Steady-state kinetic mechanism of LodA, a novel cysteine tryptophylquinone-dependent oxidase

Abstract

LodA is a novel lysine-ε-oxidase which possesses a cysteine tryptophylquinone cofactor. It is the first tryptophylquinone enzyme known to function as an oxidase. A steady-state kinetic analysis shows that LodA obeys a ping-pong kinetic mechanism with values of k_cat of 0.22±0.04 s⁻¹, K_lysine of 3.2±0.5 µM and K_O2 of 37.2±6.1 µM. The k_cat exhibited a pH optimum at 7.5 while k_cat/K_lysine peaked at 7.0 and remained constant to pH 8.5. Alternative electron acceptors could not effectively substitute for O₂ in the reaction. A mechanism for the reductive half reaction of LodA is proposed that is consistent with the ping-pong kinetics.

1. Introduction

LodA is a novel lysine-ε-oxidase which was recently found in the melanogenic marine bacterium, Marinomonas mediterranea [1–3]. This enzyme was originally called marinocine but is now referred to as LodA, the gene product of lodA. LodA exhibits antimicrobial properties in vivo that result from its production of H₂O₂. It is secreted to the external medium in the biofilms in which the bacterium resides. This causes death of a subpopulation of cells within the M. mediterranea biofilm that is accompanied with differentiation, dispersal and phenotypic variation among dispersal cells [4].

Typical amino acid oxidases utilize a flavin cofactor for catalysis and act upon the α-amino group. LodA removes the ε-amino group of lysine and the recent crystal structure of LodA reveals that it contains a cysteine tryptophylquinone (CTQ, Figure 1) cofactor [5]. CTQ is a protein-derived cofactor [6], which is generated by the posttranslational modification of cysteine 516 and tryptophan 581 of LodA. Analysis of LodA by mass spectrometry yields a molecular mass that is also consistent with the presence of CTQ [7]. Another protein, LodB, is required for the posttranslational formation of CTQ [7,8]. All other tryptophylquinone cofactor-containing enzymes are dehydrogenases that use other redox proteins as their electron acceptors [9]; this is the first time that one has been shown to function as an oxidase. The best known tryptophylquinone enzymes possess tryptophan tryptophylquinone (TTQ, Fig. 1) where the modified Trp is cross-linked to another Trp rather than a Cys [10]. The known TTQ enzymes are primary amine dehydrogenases [6,11]. The one other CTQ-dependent enzyme that has been characterized is a quinohemoprotein amine dehydrogenase (QHNDH) which possesses two covalently attached hemes in addition to CTQ [12,13].

Figure 1.

Tryptophylquine cofactors that are formed by posttranslational modifications. CTQ is cysteine tryptophylquinone. TTQ is tryptophan tryptophylquinone. TPQ is 2,4,5-trihydroxyphenylalanine quinone or topaquinone. LTQ is lysine tryptophylquinone.

Another class of protein-derived cofactors that contain quinones derived from Tyr residues has been identified [14]. These possess the topaquinone (TPQ, Fig. 1) cofactor and are found only in primary amine oxidases. In each case, these enzymes also possess a tightly bound copper in the active site which is required for biogenesis of the cofactor and subsequently participates in catalysis. A related cofactor is lysine tyrosylquinone (LTQ) which is the catalytic center of mammalian lysyl oxidase [15]. This enzyme is also a lysine ε-oxidase but it acts only on lysyl residues of a protein substrate rather than free lysine [16]. Like the TPQ-dependent enzymes, lysyl oxidase also possesses a tightly-bound copper in the active site [17]. LodA is the first example of an amino acid or primary amine oxidase that contains neither a flavin nor a metal at its active site.

This paper presents a steady-state kinetic analysis of the reaction that is catalyzed by LodA (eq 1) to determine its steady-state reaction mechanism and kinetic parameters. The results are used to propose a reaction mechanism which is consistent with the crystal structures of a lysine-bound adduct of LodA [5].

$$\mathrm{L}-\text{lysine}+{\mathrm{O}}_2+{\mathrm{H}}_2\mathrm{O}\to 2-\text{aminoadipate}6-\text{semialdehyde}+\mathrm{N}{\mathrm{H}}_3+{\mathrm{H}}_2\mathrm{O}$$ (1)

2. Materials and Methods

Recombinant LodA was expressed in E.coli Rosetta cells which had been transformed with pET15LODAB [7] which contains lodA with an attached 6xHis tag and lodB which is required for the posttranslational modification of LodA that forms CTQ. The cells were cultured in LB media with ampicillin and chloramphenicol. Expression levels of active LodA are very sensitive to induction conditions and previously determined optimal conditions [7] were followed. When the absorbance of the culture reached a value of 0.6, the temperature was decreased to15 °C and cells were induced by addition of 1 mM IPTG. Cells were harvested after 2 hr and then broken by sonication in 50 mM potassium phosphate buffer (KPi), pH7.5. The soluble extract was applied to a Nickel-NTA affinity column and the His-tagged LodA was eluted using 100–120 mM imidazole in 50 mM KPi, pH 7.5. LodA was then further purified by ion exchange chromatography with DEAE cellulose in 50 mM KPi, pH 7.5. The His-tagged LodA bound to the resin and was eluted using 180–270 mM NaCl in the same buffer. The protein was judged pure by SDS-PAGE and its UV-visible absorption spectrum exhibited a broad peak centered at 400 nm with an absorbance at 400 nm which was about 13-fold less intense than the absorbance at 280 nm. Methylamine dehydrogenase (MADH) [18], amicyanin [19] and cytochrome c-550 [20] were purified from Paracoccus denitrificans as described previously.

LodA activity had previously been assayed using a standard coupled assay for oxidases in which horseradish peroxidase (HRP) is also present and uses the H₂O₂ product of the oxidase reaction to oxidize Amplex Red to resorufin which leads to a change in fluorescence [21]. This assay was modified to instead follow the change in absorbance associated with production of resorufin which has an extinction coefficient at of 54000 M⁻¹ cm⁻¹ at its absorbance maximum of 570 nm. The assay mixture contained 0.05 mM Amplex Red, 0.1U/ml of HRP and 0.4 µM LodA in 50 mM KPi, pH 7.5 at 25 °C. The concentration of LodA was determined using an ε₂₈₀=125,180 M⁻¹cm⁻¹ which was calculated from its amino acid sequence [22]. Experiments were performed in which [lysine] was varied in the presence of fixed concentrations [O₂] and vice versa. To vary the concentration of O₂, an appropriate amount of a stock solution of air-saturated buffer ([O₂]=252 µM) was mixed with the reaction mixture which had been made anaerobic by repeated cycles of vacuum and purging with argon. The reactions were initiated by the addition of lysine. In order to account for a small background reaction in which Amplex Red is spontaneously converted to resorufin, control experiments were performed in the absence of LodA to determine the background rate which was subtracted from the experimentally-determined LodA-dependent reaction rates. The reaction was also assayed by monitoring the release of the ammonia product using a coupled assay. The assay mixture contained 20 U/ml glutamate dehydrogenase, 5 mM α-ketoglutarate, 0.4 µM LodA and 93.7 µM NADH in 50 mM KPi, pH 7.5 at 25 °C and the reactions were initiated by the addition of lysine. In assays to test the reactivity of artificial electron acceptors the assay mixture contained 0.4 µM LodA in 50 mM KPi, pH 7.5 at 25 °C under anaerobic conditions. The reaction was initiated by addition of lysine and monitored by the spectral change associated with the reduction of each electron acceptor, phenazine ethosulfate (PES) plus 2,6-dichloroinophenol (DCIP) (ε₆₀₀=21,500 M⁻¹cm⁻¹), ferricyanide (ε₄₂₀=1,040 M⁻¹cm⁻¹), or NAD⁺ (ε₃₄₀=6,220 M⁻¹cm⁻¹), amicyanin (ε₅₉₅=4,600 M⁻¹cm⁻¹), cytochrome c-550 (ε₅₅₀=30,200 M⁻¹cm⁻¹).

3. Results and Discussion

3.1. Steady-state reaction mechanism and parameters

Steady-state assays of lysine oxidation by LodA were performed in order to determine the kinetic mechanism of this reaction and the kinetic parameters that describe this reaction. The initial rates of reaction were determined at different concentrations of lysine in the presence of a set of fixed concentrations of O₂ (Fig. 2A). Double reciprocal plots of these data exhibited a set of lines which are clearly more parallel than intersecting (Fig. 2B). Similarly, when the initial rates of reaction were determined at different concentrations of O₂ in the presence of a set of fixed concentrations of lysine a set of parallel lines was obtained in the reciprocal plots. These data are consistent with a ping-pong kinetic mechanism rather than an ordered or random sequential mechanism and thus the reaction is described by eq 2 [23]. Secondary plots of the intercept (1/k_cat apparent) versus 1/[O₂] (Fig. 2C), intercept/slope (1/K_lysine apparent) versus 1/[O₂] (Fig. 2D), intercept (1/k_cat apparent) versus 1/[lysine] (Fig. 2E) and intercept/slope (1/K_O2 apparent) versus 1/[lysine] (Fig. 2E) were each linear. These plots are described by eqs 3 and 4 [23] where A is the varied substrate and B is the fixed substrate. Analysis of these data yielded values of k_cat of 0.22±0.04 s⁻¹, K_lysine of 3.2±0.5 µM and K_O2 of 37.2±6.1 µM.

$$\mathrm{v}/\mathrm{E}=\frac{{\mathrm{k}}_{\text{cat}}[\text{lysine}][{\mathrm{O}}_2]}{{\mathrm{K}}_{\text{lysine}}[{\mathrm{O}}_2]+{\mathrm{K}}_{{\mathrm{O}}_2}[\text{lysine}]+[\text{lysine}][{\mathrm{O}}_2]}$$ (2)

$$\text{intercept}=({\mathrm{K}}_{\mathrm{B}}/{\mathrm{k}}_{\text{cat}})(1/[\mathrm{B}])+(1/{\mathrm{k}}_{\text{cat}})$$ (3)

$$\text{intercept}/\text{slope}=({\mathrm{K}}_{\mathrm{B}}/{\mathrm{K}}_{\mathrm{A}})(1/[\mathrm{B}])+(1/{\mathrm{K}}_{\mathrm{A}})$$ (4)

Figure 2.

Steady-state kinetic analysis of LodA. A. Initial rates were determined at different concentrations of lysine in the presence of a fixed concentration of 252 (○), 129 (♦), 54 (▼), 11 (▲) or 5 (●) µM O₂. B. Double reciprocal plot of the data shown in panel A. C. Secondary plots of the y-intercept (1/apparent k_cat) from panel B versus 1/[O₂]. Data are fit by eq 3. D. Secondary plots of the y-intercept/slope (1/apparent K_lysine) from panel B versus 1/[O₂]. Data are fit by eq 4. E. Secondary plots of the y-intercept (1/apparent k_cat) from plots of v/E versus [O₂] against 1/[lysine]. Data are fit by eq 3. F. Secondary plots of the y-intercept/slope (1/apparent K_O2) from plots of v/E versus [O₂] against 1/[lysine]. Data are fit by eq 4. Error bars are indicated in the secondary plots (C–F).

This reaction was also studied using another coupled enzyme assay in which the rate of formation of the ammonia product was coupled to NADH oxidation by glutamate dehydrogenase in the presence of α-ketoglutarate. When the assay was performed using saturating concentrations of 100 µM lysine and 252 µM O₂ the rate of reaction was 0.25±0.07 s⁻¹ which is within experimental error of the k_cat value obtained in the assay which monitored H₂O₂ production.

3.2. Dependence of LodA activity on pH

The pH dependence of k_cat and k_cat/K_m for lysine were determined at 252 µM O₂ (Fig. 3). The k_cat value exhibited a bell-shaped pH profile with a maximum at pH 7.5. This suggests that there are at least two residues that are required and that one must be protonated and the other must be unprotonated for catalysis. In contrast the k_cat/K_m values reached a maximum at pH 7.0 and stayed relatively constant up to pH 8.5. Thus, the catalytic efficiency of LodA is retained at the higher pH as a consequence of a decrease in K_lysine.

Figure 3.

Dependence on pH of k_cat (A) and k_cat/K_m for lysine at saturating concentration of O₂. Error bars are indicated.

3.3. Reactivity towards alternative electron acceptors

The distinguishing property between an oxidase and a dehydrogenase is that the oxidases use O₂ as their final electron acceptor whereas dehydrogenases use other redox proteins or non-protein electron acceptors other than O₂. Since all other know tryptophylquinone enzymes are dehydrogenases the ability of LodA to function as a dehydrogenase was tested using alternative electron acceptors. The most widely used assay for quinoprotein dehydrogenases uses PES coupled to DCIP. For comparison the assay was also performed on MADH which possesses the TTQ cofactor. When assayed in the presence of saturating methylamine with 5 mM PES and 100 µM DCIP MADH exhibited a rate of 36 s⁻¹ while LodA when assayed in the presence of saturating lysine under these conditions exhibited a rate of 0.06 s⁻¹. Another commonly used artificial electron acceptor is ferricyanide. When similarly assayed in the presence of 1 mM potassium ferricyanide with saturating lysine LodA exhibited a rate of 0.10 s⁻¹. For a more direct comparison of the efficiency of the artificial electron acceptors, the assays were repeated using 252 µM PES or ferricyanide, a concentration equivalent to that used for O₂ (K_O2 = 37.2±6.1 µM). At this concentration the rates with PES/DCIP and ferricyanide were each 0.01 s⁻¹ which is barely above background, compared 0.22±0.04 s⁻¹ for O₂. NAD⁺ (1 mM) was also tested in the presence of 100 µM lysine and no detectable reaction was observed. Two protein electron acceptors were also tested. The copper protein amicyanin [19] is the natural electron acceptor for MADH. In the steady state reaction of methylamine dependent amicyanin reduction by MADH, amicyanin exhibits a k_cat of 48 s⁻¹ and K_m of 6 µM [24]. When LodA was assayed with 100 µM amicyanin and saturating lysine the rate was only 0.008 s⁻¹. Cytochrome c-550 from P. denitrificans is a relatively promiscuous electron acceptor that shares structural features with many bacterial and mammalian cytochromes c [25]. When LodA was assayed with 100 µM cytochrome and saturating lysine the rate was 0.002 s⁻¹.

One cannot rule out the possibility that in vivo there is an unidentified redox-active molecule or protein that could serve as an electron acceptor for LodA instead of O₂. However, no significant dehydrogenase activity of LodA could be detected in this study. These results support the proposal that the generation of H₂O₂ and consequent antimicrobial activity [1] is the main physiological function of LodA, which is secreted from cell to perform this function.

3.4. Proposed mechanism

The ping-pong kinetic mechanism requires that an intermediate form of the enzyme exist after binding of substrate and release of the first product prior to binding of the second substrate. This finding is consistent with the reported crystal structure of LodA with a covalent lysine adduct with CTQ [5] and suggests the following reaction mechanism (Fig. 4). The ε-amino group of lysine must first be deprotonated, likely by an active site residue, so that the resulting nucleophilic NH₂ group can attack the electrophilic carbonyl carbon to displace the bound oxygen and yield a Schiff base intermediate, as seen in the crystal structure. An active site residue then abstracts a proton from the ε-carbon reducing the cofactor and resulting in the Schiff base now involving the ε-carbon. Inspection of the crystal structure of the LodA-lysine adduct [5] reveals that the S of Cys448 is approximately 5.5 Å from this carbon and could possibly be involved in this step. This covalent adduct is then hydrolyzed to release the aldehyde product prior to O₂ binding. This yields an aminoquinol intermediate form of the enzyme consistent with a ping-pong mechanism. This proposed mechanism for this reductive half-reaction is essentially identical to those of QHNDH [26] and MADH [24], which possess CTQ and TTQ, respectively. It is also the same mechanism proposed in the reductive half-reactions of the TPQ-dependent amine oxidases [27]. With the TTQ-dependent MADH it was shown that the ammonia product was not released until after the quinol was oxidized to the quinone by its electron acceptor, and that in the steady-state this occurred by displacement of the amino group from the cofactor by the next amine substrate molecule [26]. This may be true for LodA as well. Alternatively, it could be hydrolyzed by water. A key question which remains unanswered is how the quinol or aminoquinol in LodA is reoxidized by O₂. In the TPQ-dependent oxidases Cu²⁺ is also present at the active site. It has been proposed that this interacts with the quinol intermediate to yield a Cu⁺-semiquinone intermediate which would allow reaction of the Cu⁺ with oxygen to generate a superoxide species in the first step of the reoxidation [28,29]. LodA, however, does not have a bound metal. Future characterization of the mechanism of the oxidative half-reaction of LodA should be of great interest.

Figure 4.

Proposed mechanism for LodA-catalyzed oxidative deamination of lysine. B represents an active site residue which participates in acid-base chemistry.

Highlights.

LodA has a tryptophylquinone cofactor but functions as an oxidase not a dehydrogenase

LodA exhibits a ping-pong kinetic mechanism with lysine and O₂ as substrates

The kinetic mechanism of LodA is consistent with a covalent lysine-CTQ intermediate