro and in trypanosomal cells in this study include azoles [best results being 50% cell growth inhibition at <1 and at 1.3 µM for Trypanosoma cruzi (TC) and Trypanosoma brucei (TB), respectively], non-azole compounds (50% TC cell growth inhibition at 5 µM) and substrate analogs of the 14|*alpha*|-demethylase reaction. 32-Methylene cyclopropyl lanost-7-enol exhibited selectivity toward TC with 50% cell growth inhibition at 3 µM.

Introduction

Sterol biosynthesis is an ancient metabolic pathway, which is believed to arise during the latter stages of prokaryotic evolution [1, 2]. Once molecular oxygen was in sufficient quantity in the atmosphere to permit squalene 2,3-epoxide production, a switching mechanism became operative to divert squalene from conversion into pentacyclic triterpenoids (sterol-like compounds formed in the Precambrian anaerobic environment [3, 4]) to sterol synthesis [5]. Incorporation of sterols in cellular membranes made it possible for eukaryotic organisms to appear. Presently the sterol biosynthetic pathway exists in all eukaryotic kingdoms and is even found in some bacteria [6]. At the stage of squalene 2,3-epoxide cyclase the pathway bifurcates into two branches (Fig. 1a). The cycloartenol route is typical for photosynthetic organisms, while nonphotosynthetic progenitors and their descendants possess a lanosterol-based pathway [7]. Further conversions of sterol precursors include C14 and C4 demethylations, introduction of double bond(s) into the B ring and, in plants, fungi and protists, modification of the side chain to form cholesterol (animals), ergosterol (fungi) and a variety of 24 alkylated and olefinated products in plants and protists. Sterols stabilize the membranes, determine their fluidity and permeability and also serve as precursors for biologically active molecules essential for regulation of growth and development.

Fig. 1.

Divergence in sterol biosynthesis: a in photosynthetic (cycloartenol) and non-photosynthetic (lanosterol) eukaryotes; b in Trypanosomatidae

In view that humans consume cholesterol from the diet thereby downregulating sterol biosynthesis, the microbial ergosterol biosynthetic pathway can serve as a potential selective target for the treatment of human infections with ergosterol biosynthetic inhibitors such as itraconazole. The majority of currently used clinical antifungal drugs (azoles) are aimed at inhibiting the fungal sterol 14α-demethylase (CYP51). Many of them have also been proven experimentally to be effective against human infections involving protozoa [8–11]. However, due to low amino acid identity of CYP51s across these two biological kingdoms (about 22–24% in average) without direct characterization of protozoan sterol 14α-demethylases it has been difficult to predict the potential efficiency of the antifungal drugs against protozoan infections and impossible to conduct directed search for new more potent compounds.

Using heterologously expressed (E. coli) and highly purified CYP51 orthologs from Trypanosoma cruzi (TC) and Trypanosoma brucei (TB), we studied inhibition of their activity in vitro, in reconstituted enzyme reactions, followed by testing the effects of the most promising compounds on trypanosomal cells. Both of these protozoan parasites are human pathogens, the origin of endemic diseases that are deadly and so far very difficult to cure. TC is the causative agent of Chagas disease, or American trypanosomiasis; while TB species cause sleeping sickness, or African trypanosomiasis. The organisms belong to the same family of lower eukaryotes (Trypanosomatidae), both having complex life-cycles that include insect and mammalian stages which are significantly different between the trypanosomes. TB and TC differ in their insect vectors, preferred mammalian hosts (though both infect humans), localization in the host organism, and symptomatic manifestation. From our studies, the most interesting differences between the two trypanosomes lie in their sterol requirements. While TC produces endogenous sterols at all life stages, the TB in the mammalian (bloodstream) stage was shown to be able to use host cholesterol. Since sterols are known to play a dual role in eukaryotic organisms [12], we used TC as a model to study the effect of inhibition of the production of structural (most abundant) sterols, while TB was a model to test the possible importance of so called functional (sparking) sterols, which are synthesized at hormonal levels, and might be important for the parasite development and life cycle regulation. In both cases direct correlation between CYP51 inhibition and antiparasiticeffect in trypanosomal cells has been observed.

Experimental Procedures

Expression and purification of TB, TC, human and Candida albicans CYP51 and use of cytochrome P450 reductase as their electron donor partner, reconstitution of enzymatic activity in vitro, and inhibition of enzymatic reaction were performed as described previously [13–15]. Inhibitory potencies of the tested compounds were compared as molar ratio inhibitor/enzyme which causes twofold decrease in the enzyme activity (I/E₂) expressed either as a turnover number (initial rate, I/E₂, 5 min) or, for the most potent inhibitors, as percentage of substrate conversion per 1 h reaction (I/E₂, 60 min) [16]. Imidazole derivatives were from Novartis Research Institute (Vienna, Austria). Optical high throughput screening (HTS) of a 20,000 compound library of bioactive molecules (http://www.fmp-berlin.de/kries/index1.html) against TB and TC CYP51s took advantage of the P450 property to change the Soret band absorbance maximum upon ligand binding [16]. It was carried out at the Screening Unit, Leibniz Institute for Molecular Pharmacology (Berlin, Germany) and the details of screening will be described elsewhere. Web search for structural similarity to reactive molecules from the screening was done using ChemDiv data base (http://chemdiv.emolecules.com). Structure of substrate analogs, YNE and MCP, was confirmed by MS and NMR data. TB and TC cellular growth and inhibition experiments, extraction of TC cellular sterols and TLC analysis were carried out as described [16, 17]. Scanning electron microscopy of TC amastigotes grown in Brain Heart Infusion Medium +10% FBS exposed or not to CYP51 inhibitors was performed using a Hitachi 2700 scanning electron microscope [18].

Results and Discussion

TB and TC CYP51s Differ in their Substrate Preferences

Trypanosomatidae, like other non-photosynthetic eukaryotes (Fig. 1a), cyclize squalene 2,3-epoxide into lanosterol and not into cycloartenol [19, 20]. In TC, prior to the CYP51 reaction, lanosterol is converted to form 24-methylenedihydrolanosterol (Fig. 1b), suggesting that the postsqualene portion of sterol biosynthesis in this protozoa is similar to the pathway in filamentous fungi except that the major sterol products in TC are not limited to ergosterol [21]. TBCYP51, on the contrary, has strict preference toward the C4 monomethylated sterols as substrates [14]. In the reconstituted enzyme reaction in vitro the TBCYP51 turnover of obtusifoliol (the substrate of CYP51s from plants) and norlanosterol (C4-monomethylated analog of lanosterol) are more than 100-fold faster than the 14α-demethylation of lanosterol, 24-dihydrolanosterol or 24-methylenedihydrolanosterol. We have shown earlier that the differences in the substrate preferences of TB and TC CYP51 are connected with a phyla-specific residue in the B′-helix (plant-specific F105 in TB and other CYP51s from Trypanosomatidae as opposed to animal/fungi-like I105 in TCCYP51) [15]. The same TB-like preference for C4-mononmethylated sterol substrates have been observed for CYP51 from Leishmania infantum (NCBI accession number EF192938) (unpublished). Based on the catalytic properties of TB sterol methyl transferase [22], detailed analysis of TB sterols [23] and information from the literature on sterol composition of Leishmania species [12], 31-norlanosterol rather than obtusifoliol appear to be the natural substrate of TBCYP51 and most likely of all other CYP51s from Trypanosomatidae except for TC (Fig. 1b). This finding strongly suggests that Trypanosomatidae developed a unique postsqualene portion of the sterol biosynthetic pathway.

Functionally Irreversible Inhibitory Effect of β-Phenyl Imidazoles is Specific for Trypanosomal CYP51s

Since TB and TC CYP51 share 83% amino acid sequence identity, it is not surprising that, regardless of the differences in their substrate preferences, they both were found to be strongly inhibited by the same groups of imidazole derivatives (example structures are shown in Table 1). The strongest inhibition was observed for the β-phenyl imidazoles. Several such compounds produced a functionally irreversible effect with complete inhibition of TB and TC CYP51 activity at equimolar ratio inhibitor/enzyme [16]. At the same binding parameters (K_d <0.1 µM) the inhibitory effect of the α-phenyl azoles on the initial rate of TB and TC CYP51 reaction was similarly strong (I/E₂ <1), but reversible, the compounds being easily replaced by the excess of substrate upon long-term enzymatic reaction. Thus the difference in the I/E₂ values per 5 min and per 1 h for the α-phenyl azole shown in Table 1 is more than 10-fold for TCCYP51 and more than 35-fold in the case of TBCYP51. Finally, imidazoles without an aromatic group between the azole ring and the amide portion of the molecule were weakly binding ligands and weak inhibitors.

Table 1.

Inhibitory effect of azoles on CYP51 from different organisms

CYP51
	α-Phenyl (SDZ285428)		β-Phenyl (SDZ285604)
	I/E2a (5 min)	I/E2 (60 min)	I/E2 (5 min)	I/E2 (60 min)
T. cruzi	<1	9	<1	<1
T. brucei	<1	35	<1	<1
H. sapiens	>100	>100	100	>100
C. albicans	>100	>100	<1	8

Numbers are averages of duplicate determinations and varied by not more than 10%

^aMolar ratio inhibitor/enzyme which causes a twofold decrease in the activity

Proving the importance of detailed preliminary knowledge about target enzyme inhibition for further drug development, direct correlation between the α- and β-phenyl azole potencies as CYP51 inhibitors and their antiparasitic effects in trypanosomal cells was observed [16]. The β-phenyl azole inhibitor shown in Table 1 produces EC₅₀ values (effective concentration, inhibitor concentration in growth media which causes 50% cellular growth inhibition) of 7 and 1.3 µM, respectively, in procyclic (insect stage) and bloodstream (mammalian stage) TB and EC₅₀ <1 µM in trypomastigotes (extracellular form) and amastigotes (intracellular form) of TC, while the EC₅₀ values for the strongest α-phenyl azole inhibitor, are 16 and 14 µM in procyclic and bloodstream TB; 8 and 20 µM in trypomastigotes and amastigotes of TC. Though the effect in TB is weaker than in TC, it strongly suggests that functional sterols must be important in the bloodstream form of the parasite and that inhibitors of sterol biosynthesis should be envisaged as potential anti-sleeping sickness drugs. For the comparison, published EC₅₀ values in TC cells for benznidazole, one of the two drugs currently used for clinical treatment of TC infections, were reported to be within 25–50 µM [24–26]. Much lower values from our work suggest using the β-phenyl azoles as lead structures for anti-trypanosomal chemotherapy.

Scanning electron microscopy of the TC cells shows that treatment with 1 µM CYP51 inhibitor affects the topology of the TC plasma membrane (Fig. 2a). Ultrastructural alterations include the formation of blebs at the plasma membrane, membrane disorganization and alterations in the parasite shape. Analysis of the sterol composition of TC amastigotes treated with this azole confirmed that the parasite membranes are lacking C14-demethylated sterols (Fig. 2b; TLC of TC epimastigotes and GC–MS analysis of their cellular sterols can be found in [16]).

Fig. 2.

Cellular effects of SDZ284692 (1 µM) on TC amastigotes. a Scanning electron microscopy. b TLC analysis of extracted sterols. Sterol standards: 24-methylenedihydrolanosterol (M), lanosterol (L), obtusifoliol (O), cholesterol (C) and ergosterol (E), 5 nmol each (lane 1). Sterols from untreated amastigotes (lane 2) and amastigotes incubated for 120 h with SDZ284692 (lane 3). c Structural formula of the inhibitor

Interestingly, extremely potent, functionally irreversible inhibition by the β-phenyl imidazoles demonstrated by TB and TC CYP51 appears to be specific for trypanosomal sterol 14α-demethylases (Table 1). The effect of SDZ 285604 on the fungal CYP51 ortholog (from C. albicans) is still strong but reversible (I/E₂ per 60 min = 8), while the activity of human CYP51 is only slightly affected by the presence of up to 100-fold molar excess of both the α-and β-phenyl azoles. The data indicate that development of species-specific drugs based on the comparative analysis of the drug target enzyme inhibition in vitro is possible.

Benznidazole as CYP51 Inhibitor

Despite >40 years of its clinical use as an anti-chagasic drug, the mechanism(s) of action of benznidazole have remained elusive. It has been proposed that in trypanosomes, the drug mediates parasite killing by inducing oxidative stress [27–29] or that it works as a prodrug activated by TC nitroheterocycle reductases which reduce the nitro-group leading to covalent binding or other types of interaction between nitroreduction intermediates and various cellular components, like DNA, lipids and proteins [30–32]. It has been shown experimentally that nitroaromatic compounds can be reduced by lipoamide dehydrogenase and trypanothione reductase from TC in vitro, as well as by other known electron transfer proteins including cytochrome P450 reductase [33].

Since benznidazole is an azole derivative, we tested it in our reconstituted TCCYP51 reaction to see whether the anti-chagasic activity of the drug might involve a sterol biosynthesis inhibitory component. The results show that benznidazole inhibits TCCYP51 with I/E₂ = 42 (5 min) and, surprisingly, I/E₂ = 45 (1 h) (Fig. 3). The rather high I/E₂ value for inhibition of the initial rate of the reaction is most likely due to low affinity of the binding (benznidazole does not induce spectral changes in TCCYP51 so the binding parameters can not be estimated by spectral titration). The comparable I/E₂ for the long-term inhibitory effect suggests that the drug might act as an irreversible CYP51 inhibitor, the nitro-group being reduced by cytochrome P450 reductase and then covalently bound to the P450 protein moiety. This hypothesis needs to be tested in more detail but it is not excluded that introduction of a nitro-group into the structures of some known CYP51 inhibitors might lead to the development of new, more potent drugs.

Fig. 3.

Inhibitory effect of benznidazole on TCCYP51 activity. I/E inhibitor enzyme ratio

Search for CYP51 Inhibitors other than Azoles

It is known that azoles, at least those currently used as clinical and agricultural fungicides, upon long-term treatment often can cause resistance. The mechanism(s) for the resistance remain unclear, but the three most often suggested reasons are: increase in the sterol flow; mutations in the target enzyme, or accelerated azole efflux from the cells [34]. In order to investigate alternative options for the development of new sets of CYP51 targeted anti-trypanosomal drugs, optical HTS of TB and TC CYP51 for binding ligands other than azoles has been undertaken and several compounds producing type 1 (substrate-like) or type 2 (azole-like) spectral responses in the cytochrome P450 Soret band were identified [best HTS hits are shown in Table 2(A)]. The most potent inhibitor, N-(4-pyridyl)-formamide, compound 3, has K_ds of 0.8 and 0.5 µM and I/E₂ (5 min) of 7 and 10 for TB and TC CYP51, respectively. It probably coordinates to the heme iron through the pyridyl nitrogen. This has been supported by results of a web-database search for structural similarity using compound 3 as template. The strongest ligands from the web search [Table 2(B)] have bulky structures attached to the N-(4-pyridyl)-amide moiety, suggesting influence of the configuration of the non-coordinated portion of the ligand molecule, which (similar to azole inhibitors) might enhance the inhibitory effect by forming interactions with amino acid residues in the CYP51 substrate binding cavity. Having strong but reversible inhibitory effect on TCCYP51, comparable with the effects of the α-phenyl azoles, compound 5 has potential to serve as a lead for the design of novel non-azole CYP51 inhibitors. It has been tested in TC cells and produced the EC₅₀ values of 5 µM in amastigotes and 8 µM in trypomastigotes.

Table 2.

Binding and inhibition of trypanosomal CYP51 activity with the HTS hits (A), web-search findings (B) and substrate analogs (C) (Numbers are averages of duplicate varied by not more than 10% determinations and varied by not more than 10%)

Inhibitor structure			TBCYP51			TBCYP51
			Kd (µM)	I/E2 (5 min)	I/E2 (60 min)	Kd (µM)	I/E2 (5 min)	I/E2 (60 min)
A	1		14	45	>100	9	28	>100
	2		45a	>50	ND	38a	>50	ND
	3		0.8	7	>100	0.5	10	>100
B	4		0.4	<1	40	0.8	14	65
	5		0.6	<1	25	0.2	<1	20
C	YNE		1.2	46	>100	1.3	4	25
	MCP		0.5	7	28	0.5	2	6

ND not determined

^aType 1 spectral response

Substrate Analogs

The idea of using substrate analogs as inhibitors of sterol 14α-demethylase has been explored as an attempt to develop new hypocholesterolemic agents. It has been found that 7-oxo, 15-keto 15-keto, 15-oxime, 15-hydroxy, 26-oxo derivatives of lanosterol have potential to inhibit cholesterol biosynthesis in humans [35–37]. The advantages of use of substrate analogs as CYP51 inhibitors arise from their strict specificity for the target enzyme, better cellular permeability, longer lifetime in aqueous solutions in comparison to many azoles, and it is quite likely that being similar to endogenous sterols they would not cause resistance. The major problem, however, appears to be connected with very strict requirements that CYP51 enzymes have towards their substrates so that minor alterations in the structure of the sterol molecule can affect binding and inhibitory potency. Recently we have shown that two 14α-methylamino derivatives of lanosterol [4,4-dimethyl-14α-aminomethyl-cholest-7-en-3β-ol (AL7) and 4,4-dimethyl-14α-aminomethyl-cholest-8-en-3β-ol (AL8)] show inhibitory effect on CYP51 from TC and C. albicans, the I/E₂ values for the initial rate of TCCYP51 catalysis being 11 and 14 for AL7 and AL8, respectively [15].

In this study two other 14α-derivatized sterols, Δ7-14α-methylene cyclopropyl didydrolanosterol (MCP) and Δ7, 14α-yne-dihydrolanosterol (YNE), were studied as potential TC and TB CYP51 inhibitors (Table 2C). Contrary to the amino-derivatized AL7 and AL8 which induce type 2 spectral responses in TCCYP51, binding of MCP and YNE cause typical type 1 spectral response (Soret band maximum 393 nm) (Fig. 4a). Titration with YNE leads to 35–40% low to high spin transition in the heme iron of TB and TC CYP51 (not shown) while upon addition of MCP the high spin content in TBCYP51 reaches 85%, the highest currently observed for a CYP51 enzyme. Under the same conditions maximal high spin content in TCCYP51 is only 41%. Since TBCYP51 does not respond spectrally to the C4-double-methylated CYP51 sterol substrates, including dihydrolanosterol [14], the fact that structures longer than the methyl group at the 14α-position (methyl-yne- and especially methylcyclopropyl) having such a profound influence on the binding parameters was quite unexpected. To test whether these differences in the maximum spectral response are connected with the phyla-specific residue in the trypanosomal CYP51s described above, binding of MCP to the TBCYP51-like I105F mutant of TCCYP51 [15] was tested and 87% high spin form was reached.

Fig. 4.

MCP as a CYP51 inhibitor. a MCP induced type 1 spectral responses in TB and TC CYP51 and in the I105F mutant of TCCYP51. b Dose–response curves of trypanosomal cell growth inhibition. c Microscopic observation of the inhibition of TC multiplication by 30 µM MCP within cardiomyoblasts at 72 h. TC pre-treated with control HPCD (used to dissolve MCP) showed high levels of parasite multiplication, whereas cells exposed to trypanosomes pre-incubated with MCP showed dramatic decrease in the number of intracellular parasites

Despite of the amplitude of the spectral responses and the calculated apparent binding parameters, both MCP and YNE were found to inhibit TCCYP51 activity much more strongly than they inhibit TBCYP51 (Table 2C). Based on the fact that in the case of MCP the I/E₂ values for the initial rate of the reaction and long-term inhibitory effect are comparable (2 and 6, respectively), it is not excluded that MCP might act as a mechanism based inhibitor. Thus, the bulky phenyl residue in the B’-helix portion of the CYP51 substrate binding cavity promotes displacement of the water molecule from the heme iron coordination sphere upon binding of MCP or YNE. However, it does not strengthen the enzyme-ligand interaction and both C4-double-methylated inhibitors, similar to the preferred substrates, remain selective for TCCYP51. This conclusion was confirmed by the results from MCP-treated trypanosomal cells: while EC₅₀ for TB was higher than 50 µM, 50% cell growth inhibition of TC was reached at 3 µM (Fig. 4b), more than 90% of the parasite cells being cleared from TC infected cardiomyoblasts upon treatment with 30 µM MCP (Fig. 4c).

Conclusions

• First, the fact that the β-phenyl imidazoles are strong, functionally irreversible inhibitors of trypanosomal but not fungal or human CYP51 opens an opportunity for the development of species specific CYP51 inhibitors

• Second, introduction of additional chemically reactive functional groups (e.g. nitro group) into the structure of known CYP51 inhibitors might strengthen their potency

• Third, optical HTS followed by web-search for structurally similar compounds is a useful method to find new CYP51 inhibitors

• Fourth, since so far TCCYP51 is the only trypanosomal sterol 14α-demethylase with animal/fungal-like substrate preferences, C4-monomethylated substrate based inhibitors are likely to demonstrate selectivity towards CYP51 s from TB and all other Trypanosomatidae sequenced to date

• We surmise that detailed knowledge of specific inhibition toward CYP51 might provide treatment for deadly human infections with protozoan parasites

Abbreviations

CYP

Cytochrome P450

CYP51

Sterol 14α-demethylase

TC

Trypanosoma cruzi

TB

Trypanosoma brucei

MCP

32-Methylene cyclopropyl-lanost-7-enol

YNE

32-Yne-lanost-7-enol