Electron transfer during the oxidation of ammonia by thechemolithotrophic bacterium Nitrosomonas europaea

Mark Whittaker, David Bergmann, David Arciero, Alan B. Hooper *Department of Biochemistry, Molecular Biology and Biophysics, University of Minnesota, St. Paul, MN 55108, USA

Abstract

The combined action of ammonia monooxygenase, AMO, (NH3+2e3+O2CNH2OH) and hydroxylamine oxidoreduc-tase, HAO, (NH2OH+H2OCHNO2+4e3+4H�) accounts for ammonia oxidation in Nitrosomonas europaea. Pathways forelectrons from HAO to O2, nitrite, NO, H2O2 or AMO are reviewed and some recent advances described. The membranecytochrome cM552 is proposed to participate in the path between HAO and ubiquinone. A bc1 complex is shown to mediatebetween ubiquinol and the terminal oxidase and is shown to be downstream of HAO. A novel, red, low-potential,periplasmic copper protein, nitrosocyanin, is introduced. Possible mechanisms for the inhibition of ammonia oxidation incells by protonophores are summarized. Genes for nitrite- and NO-reductase but not N2O or nitrate reductase are present inthe genome of Nitrosomonas. Nitrite reductase is not repressed by growth on O2 ; the flux of nitrite reduction is controlled atthe substrate level. ß 2000 Elsevier Science B.V. All rights reserved.

1. Introduction

Ammonia oxidation, electron transfer and en-ergy transduction has been reviewed in [1^4]. Theobligate aerobic chemolithotrophic bacterium Nitro-somonas derives energy from the oxidation of ammo-nia to nitrite (Fig. 1). Ammonia is oxidized to hy-droxylamine by ammonia monoxygenase, AMO,

(NH3+2e3+O2CNH2OH) [5]. Hydroxylamine isoxidized to nitrite by hydroxylamine oxidoreductase,HAO (NH2OH+H2OCHNO2+4e3+4H�) [6]. HAOwill also rapidly oxidize hydrazine to dinitrogen, anexperimentally useful reaction which may be of bio-logical signi¢cance in the microbial anaerobic trans-formation of ammonium and nitrite to dinitrogen [7].Cytochrome P460, a monoheme c-cytochrome withhydroxylamine oxidizing activity is also found insmall quantity and in the periplasm [8]. The contri-bution of this protein to inorganic nitrogen metabo-lism is not known.

A signi¢cant amount of NO and N2O [9,10] and atrace of N2 [11] are produced during oxidation ofammonia by Nitrosomonas at low concentrations ofoxygen. In vitro, NO and N2O can be products ofthe oxidation of hydroxylamine by HAO [12], how-ever, production of N2O in vivo appears to occur bythe reduction of nitrite [13^15]. A soluble enzymewith nitrite reductase and dye oxidase activity hasbeen isolated by several workers [16^18]. The enzyme

0005-2728 / 00 / $ ^ see front matter ß 2000 Elsevier Science B.V. All rights reserved.PII: S 0 0 0 5 - 2 7 2 8 ( 0 0 ) 0 0 1 7 1 - 7

Abbreviations: AMO, ammonia monooxygenase; HAO, hy-droxylamine oxidoreductase; DCCD, N,NP-dicyclohexylcarbodi-imide; DMSO, dimethylsulfoxide; CCCP, chlorocarbonylcya-nidephenyl hydrazone; ASN, ammonium sulfate supernate;ASP, ammonium sulfate precipitate ; PCR, polymerase chain re-action; EPR, electron paramagnetic resonance spectroscopy; ML,low-potential metal center; MH, high-potential metal center;pMMO, particulate methane monooxygenase; N, cytoplasmic(negative) side of membrane; P, periplasmic (positive) side ofmembrane

* Corresponding author. Fax: +1-612-625-5780;E-mail : hooper@biosci.umn.edu

BBABIO 44926 7-8-00

Biochimica et Biophysica Acta 1459 (2000) 346^355www.elsevier.com/locate/bba

contains type II copper [17]. In keeping with physio-logical observations, probable genes for one NO re-ductase and two dissimilatory nitrite reductase en-zymes appear in the genome whereas genes similarto a known nitrate or N2O reductase are not found[19].

On average, for every four electrons removed fromhydroxylamine by HAO and its electron acceptor,cytochrome c554 [20,21], 2 electrons must be re-turned to the AMO reaction, 1.65 electrons [1] passto the terminal oxidase (or to a nitrite- or NO-reduc-tase or diheme cytochrome c peroxidase, [22]) and0.35 pass to NAD. It has been proposed [1] thatNADH dehydrogenase, which is found in the ge-nome [19], catalyzes `reverse electron transfer' fromubiquinol to NAD. Cells contain ubiquinone-8 [23]in 13-fold excess relative to HAO and the electroncarriers cytochromes c554, c552 and aa3 [24]. Ubiqui-nol is thought to be an electron donor to the AMOsystem.

ATP synthase is driven by a proton gradient [25]which is generated by oxidation of hydroxylamine[26]. A value of vp of 173 mV (inside negative) hasbeen estimated for Nitrosomonas [27]. Invaginationof the cytoplasmic membrane result in an extensive`periplasmic' or `P-side' compartment of this Gram-negative bacterium [28]. This compartment containsthe HAO reaction [29] which thus contributes four

protons to the membrane gradient. Assuming thatthe terminal oxidase is a proton pump, re-oxidationof ubiquinol by a bc1 complex and cytochrome oxi-dase would yield the net translocation of six protonsfrom N to P side.

Here we address several recent advances related toelectron transfer in Nitrosomonas : (a) a survey ofelectron transfer components; (b) the role of the cy-tochrome bc1 complex; (c) the role of cytochromecM552; (d) The inhibition by protonophores of am-monia oxidation in cells ; and (e) control of the £uxof nitrite oxidation.

2. Materials and methods

2.1. Growth and harvesting of Nitrosomonas europaea

N. europaea (Schmidt strain) was grown in contin-uous culture as recorded previously [30] utilizing a15-liter stainless steel and glass reactor chamberunder high aeration. Cells were harvested by ¢ltra-tion through a Millipore Pellicon cassette using a0.45-Wm ¢lter and resuspended to 20% w/v in 50mM potassium phosphate at pH 7.8. The cell sus-pension was stored at 4³C and used within 24 h forwhole cell experiments or frozen at 320³C for pro-tein puri¢cation.

Fig. 1. Electron transport in Nitrosomonas europaea Established electron transport pathways are represented with a solid line and hy-pothetical pathways are represented with a dashed line. AMO, ammonia monooxygenase; HAO, hydroxylamine oxidoreductase; C-P460, cytochrome P460; Q, Ubiquinone-8; Cu NiR, copper-containing nitrite reductase; CCP, cytochrome c peroxidase; NOR, nitricoxide reductase, Cu aa3-type cytochrome c oxidase; CcM552, membrane cytochrome c552. `Alternative substrates' are substrates otherthan ammonia [4] and do not produce hydroxylamine.

BBABIO 44926 7-8-00

M. Whittaker et al. / Biochimica et Biophysica Acta 1459 (2000) 346^355 347

2.2. Preparation of Nitrosomonas for electronmicroscopy

To a 10-ml glass hypovial was added 1 ml Nitro-somonas which had been removed directly fromgrowth tank, washed once and resuspended in 50mM potassium phosphate at pH 7.8, and the con-tainer closed with an aluminum crimp cap and butylrubber septum. In triplicate, 1-Wl aliquots of DMSO,phenylacetylene in DMSO, or CCCP in DMSO wereadded to vials such that the ¢nal concentration ofphenylacetylene or CCCP was 10 WM. Cells wereincubated at room temperature with orbital shakingfor 20 min, after which ammonium sulfate was addedto a ¢nal concentration of 10 mM from a 1.25 Maqueous stock solution. Cells were incubated for anadditional 1.5 h, after which samples were ¢xed forelectron microscopy as described previously [28,31].Inhibition of CCCP and phenylacetylene-treated cellswas assayed by monitoring nitrite concentration aspublished previously [32]. In order to preserve os-motic balance, the bu¡er for the ¢xation procedurewas replaced with the bu¡er used for cell resuspen-sion.

2.3. General procedure for whole cell inhibitionexperiments

To 5 ml 50 mM potassium phosphate (pH 7.8) in a10-ml hypovial, equipped with an aluminum crimp

cap with Te£on-backed butyl rubber septum, wasadded inhibitor in DMSO stock solution to ¢nalconcentration followed by 20 Wl (for ammonia assay)or 5 Wl 20% w/v N. europaea cells (for hydroxylamineassay) from the above procedure for cell harvest,shaken orbitally for 20 min then the reaction initi-ated with the addition of aqueous solution of ammo-nium sulfate or hydroxylamine to ¢nal concentrationof 10 mM or 100 WM, respectively. Nitrite produc-tion was monitored every minute for the ¢rst 10 minand subsequently every 10 min over 50 min.

3. Results and discussion

3.1. A survey of electron transfer components

A sample of 13.5 g wet cells was analyzed. Anaccounting of the identity and quantity of c-cyto-chromes in membranes was complicated by the pres-ence of small amounts of HAO (thought to be inprocess of secretion) and large amounts of cyto-chrome c554 (thought to be bound by electrostaticinteraction to promote electron transfer [33], sincethey contain 21 and 4 c-hemes per molecule, respec-tively. Cytochrome c554 was removed by salt wash-ing of the membrane. Most heme c is found in thesoluble fraction (Table 1). HAO and cytochromec552 account for nearly all of the c-heme in the60^100% ammonium sulfate pellets (ASP) and are

Table 1Heme distribution in Nitrosomonas europaea

Heme c Heme b heme a

Soluble fraction 9.5540 ASP 1.49 0 060 ASP 2.25 0 0100 ASP 3.88 0 0100 ASN 1.48 0 0

Membrane fractionKCl Wash 1.36 0 0HCl/Acetone wash 0 0.89 0.91SDS wash 2.03 ^ ^

All values are given in Wmol. Cells (13.5 g wet weight) of N. europaea were disrupted by a freeze^thaw procedure [34]. After pelletingof the membrane fraction, the soluble fraction was fractionated with ammonium sulfate at 40%, 60%, and 100% of saturation. Themembranes were washed two times with bu¡er containing 1 M KCl [33]. After washing two times with bu¡er to remove KCl, themembranes were extracted two times with acidic acetone [35] to remove hemes a and b. The membrane pellet was then extracted twotimes with 0.15 M Tris/Cl, pH 8.5 bu¡er containing 2% SDS. The amount of heme was determined in alkaline pyridine [36].

BBABIO 44926 7-8-00

M. Whittaker et al. / Biochimica et Biophysica Acta 1459 (2000) 346^355348

present in a molar ratio of about 1:10 [18] (based onthe HAO monomer and minimum catalytic unit(HAO(K)), the ratio is 1:3.5). On a heme basis, theratio is about 2:1. Thus, based on the quantity ofheme c in the 60 and 100% ammonium sulfate pre-cipitates, the amount of HAO(K) and cytochromec552 present in the sample is 0.58 and 2.04 Wmol,respectively. The 100% ammonium sulfate super-natant (ASN) and the KCl wash contain almostexclusively the tetraheme cytochrome c554 hencethe total amount of cytochrome c554 present is0.71 Wmol.

All of the heme a and b is found associated withthe membrane. The membranes do not contain hemeo or other heme types. The 0.91 Wmol of heme aindicate the presence of 0.45 Wmol of aa3 oxidase.The 0.89 Wmol of heme b may be associated with atotal of 0.45 Wmol of bc1 complex and NO reductase,if it is expressed, though the presence of other b-heme containing proteins in the membrane cannotbe excluded. Of the 2.03 Wmol of c-heme in the mem-brane 0.45 Wmol are presumably distributed betweenthe bc1 complex and NO reductase, if present, andmembrane HAO and cytochrome c4 might accountfor as much as 0.2 Wmol. The remaining c-hemecould account for as much as 0.35 Wmol of thetetraheme cytochrome cM552. The relative amountsof electron transfer components are approximate-ly 1.0 HAO K subunit:1.3 cytochrome c554:0.6cytochrome cM552:13 ubiquinone-8:0.78 bc1 com-plex (together with NO reductase):3.5 cytochromec552:0.78 cytochrome aa3.

Based on sequencing of the N-terminus (ESPA-AGDVEKGKEIAAGICAGCHNADGN) and twointernal heme-containing tryptic peptides (HemePep-tide I: IAAGICAGCHNADGNSAIPLYPILAGQY-PGYIA; HemePeptide II: LYQGGNLENSIPAC-SSCHSPNGQGIPPHYTR), a cytochrome isolatedfrom the membrane fraction of N. europaea by wash-ing with H2O was identi¢ed as a cytochrome c4 andthis was con¢rmed by the complete sequence fromthe Nitrosomonas genome [19]. Cytochrome c4s havebeen reported in several Pseudomonas species [37],Azotobacter vinlandii [38], and Thiobacillus [39].These are well characterized diheme cytochromeswhose function may be as an immediate electrondonor to terminal oxidation systems, possibly actingin parallel with cytochrome c5 in donating electrons

to a cytochrome o oxidase [40]. The cytochrome c4

from Nitrosomonas europaea has similar spectroscop-ic and electrochemical properties to other membersof the cytochrome c4 family. However, a cytochromeo terminal oxidase does not appear to be present inthe organism. HPLC analysis for heme o has shownit is not present, and gene sequences for the polypep-tides have not been found in the genome. Thus, analternate role for cytochrome c4 in Nitrosomonasmust be contemplated. Since it is present at molarlevels about 5^10% of the main proteins involved innitri¢cation, it is unlikely to play a major role in theprocess.

Two other c-heme containing polypeptides haverecently been identi¢ed in minor amounts from thesoluble fraction of Nitrosomonas europaea. They co-purify during all puri¢cation steps and may representsubunits of a single protein. N-terminal sequencing(Polypeptide 1: ASDDSFEHAERLYDTYCTQCH-GVNRDGNGVNS; Polypeptide 2: ADVPAVLQ-SKCASCHALTKPESN) and a subsequent BLASTsearch of the Nitrosomonas genome database [19]reveals that their genes are in the same operon andmay also share an operon with an apparent outermembrane nirK, Cu-containing nitrite reductase.Although a soluble blue-copper nitrite reductasehas been isolated from N. europaea [18], the N-ter-minal sequence that we have obtained does notmatch this nirK gene.

3.2. Nitrosocyanin

A new monocopper cupredoxin has recently beenisolated and named nitrosocyanin [41]. It is a veryunusual red copper protein whose concentration isequivalent to the other components of the ammoniaoxidizing system (AMO, HAO and other c-cyto-chromes). The gene sequence has been obtained byprotein chemistry and PCR (D. Arciero, B. Tessema,A.B. Hooper, unpublished data). Based on the ge-nome sequence [19], the protein contains a leadersequence, thus the protein is periplasmic. The proteinsequence exhibits some homology to plastocyanin,the electron donor copper region of N2O reductase.The absorption band of the cupric form is at 390 nmrather than 600 or 450 nm peaks found in blue cop-per proteins or perturbed blue copper proteins, re-spectively. The redox potential (+85 mV) is low rel-

BBABIO 44926 7-8-00

M. Whittaker et al. / Biochimica et Biophysica Acta 1459 (2000) 346^355 349

ative to other members of this family (whose valuesare in the +300 mV range). The role of nitrosocyaninin Nitrosomonas has not been determined.

3.3. A cytochrome bc1 complex is `downstream' fromHAO

Genes for all proteins of a cytochrome bc1 arefound in the genome of Nitrosomonas [19], howeverthe expression of the cytochrome bc1 complex and itsparticipation in electron transfer between cyto-chrome c554 and the terminal oxidase has not beendocumented. Since the terminal oxidase is a CuAtype [42], the electron donor is presumably a peri-plasmic c-cytochrome, possibly cytochrome c552[20], rather than a quinol. Although the oxidation/reduction potential for cytochrome c552 (+220 mV,[20]) is in keeping with this role, genetic or kineticevidence has not established its signi¢cance. Sincetwo of the four electrons from hydroxylamine/HAO/cytochrome c554 may reduce ubiquinone andthen pass to AMO, it is logical that the two electronsbound for cytochrome c oxidase should also reduceubiquinone and pass through a bc1 complex to cyto-chrome oxidase. Acceptance of that route has beencomplicated by the readily reproducible observationthat, in vitro, electron £ow from substrate quantitiesof hydroxylamine to substrate quantities of the solu-ble cytochrome c552 is extremely rapid in the pres-ence of catalytic amounts of HAO and cytochromec554 [20,21]. This observation called for the sugges-tion that electron £ow from cytochrome c554 to the

terminal oxidase did not involve the cytochrome bc1

complex [2,3].Here we present evidence that, in vivo, the bc1

complex is, in fact, a component of the electrontransfer path, mediating electrons downstream ofHAO. As shown in Table 2, the aerobic oxidationof either hydroxylamine or ammonia by suspensionsof cells of Nitrosomonas is inhibited by antimycin A(50% inhibition at 1034 M), myxothiazol (1 mM),stigmatellin (1034 M), all of which are known inhib-itors of the bc1 complex [43]. The data indicate thatthe bc1 complex is expressed and functions down-stream from the HAO/cytochrome c554 pair. Thissimpli¢es the model pathway for electron transfer(Fig. 1). It also ¢ts well with the structures of cyto-chromes c554 and c552 and HAO; whereas a veryreasonable docking site for an HAO/cytochromec554 complex is observed, cytochromes c554 andc552 do not appear to share a complementary sur-face [44].

Cytochrome c552 is found in much higher concen-trations than the other redox components of the N-oxidation pathway. Further, its gene is present in asingle copy which is not in any of the three geneclusters encoding HAO and cytochrome c554 [45].Cytochrome c552 is also able to reduce cytochromec peroxidase [22] and possibly the copper nitrite re-ductase [17]. Thus, distribution of electrons to alter-nate carriers may be a function of that protein. It ispossible that, certain under circumstances in vivo,the bc1 complex can be circumvented. As seen here,inhibition by antimycin A, myxothiazol or stigmatel-

Table 2E¡ect of inhibitors of the bc1 complex on ammonia and hydroxylamine oxidation

Inhibitor Concentration (M) Rate of oxidation

NH3 (% of control) NH2OH (% of control)

Antimycin A 1035 41 45Myxothiazol 1034 67 75Stigmatellin 1034 23 18

1035 34 461036 41 49

Diphenylamine 1034 5 441035 87 691036 100 77

Known inhibitors of the bc1 ubiquinone reducing site (QN, antimycin A) and ubihydroquinone oxidizing site (QP, myxothiazol, stig-matellin, diphenylamine). The data shown are the percent remaining ammonia or hydroxylamine oxidizing capability relative to con-trol experiments containing no inhibitors. Hydroxylamine and cell concentrations were adjusted as described in Section 2 to preventinhibition of hydroxylamine oxidation in cells (observed at greater than 100 WM NH2OH).

BBABIO 44926 7-8-00

M. Whittaker et al. / Biochimica et Biophysica Acta 1459 (2000) 346^355350

lin is not complete. The reasons are not understoodbut, in this case, could include the utilization of analternate pathway such as the cytochrome c552 by-pass.

3.4. Cytochrome cM552; a possible electron donor toubiquinone

The gene encoding cytochrome cM552 is immedi-ately downstream of the gene for cytochrome c554(part of a gene cluster containing HAO) [46] and isco-transcribed with the latter gene [47]. The proteinis predicted to have four c-heme attachment sites.The N-terminal sequence of the puri¢ed detergent-solubilized protein appears to be the only transmem-brane domain [3]. Based on its amino acid sequence,this protein is a member of the `NirT/NapC' familyof tetraheme membrane-anchored cytochromes im-plicated by genetic experiments as electron transferagents from membrane quinols to periplasmic termi-nal electron acceptors (nitrate reductase, nitrite re-ductase, DMSO reductase, etc.) [46,48]. However,we and others [49] suggest that cytochrome cM552of Nitrosomonas carries electrons from the oxidationof hydroxylamine by HAO/c554 in the periplasm toubiquinone in the membrane. Although all membersof the NirT/NapC family have similar N-terminalportions with four c-heme binding motifs [46], alldi¡er in their C-terminal domains. Since they areall thought to react with membrane quinones andeach reacts with a unique periplasmic electron accep-tor (or donor), it is likely that the conserved portionsof these cytochromes all interact with quinones in themembrane whereas the variable C-terminal regionsinteract with di¡ering periplasmic electron transportproteins. The C-terminal domain of Nitrosomonas isparticularly unique; 24 of the 60 amino acids areaspartate or glutamate.

The optical spectrum of this protein is shown be-low (Fig. 2).

The only other member of this family of proteinswhich has had substantial biochemical analysis isNapC, the putative mediator to the periplasmic ni-trate reductase [48]. A contrast of properties andfunction between this protein and cytochromecM552 of Nitrosomonas is likely to prove interesting.

One property of this protein may be relevant tothe vP economy of Nitrosomonas. Since all of the

protein except the transmembrane domain is likelyto be periplasmic (or on the P face of the membrane),reductive protonation of ubiquinol anion would pre-sumably occur at the P face of the membrane. If thisis true, the reduction of two molecules of ubiquinonewould be promoted by vP and would, in e¡ect, con-sume the four protons yielded in the HAO reaction.This e¡ect would aid in overcoming the thermody-namically unfavorable values of the oxidation-reduc-tion potential for the 2 HAO-reducible hemes of cy-tochrome c554 (+47 mV, [50]) and ubiquinone (390mV [36]).

3.5. Inhibition of ammonia oxidation byprotonophores and inhibitors of proton transport

Ammonia dependent- but not hydroxylamine de-pendent- oxygen uptake and nitrite production bycells of Nitrosomonas is reversibly inhibited by theprotonophores chlorocarbonylcyanidephenyl hydra-zone (CCCP), tetrachlorosalicylanilide, and KKP-bis(hexa£uoroacetonyl)cyclohexanone (duPont 1799)in the 1035 M range and dinitrophenol in the 1034 Mrange (Table 3, [51]). The inhibitory e¡ect of CCCPis apparently on AMO or its electron donor systemand not the HAO-to-terminal oxidase reaction andelectron transfer pathway which is, in fact, stimu-lated by CCCP. The concentration dependence of

Fig. 2. Optical spectra of cM552. An aliquot of cM552 puri¢edthrough DEAE-Sepharose and gel ¢ltration chromatographywas diluted to 1.4 WM, based on heme absorbance at 408 nm,in 50 mM potassium phosphate (pH 7.0). The UV-visible spec-trum of the resting oxidized (solid line) and potassium dithion-ite-reduced (dashed line) protein was recorded in a 1-cm path-length quartz cuvette.

BBABIO 44926 7-8-00

M. Whittaker et al. / Biochimica et Biophysica Acta 1459 (2000) 346^355 351

the inhibition of ammonia oxidation and stimulationof hydroxylamine oxidation by CCCP are very sim-ilar, suggesting that the two e¡ects of CCCP share acommon mechanistic basis. The ionophores valino-mycin or gramicidin A are not inhibitory, hence ane¡ect on a sodium or potassium gradient is not in-volved.

As suggested by Wood [1], a simple mechanism forinhibition of ammonia oxidation by CCCP is thatthe dissipation of the proton gradient and the result-ing increased £ux of electrons through the bc1/cyto-chrome oxidase pathway to dioxygen (Fig. 1) (re-£ected in increased rate of hydroxylamineoxidation) decreases the £ux of electrons to AMO.The inhibition of ammonia oxidation by catalyticconcentrations of phenazine methosulfate [50] is anexample of this general mode of inhibition: PMS israpidly reduced by HAO and reoxidized by O2 [52],hence electrons would not reach AMO. A secondmechanism is based on the apparent ability ofCCCP to catalyze the decay of a reduced state ofthe reaction center in Photosystem II, `ADRY' (Ac-celeration of Deactivation Reactions of the watersplitting enzyme systems Y) [53]; in this way CCCPmight cause electrons to bypass AMO.

Inhibition by protonophores also raises the inter-esting possibility that ammonia oxidation is depen-dent on a proton gradient, possibly in order to trans-fer electrons from a redox carrier through a lowerpotential electron donor to AMO. Electron donorsto the active site of many monooxygenase enzymes

have values of midpoint potentials which arelower than the redox potential of ubiquinol, e.g.,£avins (3190 and 3328 mV) of microsomal cyto-chrome P450 reductase [54,55], putidaredoxin(3230 mV) of cytochrome P 450cam [56] or an Fe2S2

center (3220 mV) of soluble MMO [57]. Fig. 3illustrates a plausible mechanism for reverse elec-tron transfer within the AMO reaction (ubiqui-nol+NH3+O2C NH2OH+H2O+ubiquinone). If ap-propriately oriented in the membrane, the partialreactions (involving hypothetical low and highpotential metal centers, ML and MH, respectively):(a) ubiquinol+2M��

H Cubiquinone+2M��H � 2H�

on the N face; (b) 2M��L � 2M�

H ! 2M�L � 2M��

Hoccurring vectorally from N to P; and (c)2H� �O2 � 2M�

L �NH3 ! NH2OH� 2M��L on the

P face, would lead to utilization of two P protons,production of two N protons and vectoral movementof an electron from N to P side and could require acharge gradient for turnover of the enzyme. A pair ofb-cytochromes similar to those which have been pro-posed to act as electron donors to the pMMO, par-ticulate methane monooxygenase, of Methylococcuscapsulatus [58] could ¢ll the role of ML and MH. Theneed for proton gradient coupling would explain theobserved di¤culty in obtaining high levels of AMO(or pMMO) activity in crude extracts. Mechanisticcoupling of ammonia oxidation to transmembranemovement of protons and/or electrons would alsoexplain the presence of multiple transmembrane re-gions in AMO [59]. Possible precedent for reactions

Table 3E¡ect of protonophores, ionophores, and inhibitors of proton transport on ammonia and hydroxylamine oxidation

Inhibitor Concentration (WM) Rate of oxidation

NH3 (% of control) NH2OH (% of control)

CCCP 10 5 128FCCP 10 6 ND

10 19 ND10 91 ND

DCCD 10 0 29Valinomycin 10 95 82

10 100 95Gramicidin A 10 109 ND

10 115 ND10 97 ND

Assays were performed using the procedure outlined in Table 1, with stock solutions of inhibitors dissolved in DMSO or ethanol.ND, not determined.

BBABIO 44926 7-8-00

M. Whittaker et al. / Biochimica et Biophysica Acta 1459 (2000) 346^355352

driven by an ion gradient include the reduction ofCO2 to formaldehyde [60], the oxidation of nitrite[61] and the reduction of menaquinone by succinate[62].

The presence of large amounts of closely-opposedmembranes in ammonia- oxidizing bacteria [28] isconsistent with the need for an ion gradient butmight, alternatively, be needed for precise juxtaposi-tion of proteins in adjacent membranes. If so, mem-brane movement caused by osmotic changes follow-ing collapse of the proton gradient might inhibitAMO activity. With this in mind, possible majorchanges in membrane arrangements induced by com-plete inhibition of ammonia or hydroxylamine oxi-dation activity by CCCP in N. europaea and Nitro-sococcus oceanus were determined by examination ofthin sections by transmission electron microscopy.Nitrosococcus was chosen because the detection ofmajor changes was easier in the very large stacks ofmembrane discs. Photographs containing approxi-mately 10 control cells typically showed a varietyof membrane morphologies with the majority repre-sented by the images that have been published [28].Control and CCCP-treated cells were not noticeablydi¡erent with respect to the amount of closely-op-posed membranes. Hence the e¡ect on the oxidation

of ammonia- or hydroxylamine cannot be attributedto major disruption of cell ultrastructure.

Inhibition at 1034 M by N,NP-dicyclohexylcarbo-diimide (DCCD), known to inhibit proton transloca-tion of the F1F0-ATPase, cytochrome oxidases,anion- or divalent cation- translocases by covalentmodi¢cation of the proton translocation channel, iscomplete and irreversible (Table 3; Ref. [51]). Theoxidation of both ammonia and hydroxylamine areinhibited in an overlapping concentration range andcan thus probably be accounted for by an e¡ect onthe terminal oxidase.

3.6. Control of nitrite reduction in Nitrosomonas

Some elements of the control of the £ux of nitritereduction are known. First, the fact that the nitritereductase can easily be isolated from aerobically-grown cells indicates that expression the enzyme isnot completely repressed by O2. Given the low den-sity to which N. europaea grows, we wished to ruleout the slight possibility that a nitrite reductase couldhave been induced under anaerobic conditions late inthe culture cycle or during the accumulation of cellsprior to use. Cells from a chemostat culture wererapidly harvested and frozen in liquid N2 prior to

Fig. 3. Hypothetical mechanism for proton gradient-driven ammonia oxidation activity. Electrons generated by HAO through oxida-tion of hydroxylamine are passed through c554 and to ubiquinone-8 via cM552 (left hand dashed box). Electrons from ubiquinol arepassed either to the bc1 complex (and cytochrome c oxidase) or to AMO. AMO is represented in the right-hand tripartite dashed box.ML and MH are low and high potential metal centers, respectively. The AMO reaction is represented as containing (a) an N-side Ubi-quinol^MH reductase, (b) a transmembrane MH^ML reductase, and (c) an ML-dependent oxygen activating center.

BBABIO 44926 7-8-00

M. Whittaker et al. / Biochimica et Biophysica Acta 1459 (2000) 346^355 353

use, when they were broken in a French pressure celland soluble extracts prepared by ultracentrifugation.We observed comparable nitrite reductase activity,assayed as p-phenylaminediamine oxidation [24] insoluble extracts from cells grown at high levels ofoxygen (100^40% saturation) and in cells grown atlow levels of oxygen (about 1.0% saturation). Thecontrasting report of Miller and Nicholas [17], whopuri¢ed substantial quantities of nitrite reductasefrom cultures grown at high aeration but obtained10-fold greater quantities of the reductase from cul-tures grown at low rates of aeration, indicates thatthe situation is complex.

Secondly, nitrite reductase, though present, doesnot transform nitrite at concentrations up to 40mM when the oxygen concentration is high: culturesroutinely convert ammonia stoichiometrically intobiomass and nitrite. When the oxygen concentrationdecreases to the 1% range, N2O will account for amajor fraction of the ammonia-N utilized [10].Hence, in the presence of nitrite, nitrite reductase ispresent but reacts only when oxygen is diminished.This may account for the observation that in a ni-trite-containing chemostat culture, N2O productionwas initiated within 15 min after a decrease in oxy-gen concentration [63]. Hence, although increasedN2O production resulting from de novo synthesisof denitrifying enzymes cannot be ruled out, avail-able evidence suggests that the catalytic turnover ofthe enzyme is apparently modulated at the metabo-lite level. Although dioxygen reacts with nitrite re-ductase and might outcompete nitrite [17], this £uxof electrons would represent the loss to the cell of atleast the proton gradient generated by cytochromeoxidase. Thus a more attractive hypothesis involvesthe obligatory participation of a low-potential elec-tron donor in the electron transport pathway leadingto nitrite reductase. The £ux of electrons to nitritereductase would be zero as long as this donor re-mained oxidized and re-initiate when the redox po-tential of the cell was decreased by limitation of oxy-gen.

Acknowledgements

Supported by Research Grants from the National

Science Foundation (MCB9723608) and Departmentof Energy (DE-FG02-95FR20191).

References

[1] P.M. Wood, in: J.I. Prosser (Ed.), Nitri¢cation, Spec. Pub.Soc. Gen. Microbiol. vol. 20, 1986, pp. 39^62.

[2] A.B. Hooper, in: H.G. Schlegel, B. Bowien (Eds.), Autotro-phic Bacteria, Sci. Tech. Publishers, Madison, WI, 1989, pp.239^265.

[3] T. Vannelli, D. Bergmann, D.M. Arciero, A.B. Hooper, in:M.E. Lidstrom, F.R. Tabita (Eds.), Proceedings of the 8thInternational Symposium on Microbial Growth on C1 Com-pounds, Kluwer, Dordrecht, 1996, pp. 80^87.

[4] A.B. Hooper, T. Vannelli, D.J. Bergmann, D.M. Arciero,Antonie van Leeuwenhoek 71 (1996) 59^67.

[5] T.C. Hollocher, M.E. Tate, D.J.D. Nicholas, J. Biol. Chem.256 (1981) 10834^10836.

[6] K.K. Andersson, A.B. Hooper, FEBS Lett. 164 (1983) 236^240.

[7] J. Schalk, S. de Vries, G. Kuenen, M.S.M. Jetten, Biochem-istry 39 (2000) 5405^5412.

[8] D. Bergmann, A.B. Hooper, FEBS Lett. 353 (1994) 324^326.[9] I.C. Anderson, J.S. Levine, Appl. Environ. Microbiol. 51

(1986) 938^945.[10] T.J. Goreau, W.A. Kaplan, S.C. Wofsy, M.B. McElroy,

F.W. Valois, S.W. Watson, Appl. Environ. Microbiol. 40(1980) 526^532.

[11] M. Poth, Appl. Environ. Microbiol. 52 (1986) 957^959.[12] A.B. Hooper, K.R. Terry, Biochim. Biophys. Acta 571

(1979) 12^20.[13] M. Poth, D.D. Focht, Appl. Environ. Microbiol. 49 (1985)

1134^1141.[14] A.B. Hooper, D.M. Arciero, A.A. DiSpirito, J. Fuchs, M.

Johnson, F. LaQuier, G. Mundfrom, H. McTavish, in: P.M.Gresshof, W.E. Newton, W.E. Roth, G. Stacey (Eds.), Ni-trogen Fixation: Achievements and Objectives, Chapman-Hall, New York, 1990, pp. 387^391.

[15] A. Remde, R. Conrad, Arch. Microbiol. 154 (1990) 187^191.[16] D.J. Miller, P.M. Wood, J. Gen. Microbiol. 129 (1983)

1645^1650.[17] D.J. Miller, D.J.D. Nicholas, J. Gen. Microbiol. 131 (1985)

2851^2854.[18] A.A. Dispirito, L.R. Taa¡e, J.D. Lipscomb, A.B. Hooper,

Biochim. Biophys. Acta 827 (1985) 320^326.[19] J. Lamerdin, F. Larimer, D. Arp, A. Hooper, J. Norton, M.

Klotz, L. Sayavedra-Soto, D. Arciero, N. Hommes, M.Whittaker, American Society of Microbiology Annual Meet-ing, Los Angeles, CA, May 24, 2000, abstract H-141.

[20] T. Yamanaka, M. Shinra, J. Biochem. 75 (1974) 1265^1273.[21] D. Arciero, C. Balny, A.B. Hooper, Biochemistry 30 (1992)

11466^11472.[22] D.M. Arciero, A.B. Hooper, J. Biol. Chem. 269 (1994)

11878^11886.

BBABIO 44926 7-8-00

M. Whittaker et al. / Biochimica et Biophysica Acta 1459 (2000) 346^355354

[23] A.B. Hooper, R.H. Erickson, K.R. Terry, J. Bacteriol. 110(1972) 430^438.

[24] A.A. DiSpirito, L.R. Taa¡e, A.B. Hooper, Biochim. Bio-phys. Acta 806 (1985) 320^330.

[25] S. Kumar, D.J.D. Nicholas, FEMS Microbiol. Lett. 14(1982) 21^25.

[26] T.C. Hollocher, S. Kumar, D.J.D. Nicholas, J. Bacteriol. 149(1982) 1013^1020.

[27] S. Kumar, D.J.D. Nicholas, J. Bacteriol. 154 (1983) 65^71.[28] R.G.E. Murray, S.W. Watson, J. Bacteriol. 89 (1965) 1594^

1609.[29] A.B. Hooper, A.A. DiSpirito, T.C. Olson, K.K. Andersson,

W. Cunningham, L. Taa¡e, in: R. Crawford, R. Hanson(Eds.), Microbial Growth on C1 Compounds, 1984, pp.53^58.

[30] M.S.P. Logan, A.B. Hooper, Biochemistry 34 (1995) 9257^9264.

[31] E. Kellenberger, A. Ryter, J. Sechaud, J. Biophys. Biochem.Cytol. 4 (1958) 671^676.

[32] D. Arciero, T. Vannelli, M. Logan, A.B. Hooper, Biochem.Biophys. Res. Commun. 159 (1989) 640^643.

[33] H. McTavish, D.M. Arciero, A.B. Hooper, Arch. Biochem.Biophys. 324 (1995) 53^58.

[34] A.B. Hooper, P.C. Maxwell, K. Terry, Biochemistry 17(1978) 2984^2989.

[35] J.-H. Fuhrup, K.S. Smith, in: Laboratory Methods in Por-phyrin and Metalloprotein Research, Elsevier, Amsterdam,1975.

[36] E.A. Berry, B.L. Trumpower, Anal. Biochem. 161 (1987) 1^15.

[37] G.W. Pettigrew, K.R. Brown, Biochem. J. 252 (1988) 427^435.

[38] R.T. Swank, R.H. Burris, Biochim. Biophys. Acta 180(1969) 473^489.

[39] C. Cavazza, M.T. Giudici-Orticoni, W. Nitschke, C. Appia,V. Bonnefoy, M. Bruschi, Eur. J. Biochem. 242 (1996) 308^314.

[40] T.C.N. Ng, A.N. Laheri, R.J. Maier, Biochim. Biophys.Acta 1230 (1995) 119^120.

[41] A.B. Hooper, D.M. Arciero, J. Inorgan. Biochem. 74 (1999)166.

[42] A.A. DiSpirito, J.D. Lipscomb, A.B. Hooper, J. Biol. Chem.261 (1986) 17048^17056.

[43] B.M. Geier, U. Haase, G. van Jagow, Biochem. Soc. Trans.22 (1994) 203^209.

[44] T.M. Iverson, D.M. Arciero, B.T. Hsu, M.S.P. Logan, A.B.Hooper, D.C. Rees, Nat. Struct. Biol. 5 (1998) 1005^1012.

[45] H. McTavish, F. LaQuier, D. Arciero, M. Logan, G. Mund-from, J. Fuchs, A.B. Hooper, J. Bacteriol. 175 (1993) 2445^2447.

[46] D. Bergmann, D. Arciero, A.B. Hooper, J. Bacteriol. 176(1994) 3148^3153.

[47] D. Bergmann, Ph.D. dissertation, University of Minnesota,1996.

[48] M.D. Roldan, H.J. Sears, M.R. Cheesman, S.J. Ferguson,A.J. Thomson, B.C. Berks, D.J. Richardson, J. Biol. Chem.44 (1998) 28785^28790.

[49] D.J. Richardson, N.J. Watmough, Curr. Opin. Chem. Biol.3 (1999) 207^219.

[50] D.M. Arciero, M. Collins, J. Haladjian, P. Bianco, A.B.Hooper, Biochemistry 30 (1991) 11459^11465.

[51] A. Hooper, K. Terry, J. Bacteriol. 115 (1973) 480^485.[52] A. Hooper, A. Nason, J. Biol. Chem. 240 (1965) 4044.[53] B. Hanssum, G. Dohnt, G. Renger, Biochim. Biophys. Acta

806 (1985) 210^220.[54] T. Iyanagi, N. Makino, H.S. Mason, Biochemistry 13 (1974)

1701^1710.[55] J.L. Vermilion, M.J. Coon, J. Biol. Chem. 253 (1978) 8812^

8819.[56] M. Sono, M.P. Roach, E.D. Coulter, J.H. Dawson, Chem.

Rev. 96 (1996) 2841^2887.[57] J. Lund, H. Dalton, Eur. J. Biochem. 147 (1985) 291^296.[58] J.A. Zahn, A.D. DiSpirito, J. Bacteriol. 178 (1996) 1018^

1029.[59] H. McTavish, J. Fuchs, A.B. Hooper, J. Bacteriol. 175

(1993) 2436^2444.[60] B. Kaesler, P. Schonheit, Eur. J. Biochem. 184 (1989) 223^

232.[61] J.G. Cobley, Biochem. J. 156 (1976) 493^498.[62] J. Schirawski, G. Unden, Eur. J. Biochem. 257 (1998) 210^

215.[63] R.A. Kester, R. Laanbroek, Appl. Environ. Microbiol. 63

(1997) 3872^3877.

BBABIO 44926 7-8-00

M. Whittaker et al. / Biochimica et Biophysica Acta 1459 (2000) 346^355 355