pH and Deuterium Isotope Effects on the Reaction of Trimethylamine Dehydrogenase with Dimethylamine

Abstract

The flavoprotein trimethylamine dehydrogenase is a member of a small class of flavoproteins that catalyze amine oxidation and transfer the electrons through an Fe/S center to an external oxidant. The mechanism of amine oxidation by this family of enzymes has not been established. Here, we describe the use of pH and kinetic isotope effects with the slow substrate dimethylamine to study the mechanism. The data are consistent with the neutral amine being the form of the substrate that binds productively at the pH optimum, since the pK_a seen in the k_cat/K_amine pH profile for a group that must be unprotonated matches the pK_a of dimethylamine. The ^D(k_cat/K_amine) value decreases to unity as the pH decreases. This suggests the presence of an alternative pathway at low pH, in which the protonated substrate binds and is then deprotonated by an active-site residue prior to oxidation. The k_cat and ^Dk_cat values both decrease to limiting values at low pH with similar pK_a values. This is consistent with a step other than amine oxidation becoming rate-limiting for turnover.

Trimethylamine dehydrogenase (TMDH), dimethylamine dehydrogenase, and histamine dehydrogenase make up a small family of amine-oxidizing flavoproteins [1]. The FMN in each is covalently bound to a cysteinyl residue through a 6-S-linkage. In addition to the flavin, these enzymes contain a 4Fe-4S center and a tightly bound ADP. The amine substrate initially reacts with the flavin cofactor, generating the reduced flavin and the oxidized amine. Reoxidation of the reduced flavin occurs by transfer of electrons first to the 4Fe-4S center and then to an electron transferring flavoprotein [2]. The role of the bound ADP, which is in a separate domain from the redox-active cofactors [3], is not known.

The mechanism of amine oxidation by TMDH has not been established. As was the case with other amine-oxidizing flavoproteins[4, 5], the initial focus was on a mechanism involving formation of the amine anion. However, the minor effects on activity upon mutation of potential active-site bases[6–9] provided strong evidence against such a mechanism. Alternative mechanisms have since been suggested involving flavin adducts[5] and substrate radicals as intermediates; less emphasis has been given to a mechanism involving direct hydride transfer to the flavin, although this mechanism is generally accepted for other families of amine-oxidizing flavoproteins[10].

A key mechanistic question for TMDH is the form of the amine substrate that binds productively to the enzyme. Previous studies of the pH dependence of TMDH have relied on the effect of pH on rapid-reaction kinetics. The results have been taken as evidence both for the protonated form of the amine (Scheme 1, lower path) [6, 11] and for the unprotonated form (Scheme 1, uppoer path) being the preferred substrate [8]. The effects of pH on enzyme kinetics with physiological substrates are frequently complicated by high commitments to catalysis, with the bound substrate reacting further faster than it dissociates[12]. This will perturb the pK_a values seen in pH profiles from their intrinsic values, making assignment of pK_a values to specific species complicated. Indeed, this is likely to be the case with TMDH given the high rates seen with triethylamine as substrate. One solution to this problem is to analyze the kinetics using a slow substrate for which binding is much more likely to be in equilibrium, allowing intrinsic pK_a values to be determined.. We describe here the use of the slow substrate dimethylamine to probe the mechanism of TMDH.

Scheme 1.

Materials and Methods

Materials

Dimethylamine and ferrocenium hexafluorophosphate were obtained from Aldrich. d₆-Dimethylamine was from Cambridge Isotopes. Purified TMDH from Methylophilus methylotrophus (sp. W3A1) was a generous gift from Dr. Russ Hille of the University of California-Riverside. All other chemicals were from Sigma or Fisher.

Assays

The initial rate of dimethylamine oxidation by TMDH was determined using the continuous spectrophotometric assay of Lehmen et al.[13], with ferrocenium as the electron acceptor at 30 °C. Assays contained 0.0625–100 mM dimethylamine/d₆-dimethylamine, 0.2 mM ferrocenium hexafluorophosphate, and 200 mM HEPES at pH 7.5–8.5, 200 mM CHES at pH 8.5–9.5, or 200 mM CAPS at pH 10–11. The reaction was started by addition of TMDH to the mixture after following the reaction for ~30 s to determine the background rate. Reactions were then followed for up to 100 s to determine the initial rates. When the concentration of dimethylamine was varied at a fixed concentration of ferrocenium, the data were fit to the Michaelis-Menten equation to determine k_cat/K_m values. To determine deuterium kinetic isotope effects, the concentrations of both dimethylamine and d₆-dimethylamine were varied in the same experiment and the data fit to eq 1, where F is the fractional deuterium content, E_vk is the isotope effect on k_cat/K_m for dimethylamine (DMA) and E_v is the isotope effect on k_cat. The program KaleidaGraph (Synergy Software) was used for fitting.

$$\mathrm{v}/\mathrm{e}=\left[\mathrm{DMA}\right]{\mathrm{k}}_{\mathrm{cat}}/\left\{{\mathrm{K}}_{\mathrm{DMA}}\left[1+{\mathrm{F}}_{\mathrm{i}}\left({\mathrm{E}}_{\mathrm{vk}}-1\right)\right]+\left[\mathrm{DMA}\right]\left[1+{\mathrm{F}}_{\mathrm{i}}\left({\mathrm{E}}_{\mathrm{v}}-1\right)\right]\right\}$$ (1)

Results

The initial rate of oxidation of dimethylamine by TMDH was determined as a function of pH with ferrocenium as the electron acceptor. Controls showed that the concentration of ferrocenium was saturating across the pH range. Above pH 11.5 the nonenzymatic background was too high for reliable measurements, so the analyses were limited to pH 6.5–11.5. The k_cat/K_m pH profile for dimethylamine is shown in Figure 1A. The activity increase with increasing pH to a limiting value. Fitting of the data to eq 2 yields a pK_a of 10.8 ± 0.2 for a single group that must be unprotonated for catalysis. This agrees well with the pKa of dimethylamine of 10.64[14]. The k_cat value also decreases from a maximum at high pH, but reaches a limiting value below pH 8.8 (Figure 1B); fitting these data to eq 3 yields a pK_a value of 9.9 ± 0.3 for the transition.

Figure 1.

Effect of pH on the (A) k_cat/K_m and (B) k_cat values for dimethylamine as a substrate for trimethylamine dehydrogenase at 25 °C. The lines are from fits of the data to eq 2 (A) or 3 (B).

$$\log \left({\mathrm{k}}_{\mathrm{cat}}/{\mathrm{K}}_{\mathrm{m}}\right)=\log \left[\mathrm{y}/\left(1+\left[\mathrm{H}\right]/{\mathrm{K}}_{\mathrm{a}}\right)\right]$$ (2)

$$\mathrm{y}=\left({\mathrm{k}}_{\mathrm{low}}+{\mathrm{k}}_{\mathrm{high}}{\mathrm{K}}_{\mathrm{a}}/\left[\mathrm{H}\right]\right)/\left(1+{\mathrm{K}}_{\mathrm{a}}/\left[\mathrm{H}\right]\right)$$ (3)

Initial rates were also determined with d₆-dimethylamine from pH 7.5–10.5. The isotope effect on the k_cat/K_m value, ^D(k_cat/K_m), has an average value of 8.6 ± 0.6 at pH 9 and above, decreasing to unity at lower pH (Figure 2A). Fitting the data to eq 4 yields a pK_a for the transition of 8.5 ± 0.2. The isotope effect on k_cat, ^Dk_cat, is also pH-dependent, with a value of 2.6 ± 0.2 at high pH that decreases to unity at low pH (Figure 2B). Fitting these data to eq 4 yields a pK_a value of 9.3 ± 0.2.

Figure 2.

Effect of pH on the deuterium isotope effects for dimethylamine as a substrate for trimethylamine dehydrogenase at 25 °C. The lines are from fits of the data to eq 4.

$$\mathrm{y}=\left(1+{\mathrm{kie}}_{\max }{\mathrm{K}}_{\mathrm{a}}/\left[\mathrm{H}\right]\right)/\left(1+{\mathrm{K}}_{\mathrm{a}}/\left[\mathrm{H}\right]\right)$$ (4)

Discussion

The most detailed previous analyses of the pH-dependence of TMDH have relied on rapid-reaction analyses of the kinetics of flavin reduction by amines. Such analyses yield k_lim and k_lim/K_d values. The former is the rate constant for flavin reduction at saturating substrate concentrations, while the latter is equivalent to the k_cat/K_m value in steady-state kinetics and similarly reflects the behavior of the interaction of the free enzyme and substrate through the first irreversible step. With trimethylamine as substrate, the k_lim/K_d pH profile has been described as bell-shaped, with the two pK_a values assigned as ~9.3 and ~10.0; the latter pK_a was assigned to the substrate, which has a pK_a of 9.8[11]. Similar bell-shaped k_lim/K_d pH profiles were reported for the H172Q[6] and Y174F variants[8]. In contrast, the Y169F variant yielded pK_a values of 9.7 and 11.0, although the higher value was not well-defined [7]. When deuterated trimethylamine was used as the substrate the pK_a values for the wild-type enzyme were 9.4 and 10.5; again the higher pK_a value was attributed to the substrate, which must then be converted to the neutral form in the active site before amine oxidation can occur[6]. The change in the pK_a value of ~0.5 was attributed to a change in the pK_a of the amine upon deuteration[5]. The H174F variant gave an identical result with the deuterated substrate, but in that case it was concluded that the neutral form binds[8]. Finally, the pK_a values for the H172Q variant were reported as 11.3 and 9.7 with deuterated trimethylamine, with the latter value assigned to the substrate[9]. With the slower substrate diethylmethylamine (pK_a = 10.3) Rohlfs and Hille[15] reported that the k_lim/K_d pH profile decreased with decreasing pH with a pK_a value of 8.7. This value was attributed to an active site residue, with the form of the substrate that binds the enzyme again assigned as the ammonium form. In contrast, Jang et al. subsequently described the k_lim/K_d pH profile as bell-shaped with pK_a values of ~9.4 and ~ 9.7[11] and attributed the higher pK_a value to the substrate.

When the two pK_a values in a bell-shaped profile are within ≤ 1 unit of each other, they cannot successfully be resolved with data of typical precision [12, 16]. It is not even valid in such cases to assign the lower pK_a to a moiety that must be unprotonated and the higher pK_a to one that must be protonated, since the alternate assignment will yield an identical profile. In all cases in which a bell-shaped profile has been described for TMDH, the two pK_a values are too close together to separate or even to confidently assign one to a moiety that must be protonated and one that must be unprotonated. While the shift in the basic pK_a value of 0.5 upon deuteration was attributed to a change in the pK_a of the substrate, deuteration increases the pK_a of trimethylamine by less than 0.1[17] rather than by 0.5–0.8. Fast substrates such as trimethylamine frequently have high commitments to catalysis; this will perturb the observed pK_a value(s) from their intrinsic value. The use of a deuterated substrate can result in a kinetic isotope effect that will decrease the commitment and the resulting perturbation in the pK_a value. This likely explains the change in the pK_a values in the k_lim/K_d pH profile for TMDH when deuterated trimethylamine is used as a substrate.

The use of a slow substrate such as dimethylamine avoids complications due to a high commitment to catalysis, allowing the intrinsic pK_a to be seen in the k_cat/K_a pH profile. The data in Figure 1A shows a pK_a for a group that must be unprotonated in the free enzyme or substrate with a value that matches that of the substrate. This is clear evidence for the neutral amine being the form that binds. This is consistent with the precedent from other amine-oxidizing flavoproteins[18, 19] and the conclusion of Basran et al.[8]

For a substrate such as dimethylamine that has a low commitment to catalysis, the deuterium isotope effect on the k_cat/K_amine value should be pH-independent. If there is an external commitment, the intrinsic isotope effect will only seen off the pH optimum[20]. For example, in the case of the flavoprotein D-amino acid oxidase, for which the substrate nitrogen must be neutral for oxidation, the ^D(k_cat/K_amine) value with the sticky substrate D-alanine is 1.3 at pH 9.5, increasing to the intrinsic value of 5.1 at pH 4.[21] In contrast, the ^D(k_cat/K_amine) value with TMDH shows the opposite pattern, with an average value of 8.6 at high pH that is likely to be the intrinsic isotope effect that decreases to one below an apparent pK_a value of 8.5. One explanation for the pH behavior of the ^D(k_cat/K_amine) value for TMDH would be the presence of a pH-dependent commitment that increases at low pH, since the intrinsic isotope effect is unlikely to be pH-dependent. This could occur if the predominant pathway at low pH involves binding of the charged amine followed by loss of a proton from the enzyme-bound substrate prior to oxidation. If the oxidation of the neutral amine is significantly faster that its reprotonation, the resulting commitment will decrease or eliminate the observed isotope effect. The proton acceptor is most likely an active-site residue, since the increasing concentration of the hydronium ion as the pH decreases would result in an increase in the rate of amine deprotonation, while the rate constant for deprotonation by an active-site residue would only depend on its own protonation state. A reasonable candidate for the base is His 172, one of two pK_a values seen in the k_lim profiles for TMDH with trimethylamine as the substrate[6].

The k_cat pH profile (Figure 1B) is sigmoidal, decreasing from a maximum at high pH to a smaller value at low pH[11]. Similar results have been obtained for the k_lim pH profile with a number of active site variants of the enzyme[6–8]. The ^Dk_cat value similarly decreases from a maximum at pH ≥ 10 to a smaller value at lower pH, with a pK_a value very close to that seen in the k_cat profile. Both of these results are consistent with a change in the identity of the rate-limiting step in the reaction as the pH decreases. Rohlfs and Hille[15] have reported that the rate constant for intramolecular electron transfer from the reduced flavin is pH-dependent with the substrate diethylmethylamine, decreasing below a pK_a value of 7.3. This provides a reasonable explanation for the pH-dependence of the k_cat value. Above pH 10, where both the ^Dk_cat and ^D(k_cat/K_m) values are at their maxima, the relative values of the two isotope effects establishes that amine oxidation is ~4 times faster than a subsequent step or steps.

The mechanism of oxidation of dimethylamine by flavins has been examined by ab initio computational analyses. The results indicated that a mechanism involving direct hydride transfer from the neutral amine to the flavin would be of significantly lower energy than either a radical or nucleophilic reaction [22]. This is consistent with the need for presence of the neutral amine in the active site of TMDH prior to substrate oxidation and the overwhelming evidence for hydride transfer as the mechanism for the majority of amine-oxidizing flavoproteins. However, there does not appear to have been significant consideration of this possibility for the TMDH family [5].