Study of IspH, A Key Enzyme in the Methylerythritol Phosphate Pathway Using Fluoro-substituted Substrate Analogues

Abstract

IspH, a [4Fe-4S]-cluster-containing enzyme, catalyzes the reductive dehydroxylation of 4-hydroxy-3-methyl-butenyl diphosphate (HMBPP) to isopentenyl diphosphate (IPP) and dimethylallyl diphosphate (DMAPP) in the methylerythritol phosphate pathway. Studies of IspH using fluoro-substituted substrate analogues to dissect the contributions of several factors to IspH catalysis, including the coordination of the HMBPP C₄-OH group to the iron-sulfur cluster, the H-bonding network in the active site, and the electronic properties of the substrates are reported.

IspH is a [4Fe-4S]-cluster-containing enzyme found in the newly discovered isoprene biosynthetic pathway known as the methylerythritol phosphate (MEP) pathway.¹ It catalyzes the last step of the MEP pathway by reductive dehydroxylation of the C₄-OH group of 4-hydroxy-3-methyl-butenyl diphosphate (HMBPP, 1) to isopentenyl diphosphate (IPP, 2) and dimethylallyl diphosphate (DMAPP, 3, Figure 1), which are the two building blocks for all isoprenoids.^2,3 While the MEP pathway is not present in humans, it is essential for many pathogenic microorganisms. Therefore, the enzymes involved in this pathway are attractive new targets for antimicrobial drugs.^3,4 Early results showed that its labile [4Fe-4S] cluster contains a unique iron site,^5,6,7,8,9,10 to which the C₄-OH group of HMBPP binds (4 in Scheme 1).^6,7,11 The bound HMBPP adopts a hairpin conformation in the active site with the carbon chain skeleton sandwiched between its pyrophosphate group and the [4Fe-4S] cluster (Figure 1).¹² Due to the lack of a nearby proton source, protonation of the allylic anion intermediate (8, Scheme 1) during turnover has been proposed to be mediated by the terminal phosphate group of HMBPP.^8,11,12

Figure 1.

Reaction catalyzed by IspH, and the IspH-HMBPP (1) complex in the active site. Due to the short distance (~ 2.9–3.0 Å) between the olefinic moiety (C₂–C₃) of 1 and the unique apical iron site, some interactions between the double bond of 1 and the iron-sulfur cluster may exist.

Scheme 1.

Proposed mechanisms for IspH-catalyzed reaction

Studies in recent years have led to three mechanistic models for IspH catalysis.^{5,8,11,13,14,15,16} Mechanism A proposed by Rohdich et al. (Scheme 1, route A) resembles a Birch reduction in which a one-electron transfer from the reduced iron-sulfur cluster to HMBPP (5 → 6) triggers the C₄-deoxygenation from the nascent radical anion species (6) to form an allyl radical intermediate (7).^11,14,15 Subsequent one-electron reduction (7 → 8) followed by protonation produces IPP (2) or DMAPP (3) depending on whether protonation occurs at C₂ or C₄ of 8. Mechanism B proposed by Wang et al. (Scheme 1, route B), is based mainly on the biophysical characterization of a paramagnetic species trapped in the IspH E126A mutant.¹⁷ In this organometallic model, formation of a substrate/metal π complex or an π²-alkenyl/metallacycle (9) followed by dehydration gives an π¹-allyl intermediate (10). Upon the second one-electron transfer, 10 is reduced to an π³-allyl complex (11), which is then protonated to form IPP or DMAPP. Alternatively, in mechanism C (Scheme 1, route C) proposed by Altincicek et al.,¹³ the Lewis acidity of the metallo-center may facilitate the heterolytic cleavage of the C₄-OH bond to form an allylic cation intermediate (12), which then undergoes two-electron reduction and protonation to yield the final products. Although coordination of the C₄-OH group of HMBPP to the [4Fe-4S] cluster is important for all three proposed mechanisms, formation of metallocycle intermediates using the HMBPP olefinic moiety is specific for mechanism B, and the involvement of a cation intermediate is unique for mechanism C.

To learn more about the substrate specificity and to assess the energetic contributions of various interactions in the active site during IspH catalysis, a series of fluoro-substituted substrate analogues were synthesized and their competence as IspH substrates was examined (Scheme 2).¹⁸ The results and insights gained from these studies are reported herein.

Scheme 2.

Products produced from incubations of substrate analogues with IspH (please refer to Figure 1 for the distances between various atoms)

Previously, we reported that IspH can process the monofluoro analogue 13 to produce both IPP (2) and DMAPP (3) in a ratio of ~ 5:1.¹¹ The kinetics of 13 as an IspH substrate was re-examined in this work by a more sensitive assay using methyl viologen, instead of flavodoxin-flavodoxin reductase, as the electron mediator.⁷ In the new assay, the IspH activity is nearly 100-fold higher than that of the NADPH-flavodoxin-flavodoxin reductase system.¹¹ As listed in Table 1, the K_m for 13 is 5.3-fold higher and k_cat is 21.8-fold lower with a total of ~115-fold reduction of k_cat/K_m relative to that of HMBPP (1). Clearly, replacing the C₄-OH group by a fluorine substituent affects overall catalysis. The reduction in catalysis may result from a combination of several factors. First, substitution of the C₄-OH group by a fluorine atom will dramatically reduce the interaction between the substrate and the [4Fe-4S] cluster apical iron site because fluorine is a poor metal ligand. Second, the electronegativity difference between a fluoro and a hydroxyl group will change the electronic properties of the double bond, which in turn, may affect IspH activity. Third, the HMBPP C₄ hydroxyl group is part of the H-bonding network shown in Figure 1. To dissect the relative contributions of the above factors to IspH catalysis, four additional substrate analogues, 14, 15, 17, and 18, were prepared and examined (Scheme 3 and Table 1).

Table 1.

Summary of activity using various mechanistic probes

probes	kcat (min−1)	Km (μM	kcat/Kms (μM−1 · min−1)
1	604 ± 17	19.7 ± 2.4	30.6
13	27.7 ± 2.2	104 ± 31	0.27
14	8.4 ± 1.9	489 ± 170	0.017
15	91.6 ± 2.2	447 ± 32	0.21
17	No detectable activity
18	No detectable activity

Scheme 3.

Mode of protonation of the allylic anion intermediate generated during turnover

The substrate analogue, 14, has the fluorine group at the C₅ position instead of C₄. Such a change should have little impact on the electronic properties of the olefin, but will affect the hydrogen-bond network with the active site residues T167 and E126. Thus, a comparison with analogue 13 may allow us to better assess the effect of the hydrogen-bonding network in IspH catalysis. As shown in Table 1, incubation of 14 with IspH showed that IspH can accept 14 as a substrate and produce IPP (2) as the sole product. In comparison with HMBPP (1), the K_m for 14 is 25-fold higher, the k_cat is 72-fold lower, and the k_cat/K_m is ~1783-fold reduced. The significant reduction in k_cat/K_m for 14 relative to 1 is not surprising and can be attributed to a combination of the lack of direct coordination of 14 to the apical iron site of the iron-sulfur cluster and the change in the electrophilic nature of the double bond. Notably, comparing 13 with 14, the K_m for 14 is 4.7-fold higher, the k_cat is 3.3-fold lower, and the k_cat/K_m is ~15.5-fold reduced. The observed activity differences between 13 and 14 suggest that the C₄-fluoro group of 13 may still maintain partial hydrogen-bond interaction with the T167 and E126 side-chain rendering it a better substrate for IspH than 14.

To further probe the effects of the substrate electronic properties on IspH catalysis, (Z)-4,4,4-trifluoro-3-(hydroxymethyl)-but-2-en-1-yl diphosphate (15) was prepared. The strong electron withdrawing nature of the trifluoromethyl substituent will destablize the proposed allylic carbocation intermediate (12) which will in turn reduce the reaction rate to a great extent if mechanism C is operative. When 15 was incubated with IspH, 3-(trifluoromethyl)but-3-en-1-yl diphosphate (16) was produced as the only product (see Scheme 3).²⁰ Clearly, IspH is capable of catalyzing the dehydroxylation of 15, but not the subsequent dehalogenation reaction. A similar outcome was also observed when (Z)-4,4,4-trifluoro-3-methyl-but-2-en-1-yl diphosphate (17) or (E)-4,4,4-trifluoro-3-methyl-but-2-en-1-yl diphosphate (18) was incubated with IspH and no product could be detected (Scheme 2). Although the k_cat/K_m value is reduced ~150-fold for 15 (compared to 1), the k_cat is only 6.6-fold lower. This data suggests that route C is less likely as the mechanism for IspH. A similar conclusion was also reached in a recent report in which incubation of IspH with several proposed carbocationic intermediates caused little inhibition against IspH.¹⁹ The fact that the k_cat for 15 is ~3.3-fold greater than that of the monofluoro analogue 13 and ~10.9-fold higher than that of the monofluoro analogue 14 again underscores the important role played by the C₄ hydroxyl group in IspH catalysis.

The fact that only IPP (2) is produced in the reaction with 14 is intriguing since both IPP (2) and DMAPP (3) are formed when HMBPP (1) is used as the substrate. As shown in Scheme 3A, generation of 2 and 3 from 1 results from protonation of the proposed allylic anion intermediate 8 at C₂ and C₄, respectively. The proton donor is believed to be the terminal phosphate group of HMBPP, which is within ~3.4–3.5 Å of the C₂ and C₄ position of HMBPP (see Figure 1).^12,20 In the case of 14, the negative charge of the proposed allylic anion intermediate (e.g., 19, Scheme 3B) is delocalized through C₂, C₃, and C₅ instead of C₂, C₃, and C₄ as shown in intermediate 8 (see Scheme 3A) The fact that protonation takes place only at C₂, but not at C₅ in 19, likely reflects the distance and unfavorable orientation of C₅ of 19 relative to the phosphate group (e.g., O_B) in the active site (e.g., ~4.6 Å between C₅ and O_B, Figure 1).

The conclusions reached based on the outcomes of these fluoro-analogues are also consistent with that derived from the incubation of IspH with a [¹³C]-labeled substrate analogue, 3-(hydroxymethyl)but-3-en-1-yl diphosphate ([5-¹³C]-20, Scheme 3C),²¹ which is expected to bind to [4Fe-4S]⁺ cluster in two different orientations depending on whether the reaction follows mechanism A or B. If the C₄-OH of 20 is the anchor that positions the substrate in the active site (21 in Scheme 3D), protonation mediated by O_B at C₄ of the allylic anion intermediate (23, Scheme 3D, route a) would yield [¹³C]-IPP (2b) as the product (also see Figure 1). In contrast, if the reaction proceeds via an π²-alkenyl intermediate (22 in Scheme 3D), protonation of the allylic anion at the carbon closer to O_B (now C₅, Scheme 3D, route b) should afford [¹³C]-IPP (2c). The observation that 2b is the sole product after incubation strongly suggests that the C₄-OH group, rather than the olefinic π-system, plays the dominant role in orienting the substrate in the active site, and thus favors the Birch reduction model. Together, these observations strongly indicate that protonation of the allylic anion intermediate (8, 19 or 23) occurs at C₂ and/or C₄, but not at C₅.

In conclusion, the results obtained from these studies provide important insights into the mode of substrate binding and the mechanism of IspH. First, in view of the relatively minor influence of electron-withdrawing substituents on the reaction rate (e.g., 15), a mechanism involving a carbocationic intermediate/transition state is less likely. Second, coordination of the C₄-OH group of HMBPP to the unique iron site of the [4Fe-4S] cluster and its involvement in the hydrogen-bond network (e.g., with T167 and E126) is crucial for efficient catalysis (13/14 versus 1, and 15 versus 17). The adverse effects (both steric and electronic) imposed by a trifluoromethyl substituent in 17 can be balanced by the introduction of a C₄-OH group (15 versus 17). Third, protonation of the allylic anion intermediate is regiospecific and takes place only at C₂ and/or C₄ (see 8 → 2/3, 19 → 2, and 23 → 2b), but not at C₅ (Scheme 3). These findings support the assigned role of the terminal phosphate group of HMBPP in mediating the final protonation step due to its close proximity to the C₂ and C₄ positions of the allylic anion intermediate (8, Schemes 1 and 3A). More experiments are in progress to further investigate this interesting example of an enzymatic reductive dehydroxylation reaction.