Lipids (2016) 51:335–347 DOI 10.1007/s11745-016-4127-z

ORIGINAL ARTICLE

Kinetics of Bis‑Allylic Hydroperoxide Synthesis in the Iron‑Containing Lipoxygenase 2 from Cyanothece and the Effects of Manganese Substitution Julia Newie1 · Müge Kasanmascheff2,3 · Marina Bennati2,3 · Ivo Feussner1,4,5 

Received: 13 November 2015 / Accepted: 19 January 2016 / Published online: 2 February 2016 © AOCS 2016

Abstract  Lipoxygenases (LOX) catalyze the regio- and stereospecific insertion of dioxygen into polyunsaturated fatty acids. While the catalytic metal of LOX is typically a non-heme iron, some fungal LOX contain manganese as catalytic metal (MnLOX). In general, LOX insert dioxygen at C9 or C13 of linoleic acid leading to the formation of conjugated hydroperoxides. MnLOX (EC 1.13.11.45), however, catalyze the oxygen insertion also at C11, resulting in bis-allylic hydroperoxides. Interestingly, the ironcontaining CspLOX2 (EC 1.13.11.B6) from Cyanothece PCC8801 also produces bis-allylic hydroperoxides. What role the catalytic metal plays and how this unusual

Electronic supplementary material  The online version of this article (doi:10.1007/s11745-016-4127-z) contains supplementary material, which is available to authorized users. * Ivo Feussner ifeussn@uni‑goettingen.de 1

Department of Plant Biochemistry, Albrecht-von-Haller Institute for Plant Sciences, Georg-August-University, Justus‑von‑Liebig Weg 11, 37077 Göttingen, Germany

2

Max Planck Institute for Biophysical Chemistry, Electron Paramagnetic Resonance Spectroscopy Group, Am Fassberg 11, 37077 Göttingen, Germany





3

Institute for Organic and Biomolecular Chemistry, Georg-August-University, Tammanstrasse 4, 37077 Göttingen, Germany

4

Department of Plant Biochemistry, Goettingen Center for Molecular Biosciences (GZMB), Georg-AugustUniversity, Justus‑von‑Liebig Weg 11, 37077 Göttingen, Germany

5

Department of Plant Biochemistry, Goettingen International Center for Advanced Studies of Energy Conversion (ICASEC), Georg-August-University, Justus‑von‑Liebig Weg 11, 37077 Göttingen, Germany







reaction is catalyzed by either MnLOX or CspLOX2 is not understood. Our findings suggest that only iron is the catalytically active metal in CspLOX2. The enzyme loses its catalytic activity almost completely when iron is substituted with manganese, suggesting that the catalytic metal is not interchangeable. Using kinetic and spectroscopic approaches, we further found that first a mixture of bis-allylic and conjugated hydroperoxy products is formed. This is followed by the isomerization of the bisallylic product to conjugated products at a slower rate. These results suggest that MnLOX and CspLOX2 share a very similar reaction mechanism and that LOX with a Fe or Mn cofactor have the potential to form bis-allylic products. Therefore, steric factors are probably responsible for this unusual specificity. As CspLOX2 is the LOX with the highest proportion of the bis-allylic product known so far, it will be an ideal candidate for further structural analysis to understand the molecular basis of the formation of bisallylic hydroperoxides. Keywords  Cyanobacteria · Electron paramagnetic resonance spectroscopy · Enzyme kinetics · Lipid peroxidation · Metalloenzyme · Octadecanoid pathway Abbreviations H(P)ODE Hydroperoxy-octadecadienoic acid(s) ICP-AES Inductively coupled plasma atomic emission spectroscopy kcat Turnover number Km Michaelis Menten constant LOX Lipoxygenase(s) MnLOX Manganese lipoxygenase(s) SDS Sodium dodecylsulfate

13

336

Lipids (2016) 51:335–347

Introduction Lipoxygenases (LOX, EC 1.13.11.-) catalyze the stereoand regiospecific insertion of molecular oxygen into polyunsaturated fatty acids that contain one or more (1Z,4Z)pentadiene systems [1]. These non-heme dioxygenases are widely distributed in plants and mammals, but have also been found in a number of fungi and cyanobacteria [2–4]. The hydroperoxy fatty acid products of the LOX reaction are precursors for a variety of biological mediators, including aldehydes and jasmonates in plants and leukotrienes, resolvins and lipoxins in mammals. These signaling molecules are of biological and medical interest, as they play important roles in defense mechanisms and developmental processes in plants and inflammation and cancer development in mammals [5, 6]. While almost all LOX contain iron as catalytic metal in the active site, some fungal LOX have been characterized as manganese-containing LOX (MnLOX). These MnLOX include the 13R-MnLOX from Gaeumannomyces graminis [7], 9S-MnLOX from Magnaporthe salvinii [8], Fo-MnLOX from Fusarium oxysporum and Cg-MnLOX from Colletotrichum gloeosporioides [9]. All of these LOX originate from fungi, however, not all fungal LOX contain manganese in the active site, as FoxLOX from F. oxysporum has been characterized as iron-containing LOX [10]. During the LOX reaction, the inactive Fe(II) is first oxidized to the active Fe(III), or in MnLOX from Mn(II) to Mn(III) [11, 12]. Hydrogen is then abstracted from the central carbon atom C3 of the (1Z,4Z)-pentadiene system to yield a carbon centered radical that is likely delocalized over all five carbon atoms of the pentadiene. Usually, dioxygen is then inserted at position n + 2 or n − 2 relative to the hydrogen abstraction to form conjugated hydroperoxides [1]. As exception to this, all of the characterized MnLOX and one iron-containing LOX from the cyanobacterium Cyanothece (CspLOX2) have been shown to insert dioxygen also at the central position of the pentadiene system to form considerable amounts of a bis-allylic hydroperoxide (Fig. 1) [7, 8, 13, 14]. Of all LOX with bis-allylic products, 13R-MnLOX is the best characterized enzyme. This enzyme not only produces the bis-allylic 11S-hydroperoxy octadecadienoic acid (11S-HPODE) from linoleic acid, but also isomerizes it subsequently to 13R-HPODE [7]. In contrast to other LOX reactions, dioxygen is inserted from the same side from which hydrogen was initially abstracted, in a so-called suprafacial way [7]. Experiments with 18O labeled 11S-HPODE have further suggested that the isomerization of 11S-HPODE to 13R-HPODE occurs via deoxygenation into a linoleoyl radical and subsequent reoxygenation [7, 15]. The cyanobacterial LOX CspLOX2 is so far the only known iron-containing LOX that produces considerable

13

abstraction of the pro-S hydrogen H

H

9

13 H

9 suprafacial oxygen insertion: MnLOX HOO

13 antarafacial oxygen insertion: CspLOX2

H

H

9

13 11S-HPODE

HOO

linoleyl radical

OOH

9

13 11R-HPODE

isomerization to conjugated HPODE (9- and 13-HPODE)

OOH

Fig. 1  Reaction of MnLOX and CspLOX2 in comparison. Both enzymes produce the bis-allylic 11-HPODE, but while MnLOX inserts oxygen in a suprafacial way, CspLOX2 catalyzes the oxygenation in an antarafacial way. Finally, both enzymes isomerize 11-HPODE to the conjugated products

amounts of the bis-allylic 11-HPODE [13]. Although the coral 8R-LOX has also been shown to produce minor amounts of the bis-allylic 10-hydroperoxy eicosatetraenoic acid (~5 %) from arachidonic acid [16], CspLOX2 produces 11-HPODE from linoleic acid as major product [13]. In contrast to the MnLOX reaction, CspLOX2 inserts oxygen in an antarafacial way, i.e. at the opposite side of the hydrogen abstraction (Fig. 1). But similar to MnLOX, CspLOX2 also isomerizes 11-HPODE to conjugated hydroperoxides [13]. In order to understand which factors influence the formation of the bis-allylic product and its isomerization to the conjugated hydroperoxides, we focused on the biphasic reaction and the role of the metal cofactor. We studied the individual reaction steps and kinetics of the CspLOX2 reaction in more detail. Interestingly, we found a number

Lipids (2016) 51:335–347

of similarities between the MnLOX and CspLOX2 reaction, suggesting that both reactions are catalyzed in a very similar way and thus a similar molecular basis might be responsible for 11-HPODE formation in MnLOX and CspLOX2. So far, it was assumed that both metal cofactors function in a similar way, as the metal-coordinating residues are largely conserved in iron- and manganesecontaining enzymes. By substituting the catalytic iron with manganese, we found that both catalytic metals are not simply interchangeable in CspLOX2. These results suggest that not the type of metal cofactor, but instead structural factors of the amino acid environment are important for 11-HPODE formation. As CspLOX2 produces the highest relative amount of 11-HPODE, it will be a good model for elucidating these structural factors involved in formation of bis-allylic products.

Materials and methods Chemicals Chemicals were obtained from Sigma and Carl Roth & Co. Linoleic acid (LA) was from Sigma. Acetonitrile was from Fisher Scientific.

337

equal to the 11-HPODE amount utilized, we could assign a concentration for the solution. To measure the formation of all HPODE, an oxygen electrode (Oxygraph, Hansatech Instruments) was applied with same buffers and conditions. Preparation of 11‑HPODE For the preparation of 11-HPODE, 1 mg linoleic acid was converted with 15 µg CspLOX2 in 200 mM sodium borate buffer pH 9 at 25 °C for 8 min. The reaction was stopped by addition of diethylether and cooling it down on ice. The hydroperoxides were extracted twice with diethylether. From the mixture of hydroperoxides, 11-HPODE was separated from 9- and 13-HPODE by RP-HPLC using a EC250/2 Nucleosil 120–5 C18 column (250 × 2.1 mm, 5 µm particle size, Macherey–Nagel) as described previously [13]. After evaporation of the solvent, 11-HPODE was dissolved in 200 mM sodium borate buffer pH 9 for isomerization experiments. For the direct quantification of [1-14C]-labeled substrates and products, compounds eluting from the RP-HPLC were detected by a Raytest radiodetector. Of the detected substances, 11-HPODE elutes first, followed by the coelution of 9- and 13-HPODE and subsequently linoleic acid. SP‑HPLC Analysis of 11‑HPODE Isomerization

Expression and purification of LOX The LOX CspLOX1 (WP_012595715.1), CspLOX2 (WP_012596348.1), sLOX1 (NP_001236153.1), FoxLOX (EXK38530), CsLbLOX (NP_001292659.1) and AtLOX2 (NP_566875.1) were heterologously produced in E. coli BL21 star (Thermo Scientific). The proteins were purified via the N-terminal His-Tag by affinity chromatography. Kinetic Analysis of Formation of Conjugated HPODE at 234 nm and with an Oxygen Electrode The formation of conjugated double bonds was monitored in a UV/vis spectrophotometer (Varian Cary 100 Bio) at 234 nm. This allows the detection of 9- and 13-HPODE formation from linoleic acid, while 11-HPODE cannot be detected due to the lack of conjugated double bonds. The measurement was performed at 30 °C in 200 mM sodium borate buffer pH 9 with linoleic acid concentrations ranging from 5 to 100 µM. The reaction was initiated by addition of 15 µg CspLOX2. For the calculation of kcat, the extinction coefficient of 25,000 M−1 cm−1 was used for conjugated hydroperoxy fatty acids [17]. In order to determine the concentration of 11-HPODE to measure kinetics, we fully converted low amounts of 11-HPODE with CspLOX2 and calculated from the difference in absorbance at 234 nm the amount of conjugated HPODE produced. As this should be

11-HPODE was incubated with 600 U (as activity of linoleic acid oxidation) of different LOX. The reaction was incubated for 15 min at 23 °C and the products formed were extracted with diethylether and directly subjected to SPHPLC on a Zorbax Rx-SIL column (150 × 2.1 mm, 5 µm particle size, Agilent) with a solvent system of n-hexane/2propanol/acetic acid (100:1:0.05, by vol.) at a flow rate of 0.2 ml/min. 11-HPODE was quantified as peak area at 202 nm and the conjugated HPODE (9- and 13-HPODE) at 234 nm. Because of the different extinction coefficients, the relative scaling factors were assigned by the ratio of absorption at the respective wavelengths and the radioactive response of [1-14C]-labeled compounds measured by HPLC (Supplemental Fig. 1). At least 6 HPLC runs were evaluated. As a result, the following factors were determined: 0.79 for 9-HPODE at 234 nm, 0.64 for 13-HPODE at 234 nm and 1.57 for 11-HPODE at 202 nm. Determination of the Metal Cofactor The metal content of the enzymes was determined by inductively coupled plasma atomic emission spectroscopy (ICP-AES) using an Optima 5500 DV (Perkin Elmer Precisely) as described before [10]. To obtain a manganeseand an iron-containing version of CspLOX2, cells expressing CspLOX2 were grown in the minimal auto-induction

13

338

Lipids (2016) 51:335–347

Results

products, we added another 100 µM of linoleic acid substrate when the second slower phase of the reaction was reached. The restart of a fast reaction similar to the initial one suggests that a slowing down of the reaction is caused by a depletion of the substrate (Fig. 2c). We hypothesized that the specific features of the CspLOX2 progression curve might be related to the formation of the bis-allylic 11-HPODE from linoleic acid. First it seemed possible that the long lag phase detected at 234 nm is a result of high amounts of 11-HPODE being produced at the beginning of the reaction. This product does not absorb at 234 nm and is therefore invisible with this method. Furthermore, it seemed possible that the second small increase is a result of the isomerization of 11-HPODE to the conjugated products. To test these hypotheses, we analyzed the relative amounts of conjugated and bis-allylic HPODE at different stages of the reaction. We stopped the reaction of [1-14C]-labeled linoleic acid with CspLOX2 after different points and analyzed the amounts of the remaining substrate, 11-HPODE and the conjugated HPODE by radio RP-HPLC. The results strongly suggest that the substrate is completely used up after 5 min (Fig. 2d). At this point, 11-HPODE reaches its highest concentration and is then slowly declining again. The conjugated 9-HPODE and 13-HPODE, which cannot be separated by RP-HPLC, are first increasing quickly and then slightly, nicely resembling the progression curve observed at 234 nm (Fig. 2a). This suggests that during the first reaction phase the substrate is primarily converted into 11-HPODE. When the substrate is used up, an isomerization of the formed 11-HPODE to the conjugated products takes place, explaining the second slower reaction rate observed at 234 nm.

11‑HPODE Is only Transiently Formed

The Lag Phase is Removed by Addition of HPODE

In order to understand the formation of the bis-allylic product by CspLOX2 in more detail, we expressed the protein in E. coli and purified it by affinity chromatography as described before [13]. When following the reaction of CspLOX2 at 234 nm with a spectrophotometer, two characteristic features distinguish the CspLOX2 reaction from most other LOX reactions. First, the reaction has a long kinetic lag phase before the reaction starts. Second, the reaction comprises two linear phases, a first steep increase of absorbance and a following flat increase (Fig. 2a). In contrast to this, the reactions of sLOX1 from soybean, CsLbLOX from cucumber and FoxLOX from F. oxysporum have simply one initial linear phase and considerably shorter lag phases. Only CspLOX1, the isozyme from Cyanothece sp. [19] has a comparably long lag phase (Fig.  2b). To address the question, whether the transition from the first reaction to the second reaction rate is caused by a consumed substrate or by inhibitory effects of the

Since the lag phase might be an indication for an early formation of 11-HPODE, we also studied this phase of the reaction. It was shown before that the comparably short kinetic lag phase of the sLOX1 reaction could be removed, if reaction products, namely HPODE, were added. These HPODE probably oxidize the catalytic iron from the inactive Fe(II) to the active Fe(III) to initiate the reaction [20, 21]. To elucidate whether the long lag phase is simply caused by inactive Fe(II) as in other LOX, we added a small amount of reaction product to the CspLOX2 assay. As little as 1 µM of reaction products was enough to remove the lag phase completely and immediately start the increase at 234 nm upon addition of CspLOX2 (Fig. 3). Although the lag phase is remarkably longer in the CspLOX2 catalyzed reaction compared to sLOX1 and other LOX (Fig. 2b), the reason for the phenomenon seems to be the same: Activation of Fe(II) to Fe(III) by HPODE. As the lag phase is apparently not

medium PA-5052 [18] that was supplemented with either 50 µM MnCl2 or 50 µM FeCl3, but without further metals. The proteins were purified using His-Trap affinity columns (GE healthcare) [13] and the size exclusion chromatography column S200 pg 26/60 (GE-Healthcare). Size exclusion chromatography was performed in buffer (50 mM Tris/HCl pH 8, 100 mM NaCl) treated with Chelex100 (BioRad). For the determination of the metal content, proteins samples were diluted in the same buffer to a final concentration of 5–25 µM. The analysis was performed in triplicate. Electron Paramagnetic Resonance (EPR) Spectroscopy The 9.4 GHz (X-band) continuous wave (CW)-EPR spectra were recorded on a Bruker Elexsys E500 CW-EPR spectrometer equipped with a standard Bruker X-band ER4119-SHQE cavity and a liquid helium cryostat (Oxford Instruments). The substituted versions of CspLOX2 were mixed with linoleic acid to final concentrations of 30 µM enzyme and 700 µM substrate, respectively. Furthermore, CspLOX2 and HPODE were mixed to a final concentration of 30 µM enzyme and 70 µM substrate, respectively. The reactions were started at room temperature and manually freeze-quenched after a minute in liquid N2. Experimental parameters for all recorded spectra were as follows: T  = 10 K, microwave power = 1 mW, conversion time  = 80 ms, modulation amplitude = 5 G, modulation frequency = 100 kHz, 40 min, signal averaging.

13

Lipids (2016) 51:335–347

absorbance at 234 nm

2.0

b

second reaction rate

first lag reaction phase rate

sLOX1 2.0

FoxLOX CspLOX1

absorbance at 234 nm

a

339

1.5 1.0 0.5

1.5

CspLOX2 CsLbLOX

1.0 0.5 0.0

0.0 0

2

4

6

8

10

12

0

14

2

4

time (min)

d

2.5

rel. abundance

c

1.0 0.8 relative abundance

absorbance at 234 nm

2.0 1.5 1.0 addition of 100 µM linoleic acid

0.5 0.0 2

4

6

8

10

8

10

time (min)

1.0

12

14

11

0.5

9 and 13 linoleic acid

0.0 0

0.6

5 10 time (min)

11-HPODE

0.4 9- and 13-HPODE

0.2

linoleic acid

0.0 0

6

time (min)

0

10

20

30

40

50

time (min)

Fig. 2  Phases of the CspLOX2 reaction. The LOX enzymes were heterologously expressed in E. coli and purified by affinity chromatography. The LOX activity assay was performed in 200 mM sodium borate buffer pH 9 containing 100 µM linoleic acid and started by adding LOX. a The CspLOX2 reaction shows 3 distinct phases at 234 nm. b These phases are not present in other LOX reactions. Different amounts of enzymes were used to achieve an increase in the same range. c Addition of new substrate during the second phase restarts the fast first phase of the CspLOX2 reaction. The same

amount of linoleic acid present at the start of the reaction (100 µM) was added during the second reaction phase. d Relative amounts of the [1-14C]-labeled substrate and products during the course of the CspLOX2 reaction. Samples were taken after different points, quickly extracted and analyzed by radio-RP-HPLC. The graph shows the progression of relative amounts. The values presented here are the mean values of two independent experiments. The inset shows the data in the time scale of 0–10 min for an easier comparison with the other panels

related to 11-HPODE formation, it should behave similar when the oxygen consumption is monitored. As one molecule of dioxygen is needed to form one molecule of hydroperoxy product, the peroxidation at all possible positions of linoleic acid (C9, C11 and C13) is detected. As expected, the long lag phase is present as well and is shortened upon activation with HPODE. Notably, the progression curve observed with the oxygen electrode shows only the first reaction phase. This is in good agreement with the assumption that the second rate is a result of the isomerization as this would not consume further oxygen.

Kinetics of the Isomerization Reaction We further addressed the question of whether the reaction of CspLOX2 proceeds as a sequence in which 11-HPODE needs to be formed first and is then converted to the conjugated hydroperoxides (Fig. 9a). We therefore investigated the kinetics of the isomerization reaction. We produced 11-HPODE enzymatically with CspLOX2 and purified it by RP-HPLC. Different concentrations of 11-HPODE were then incubated with CspLOX2 and the reaction was monitored at 234 nm (Fig. 4a). The isomerization reaction has a

13

340

Lipids (2016) 51:335–347

b

1.4

230

1.2

220 210

1.0 oxygen (nmol/ml)

absorbance at 234 nm

a

activated

0.8 0.6 0.4

not activated

0.2

200 190

170 160

0

not activated

180

activated

150 0

1

2

3

4

0

5

0.5

1.0

1.5

time (min)

Fig. 3  The lag phase is eliminated, when the enzyme is activated. a Effect of HPODE-dependent activation was observed at 234 nm. Freshly purified CspLOX2 was incubated with 100 µM linoleic acid and the reaction was monitored at 234 nm (not activated). If 1 µM

2.5

3.0

3.5

4.0

reaction product, i.e. HPODE, was added to the assay, the reaction starts directly without lag phase (activated). b The same effect can be observed when the oxygen decrease is monitored

a

b

1.4

5

1.2

4

1.0

reaction rate (s-1)

absorbance at 234 nm

2.0

time (min)

0.8 0.6 0.4

3

first reaction rate (linoleic acid) (kcat = 5.5 s-1, Km = 33.5 µM)

2 1

isomerization of 11-HPODE (kcat = 0.58 s-1, Km = 4.0 µM)

0

second reaction rate (linoleic acid) (kcat = 0.50 s-1)

0.2 0 0

2

4

6

8

10

12

14

16

time (min)

0

20

40

60

80

100

120

substrate (µM)

Fig. 4  Kinetics of the isomerization reaction. 11R-HPODE was enzymatically produced with CspLOX2 and purified by RP-HPLC. a 11-HPODE was incubated with CspLOX2. The progression curve at 234 nm shows only one phase and the reaction starts immediately without lag phase. b Plot of the Michaelis–Menten kinetics of the

isomerization reaction in comparison to the first and second reaction rate of the CspLOX2 reaction with linoleic acid. Different amounts of linoleic acid and 11-HPODE were used and the reaction rate measured at 234 nm was plotted against the substrate concentration

turnover number kcat of 0.58 s−1 and a Michaelis–Menten constant Km of 4-µM 11-HPODE. Interestingly, the second reaction rate of the linoleic acid conversion measured at 234 nm exhibits a kcat of 0.5 s−1 which is in the similar range as the directly measured isomerization reaction (Fig.  4b). This strongly suggests that the second reaction rate of linoleic acid conversion represents the isomerization reaction. Moreover, it is evident that the reaction cannot proceed in a sequence in which 11-HPODE is a mandatory

intermediate for the formation of 9- and 13-HPODE (Fig.  9a). Instead, a mixture of 9-, 11- and 13-HPODE has to be formed during the first reaction phase (Fig. 9b). If conjugated HPODE would only result from the isomerization of 11-HPODE (Fig. 9a), it would be impossible to explain the fast initial reaction rate, in which conjugated hydroperoxides are formed at a 10-fold higher rate than the isomerization rate (first reaction rate: 5.5 s−1 vs isomerization: 0.58 s−1, Fig. 4).

13

Lipids (2016) 51:335–347

341

Rates of Bis‑Allylic and Conjugated HPODE Formation

Formation of 11‑HPODE and Isomerization of 11‑HPODE by Other LOX

To separately analyze the rate of 11-HPODE formation and the rate of conjugated HPODE formation, we compared the kinetics determined with the oxygen electrode with the kinetics measured at 234 nm with the spectrophotometer (Fig. 5). As the oxygen electrode detects all oxygenations and the photometer only detects the formation of conjugated products, the difference of both methods should reflect the velocity of the 11-HPODE formation. For CspLOX2, this difference is 23 s−1, as 28.5 s−1 was determined with the oxygen electrode and 5.5 s−1 with the photometer at 234 nm (Table 1; Fig.  4b). This implies that during the first reaction about 80 % of the catalytic activity account for the 11-HPODE formation, while the remaining 20 % constitute the formation of 9- and 13-HPODE. This is in agreement with results obtained from a full conversion of linoleic acid with CspLOX2: To quantify the portion of the first reaction relative to the second reaction, the increase in absorbance was monitored until a plateau was reached (Fig. 5a). When assuming that this end point of this reaction marks almost 100 % of the full conversion, then the transition of the first to the second reaction occurs at ~20 %. This is supported by the product ratio determined at the end of the first reaction by RP-HPLC with [1-14C]-labeled linoleic acid. The ratio of 75:25 of 11-HPODE to 9- and 13-HPODE was again close to 80:20 (Fig. 5b).

During our studies we found that 11-HPODE is not stable at low pH values as already reported by Hamberg et al. [7]. Furthermore, it is only produced transiently, as shown in Fig. 2d, and it is also isomerized chemically to the conjugated HPODE without addition of any enzyme at a low rate (Fig. 6). Consequently, it was necessary to avoid acidification of the reaction during the extraction process, to stop the reaction early enough and analyze the products as soon as possible by HPLC to observe the bisallylic product. As it might be possible, that other LOX also produce 11-HPODE under these optimized conditions, we analyzed LOX from various phylogenetic groups for 11-HPODE formation using HPLC analysis. However, none of the tested LOX (FoxLOX, CspLOX1, CsLbLOX, sLOX1) produced detectable amounts of 11-HPODE under Table 1  Kinetics measured with the oxygen electrode and the spectrophotometer at 234 nm kcat (s−1)

Relative rate (%)

Oxygen electrode (all HPODE) Photometer (only 9- and 13-HPODE)

28.5 ± 0.7 5.5 ± 0.3

100 19.3

Difference (11-HPODE)

23

80.7

Using the oxygen electrode, the formation of 9-, 11- and 13-HPODE can be detected, while only the formation of 9- and 13-HPODE is detected at 234 nm with a spectrophotometer. The difference of both measurements should therefore account for the 11-HPODE formation

b

100

100 second reaction rate

80 60

relative abundance

relative absorbance at 234 nm

a

~80%

40 20

first reaction rate

0 0

10

~20% 20

30 time (min)

Fig. 5  Ratio of bis-allylic to conjugated products after the first reaction of CspLOX2. a Full reaction of CspLOX2 observed at 234 nm. CspLOX2 was incubated with 100 µM linoleic acid and the reaction was monitored until almost no changes in absorbance were detected anymore. The end point of the reaction reflects almost a complete conversion to conjugated HPODE from linoleic acid and was set to 100 %. About 20 % of the conjugated HPODE are formed during the

80 60 40 20 0

40

50

11-HPODE: 75%

9- and 13-HPODE: 25% radio HPLC

first reaction phase, while 80 % are formed during the second reaction phase. b HPLC analysis of [1-14C] linoleic acid conversion after the first reaction phase. CspLOX2 was incubated with [1-14C]-labeled linoleic acid and the products were quickly extracted after the first reaction phase and analyzed by HPLC equipped with a radio-detector. The products were quantified as area below the peaks. The chart shows mean values of 3 experiments

13

342

Lipids (2016) 51:335–347

1500 1000 500 0

-500 10

15

20 time (min)

25

after 1 day

800

absorbance (mAU)

2000

500

1000

freshly purified 11-HPODE

absorbance (mAU)

absorbance (mAU)

2500

600 400 200

0 15

20 time (min)

CspLOX2

absorbance at 234 nm

0.4 0.3 0.2

CspLOX1

0.1

FoxLOX CsLbLOX

0 1

2 time (min)

3

100 0 10

25

15

20 time (min)

25

In a time frame of 1–5 days, a significant portion of 11-HPODE was already chemically isomerized to 9- and 13-HPODE

b

11-HPODE 9-HPODE 13-HPODE

100 80 60 40 20

sLOX1 0

200

at 234 nm (9- and 13-HPODE)

relative abundance

Fig. 6  Degradation of 11-HPODE over time. 11-HPODE was enzymatically produced with CspLOX2 from linoleic acid. The bis-allylic product was purified by RP-HPLC and stored at −80 °C in methanol.

0.5

300

-100

-200 10

at 202 nm (11-HPODE)

a

after 5 days

400

0

4 w/

e

zym

n oe

OX

sL

1

OX

bL

L Cs

X1

LO

p Cs

OX

xL

Fo

X2

LO

p Cs

Fig. 7  Isomerization products from 11-HPODE from reactions of different LOX. a 11R-HPODE was purified from the CspLOX2 reaction and incubated with 600 U of different LOX. While the reaction with CspLOX2, CspLOX1 and FoxLOX resulted in a slight increase at 234 nm, CsLbLOX and sLOX1 showed virtually no increase. b The reaction products were extracted after 15 min and analyzed by

SP-HPLC. The remaining 11-HPODE was quantified at 202 nm and 9- and 13-HPODE at 234 nm. The scaling factor determined with [1-14C]-labeled compounds was included in the quantification. As a negative control, only buffer was incubated with 11-HPODE (w/o enzyme)

the applied conditions (data not shown). However, even if these LOX are unable to produce 11-HPODE, they might be able to isomerize it to the conjugated HPODE. We therefore incubated the same LOX with purified 11-HPODE and monitored the reaction at 234 nm to detect the formation of conjugated HPODE. Although FoxLOX and CspLOX1 were unable to form 11-HPODE, they were able to isomerize it (Fig. 6). Analysis of the products revealed that the positional specificities were the same as for linoleic acid, as FoxLOX, which is characterized as 13-LOX [10], produced mainly 13-HPODE and the 9-LOX CspLOX1 [19] produced mainly 9-HPODE (Fig. 7b). Also the isomerization products of CspLOX2 are the same as those produced from

linoleic acid after 2 h. In contrast, the plant LOX sLOX1 and CsLbLOX did not isomerize 11-HPODE at a significant rate.

13

Effect of the Catalytic Metal in the CspLOX2 Active Site To study the catalytic metal in the active site of CspLOX2, we recorded EPR spectra at 9.4 GHz. We compared the asisolated state of CspLOX2 with the activated state upon addition of HPODE and linoleic acid. Unexpectedly, we not only identified the characteristic iron signal of ironcontaining LOX [22, 23] that changed from Fe(II), which

Lipids (2016) 51:335–347

343

a CspLOX2 (mix) CspLOX2 (mix) + HPODE

CspLOX2 (mix) + linoleic acid

Fe-CspLOX2 Fe-CspLOX2 + HPODE

*

Fe-CspLOX2 + linoleic acid

Mn-CspLOX2

Mn-CspLOX2 + HPODE Mn-CspLOX2 + linoleic acid

buffer

7

* 6

5

4

3

2

4

3

2

g value

b MnCl2 MnCl2 + HPODE MnCl2 + linoleic acid

is EPR silent at this frequency, in the as-isolated state to Fe(III) in the activated state, but we could also detect a sharp manganese six line signal that disappeared during activation and reaction (Fig. 8a, upper part). This suggested that some enzyme species contained iron and some of them contained manganese. ICP-AES measurements confirmed that overall 58 mol% Fe and 18 mol% Mn is bound per enzyme (Table 2). We therefore asked if manganese could also fulfill catalytic functions in CspLOX2, similar to MnLOX that can produce 11-HPODE as well. Similar manganese signals have been reported for them [14, 24, 25]. In order to elucidate whether iron, manganese or both metals can fulfill catalytic functions in the CspLOX2 active site, we aimed at obtaining enzyme variants that only contain iron or manganese. Since initial trials failed to extract the active site metal with different chelators and to substitute it with the desired metal, we tried to incorporate the selected metal by adjusting the expression conditions. We finally isolated CspLOX2 from cells grown in a defined minimal auto-induction medium that was supplemented with either manganese or iron. Interestingly, the iron-containing version of CspLOX2 (Fe-CspLOX2) was slightly more active than the enzyme expressed in complex auto-induction medium, while the manganese-containing version (Mn-CspLOX2) almost completely lost its activity (Table 2). To elucidate whether this effect on catalytic activity is indeed caused by a substituted metal center, we applied ICP-AES with the proteins purified by size exclusion chromatography. As shown in Table 2, Fe-CspLOX2 contains 97 mol% Fe and only 1 mol% Mn. In contrast, Mn-CspLOX2 contains 62 mol% Mn and only 2 mol% Fe. The low Fe content might explain the residual activity of Mn-CspLOX2. Additionally, we cannot exclude that the manganese cofactor in Mn-CspLOX2 may also contribute to the observed traces of catalytic activity since a kcat of 0.8 s−1 (2 % of 35.5) instead of 1.9 s−1 (5.5 % of 35.5) would be expected, if the residual activity only results from

MnCl2 + HPODE + linoleic acid

7

6

5

Table 2  Metal occupancy and catalytic activity of the substituted LOX versions

g value

Fig. 8  Comparison of the 9.4 GHz CW-EPR spectra of CspLOX2, which was grown in a complex auto-induction medium [CspLOX2 (mix)] or in a defined minimal auto-induction medium supplemented either with iron (Fe-CspLOX2) or manganese (Mn-CspLOX2), before and after activation with HPODE and/or linoleic acid (a). The experimental conditions are given in Materials and methods. The EPR spectrum of the buffer control is also shown. The iron impurity at g′  = 4.3 and the signal caused by the glass impurity are shown with an asterisk. b EPR control with MnCl2. The typical manganese signal that was also detected with Mn-CspLOX2 and disappears upon incubation with linoleic acid is also detected for MnCl2 without enzyme

Fe-CspLOX2 Mn-CspLOX2 CspLOX2 (mix) Fe occupancy (mol%)

96.6 ± 2.6

2.1 ± 0.2

58.4 ± 0.6

Mn occupancy (mol%)

1.3 ± 0.1

61.8 ± 5.1

18.1 ± 0.1

35.3 ± 3.1

1.9 ± 0.6

28.5 ± 0.7

kcat (s−1)

CspLOX2 was expressed in a minimal medium containing either iron or Mn to produce Fe-CspLOX2 and Mn-CspLOX2, respectively, or in complex medium [CspLOX2(mix)]. The metal content of the CspLOX2 variants was determined by ICP-AES and is shown as a molar ratio of metal per enzyme. The kcat was determined using the oxygen electrode

13

344

Lipids (2016) 51:335–347

2 % iron (96.6 mol% vs 2.1 mol%, Table 2) in the active site as determined by ICP-AES. The only EPR signal detected in Fe-CspLOX2 upon activation with HPODE was the characteristic Fe(III) signal at gxyz  ≈ 6.0, 5.8 and 2.0 (Fig. 8, panels in the middle). However, the Fe(III) signal was not observed when Fe-CspLOX2 was incubated with linoleic acid for a minute. This result was not unexpected, as the catalytic reaction is completed in this time scale and the enzyme harbors Fe(II) that is EPR silent at 9.4 GHz. We note that the small amount of iron signal visible at g′ = 4.3 in all EPR spectra of the Fe-CspLOX2 samples was attributed to contaminating iron. On the other hand Mn-CspLOX demonstrated only the typical Mn(II) EPR six lines signal in the region near g = 2. Addition of HPODE and/or linoleic acid led to a decrease in Mn(II) signal, suggesting that Mn-CspLOX oxidizes Mn(II) to Mn(III), which is EPR silent, although it retains traces of oxygenation activity (Fig. 8, lower panels). The same oxidation reaction is observed when MnCl2 is incubated with HPODE and/or linoleic acid (Fig. 9), suggesting that the oxidation of Mn(II) to Mn(III) can even take place under these conditions without an enzyme. To confirm that manganese is indeed bound in the active site of Mn-CspLOX2, crystal structures of Mn-CspLOX2 and FeCspLOX2 were obtained. The Mn-LOX2 preparation was the same as used for the EPR and ICP-AES measurements. Preliminary analysis of the diffraction data revealed defined electron density in the metal binding site of Mn-CspLOX2. The electron density likely corresponds to Mn that is coordinated in the same geometry as Fe in Fe-CspLOX2. The occupancy of the metal is 94 % in Mn-CspLOX2 and 95 % in Fe-CspLOX2 (data not shown). The occupancy of the metal cofactor determined by crystallography is Fig. 9  Model of the reaction steps. Two modes seem possible: a The CspLOX2 reaction proceeds as a reaction sequence in which 11-HPODE is a required intermediate or b a mixture of all HPODE is formed in a first peroxidation step. Of this mixture, 11-HPODE is further isomerized to 9- and 13-HPODE. Our results are clearly in favor of model b

13

a

higher compared to the ICP-AES measurements since apoCspLOX2 molecules might not crystalize. These results together with the ICP-AES (Table 2) and EPR data (Fig. 8) suggest that manganese is indeed integrated into the active site of Mn-CspLOX2, but cannot catalyze the dioxygenation of linoleic acid.

Discussion This study aimed at characterizing the 11-HPODE formation by the iron-containing enzyme CspLOX2 and elucidating the role of the metal cofactor for this reaction. Kinetic and spectroscopic data, as well as the analysis of reaction products gave a detailed image of the distinct reaction phases. Our results suggest that the reaction starts after a lag phase during which CspLOX2 is fully activated from Fe(II) by HPODE to the catalytically active Fe(III) as shown by kinetic (Fig. 3) and EPR experiments (Fig. 8). During the peroxidation phase, which can be detected as first fast reaction phase at 234 nm, a mixture of 9-, 11and 13-HPODE is formed in parallel (Fig. 2). During this phase, the bis-allylic product is formed at a ~4 times higher rate than the conjugated products and therefore ~75 % of 11-HPODE and only ~25 % of the conjugated 9- and 13-HPODE are formed (Fig. 5). The first reaction phase comes to an end, when the substrate is depleted. At this point, the isomerization takes over which can be identified as slow second increase in the progression curve at 234 nm. In this second phase, the formed 11-HPODE is isomerized to 9- and 13-HPODE at a very low rate (Figs. 4, 9b). Although the reaction of 13R-MnLOX has been studied before, these are to our knowledge the first results focusing

b

Lipids (2016) 51:335–347

345

Table 3  Comparison of CspLOX2 and 13R-MnLOX features and their reactions with linoleic acid

Abstracted hydrogen Site of oxygen insertion relative to hydrogen abstraction Ratio of bis-allylic to conjugated products from linoleic acid during peroxidation phase

CspLOX2

13R-MnLOX

Pro-S [13] Antarafacial [13] ~75:25

Pro-S [7] Suprafacial [7] 29:71 [25]

kcat of peroxidation rate

28.5

26 [25]

kcat of 11-HPODE formation from linoleic acid

23

7 [25]

kcat of 11-HPODE isomerization to conjugated products 11-HPODE isomerization has a kinetic lag phase Number of reaction phases with linoleic acid visible at 234 nm Parallel formation of bis-allylic and conjugated HPODE

0.6

9 [25]

No 2 Yes

Yes [25, 26] Only 1 [7, 25] Yes [7, 25]

Active site metal

Iron: Fe(II) and Fe(III)

Manganese: Mn(II) and Mn(III) [25]

on the formation and isomerization of 11-HPODE by an iron-containing LOX. Having data from both classes of enzymes forming bis-allylic products together allows us to address the question what both reactions have in common and thus to draw conclusions about a common mechanism. Table 3 compares their reaction features and kinetics. Aside of the known facts that CspLOX2 contains iron instead of manganese and catalyzes the dioxygen insertion in an antarafacial way like all other characterized ironcontaining LOX, we could identify a number of similarities in the reaction modes of CspLOX2 and the 13R-MnLOX. First, the formation of bis-allylic and conjugated products occurs in a parallel instead of a sequential reaction [7] (Fig. 9). However, while 13R-MnLOX produces only up to 20–30 % of the bis-allylic product from linoleic acid [25], 11-HPODE is with ~75 % clearly the major product of the initial CspLOX2 reaction. The reaction rates of the linoleic acid peroxidation are comparable in both enzymes with 26 and 28.5 s−1 in 13R-MnLOX and CspLOX2, respectively. However, while a slower rate for the 11-HPODE formation than for the isomerization to 13-HPODE was reported for the 13R-MnLOX [25], the CspLOX2 reaction shows a fast formation of 11-HPODE and only a slow isomerization to the conjugated products. This is probably the reason why CspLOX2 shows two linear reaction phases (Fig. 2) whereas 13R-MnLOX shows only one [7, 25]. As the reactions of both enzymes are similar and exhibit a peroxidation and an isomerization phase, it is likely that the underlying molecular factors for 11-HPODE formation are the same. In addition, we asked, why only relatively few LOX have been reported to form bis-allylic products. We selected four different iron-containing LOX from algae, plants and fungi (CspLOX1, CsLbLOX, sLOX1 and FoxLOX) to test whether the inability to detect bis-allylic

products might be due to suboptimal experimental conditions. Our results could however confirm that the formation of bis-allylic products seems to be a unique feature of a small set of microbial LOX, as none of the selected LOX was able to catalyze this reaction. The capability to isomerize 11R-HPODE, on the other hand, seems to be possible for some LOX (Fig. 7). This might be due to the fact that the isomerization probably occurs via a deoxygenation and subsequent reoxygenation [7, 15] which is similar to the classical LOX peroxidation reaction. Therefore, it also makes sense that the specificity of the isomerization is similar to the product specificity observed with linoleic acid. Interestingly, it was reported that 11S-HPODE can be isomerized by sLOX1 to mainly 13R-HPODE, and the racemic 11-HPODE to racemic 13-HPODE [26]. We could however not observe a detectable isomerization of 11R-HPODE by sLOX1, which may be due to a much lower enzyme concentration in our experiments. Since our results confirm that the reactions catalyzed by MnLOX and CspLOX2 are very similar, we assume that manganese- and iron-containing LOX form bisallylic products in an analogous way and that the ability to form the bis-allylic product is not restricted to MnLOX. When iron was substituted by manganese, CspLOX2 lost its activity almost completely, suggesting that manganese cannot fulfil the function of a catalytic metal in CspLOX2 (Fig. 8). It has been proposed that the coordination of manganese and its catalytic properties are very similar to those of iron in LOX [11, 27]. However, when the metal was extracted from a truncated version of soybean LOX1, it could regain activity with external iron, but not with manganese, which is in line with our data [28]. Our results clearly show that manganese and iron centers in LOX are not interchangeable even though they share many features. When CspLOX2 expressing E.  coli cells

13

346

were grown in a complex medium, mainly iron, but also a considerable portion of manganese was inserted into the active site. The selection for a certain metal thus also depends on the host organism and the metals that are available, which is also known from other enzymes like superoxide dismutase or ribonucleotide reductase [29]. In fact, Mn(II) and Fe(II) prefer a similar coordination sphere and the selection of one of both metals is probably not exclusively based on the structure, but rather on the metal availability [29, 30]. However, the midpoint potentials of Mn(II)/Mn(III) and Fe(II)/Fe(III) of the hexaquo ions are quite different with 1.5 and 0.8 V, respectively, which might explain why the redox chemistry of both metal cofactors is not interchangeable [31]. In superoxide dismutase, iron- and manganese-containing enzymes show similar, but not identical active sites, which might be necessary to adjust the redox potential to allow reduction as well as oxidation of O2 [29, 32, 33]. A crystal structure of a MnLOX will be required to elucidate whether a slightly different coordination environment of the metal cofactor also explains why manganese and iron are not interchangeable in LOX. Interestingly, our EPR data show that the oxidation of Mn(II) to Mn(III) was also observed with Mn-CspLOX2 in the presence of HPODE or linoleic acid, even though this variant was unable to form the hydroperoxide product. In comparison, Fe(II) was not only converted to Fe(III) in Fe-CspLOX2 in the presence of HPODE, but also back to Fe(II) after 1 min with linoleic acid (Fig. 8). It might thus be possible that due to the different redox properties, iron can complete the full catalytic cycle in the active site of CspLOX2, while manganese cannot. Although it would also be interesting to know if an iron-substituted version of MnLOX is active, trials of other groups to obtain such a variant were not successful in the past [34]. Our data suggest that the critical factors for the specificity towards bis-allylic hydroperoxide products probably are structural features of the substrate-binding site rather than differences in the catalytic metal. For a more detailed understanding which factors finally determine the formation of bis-allylic products structural data will be required. As CspLOX2 produces the largest amount of 11-HPODE known so far, it will be an ideal candidate for further structural analyses. Acknowledgments  We thank Uta Nüsse-Hahne and Dr. Dietrich Hertel for the ICP-AES measurements. We are grateful for financial support from the German Research Foundation (DFG) in frame of the International Research Training Group 1422, Metal Sites in Biomolecules: Structures, Regulation and Mechanisms. In addition, JN was supported by the Fonds of the Chemical Industry (FCI). Conflict of interest  The authors declare that they have no conflicts of interest with the contents of this article.

13

Lipids (2016) 51:335–347

References 1. Schneider C, Pratt DA, Porter NA, Brash AR (2007) Control of oxygenation in lipoxygenase and cyclooxygenase catalysis. Chem Biol 14:473–488 2. Andreou A, Brodhun F, Feussner I (2009) Biosynthesis of oxylipins in non-mammals. Prog Lipid Res 48:148–170 3. Horn T, Adel S, Schumann R, Sur S, Kakularam KR, Polamarasetty A, Redanna P, Kuhn H, Heydeck D (2015) Evolutionary aspects of lipoxygenases and genetic diversity of human leukotriene signaling. Prog Lipid Res 57:13–39 4. Kühn H, Banthiya S, van Leyen K (2015) Mammalian lipoxygenases and their biological relevance. Biochim Biophys Acta 1851:308–330 5. Mashima R, Okuyama T (2015) The role of lipoxygenases in pathophysiology; new insights and future perspectives. Redox Biology 6:297–310 6. Wasternack C, Hause B (2013) Jasmonates: biosynthesis, perception, signal transduction and action in plant stress response, growth and development. An update to the 2007 review in Annals of Botany. Ann Bot 111:1021–1058 7. Hamberg M, Su C, Oliw E (1998) Manganese lipoxygenase— discovery of a bis-allylic hydroperoxide as product and intermediate in a lipoxygenase reaction. J Biol Chem 273:13080–13088 8. Wennman A, Oliw EH (2013) Secretion of two novel enzymes, manganese 9S-lipoxygenase and epoxy alcohol synthase, by the rice pathogen Magnaporthe salvinii. J Lipid Res 54:762–775 9. Wennman A, Jernerén F, Magnuson A, Oliw EH (2015) Expression and characterization of manganese lipoxygenase of the rice blast fungus reveals prominent sequential lipoxygenation of α-linolenic acid. Arch Biochem Biophys. doi:10.1016/j. abb.2015.1007.1014 10. Brodhun F, Cristobal-Sarramian A, Zabel S, Newie J, Ham berg M, Feussner I (2013) An iron 13S-lipoxygenase with an α-linolenic acid specific hydroperoxidase activity from Fusarium oxysporum. PLoS One 8:e64919 11. Gaffney B, Su C, Oliw E (2001) Assignment of EPR transitions in a manganese-containing lipoxygenase and prediction of local structure. Appl Magn Res 21:413–424 12. Tomchick DR, Phan P, Cymborowski M, Minor W, Holman TR (2001) Structural and functional characterization of second-coordination sphere mutants of soybean lipoxygenase-1. Biochemistry 40:7509–7517 13. Andreou A, Göbel C, Hamberg M, Feussner I (2010) A bisallylic mini-lipoxygenase from cyanobacterium Cyanothece sp. that has an iron as cofactor. J Biol Chem 285:14178–14186 14. Wennman A, Magnuson A, Hamberg M, Oliw EH (2015) Manganese lipoxygenase of F. oxysporum and the structural basis for biosynthesis of distinct 11-hydroperoxy stereoisomers. J Lipid Res 56:1606–1615 15. Oliw EH (2008) Factors influencing the rearrangement of bisallylic hydroperoxides by manganese lipoxygenase. J Lipid Res 49:420–428 16. Boutaud O, Brash AR (1999) Purification and catalytic activities of the two domains of the allene oxide synthase-lipoxygenase fusion protein of the coral Plexaura homomalla. J Biol Chem 274:33764–33770 17. Graff G, Anderson LA, Jaques LW (1990) Preparation and purification of soybean lipoxygenase-derived unsaturated hydroperoxy and hydroxy fatty acids and determination of molar absorptivities of hydroxy fatty acids. Anal Biochem 188:38–47 18. Studier FW (2005) Protein production by auto-induction in highdensity shaking cultures. Protein Expr Purif 41:207–234 19. Newie J, Andreou A, Neumann P, Einsle O, Feussner I, Ficner R (2016) Crystal structure of a lipoxygenase from Cyanothece sp.

Lipids (2016) 51:335–347 may reveal novel features for substrate acquisition. J Lipid Res doi:10.1194/jlr.M064980 20. Schilstra MJ, Veldink GA, Verhagen J, Vliegenthart JFG (1992) Effect of lipid hydroperoxide on lipoxygenase kinetics. Biochemistry 31:7692–7699 21. Schilstra MJ, Veldink GA, Vliegenthart JFG (1993) Kinetic analysis of the induction period in lipoxygenase catalysis. Biochemistry 32:7686–7691 22. Gaffney BJ, Mavrophilipos DV, Doctor KS (1993) Access of ligands to the ferric center in lipoxygenase-1. Biophys J 64:773–783 23. Slappendel S, Veldink GA, Vliegenthart JFG, Aasa R, Malmström BG (1981) EPR spectroscopy of soybean lipoxygenase-1. Biochim Biophys Acta 667:77–86 24. Oliw EH, Jernerén F, Hoffmann I, Sahlin M, Garscha U (2011) Manganese lipoxygenase oxidizes bis-allylic hydroperoxides and octadecenoic acids by different mechanisms. Biochim Biophys Acta 1811:138–147 25. Su C, Sahlin M, Oliw EH (2000) Kinetics of manganese lipoxygenase with a catalytic mononuclear redox center. J Biol Chem 275:18830–18835 26. Oliw E, Cristea M, Hamberg M (2004) Biosynthesis and isomerization of 11-hydroperoxy linoleates by manganese- and irondependent lipoxygenases. Lipids 39:319–323 27. Cristea M, Engstrom A, Su C, Hornsten L, Oliw EH (2005) Expression of manganese lipoxygenase in Pichia pastoris and

347 site-directed mutagenesis of putative metal ligands. Arch Biochem Biophys 434:201–211 28. Dainese E, Angelucci CB, Sabatucci A, De Filippis V, Mei G, Maccarrone M (2010) A novel role for iron in modulating the activity and membrane-binding ability of a trimmed soybean lipoxygenase-1. FASEB J 24:1725–1736 29. Cotruvo JA Jr, Stubbe J (2012) Metallation and mismetallation of iron and manganese proteins in vitro and in vivo: the class I ribonucleotide reductases as a case study. Metallomics 4:1020–1036 30. Tu WY, Pohl S, Gray J, Robinson NJ, Harwood CR, Waldron KJ (2012) Cellular iron distribution in Bacillus anthracis. J Bacteriol 194:932–940 31. Dean JA (1985) Lange’s handbook of chemistry. McGraw-Hill, New York 32. Borgstahl GE, Pokross M, Chehab R, Sekher A, Snell EH (2000) Cryo-trapping the six-coordinate, distorted-octahedral active site of manganese superoxide dismutase. J Mol Biol 296:951–959 33. Lah MS, Dixon MM, Pattridge KA, Stallings WC, Fee JA, Ludwig ML (1995) Structure-function in Escherichia coli iron superoxide dismutase: comparisons with the manganese enzyme from Thermus thermophilus. Biochemistry 34:1646–1660 34. Cristea M, Oliw EH (2007) On the singular, dual, and multiple positional specificity of manganese lipoxygenase and its G316A mutant. J Lipid Res 48:890–903

13