Heme Biosynthesis Is Coupled to Electron Transport Chains for Energy Generation

Heme biosynthesis is coupled to electron transport chains for energy generation

Kalle Möbiusa, Rodrigo Arias-Cartinb, Daniela Breckaua, Anna-Lena Hänniga, Katrin Riedmannc, Rebekka Biedendiecka, Susanne Schrödera, Dörte Becherd, Axel Magalonb, Jürgen Mosera, Martina Jahna, and Dieter Jahna,1

aInstitute of Microbiology, University Braunschweig, Brunswick, Germany; bLaboratoire de Chimie Bactérienne, Institute de Microbiologie de la Méditerraneé, Centre National de la Recherche Scientifique, Marseille, France; cDepartment of Agrarcultural Technology, Johann Heinrich von Thünen-Institut, Brunswick, Germany; and dInstitute of Microbiology, University of Greifswald, Germany

Edited* by Dieter Söll, Yale University, New Haven, CT, and approved April 23, 2010 (received for review January 30, 2010)

Cellular energy generation uses membrane-localized electron O2 tensions in the growth medium, cytochrome bo3 (EC 1.10.3; transfer chains for ATP synthesis. Formed ATP in turn is consumed Cyo) dominates over all other terminal oxidases. At low O2 ten- for the biosynthesis of cellular building blocks. In contrast, heme sions the second cytochrome oxidase, termed bd (EC 1.10.3; cofactor biosynthesis was found driving ATP generation via Cyd), is present in the cytoplasmic membrane (10–15). Both electron transport after initial ATP consumption. The FMN enzyme enzyme systems are known to couple the four-electron reduction protoporphyrinogen IX oxidase (HemG) of Escherichia coli of O2 to two H2O molecules with the formation of a proton po- abstracts six electrons from its substrate and transfers them via tential via the membrane (13, 14). When O2 is absent, E. coli is bo bd ubiquinone, cytochrome 3 (Cyo) and cytochrome (Cyd) capable of utilizing nitrate, nitrite, TMAO (trimethylamine N- oxidase to oxygen. Under anaerobic conditions electrons are trans- oxide), DMSO (dimethylsulfoxide), and fumarate as alternative ferred via menaquinone, fumarate (Frd) and nitrate reductase terminal electron acceptors (9). The dissimilatory nitrate reduc- (Nar). Cyo, Cyd and Nar contribute to the proton motive force that tase (EC 1.6.6.1) NarGHI is a membrane-bound enzyme complex drives ATP formation. Four electron transport chains from HemG consisting of the nitrate-reducing subunit NarG, the electron- via diverse quinones to Cyo, Cyd, Nar, and Frd were reconstituted transfer subunit NarH, and the quinol-oxidizing subunit NarI in vitro from purified components. Characterization of E. coli mu- (15–18). Fumarate reductase (EC 1.3.1.6; Frd) of E. coli is a tants deficient in nar, frd, cyo, cyd provided in vivo evidence for a membrane-bound flavoprotein catalyzing the cytoplasmic reduc- BIOCHEMISTRY detailed model of heme biosynthesis coupled energy generation. tion of fumarate to succinate, as well as the oxidation of quinols in the membrane (9, 19). The biochemical principles of coupling anabolism coupled catabolism ∣ protoporphyrinogen IX oxidase ∣ HemG ∣ protoporphyrinogen IX oxidation in E. coli to the outlined tetrapyrrole ∣ respiration electron transport chains were completely unknown. Results and Discussion eme is an essential cofactor of enzymes in electron transport Hchain mediated energy generation. It is synthesized using a E. coli HemG Is a Membrane Associated Protoporphyrinogen IX Oxi- highly conserved pathway (1). The penultimate step of heme bio- dase. In order to purify E. coli PPO we first established an enzyme synthesis—the conversion of protoporphyrinogen IX (proto’gen) assay using PPO active cell-free extract. However, this assay via the abstraction of six electrons into protoporphyrin IX (proto) always required the presence of fumarate reductase and mena- μ —is catalyzed by protoporphyrinogen IX oxidases (PPO; EC quinone for PPO activity. Hence, we replaced those with 2.5 M 1.1.3.4). An O2-dependent PPO, usually encoded by the hemY of the artificial electron acceptor triphenyltetrazolium chloride gene, is missing in Escherichia coli (2). Over 30 yr ago an oxy- (TTC, Fig. 1). In this context several of the tested potential elec- gen-independent PPO activity was detected in cell-free extracts tron acceptors [i.e. 2, 6-dichloroindophenol (DCIP), phenazin- of E. coli which was dependent on electron transfer to nitrate methosulfate (PMS), menadione, and vitamin K1, respectively] or fumarate (3–5). The observation indicated the coupling of were found to directly oxidize the substrate in the absence of anabolic heme biosynthesis to catabolic electron chain-driven HemG and were therefore not used in the activity assay system. ATP synthesis. Complementation of the proto accumulating TTC was the only stable electron acceptor utilized by HemG and E. coli strain SASX38 yielded the hemG gene (6). Because of consequently employed in this study. Corresponding chemical M ¼ 21;200 structures are given in Fig. 1. Using the TTC-based enzyme test the small size of the deduced HemG protein ( r ) 247 ∕ ∕ and the fact that no PPO activity has been shown for the isolated system, PPO activity of pmol proto mg protein h was found peptide, it was assumed that HemG may be a subunit of a larger associated with an isolated membrane fraction of E. coli (Table S1). Crude cell-free extract (53 pmol proto∕mg protein∕ PPO complex (2). Additionally, HemG does not share any amino 11 ∕ ∕ acid sequence homology to oxygen-dependent HemY. Very re- h) and the cytosolic fraction ( pmol proto mg protein h) were cently it was reported that purified recombinant E. coli HemG significantly less active. Membrane fraction proteins were solu- carries menadione-dependent protoporphyrinogen IX oxidase bilized using the detergent Thesit® (2-dodecoxyethanol) and activity (7). The recombinant protein was described as completely further purified via anion exchange chromatography. Obtained water soluble. A structural model based on the homology to long chain flavodoxins was proposed (7). Respiratory reactions possi- Author contributions: K.M., R.A.-C., R.B., S.S., A.M., J.M., and M.J. designed research; bly coupled to the HemG reaction are usually used to drive mem- K.M., R.A.-C., D. Breckau, A.-L.H., K.R., R.B., S.S., D. Becher, A.M., J.M., M.J., and D.J. brane-bound electron transport chain for the formation of an performed research; K.M., R.A.-C., R.B., S.S., D. Becher, and M.J. contributed new reagents/analytic tools; K.M., D. Breckau, R.B., S.S., D. Becher, A.M., M.J., and D.J. analyzed electrochemical ion gradient. This membrane-localized gradient, data; and K.M., A.M., M.J., and D.J. wrote the paper. called proton motive force, provides the energy for ATPase to The authors declare no conflict of interest. generate ATP from ADP and Pi (8, 9). Electron transport chains *This Direct Submission article had a prearranged editor. consist of numerous primary dehydrogenases and quinone-linked 1To whom correspondence should be addressed at: Institute of Microbiology, University terminal oxidoreductases. Primary dehydrogenases in E. coli in- Braunschweig, Spielmannstrasse 7, D-38106 Braunschweig, Germany. E-mail: d.jahn@ clude those for NADH, lactate, glucose, formate, and hydrogen tu-bs.de. (9). The three quinones of E. coli are ubiquinone (UQ), mena- This article contains supporting information online at www.pnas.org/lookup/suppl/ quinone (MQ), and demethylmenaquinone (DMQ) (9). At high doi:10.1073/pnas.1000956107/-/DCSupplemental.

www.pnas.org/cgi/doi/10.1073/pnas.1000956107 PNAS Early Edition ∣ 1of6 Downloaded by guest on September 29, 2021 O Cl N N + SO 2- Na+ - 4 O O N Cl O

Menadione (Provitamin K3) 2, 6- dichloroindophenol (DCIP) Phenazinmethasulfate (PMS)

o O CH3

o CH CH CH 3 3 3 O Fig. 1. Electron acceptors tested for E. coli HemG. The structures of the following electron acceptors used in this

Phylloquinone (Vitamin K1) Menaquinone (Vitamin K1) work are depicted: 2, 6-dichloroindolphenol (DCIP) with 00 00 an E of þ237 mV, menadione (provitamin K3, E ¼ 0 −205 mV), phenazinmethosulfate (PMS, E0 ¼þ80 mV), 0 O 2, 6-triphenoltetrazoliumchloride (TTC, E0 ¼ −80 mV), CH 00 N CH O 3 phylloquinone (vitamin K1, E ¼ −170 mV), menaqui- N 3 00 none (vitamin K2, E ¼ −74 mV), and ubiquinone (coen- N N CH 00 00 + 3 O H zyme A, E ¼þ110 mV). HemG cofactor FMN has E ¼ Cl- O CH −190 3 6 - 10 mV. Redox potentials are given for the acceptor- donor couples of the free compounds (40, 41). Protein environments might individually change the correspond- Ubiquinone (Coenzyme Q) 2, 6 -triphenoltetrazoliumchloride (TTC) ing values.

fractions harboring the highest specific PPO activity were analyzed for the identity of contained proteins by mass spectrometry. The most frequently detected peptides in the analyzed samples A rel M W1 W2 E1 E2 belonged to the HemG protein (Table S2). Nevertheless, the MW W3 question still remained whether the strong PPO activity of the purified membrane fraction is caused by HemG alone or whether 70 the membrane fraction contained other components contributing 55 to the 6 e− oxidation of proto’gen. The B. megaterium genome 45 does neither possess a hemG gene nor the enzymes of anaerobic 35 respiration including fumarate and nitrate reductase. Bacilli are synthesizing proto using the oxygen-dependent flavin-enzyme 25 HemY (20). In order to exclude association of E. coli HemG with HemG additional potential PPO subunits from E. coli the HemG protein 15 was produced fused to a His-tag in B. megaterium (Fig. 2). With TTC as electron acceptor B. megaterium cell-free extract containing recombinant E. coli, HemG revealed the formation of 10 1.2 nmol proto∕mg protein∕h. Purification via affinity and gel permeation chromatography yielded apparently pure protein (Fig. 2A). As expected, purified HemG solely revealed 300 enzyme activity in the presence of TTC as electron acceptor B (23 nmol proto∕mg protein∕h). Removal of the fused His-tag 250 via protease digestion did not change the catalytic properties, neither with TTC (21 nmol proto∕mg protein∕h) nor with puri- 200 fied fumarate reductase as electron acceptors (55 nmol proto∕ mg protein∕h). Consequently, all further experiments were con- 150 ducted with His-tagged HemG. The Michaelis-Menten constant K V Arbitraryunits 100 M as well as the maximal velocity max were determined with menaquinone as electron acceptor. The KM value of 17.3 μM 50 for the substrate was obtained. However, the KM of the substrate was 15 times higher compared to tobacco oxygen-dependent 0 HemY PPO (KM of 1.17 μM) (21). On the other hand, the cal- W1 W2 W3 E1 E2 V 960 μ −1 −1 culated maximal velocity max of HemG of Mh mg was Fig. 2. Recombinant production of E. coli HemG in B. megaterium.(A) E. coli −1 −1 4-fold higher than that of tobacco HemY with 256.2 μMh mg HemG after recombinant production in B. megaterium and affinity chroma- (21). Acifluorfen is a herbicide inhibitor of oxygen-dependent tographic purification was subjected to 15% SDS-PAGE. Lane 1, molecular PPOs of the HemY-type (22). Applied in concentrations of up weight standard; lanes W1–W3, Ni-NTA Superflow resin chromatography to 250 μM, the compound did not influence HemG activity with fractions. Lanes E1, E2, distinct bands of Mr ∼ 22;000 corresponding to the TTC (19 nmol proto∕mg protein∕h) and fumarate reductase size of His6x-tagged HemG was observed. (B) PPO activities of fractions from (52 nmol proto∕mg protein∕h) as electron acceptor. Fumarate the purification of HemG by affinity chromatography. PPO activities were obtained as described in Materials and Methods. Proto formation was depicted reductase subunit FrdA was reproducibly detected in PPO- with TTC as electron acceptor. W1–3: activities corresponding to the fractions activity containing fractions of the HemG purification from as shown in (A). E1–2: activities of recombinant HemG corresponding to the membrane fraction of E. coli and from recombinant E. coli. the fractions obtained in (A). Arbitrary units: relative fluorescence units with Corresponding peptides are given in Table S2. These observations t ¼ 0 min subtracted from t ¼ 60 min.

2of6 ∣ www.pnas.org/cgi/doi/10.1073/pnas.1000956107 Möbius et al. Downloaded by guest on September 29, 2021 indicated protein-protein interactions between HemG and E. coli Fumarate reductase employs menaquinone as electron carrier. FrdA. Gel permeation chromatography of purified E. coli HemG However, other terminal oxidoreductases including the cyto- in the presence of 2% Thesit® on Superdex 200 HR 10∕30 chrome bd complex couple to ubiquinone. Both quinones were revealed an identical elution point as the marker proteins yeast successfully tested as electron acceptors for catalysis with purified alcohol dehydrogenase (Mr ¼ 150;000) indicating a Mr of recombinant HemG (Table 1). ∼150;000 and a hexameric oligomerization state. A UV-Vis spectral analysis of the purified protein showed in agreement with a Purified Fumarate and Nitrate Reductase, Cytochrome bd and Cyto- recent investigation of recombinant HemG (7) a typical spectrum chrome bo are Electron Acceptors of E. coli HemG. Next, it was inves- for a flavin cofactor (Fig. 3A). The flavin was identified as non- tigated if purified E. coli HemG transfers electrons via the covalently bound FMN using HPLC analyses (Fig. 3B). identified quinones towards terminal oxidoreductases of the E. coli respiratory chains and, consecutively, to various terminal Ubiquinone and Menaquinone Are Electron Acceptors of HemG Cata- electron acceptors. For this purpose, fumarate reductase, nitrate lysis. The quinol analogue pentachlorophenol (PCP, Fig. 1) inhi- reductase, cytochrome bd oxidase, and cytochrome bo oxidase of bits fumarate reductase via blocking the quinone binding site of E. coli were recombinantly produced and affinity chromatogra- the enzyme (9). At a concentration of 1 μM it was also a strong phically purified as described before (16, 19, 23). In combination inhibitor of HemG activity when tested with fumarate reductase with the respective quinones and electron acceptors they were as electron acceptor (Table 1). In contrast, HemG tests with 1 μM analyzed for the ability to accept electrons from the catalytically TTC as electron acceptor remained unaffected (22 nmol proto∕ active E. coli HemG in vitro. Proto formation was monitored mg protein∕h). An inhibition of quinone-dependent electron spectroscopically. All four terminal oxidoreductases accepted transfer from HemG to fumarate reductase was concluded. electrons from the HemG catalysis (Table 1). These results clearly demonstrate the coupling of heme biosynthesis to respiratory electron transport and consequently to proton gradient formation. Interestingly, the addition of quinones to the enzyme complexes was not essential in most tested cases. Both purified oxygen-dependent cytochrome oxidases sustained up to 67% (Cyd) and 78% (Cyo) of HemG activity in the absence of ubiquinone, indicating a tight association of the quinones with the en-

zyme complexes during preparation. This tight quinone binding BIOCHEMISTRY by both cytochrome oxidases was described several times before (23, 24). Similar observations were made for respiratory nitrate reductase (17). In control reactions the end electron acceptors (nitrate, fumarate, and oxygen) did not directly accept electrons from HemG catalysis (Table 1).

Electron Transfer from E. coli HemG in Vivo. To confirm the results of our in vitro experiments, additional in vivo experiments using appropriate E. coli mutant strains were performed. Cell lysates of E. coli mutants deficient in the terminal oxidoreductases were investigated for their PPO activity (Table 1). First, the strains were grown under conditions allowing for efficient cell growth of the mutants. Afterwards, the cultures were shifted to growth conditions which allow for a strong expression of the phenotype of interest. The results of those in vivo experiments confirmed the results of the in vitro experiments (Table 1). All mutants deficient in terminal oxidases were found reduced in PPO activity when tested with the corresponding terminal electron acceptor. How- ever, only in the case of the cyo/cyd double mutant PPO activity was completely abolished. Mutant variants deficient in either Cyo or Cyd retained 50 and 87% of wild-type activity. It was concluded that the PPO electron flow towards oxygen remains largely intact as long as one out of the two cytochrome oxidases is present in the cell. Partial cross complementation of cyo and cyd mutants was observed before (25, 26). The clear cut HemG deficiency in the cyo/cyd double mutant revealed exclusive coupling of HemG to these two systems under aerobic growth conditions. The mutants deficient in fumarate reductase or all three nitrate reductases (narGHI, narXYZ, napFDAGHCB) were capable of sus- taining 48% of PPO activity and 65% of wild-type activity, respectively. It is possible that the remaining terminal oxidoreductases Fig. 3. Spectral analysis of E. coli HemG and the bound flavin cofactors. in the mutant strains take up electrons via their quinones and (A) A UV-Vis spectrum of purified HemG (black line) and extracted cofactor transfer them to residual amounts of terminal electron acceptors. (red line) against elution buffer was recorded. The spectra with characteristic Fumarate is a product of the TCA (tricarboxylic acid) cycle and peaks at 366 and 433 nm indicate the presence of a flavin cofactor (19). consequently synthesized in the nitrate reductase mutant. More- (B) HPLC analysis for the identification of the flavin cofactor of HemG. The retention time for the cofactor FAD (flavine adenine dinucleotide) over, E. coli possesses additional anaerobic terminal oxidoreduc- was 7.25 min (blue) and 10.1 min for FMN (flavine mononucleotide) (red). tases including TMAO- (TorA), DMSO- (DmsABC) and nitrite- The purified cofactor from the HemG fraction was detected at 10.1 min (NrfABCDEFG) reductases (9, 15). In conclusion, the genetic and thus identified as FMN (green). analysis of diverse E. coli respiratory-chain mutant strains clearly

Möbius et al. PNAS Early Edition ∣ 3of6 Downloaded by guest on September 29, 2021 Table 1. In vitro reconstitution of HemG coupled electron transport chains of E. coli with purified components and PPO activities in E. coli electron transport chain mutants E. coli HemG + proto’gen * + terminal oxidoreductase † + quinone †+ electron acceptor † Proto formed [nmol∕mg protein∕h] n. a. † n.d.‡ Cytochrome bo3 oxidase + UQ þO2 51.1 ± 7.3 † n. a. +UQþO2 28.3 ± 3.3 † ‡ n. a. þO2 n.d. n. a. † + UQ 25.5 ± 2.4 Cytochrome bo3 oxidase + UQ 25.2 ± 3.3 Cytochrome bo3 oxidase þO2 40.8 ± 5.4 Cytochrome bd oxidase + UQ þO2 86.9 ± 6.6 † n. a. +UQþO2 44.7 ± 4.4 † ‡ n. a. þO2 n.d. n. a. † + UQ 41.7± 3.2 Cytochrome bd oxidase + UQ 50.7 ± 3.1 Cytochrome bd oxidase þO2 58.9 ± 4.3 Fumarate reductase + MQ + fumarate 52.4 ± 6.6 n. a. † + MQ + fumarate 22.4 ± 2.1 n. a. † + fumarate n.d.‡ n. a. † + MQ 21.3 ± 2.3 Fumarate reductase + MQ 10.1 ± 1.2 Fumarate reductase + fumarate 23.1 ± 2.3 Fumarate reductase + PCP § + fumarate n.d.‡ Nitrate reductase + MQ + nitrate 61.7± 9.6 n. a. † + MQ + nitrate 16.2 ± 2.2 n. a. † + nitrate n.d.‡ n. a. † + MQ 15.4 ± 1.6 Nitrate reductase + MQ 12.3 ± 1.1 Nitrate reductase + nitrate 21.4 ± 2.6 Cell-free extract of E. coli strain and [relevant genotype] ¶ Proto formed [pmol∕mg protein∕h] wild-type E. coli þO2 62.5 ± 7.0 FB20172 [cyoB∷kan] þO2 31.6 ± 4.2 FB20228 [cydA∷kan] þO2 54.4 ± 5.5 ‡ DS253 [Δcyd∷cam Δcyo∷kan] þO2 n.d. ‡ G0105 [Δ(cydAB)∷cyo123] þO2 n.d. wild-type E. coli + fumarate 56.0 ± 4.0 DW35 [ΔfrdABCD] + fumarate 27.2 ± 2.0 wild-type E. coli + nitrate 51.3± 5.5 JCB4023 [narG::ery ΔnapAB narZ∷Ω] + nitrate 33.8 ± 2.0 *Thirty pmol purified E. coli HemG were incubated with 30 pmol of purified terminal oxidoreductase and 10 μM proto’gen under conditions as outlined in Materials and Methods.O2 was introduced by vigorous shaking of the reaction vessel. Where indicated 1.8 μM MQ or UQ were added, respectively. Fumarate or nitrate was employed at a concentration of 1 mM, respectively. All experiments were performed with three independent HemG and terminal oxidoreductase preparations in triplicate. †As indicated below, n:a: ¼ no further additions. ‡n:d: ¼ not detectable; the detection limit of the employed test was below 10 pmol∕mg protein∕h. §Pentachlorophenol (PCP) was first titrated from 1 nM to 10 mM for specific inhibition of fumarate reductase without direct reaction with the substrate and further used at specific concentrations of 1 μM. ¶All employed mutant strains and the wild-type strain were grown in parallel under the required conditions as detailed in Materials and Methods. For most of the E. coli mutant strains the parental strains were analyzed in parallel. Since no obvious differences between these strains and BL21-DE3 concerning their PPO-activity were observed, BL21-DE3 was used as the general wild-type strain. Cell-free extracts were prepared. Hundred μg protein were incubated with 10 μM proto’gen and appropriate electron acceptors 1 mM fumarate, 1 mM nitrate and O2 under active aeration. All experiments were performed with three independent experiments done in triplicate.

demonstrated the dependence of HemG catalysis and thus of the and consequently ATP generation via HemG coupled electron overall heme biosynthesis on respiratory electron transport. transfer might not contribute quantitatively to its energy charge. Nevertheless, a principle of an anabolic, biosynthetic pathway A Model for Electron Transport Coupled Heme Biosynthesis. The com- driven catabolic proton gradient formation and consequently plete in vitro reconstitution of potential electron transport chains ATP generation was biochemically unraveled. Because of the em- for aerobic and anaerobic HemG activity in combination with ployment of oxygen, nitrate, and fumarate as electron acceptors, genetic investigations led to a model for respiratory chain driven heme biosynthesis is ensured in aerobic as well as anaerobic en- PPO activity in E. coli (Fig. 4). Interestingly, the system couples vironments. Our model is further substantiated by the observed the biosynthesis of the cofactor heme to respiratory systems protein-protein interactions of immobilized HemG with fumarate which in turn are highly dependent on hemes. Clearly, ATP for- reductase subunit FrdA. Such protein-protein interactions might mation is connected with this biosynthetic step due to the proton provide the basis for the assembly of the electron transport chain translocation catalyzed by most of the terminal oxidoreductases machinery. Consequently, heme biosynthesis with HemG might (8–16). Heme biosynthesis in E. coli, starting from glutamyl- be part of the recently described respirazones in E. coli, mobile tRNA, requires eight ATP molecules per formed protoporphyr- patches of the plasma membrane where respiratory enzymes like inogen IX during the initial charging of tRNAGlu with glutamate cytchrome bd are concentrated (27). Moreover, the necessary (1). Consequently, the observed coupling of HemG to oxidative quinones were found associated with the terminal oxidoreduc- phosphorylation might balance this ATP consumption. Clearly, tase. In agreement, the addition of the quinol analogue, PCP, heme biosynthesis is not a dominating cellular process in E. coli abolished PPO activity. PCP is also known to inhibit nitrate

4of6 ∣ www.pnas.org/cgi/doi/10.1073/pnas.1000956107 Möbius et al. Downloaded by guest on September 29, 2021 membrane cytoplasm

H+ Cyo H H H2O or H NH NH H H+ Cyd O2 H NH NH H

H H Aerobic conditions QQH2 CO2H CO2H protoporphyrinogen IX FMN

HemG protoporphyrin IX

MQH2 MQ Anaerobic conditions NH N H H fumarate NH Frd N succinate or H

nitrate CO H CO H H+ Nar 2 2 nitrite

ADP + Pi H+ ATPase ATP

Fig. 4. Model for the E. coli PPO coupled electron transfer reactions and ATP generation. PPO enzymatically converts protógen to proto, probably via hydride BIOCHEMISTRY transfer to FMN. From this flavin, the electrons are transferred to quinones. Electron uptake from terminal oxidases recycles the quinones. Oxygen is used as electron acceptor via two different cytochrome oxidases. Under anaerobic conditions, electrons are dissipated to the terminal electron acceptors fumarate and nitrate by the respective reductases. Three of the terminal oxidoreductases couple electron transport to protein transfer via the cytoplasmic membrane. The generated proton motive force is employed for ATP generation by ATPase.

reductase and fumarate reductase by occupying their quinol bind- of proteins in the periplasm of Gram-negative bacteria (32). An ing sites (17, 28). Thus, we presume that quinones directly asso- in vitro system containing DsbA, DsbB, and either cytochrome bd ciated to the terminal oxidases are utilized for electron transfer. or cytochrome co3 was employed (33, 34). In contrast to HemG, In agreement with this assumption HemG was found membrane associated.

A HemG Model. In absence of structural data for E. coli HemG, computational models can be built by homology modeling using related structures of long chain flavodoxins to which HemG is sequence related (7). The model built in this investigation showed differences compared to the model for E. coli HemG proposed before (8). The main difference was observed for the structure deduced for the long chain insert region which is unique in HemG proteins. This region is supposed to be responsible for specifying both substrate binding (protoporphyrinogen IX and quinones) and membrane anchoring. Our bioinformatic analysis indicated that this region adopts an amphiphilic helical conformation in the predicted structure of E. coli HemG (pdb 2ark) (Fig. 5). The helix is mainly composed of nonpolar residues on one side, while the opposite side is composed of basic charged residues. Such an amphiphilic character may explain the membrane associated feature and the demonstrated ability of HemG to direct electron transfer towards lipophilic quinone molecules via interaction with membrane-bound respiratory complexes.

Quinone-Dependent Metabolic Reactions. For a second biosynthetic enzyme, namely class 2 dihydroorotate dehydrogenase, coupling of enzyme catalysis to ubiquinone was described before (29, 30). These FMN containing enzymes of eukaryotes and Gram- negative bacteria are found associated with the cytoplasmatic membrane via an N-terminal extension (31). However, no direct Fig. 5. Model structure of E. coli HemG. The structure was generated as outlined in SI Materials and Methods. These analyses predicted the presence of electron transfer to a terminal oxidoreductase was demonstrated an amphiphilic helix (red) involved in substrate binding, catalysis, and elec- for this type of enzyme. A second known example for electron tron transport chain interaction. One side of the helix is mainly composed of transfer chain coupled process is the formation of disulfide bonds nonpolar residues while the opposite side consists of basic charged residues.

Möbius et al. PNAS Early Edition ∣ 5of6 Downloaded by guest on September 29, 2021 in this case electrons are transferred from the periplasmic site of Protein Identification Using Proteomics Techniques. The peptide analysis of the the membrane from an enzyme involved in protein modification, E. coli PPO activity containing fraction after membrane solubilization and chromatographic purification was performed according to standard proce- not in classical anabolic biosynthesis. Consequently, we showed dures (36) (SI Materials and Methods). HemG was identified by the presence that an anabolic biosynthetic oxidation is directly coupled to of at least 17 different HemG specific peptides in a fraction with high PPO energy conservation via respiration. The biosynthesis of a cofac- activity prepared from E. coli membrane fraction. The identified peptides are tor is used for ATP generation. listed in Table S2. A HemG containing fraction after affinity chromatographic purification reproducibly of recombinant HemG from cell-free extract Materials and Methods yielded the FrdA specific peptides (Table S2). A highly enriched HemG Bacterial Strains and Plasmids. The strains and plasmids used in this work containing fraction prepared from E. coli membranes also contained FrdA, are listed in SI Text Table S1. Growth of the bacterial strains and expression which was identified by various peptides (Table S2). plasmid construction are outlined in SI Materials and Methods. Protoporphyrinogen IX Oxidase Activity Assays. The substrate preparation strategy was described by Kushner and coworkers (37). PPO assays were Recombinant HemG Production in E. Coli and B. Megaterium and Affinity Chro- performed and kinetic data obtained as outlined before (38). Additions of matographic Purification. E. coli BL21 (λDE3) harboring pETDuet1hemG was quinones and terminal oxidases are detetailed in SI Materials and Methods. used for the recombinant production of His6X-tagged E. coli HemG in E. coli. Purification of the recombinant protein was achieved by affinity and gel Identification of the HemG Flavin Cofactor. Spectral analysis and HPLC identi- permeation chromatography (SI Materials and Methods). fication of the HemG FMN cofactor was described before (39) with modi- fications shown in SI Materials and Methods. Production and Purification of Recombinant E. Coli Fumarate, Nitrate, and Cy- tochrome Oxidoreductases. The terminal oxidoreductases were recombinantly produced and affinity chromatographically purifed as detailed before (17, ACKNOWLEDGMENTS. This work was supported by various grants from the Deutsche Forschungsgemeinschaft and the Fonts der Chemischen industrie. 19, 23, 26, 35) with modification described in SI Materials and Methods. Several of the employed strains and expression plasmids were generous gifts from Deborah Siegele, Texas A&M University, Texas; Gary Cecchini, Veterans Purification of E. coli PPO from the Membrane Fraction. The membrane fraction Affairs Medical Center, San Francisco, California; Robert Gennis, University of was isolated using sucrose gradient centrifugation and PPO activity enriched Chicago, Illinois; and Frederick Blattner, University of Wisconsin-Madison, by DEAE-Sepharose FF chromatography as outlined in SI Materials and Wisconsin. We thank Fritz Unden, Simone Virus, and Denise Wätzlich for helpful discussions. We also thank John Phillips for assisting with the set Methods. up of the reduction of protoporphyrin.

1. Heinemann IU, Jahn M, Jahn D (2008) The biochemistry of heme biosynthesis. Arch 22. Birchfield NB, Latli B, Casida JE (1998) Human protoporphyrinogen oxidase: Relation Biochem Biophys 474:238–251. between the herbicide binding site and flavin cofactor. Biochemistry 12:6905–6910. 2. Panek H, O’Brian MR (2002) A whole genome view of prokaryotic haem biosynthesis. 23. Rumbley JN, Furlong Nickels E, Gennis RB (1997) One-step purification of histidine- Microbiology 148:2273–2282. tagged cytochrome bo3 from Escherichia coli and demonstration that associated 3. Jacobs NJ, Jacobs JM (1975) Fumarate as alternate electron acceptor for the late steps quinone is not required for the structural integrity of the oxidase. Biochim Biophys of anaerobic heme synthesis in Escherichia coli. Biochem Biophys Res Commun Acta 1340:131–142. 65:435–441. 24. Mogi T, et al. (2006) Probing the ubiquinol-binding site in cytochrome bd by site- 4. Jacobs NJ, Jacobs JM (1976) Nitrate, fumarate, and oxygen as electron acceptors for a directed mutagenesis. Biochemistry 45:7924–7930. late step in microbial heme synthesis. Biochim Biophys Acta 449:1–9. 25. Minohara S, Sakamoto J, Sone N (2002) Improved H þ ∕O ratio and cell yield of Escher- 5. Jacobs NJ, Jacobs JM (1978) Quinones as hydrogen carriers for a late step in anaerobic ichia coli with genetically altered terminal quinol oxidases. J Biosci Bioeng 93:464–469. heme biosynthesis in Escherichia coli. Biochim Biophys Acta 544:540–546. 26. Sturr MG, Krulwich TA, Hicks DB (1996) Purification of a cytochrome bd terminal 6. Sasarman A, et al. (1979) Mapping of a new hem gene in Escherichia coli K12. J Gen oxidase encoded by the Escherichia coli app locus from a delta cyo delta cyd strain Microbiol 113:297–303. complemented by genes from Bacillus firmus OF4. J Bacteriol 178:1742–1749. 7. Boynton TO, Daugherty LE, Dailey TA, Dailey HA (2009) Identification of Escherichia 27. Lenn T, Leake MC, Mullineaux CW (2008) Clustering and dynamics of cytochrome coli HemG as a novel, menadione-dependent flavodoxin with protoporphyrinogen bd-I complexes in the Escherichia coli plasma membrane in vivo. Mol Microbiol oxidase activity. Biochemistry 48:6705–6711. 70:1397–1407. 8. Vik SB (2008) ATP Synthesis by Oxidative Phosphoregulation, eds A Boeck et al. (ASM 28. Maklashina E, et al. (2006) Differences in protonation of ubiquinone and menaqui- Press, Washington, DC), http://www.ecosal.org/ Chapter 3.2.3. none in fumarate reductase from Escherichia coli. J Biol Chem 281:26655–26664. 9. Unden G, Duennwald P (2008) The Aerobic and Anaerobic Respiratory Chain of Escher- 29. Liu S, Neidhardt EA, Grossman TH, Ocain T, Clardy J (2000) Structures of human ichia Coli and Salmonella Enterica: Enzymes and Energetics, eds A Boeck et al. (ASM dihydroorotate dehydrogenase in complex with antiproliferative agents. Structure Press, Washington, DC), http://www.ecosal.org/ Chapter 3.2.2. 8:25–33. 10. Brzezinski P, Gennis RB (2008) Cytochrome c oxidase: Exciting progress and remaining 30. Norager S, Jensen KF, Bjornberg O, Larsen S (2002) E. coli dihydroorotate dehydro- mysteries. J Bioenerg Biomembr 40:521–531. genase reveals structural and functional distinctions between different classes of di- 11. Minagawa J, Mogi T, Gennis RB, Anraku Y (1992) Identification of heme and copper hydroorotate dehydrogenases. Structure 10:1211–1223. ligands in subunit I of the cytochrome bo complex in Escherichia coli. J Biol Chem 31. Couto SG, Nonato MC, Costa-Filho AJ (2008) Defects in vesicle core induced by 267:2096–2104. Escherichia coli dihydroorotate dehydrogenase. Biophys J 94:1746–1753. 12. Junemann S (1997) Cytochrome bd terminal oxidase. Biochim Biophys Acta 32. Ritz D, Beckwith J (2001) Roles of thiol-redox pathways in bacteria. Annu Rev Micro- 1321:107–127. biol 55:21–48. 13. Branden G, Gennis RB, Brzezinski P (2006) Transmembrane proton translocation by 33. Bader M, Muse W, Ballou DP, Gassner C, Bardwell JC (1999) Oxidative protein folding is cytochrome c oxidase. Biochim Biophys Acta 1757:1052–1063. driven by the electron transport system. Cell 98:217–227. 14. Gennis RB (2004) Coupled proton and electron transfer reactions in cytochrome 34. Bader M, Muse W, Zander T, Bardwell J (1998) Reconstitution of a protein disulfide oxidase. Front Biosci 9:581–591. catalytic system. J Biol Chem 273:10302–10307. 15. Cole JA, Richardson DC (2008) Respriration of nitrate and nitrite, eds A Boeck et al. 35. Kaysser TM, Ghaim JB, Georgiou C, Gennis RB (1995) Methionine-393 is an axial ligand (ASM Press, Washington, DC), http://www.ecosal.org/ Chapter 3.2.5. of the heme b558 component of the cytochrome bd ubiquinol oxidase from Escher- 16. Bertero MG, et al. (2003) Insights into the respiratory electron transfer pathway from ichia coli. Biochemistry 34:13491–13501. the structure of nitrate reductase A. Nat Struct Biol 10:681–687. 36. Wolff S, Hahne H, Hecker M, Becher D (2008) Complementary analysis of the vegeta- 17. Bertero MG, et al. (2005) Structural and biochemical characterization of a quinol tive membrane proteome of the human pathogen Staphylococcus aureus. Mol Cell binding site of Escherichia coli nitrate reductase A. J Biol Chem 280:14836–14843. Proteomics 7:1460–1468. 18. Magalon A, et al. (1997) Heme axial ligation by the highly conserved His residues in 37. Bergonia HA, Phillips JD, Kushner JP (2009) Reduction of porphyrins to porphyrino- helix II of cytochrome b (NarI) of Escherichia coli nitrate reductase A. J Biol Chem gens with palladium on carbon. Anal Biochem 384:74–78. 272:25652–25658. 38. Heinemann IU, et al. (2007) Functional definition of the tobacco protoporphyrinogen 19. Luna-Chavez C, Iverson TM, Rees DC, Cecchini G (2000) Overexpression, purification, IX oxidase substrate-binding site. Biochem J 402:575–580. and crystallization of the membrane-bound fumarate reductase from Escherichia coli. 39. Engel PC (1992) Chemistry and Biochemistry of Flavoenzymes, ed F Mueller (CRC Press, Protein Expres Purif 19:188–196. Boca Raton, FL), Vol 3, pp 597–699. 20. Qin X, et al. (2010) Structural insight into unique properties of protoporphyrinogen 40. Thauer RK, Jungermann K, Decker K (1977) Energy conservation in chemotrophic oxidase from Bacillus subtilis. J Struct Biol 170:76–82. anaerobic bacteria. Bacteriol Rev 41:100–180. 21. Koch M, et al. (2004) Crystal structure of protoporphyrinogen IX oxidase: A key 41. Wardman P (1989) Reduction potentials of one-electron couples involving free radicals enzyme in haem and chlorophyll biosynthesis. Embo J 23:1720–1728. in a quinone solution. J Phys Chem Ref Data 18:1637–1755.

6of6 ∣ www.pnas.org/cgi/doi/10.1073/pnas.1000956107 Möbius et al. Downloaded by guest on September 29, 2021