James H. Naismith1 Biomedical Sciences Research Complex, University of St Andrews, St Andrews, Scotland KY16 9ST, U.K.
Abstract From a protein structural viewpoint, tryptophan is often considered an inert structural amino acid, playing a role as a hydrophobic anchor in membrane proteins or as part of the hydrophobic core of soluble proteins. However, tryptophan is the only polyaromatic amino acid and, from a chemical viewpoint, possesses unique reactivity owing to the electron-richness of the indole system. This reactivity is seen in the area of natural products and metabolites which have exquisite modifications of the indole ring system. Enzymes have evolved multiple strategies to break or modify the indole ring; one particular class is the IDO/TDO (indoleamine/tryptophan dioxygenase) superfamily. A new member of this family, PrnB, on the surface catalyses a very different reaction, but actually shares much of the early chemistry with the tryptophan dioxygenases. Studies on PrnB have contributed to our understanding of the wider superfamily. In the present mini-review, recent developments in our understanding of how the TDO class of enzymes use activated molecular oxygen to break the indole ring are discussed.
Tryptophan, cutting the ring Tryptophan is the largest natural amino acid and has an indole side chain. An indole is a fusion of two rings: a pyrrole (fivemembered ring) and a benzene (six-membered ring) which forms a single aromatic system. The presence of the nitrogen atom with its lone pair results in a much higher reactivity to electrophilic aromatic substitution than benzene alone. Almost exclusively electrophilic substitution occurs at the C3 position (Figure 1A). The enhanced reactivity at the C3 position to electrophilic aromatic substitution results from the lone pair of the indole nitrogen, which creates increased electron density at C3, often illustrated by resonance structures. It is this extra electron density in the ring that also makes the indole particularly susceptible to oxidation by even mild oxidants such as N-bromosuccinamide which can oxidize the ring to an oxindole. The mechanism of this oxidation is revealing as it proceeds via an initial electrophilic addition of bromine at C3, followed by a second addition of water at C2, and subsequent elimination of HBr across the C2–C3 bond. Thus, although the final product is substituted at C2, the first reaction takes place at C3. TDO (tryptophan dioxygenase) and its functional (but not sequence) homologue IDO (indoleamine 2,3dioxygenase) catalyse the conversion of tryptophan into Nformylkynurenine (see [1] for a review) (Figure 1A). This is the key first step in the synthesis of nicotinamide, which goes on to make NAD + . As part of this pathway, the excitatory neurotoxin quinolinic acid is formed; reduction Key words: indole, indoleamine dioxygenase (IDO), oxygenation, PrnB, tryptophan, tryptophan dioxygenase (TDO). Abbreviations used: ENDOR, electron nuclear double resonance; IDO, indoleamine 2,3dioxygenase; TDO, tryptophan dioxygenase. 1 email
[email protected]
Biochem. Soc. Trans. (2012) 40, 509–514; doi:10.1042/BST20120073
of levels of quinolinic acid is therapeutically important, which has driven interest in the system. TDO was quickly identified as a haem protein and shown to bind molecular oxygen [1]. However, there followed an intense debate as to the most likely mechanism of the process. Early on, two principal mechanisms found favour, one involved a Criegee rearrangement and the other a dioxetane intermediate (Figure 1B). The Criegee rearrangement is the migration of carbon to an electron-deficient oxygen and is well established in organic chemistry (conceptually similar to the Baeyer– Villiger reaction). Dioxetanes are also well precedented in both organic chemistry and natural products, with such compounds being stable but highly reactive. The first structure of the human IDO enzyme revealed that it belonged to the α-helical globin gene family [2], consistent with it binding a haem prosthetic group (Figure 2A). Interestingly, a complex of the enzyme with the inhibitor 4-phenylimidazole bound at the active site was also obtained. On the basis of the structural observations, the authors proposed a Criegee mechanism in which C3 attacks the oxyferrous species, with a cyclic intermediate in which the indole nitrogen is deprotonated [2]. This modification to the original mechanism was proposed since the active site showed no convincing base and the dogma had been that a base would be required, since N-methyltryptophan was not a substrate. Furthermore, indoles are, in general, more reactive under basic conditions. However, this mechanism, at least at the level of involvement of the indole nitrogen proton, was invalidated when N-methyltryptophan was shown in fact to be a substrate of several members of the TDO/IDO superfamily [3]. At a stroke, this eliminated the need for a base or proton abstraction to be a conserved feature of the mechanism. A high-resolution apo-TDO C The
Authors Journal compilation
C 2012
Biochemical Society
Biochemical Society Annual Symposium No. 79
Tryptophan oxygenation: mechanistic considerations
509
510
Biochemical Society Transactions (2012) Volume 40, part 3
Figure 1 Indole chemistry (A) TDO (and IDO) transform tryptophan in the kynurenine family. (B) The proposed dioxetane and Criegee-type mechanisms of TDO. Without proton abstraction from the indole nitrogen, the doubly charged intermediate would seem unlikely.
compound II ferryl intermediate (Fe4 + = O2 − ) (a vibration band at 799 cm − 1 ) [6]. The presence of such a species has been confirmed independently on IDO [7]. The authors who identified the type II ferryl intermediate proposed a radical-based mechanism in which C2 is attacked first [6]. Another computational study strongly argued against the Criegee rearrangement, supporting sequential addition of oxygen across the C2 and C3 bond [8]. This opened up the mechanistic possibilities, and notably suggested the dioxetane mechanism was, in fact, more stable than previously assumed. An EPR [ENDOR (electron nuclear double resonance)] study of IDO–tryptophan complexes indicates that the protonated amine of the substrate interacts with the oxyferrous complex [9]. More recently, MS [10] and mutagenesis modelling [11,12] have favoured an epoxide-type intermediate (Figure 2C), although a radical mechanism is still favoured by one group [11].
PrnB, an unexpected addition to the family
structure from Ralstonia metallidurans was later reported in a study in which a series of substrate analogues were tested for activity to probe the basis of substrate recognition [4]. As no co-complex with this enzyme was obtained, molecular models of an R-enantiomeric C3 hydroperoxide compound with a syn and trans arrangement (compounds 7 and 9 in [4]) and an S-enantiomeric C3 hydroperoxide (compounds 8 and 10 in [4]) were docked at the haem [4]. The formation of dioxetane and Criegee intermediates require different arrangements of the molecular orbitals on C2-C3-O2-O1 (syn and trans respectively). Docking studies strongly favoured an anti-periplanar arrangement of the (S)hydroperoxide intermediate (compound 8 in [4]) (required for a Criegee mechanism) in the active site [4]. A structure of Xanthomonas campestris TDO in complex with tryptophan, however, was consistent with the C3 hydroperoxide requirement, but showed that the stereochemistry of the C3 hydroperoxide would have to be an Renantiomer [5], essentially invalidating the modelling work [4]. The position of the tryptophan is conserved in a 6fluorotryptophan co-complex reported in the same study, supporting this as a relevant conformation [5] (Figure 2B). Reevaluation of the modelling studies shows that the tryptophan had been placed in such a way that the indole nitrogen pointed away from the haem ring [4], whereas in the crystal structure, the indole nitrogen points towards the ring. Since the complex is only binary rather than tertiary (only a water sits above the haem ring), definitive testing of the Criegee compared with the dioxetane intermediate was not possible, but the authors did suggest that the Criegee-type mechanism was more likely. Spectroscopic and computational studies have sought to address the gap in our mechanistic knowledge. A study of human TDO first identified strong evidence of the C The
C 2012 Biochemical Society Authors Journal compilation
PrnB is thought to catalyse an entirely different reaction when compared with the TDO/IDO superfamily enzymes, with an indole rearrangement of 7-chlorotryptophan [13] (Figure 3A), which occurs in the pyrrolnitrin pathway [14]. The structure of PrnB from Pseudomonas revealed it to be a very close homologue of TDO and IDO. There were some differences between the active sites, most significantly a loop in PrnB which partly shields the haem from the incoming substrate, as well as a slightly different arrangement of residues above the haem [13]. The fact that PrnB was related to IDO was very interesting, since the overall reaction that it is thought to catalyse does not result in the insertion of oxygen. A mechanism involving the insertion and removal of oxygen during turnover was suggested on the basis of the multiple (D- and L-tryptophan, and 7-chloro-D- and 7-chloro-L-tryptophan) co-complex structures [13]. As a warning of the dangers of interpreting binary complexes, the L-tryptophan was bound differently from the other three substrate mimics, and, in all co-complex structures, the iron was ligated by either the amino or carboxy group of the substrate. As the authors pointed out, such co-ordination was not possible if the haem was to bind molecular oxygen. Despite repeated attempts, to date, the purified PrnB has not been turned over with its presumed substrate in vitro [15]. Activity has been detected in crude extracts only, and, for this reason, data from this system have and should be treated with caution. It remains possible that the enzyme operates on a different substrate than is thought or that there is an additional enzyme-catalysed step. Equally, it is possible that the apparent inactivity stems from the need to supply electrons at an appropriate potential to reduce molecular oxygen at various points in the catalytic cycle. All four structures of PrnB in complex with substrates position the indole ring orthogonal to the haem ring, as seen in the X. campestris IDO–tryptophan complex. Interestingly, the L-tryptophan complexes of the two proteins are quite similar, with the indole nitrogen pointing down towards
Biochemical Society Annual Symposium No. 79: Frontiers in Biological Catalysis
Figure 2 IDO and TDO (A) Structure of human IDO is shown in cartoon, and the inhibitor and haem are shown in stick. Carbons in the inhibitor are white, and carbons in the haem are yellow. (B) Structure of X. campestris TDO in complex with l-tryptophan, with the protein shown in cartoon. Carbons in tryptophan are white, and carbons in haem are yellow. The helical bundle surrounding the haem is conserved in all members of the IDO/TDO family, including PrnB. (C) An epoxide intermediate proposed for TDO/IDO. (D) Mechanism of peroxide formation by peracids and alkenes.
the haem [13] (Figure 2B). The other three PrnB substrate complexes differ from the L-tryptophan complex and have the indole ring nitrogen pointing away from the haem [13] (Figure 3B). This arrangement is actually similar to that proposed in the modelling studies of TDO with compound 8 and would form an S-enantiomeric C3 hydroperoxide [4]. If 7-chlorotryptophan is indeed the true substrate for PrnB, the binding mode observed for L-tryptophan is impossible owing to steric clashes with the chlorine atom. The authors did note, however, that the C3 atom of 7-chlorotryptophan in PrnB and the C3 atom of L-tryptophan in IDO superimposed despite the indole rings being oriented differently (and hence different chirality of the hydroperoxide). More recent studies on PrnB highlighted mimicry of the ternary complex with 7-chloro-L-tryptophan and cyanide (CN) bound [15]. The indole of 7-chloro-L-tryptophan binds in the same manner as seen previously, but the main-chain amino and carboxy groups have rotated away from the iron to allow CN to bind to it. Notably the ternary complex
resulted in ordering a loop adjacent to the active site [15]. Measurement of binding affinity suggested an obligate order of substrate binding, with CN binding first [15], which was also recently shown to be the case for TDO [16]. Unusually, when bound to PrnB, CN adopted a bent conformation (like that expected for oxygen), with the nitrogen pointing towards an oxyanion hole. The oxyanion hole is formed by the loop which is positioned differently in PrnB compared with in IDO. This structure foreshadowed the human TDO ENDOR study [9], which showed clearly a hydrogen bond between the oxygen of the oxyferrous and the amino group of L-tryptophan (Figure 3B). Assuming that the bent haem CN is a mimic for the oxyferrous complex, and that 7chlorotryptophan is correctly positioned, then the ternary complex is almost a perfect match for an S-enantiomeric C3 hydroperoxide adduct (Figure 3B). In this orientation, the C2–C3 bond of the indole is anti-periplanar to the O–O bond, which is indeed a very close match to compound 8 of [4]. C The
C 2012 Biochemical Society Authors Journal compilation
511
512
Biochemical Society Transactions (2012) Volume 40, part 3
Figure 3 PrnB (A) Pyrrolnitrin biosynthetic pathway, with the reaction proposed to be catalysed by PrnB highlighted. (B) Structure of the ternary complex of PrnB. The iron is shown as a rust sphere. 7-Chloro-l-tryptophan shown as sticks, with chlorine in purple and other atoms coloured as in Figure 2(B). CN is bound in a bent configuration. The dihedral angle C-N-C3-C2 is trans, giving an anti-periplanar arrangement of the orbitals.
PrnB has implications for the TDO/IDO mechanism Enzymes, in general, conserve features that stabilize the transition state of the rate-determining step. This can be seen in enzymes that superficially catalyse very different reactions, e.g. kinases and adenylases [17]. Thus it seems almost certain that, given the high degree of similarity between PrnB and TDO/IDO, the enzymes share a common rate-determining step. At present, no convincing mechanism that involves a Criegee or dioxetane intermediate for PrnB has been proposed, and no TDO activity has been detected for C The
C 2012 Biochemical Society Authors Journal compilation
PrnB, suggesting either that the common step occurs before formation of the intermediates or that neither mechanism is operative. Similarly, a radical mechanism for the initial steps of PrnB catalysis seems unlikely. Superimposing the PrnB ternary complex and the IDO binary complex [5] shows a very convincing syn arrangement of the atoms (Figure 4A), thus eliminating the Criegee mechanism and, if anything, favouring a dioxetane mechanism for IDO/TDO. Similar conclusions were reached by calculations [8,11,12]. The ternary complex of PrnB also rules out the dioxetane mechanism for PrnB as the C2 atom of 7-chloro-L-tryptophan is too far from the haem for this to be feasible. The close superposition of the C3 atoms in the ternary PrnB complex and the IDO–L-tryptophan complex argues that the rate-determining step is the reaction at the C3 position. Since the formation of the dioxetane is predicted to be the high-energy point of the reaction, then if it is impossible for PrnB, it would also seem unlikely for TDO. The structures of PrnB led to the proposal that the first two steps were the common steps, these being the formation of the C3 hydroperoxide which would then undergo O–O bond cleavage to give the Fe4 + = O2 − intermediate and a central C3 hydroxylated indole intermediate (Figure 4B). The presence of oxygen at C3 of the hydroxylated indole serves to activate the C2 position to nucleophilic attack. For PrnB, it was proposed that the N-terminus attacks the indole intermediate, leading, via a tricyclic intermediate (cyclic amino acetal), to the product. Strong support for this mechanism was provided by a TDO study which detected a cyclic amino acetal from an IDO reaction [10]. The presence of this compound following an IDO reaction would arise from a side reaction of the central intermediate. If the proposed PrnB mechanism is correct, then the amino acetal compound would have the opposite stereochemistry if formed in the active site of IDO to that if formed in the active site of PrnB (owing to the different indole ring orientation). Work on PrnB suggests that, in IDO/TDO, the Fe4 + = O2 − intermediate attacks at C2 to give the 1,2-dihydroxyiron-linked intermediate which decomposes to the final product. The most recent mechanistic studies of TDO/IDO have converged upon an epoxide rather than a C3 hydroxy intermediate in the first step [10–12]. The epoxide could still link between the PrnB and TDO families, but in both would have to break to give the C3 hydroxy species, otherwise it would prevent attack at C2 by either the amino group or Fe4 + = O2 − intermediate which must both occur on the same face of the tryptophan ring as the C3 hydroperoxide formation. The formation of epoxides in organic chemistry uses peracids to add across a double bond. Although this is temptingly similar to IDO/TDO, the mechanism of peracids involves a five-membered ring which helps to stabilize charge and shuttle protons (Figure 2D), and it is unclear whether the oxyferrous complex can be considered analogous to a peracid. Furthermore, it is the C3 atom of tryptophan that superimposes between PrnB and IDO co-complexes, not the C2–C3 bond, thus epoxide formation in the two enzymes is subject to different orbital arrangements, making it less likely to be conserved in both systems.
Biochemical Society Annual Symposium No. 79: Frontiers in Biological Catalysis
Figure 4 Mechanism of PrnB and inferences for IDO/TDO
Summary
(A) Model of the ternary complex of TDO on the basis of the structures of tryptophan in complex with X. campestris TDO and the ternary complex
PrnB is yet another example of where enzymes, despite apparently different substrates and products, share a key catalytic step and a common fold. Despite PrnB being problematic to assay, studies have anticipated several of the key findings of the TDO/IDO superfamily, which highlights the benefit of exploring distant members of enzyme families.
of PrnB. The colour scheme is the same as in Figure 3(B). Note the syn arrangement of the O–O bond and the C2–C3 of indole. (B) Mechanism of the TDO superfamily showing the steps common with PrnB.
Funding This work was funded by the Biotechnology and Biological Sciences Research Council [grant number BB BB/C000080/1].
References 1 Sono, M., Roach, M.P., Coulter, E.D. and Dawson, J.H. (1996) Heme-containing oxygenases. Chem. Rev. 96, 2841–2888 2 Sugimoto, H., Oda, S.-i., Otsuki, T., Hino, T., Yoshida, T. and Shiro, Y. (2006) Crystal structure of human indoleamine 2,3-dioxygenase: catalytic mechanism of O2 incorporation by a heme-containing dioxygenase. Proc. Natl. Acad. Sci. U.S.A. 103, 2611–2616 3 Chauhan, N., Thackray, S.J., Rafice, S.A., Eaton, G., Lee, M., Efimov, I., Basran, J., Jenkins, P.R., Mowat, C.G., Chapman, S.K. and Raven, E.L. (2009) Reassessment of the reaction mechanism in the heme dioxygenases. J. Am. Chem. Soc. 131, 4186–4187 4 Zhang, Y., Kang, S.A., Mukherjee, T., Bale, S., Crane, B.R., Begley, T.P. and Ealick, S.E. (2007) Crystal structure and mechanism of tryptophan 2,3-dioxygenase, a heme enzyme involved in tryptophan catabolism and in quinolinate biosynthesis. Biochemistry 46, 145–155 5 Forouhar, F., Anderson, J.L.R., Mowat, C.G., Vorobiev, S.M., Hussain, A., Abashidze, M., Bruckmann, C., Thackray, S.J., Seetharaman, J., Tucker, T. et al. (2007) Molecular insights into substrate recognition and catalysis by tryptophan 2,3-dioxygenase. Proc. Natl. Acad. Sci. U.S.A. 104, 473–478 6 Lewis-Ballester, A., Batabyal, D., Egawa, T., Lu, C., Lin, Y., Marti, M.A., Capece, L., Estrin, D.A. and Yeh, S.R. (2009) Evidence for a ferryl intermediate in a heme-based dioxygenase. Proc. Natl. Acad. Sci. U.S.A. 106, 17371–17376 7 Yanagisawa, S., Horitani, M., Sugimoto, H., Shiro, Y., Okada, N. and Ogura, T. (2011) Resonance Raman study on the oxygenated and the ferryl-oxo species of indoleamine 2,3-dioxygenase during catalytic turnover. Faraday Discuss. 148, 239–247 8 Chung, L.W., Li, X., Sugimoto, H., Shiro, Y. and Morokuma, K. (2008) Density functional theory study on a missing piece in understanding of heme chemistry: the reaction mechanism for indoleamine 2,3-dioxygenase and tryptophan 2,3-dioxygenase. J. Am. Chem. Soc. 130, 12299–12309 9 Davydov, R.M., Chauhan, N., Thackray, S.J., Anderson, J.L., Papadopoulou, N.D., Mowat, C.G., Chapman, S.K., Raven, E.L. and Hoffman, B.M. (2010) Probing the ternary complexes of indoleamine and tryptophan 2,3-dioxygenases by cryoreduction EPR and ENDOR spectroscopy. J. Am. Chem. Soc. 132, 5494–5500 10 Basran, J., Efimov, I., Chauhan, N., Thackray, S.J., Krupa, J.L., Eaton, G., Griffith, G.A., Mowat, C.G., Handa, S. and Raven, E.L. (2011) The mechanism of formation of N-formylkynurenine by heme dioxygenases. J. Am. Chem. Soc. 133, 16251–16257 11 Capece, L., Lewis-Ballester, A., Yeh, S.R., Estrin, D.A. and Marti, M.A. (2012) Complete reaction mechanism of indoleamine 2,3-dioxygenase as revealed by QM/MM simulations. J. Phys. Chem. B 116, 1401–1413 12 Capece, L., Lewis-Ballester, A., Marti, M.A., Estrin, D.A. and Yeh, S.R. (2011) Molecular basis for the substrate stereoselectivity in tryptophan dioxygenase. Biochemistry 50, 10910–10918 13 De Laurentis, W., Khim, L., Anderson, J.L., Adam, A., Johnson, K.A., Phillips, R.S., Chapman, S.K., van Pee, K.H. and Naismith, J.H. (2007) The second enzyme in pyrrolnitrin biosynthetic pathway is related to the heme-dependent dioxygenase superfamily. Biochemistry 46, 12393–12404 14 van Pee, ´ K.H. and Ligon, J.M. (2000) Biosynthesis of pyrrolnitrin and other phenylpyrrole derivatives by bacteria. Nat. Prod. Rep. 17, 157–164 C The
C 2012 Biochemical Society Authors Journal compilation
513
514
Biochemical Society Transactions (2012) Volume 40, part 3
15 Zhu, X., van Pee, ´ K.H. and Naismith, J.H. (2010) The ternary complex of PrnB (the second enzyme in the pyrrolnitrin biosynthesis pathway), tryptophan, and cyanide yields new mechanistic insights into the indolamine dioxygenase superfamily. J. Biol. Chem. 285, 21126–21133 16 Efimov, I., Basran, J., Sun, X., Chauhan, N., Chapman, S.K., Mowat, C.G. and Raven, E.L. (2012) The mechanism of substrate inhibition in human indoleamine 2,3-dioxygenase. J. Am. Chem. Soc. 134, 3034–3041
C The
C 2012 Biochemical Society Authors Journal compilation
17 Schmelz, S., Kadi, N., McMahon, S.A., Song, L., Oves-Costales, D., Oke, M., Liu, H., Johnson, K.A., Carter, L.G. and Botting, C.H. (2009) AcsD catalyzes enantioselective citrate desymmetrization in siderophore biosynthesis. Nat. Chem. Biol. 5, 174–182 Received 9 March 2012 doi:10.1042/BST20120073
