Biological Sulfide Oxidation: Sulfide-Quinone Reductase ... · Biological Sulfide Oxidation: Sulfide-Quinone Reductase (SQR), the Primary Reaction Christoph Griesbeck, Günter Hauska

Christoph Griesbeck, Günter Hauska and Michael Schütz (2000): Biological Sulfide-Oxidation: Sulfide-Quinone Reductase (SQR), the Primary Reaction.In: Pandalai, S.G. (ed): Recent Research Developments in Microbiology, Vol. 4, 179-203. Research Signpost, Trivadrum , India.

Biological Sulfide Oxidation: Sulfide-Quinone Reductase (SQR), the Primary Reaction

Christoph Griesbeck, Günter Hauska and Michael Schütz

Lehrstuhl für Zellbiologie und Pflanzenphysiologie, Universität RegensburgUniversitätsstrasse 31, 93040 Regensburg, Germany

Contents

ABSTRACT1. INTRODUCTION: PHOTO- AND CHEMOLITHOTROPHIC GROWTH ON SULFIDE2. DISCOVERY OF SQR3. PROPERTIES OF SQR3.1. PROPERTIES OF SQR BOUND TO THE MEMBRANE3.2. SUBCELLULAR LOCALIZATION OF SQR3.3. ISOLATED SQR ENZYMES4. SQR GENES AND PROTEIN SEQUENCES4.1 CLONING OF THE SQR GENE FROM R. CAPSULATUS

4.2 CLONING OF CYANOBACTERIAL SQR GENES4.3 COMPARISON OF BACTERIAL SQR SEQUENCES4.4. FAD-BINDING4.5. AMINO ACID RESIDUES THAT MIGHT BE INVOLVED IN QUINONE-BINDING4.6. AMINO ACID RESIDUES THAT MIGHT BE INVOLVED IN SULFIDE OXIDATION5. MECHANISM OF SULFIDE OXIDATION6. REGULATION OF EXPRESSION OF SQR6.1. EXPRESSION OF SQR IN R. CAPSULATUS

6.2. EXPRESSION OF SQR IN CYANOBACTERIA7. OCCURENCE OF SQR7.1. SQR IN PROKARYOTES7.2. THE HMT2 GENE FROM SCHIZOSACCHAROMYCES POMBE

8. APPLICATIONS OF SULFIDE OXIDIZING MICROORGANISMS9. OPEN QUESTIONS AND PROSPECTS

ACKNOWLEDGEMENTREFERENCES

Abbrevations used are: SQR, sulfide-quinone reductase; FCC, flavocytochrome c; ET, electrontransport; Q, quinone; UQ, ubiquinone; PQ, plastoquinone; SDS-PAGE, sodium dodecylsulfatepolyacrylamide gel electrophoresis; bp, basepairs; NQNO, nonylhydroxyquinoline-N-oxide;HQNO, heptylhydroxyquinoline-N-oxide.

Christoph Griesbeck, Günter Hauska and Michael Schütz (2000): Biological Sulfide-Oxidation: Sulfide-Quinone Reductase (SQR), the PrimaryReaction. In: Pandalai, S.G. (ed): Recent Research Developments in Microbiology, Vol. 4, 129-203. Research Signpost, Trivadrum , India.

ABSTRACT

Inorganic reduced sulfur compounds as sulfidecan serve as electron donors for phototrophicand lithotrophic growth of Bacteria andArchaea. For the initial step of sulfide oxidationmainly two enzymatic systems are discussed,flavocytochrome c and sulfide-quinonereductase. Sulfide-quinone reductase (SQR),an ancient flavoprotein, is obligatory forgrowth on sulfide as hydrogen donor in photo-and chemolithoautotrophic bacteria. During thelast ten years it has been studied in detail forthe cyanobacterium Oscillatoria limnetica andthe proteobacterium Rhodobacter capsulatus.In both cases the enzyme has been purified tohomogeneity, and the genes have been cloned,sequenced and expressed in E. coli. SQR iswidely spread among Eubacteria, asdocumented by the occurence of genes and/oractivities, the latter even including mitochondriafrom lower animals. In this review latest resultsabout SQR, its function, occurence andexpression are summarized, and implicationsfrom gene sequences and biotechnologicalapplications are discussed.

1. INTRODUCTION: PHOTO- ANDCHEMOLITHO-TROPHIC GROWTH ONSULFIDE

For more than a century, conversion ofinorganic sulfur compounds by biologicalprocesses has been well established. Hydrogensulfide, the most reduced form of inorganicsulfur, occurs in hydrothermal vents, as well asin sediments, where it is generated by sulfate-reducing bacteria [1,2]. Although hydrogensulfide is toxic for most organisms, mainlybecause of the inhibition of aerobic respiration[3], it serves as electron donor for the energygenerating systems of photo- andchemolithotrophic bacteria (for reviews see [4-

6]), of some archaea [7] and even of eukarya(reviewed in [8]).

During autotrophic growth of bacteria,hydrogen sulfide provides electrons forreduction of NAD+ that is required for CO2

fixation. The midpoint potential of theNAD+/NADH couple at pH 7 is approximately50 mV more negative than the midpointpotential of the S0/H2S couple [9]. Therefore,energy is required to transport electrons fromhydrogen sulfide upwards to NAD+ (see Fig.1). In all bacteria studied so far, the transportof electrons from sulfide to NAD+ is mediatedby membrane bound electron transport (ET).This is depicted in Fig. 1 and Fig. 2 forphotosynthetic and respiratory(chemolithotrophic) proteobacteria. Theelectrons from sulfide enter the chain either atthe level of quinone, via a sulfide:quinoneoxidoreductase (SQR) - depending on thequinone species the midpoint potential of theQ/QH2 couple is 100-300 mV more positivethan the S0/H2S couple - or at the level of c-type cytochromes via a sulfide:cytochrome coxidoreductase (flavocytochrome c, FCC). C-type cytochromes exhibit midpoint potentialsthat are 500-700 mV more positive than themidpoint potential of the sulfide couple.

Phototrophic bacteria use light for the upwardtransport of electrons from sulfide to NAD+. Inproteobacteria and Chloroflexaceae thephotosystem contains a quinone-type reactioncenter [10]. As depicted in Fig. 1, theelectrons are first transferred to quinone andhave to be transported uphill to NAD+ byreverse ET through the NADH dehydrogenase.This reverse transport requires anelectrochemical proton potential (∆µH+) acrossthe cytoplasmic membrane that is generated bycyclic ET through the photosystem. As far assufficient light is available, it makes nosignificant

difference, whether electrons from sulfide enterthe ET chain at the level of cytochrome c or atthe level of quinone, however, at low light thismay be decisive. If the electrons enter the ETchain via FCC at the level of cytochrome c,they have to be transported upwards to NAD+

by reverse ET through the cytochrome bccomplex, the quinone pool and the NADH

dehydrogenase (Fig. 2A). If the electrons enterthe ET chain at the level of quinone via SQR,then the cytochrome bc complex contributes tothe generation of ∆µH+ that is necessary for theupward transport from quinol to NAD+ throughthe NADH dehydrogenase. Thus, sulfideoxidation by SQR provides more energy thansulfide oxidation by FCC.

Cyanobacteria do not face this problem,because they are able to transfer electrons fromcytochrome c to NAD+ without the help of∆µH+ via a photosystem of the reaction centerI-type and the ferredoxin:NAD+ oxido-reductase [10]. This may also hold true forChlorobiaceae and heliobacteria, whichcontain a type-I reaction center as well.

Chemolithotrophic bacteria gain all energy frominorganic substrates by dark redox reactions.Respiratory ET from the H2S/ S0 to O2/H2O,comprising a redox potential difference ofabout 1 V, serves as the “proton pump“ anddrives the upward ET from sulfide to NAD+

(Fig. 1). This again is more efficient with SQRthan with FCC, which is important at lowoxygen pressures, when the oxidases becomelimiting (Fig. 2B). Higher efficiency of energyconservation by sulfide oxidation via SQRover FCC certainly is critical under themixotrophic und changing conditions in thenatural habitat.

Both enzymatic sulfide oxidizing systems, thesulfide:cytochrome c oxidoreductase namedflavocytochrome c (FCC) and sulfide:quinoneoxidoreductase (SQR), were isolated frombacteria. FCC is a soluble periplasmic enzyme(Fig. 2). It consists of a FAD-binding subunitand a cytochrome c subunit. The enzyme iswell studied with respect to its catalyticmechanism [11-15], and the three-dimensionalstructure of FCC from Allochromatiumvinosum has been solved to a resolution of2.53Å [16,17]. Although the enzyme was

E (mV)

P870

P870*

hν

-800

-400

0

+400

+800 O2

BPhe

Q pool

H S2NAD

+

DH

cyt c

SQR

FCC

QA

QB

cyt oxidasequinol

oxidase

∆µH+

cyt bc

PET

CET

Fig. 1: Redox Scale of Photo- andChemolithotrophic Electron Transportfrom Sulfide in Proteobacteria.Abbreviations used: PET, photosyntheticelectron transport; CET, chemotrophicelectron transport; P870, special pairchlorophyll; BPhe bacteriopheophytin, QA

and QB, quinone-binding sites of the reactioncenter; DH, NADH-dehydrogenase; cyt bc,cytochrome bc complex; cyt c, cytchrome c;cyt oxidase, cytochrome oxidase.


== KCNFCC

RC

FAD ==

==

==

I1

I2

I3

NAD + H NADH+ +

n H+

DH

== KCNFCC

FAD ==

==

==

I1

I2

I3

NAD + H NADH+ +

n H+

DH

2 H+

2 H+

quinoloxidase

A phototrophic ET

B chemotrophic ET

Fig. 2: Topography of Electron Transport from Sulfide in CytoplasmicMembranes of Proteobacteria. A, phototrophic electron transport and B,chemotrophic electron transport. Sites of inhibition of the cytochrome bc complex andof SQR by quinone-analogues are designated as I1, I2 and I3. Inhibitors acting at I1 areantimycin A, aurachin C, NQNO and HQNO, at I2 myxothiazol, stigmatellin and at I3

myxothiazol, antimycin A, stigmatellin, aurachin C, NQNO, HQNO and KCN. (DH,NADH-dehydrogenase; Cyt b, FeS and Cyt c1 stand for the cytochrome b, Rieske ironsulfur center and cytochrome c1 subunits of the cytochrome bc complex; bh and bl, highand low potential heme of cytochrome b; Qi and Qo, quinone-reduction and quinone-oxidation sites of the bc complex; QS, quinone-binding site of SQR; QB, quinone-reduction site of the photosystem; RC, photosynthetic reaction center; cyt c,cytochrome c).

found in a variety of chemotrophic andphototrophic sulfide-oxidizing bacteria, it is notobligatory in sulfide oxidation. FCC does notoccur in a variety of sulfide-oxidizing bacteria,and the enzyme seems to be confined tospecies that are also able to oxidize thiosulfate[4,6]. Moreover, mutational inactivation ofFCC did not have any significant effect on thesulfide oxidizing ability of the phototrophicsulfur bacterium A. vinosum [18].

During the last twenty years evidence hasaccumulated for the second sulfide oxidationpathway. The enzyme that catalyzes the sulfide-dependent reduction of quinones, thesulfide:quinone oxidoreductase (E.C.1.8.5.´.;SQR), is a single polypeptide with an apparentmolecular mass of 55 kDa [19,20]. It is amembrane-bound enzyme (Fig. 2) that belongsto the glutathion reductase family offlavoproteins. SQR activity has been found tobe widely distributed among prokaryotes (seeFig. 5), and protein sequence comparisonleads to the conclusion that SQR is aphylogenetically very old enzyme, that wasacquired early in evolution, as reviewedrecently [21].

In the present account, this review will beupdated, with some emphasis on SQR from R.capsulatus. After a view on the biochemicalproperties of the enzyme in the membrane andof the isolated protein, characteristics evidentfrom peptide sequence comparison of SQRsare summarized, and some mechanistic aspectsof the catalytic reaction are discussed. Finallywe focus on some aspects of the regulation ofSQR expression and give a glimpse onpossible applications of sulfide-oxidizingbacteria.

2. DISCOVERY OF SQR

First evidence for the participation of quinonesin sulfide oxidation came from the work ofKnaff and Buchanan [22]. They found that thesulfide-dependent photoreduction of NADP+ inmembranes of the green sulfur bacteriumChlorobium thiosulfatophilum (Tassarjara)was sensitive towards inhibitors of cytochromebc and b6f complexes, which are membrane-bound quinol:cytochrome c oxidoreductases.As shown in Fig. 2, these inhibitors arequinone analogues that compete with thequinone substrate for the quinone-binding sitesof these complexes (for reviews see [23,24]).More than ten years later, Brune and Trüper[25] found that membrane vesicles of thepurple "non-sulfur" bacterium Rhodobactersulfidophilus photoreduce NAD+ with sulfideas the electron donor in a reaction sensitivetowards quinone-analogues, and theydemonstrated that sulfide oxidation in the darkleads to the reduction of quinones within themembranes. On the basis of their observationsthey suggested that the electrons from sulfideenter the membrane-bound photosynthetic ETchain at the level of quinone.

The cyanobacterium Oscillatoria limneticaswitches from oxygenic plant typephotosynthesis to anoxygenic photosynthesisin presence of sulfide [26]. Sulfide-dependentformation of a proton gradient across thethylakoid membranes of O. limnetica wasobserved [27], which was again blocked byinhibitors of bc and b6f complexes.

Furthermore, Arieli et al. [28] established anassay for measuring sulfide-quinoneoxidoreductase activity by monitoring thereduction of externally added quinones in adual wavelength spectrophotometer. Theyfound that in presence of thylakoid membranesfrom sulfide-induced O. limnetica cellsexternally added quinone becomes rapidlyreduced by sulfide in the dark. This reactionwas sensitive towards quinone analogues (Tab.


1). From their results they proposed thefollowing reaction scheme for the sulfide-dependent ET in membranes. Sulfide becomesoxidized by a membrane-bound SQR, and theelectrons are transferred to quinone at thequinone-binding site of SQR. Subsequently,the cytochrome b6f complex reoxidizes quinoland reduces plastocyanine that transfers theelectrons to photosystem I. As a result, aproton gradient is formed across the membraneby the Q-cycle mechanism of the b6f complex(for a review of the Q-cycle see [24]).

Participation of the energy gaining Q-cyclemechanism in the sulfide-dependent ETpathway via SQR was first monitored inmembranes of the purple "non-sulfur"bacterium R. capsulatus by time resolvedmeasurements of the sulfide-dependentreduction of membrane-bound b- and c-typecytochromes [29]. Up to now, membrane-bound SQR activity and subsequent oxidationof quinol by the Q-cycle mechanism was foundin a variety of phototrophic and chemotrophicbacteria (see Fig. 5) [18,21,28-34]. In Fig. 2models of the sulfide oxidation pathway asevident from measurements with membranesfrom phototrophic and chemotrophicproteobacteria are shown.

3. PROPERTIES OF SQR

3.1. PROPERTIES OF SQR BOUND TO THEMEMBRANE

Properties of the best studied SQRs inmembranes are summarized in Tab. 1. Wheredetermined, Km values of SQRs for sulfide andexternally added quinone substrates are in themicromolar range. The highest specific SQR-activity was found in membranes of thehyperthermophilic, chemotrophic bacteriumAquifex aeolicus. At room temperature, it wasmore than tenfold higher than the activitiesmeasured in membranes from proteobacteria as

well as from the cyanobacterium O. limneticaand the green sulfu bacterium Chlorobium.Close to the optimal growth temperature ofAquifex (85°C [35]), the specific activity waseven more than 50 times higher. It is so farunknown, whether this high activity is due to ahigh amount of enzyme in membranes or to ahigh turnover rate. However, this high activitysuggests that hydrogen sulfide, besideshydrogen, is a main electron source for thisbacterium that was isolated from marinehydrothermal vents [35].

All SQR enzymes were sensitive towardsquinone-analogues at micromolar or nanomolarconcentrations (Tab. 1; in Fig. 2 this inhibitionis represented by the site I3, while I1 and I2

reflect the two Q-binding sites on the bc-complexes [24]). Sensitivity towards theseinhibitors is a characteristic feature of quinone-binding enzymes. Except for the enzyme fromParacoccus denitrificans, aurachin C was themost potent inhibitor of sulfide-dependentreduction of externally added quinone by SQR.In contrast to the enzyme from proteobacteriaand green sulfur bacteria, SQR of thecyanobacterium O. limnetica was insensitivetowards the inhibitors myxothiazol andantimycin A. Both are inhibitors of ubiquinol-oxidizing cytochrome bc1 complexes ofproteobacteria and mitochondria, but do notinhibit the plastoquinol-oxidizing b6f complexespresent in thylakoids of cyanobacteria andplants (for reviews see [23,24]). Themenaquinol-oxidizing bc complex ofChlorobiaceae combines properties of both[36] and is antimycin A sensitive [31]. Thissuggests that the different sensitivities of SQRstowards quinone-analogues reflect differencesof the quinone substrates naturally occurring inmembranes.

Cyanide is a potent inhibitor of manyflavoproteins in micromolar concentrations[37] and inhibits sulfide oxidation by FCC

(Fig. 2) both in phototrophic [38] andchemotrophic bacteria [33]. Cyanide was alsofound to inhibit sulfide-reduction of quinonesby SQR in membranes of phototrophicbacteria (Tab. 1). In contrast, cyanide did notinhibit SQR-activity in membranes of thechemotrophic bacteria A. aeolicus and P.denitrificans suggesting differences of SQRenzymes in phototrophic and chemotrophicbacteria. The structural basis for this differenceremains unknown and needs to be investigated.

3.2. SUBCELLULAR LOCALIZATION OF SQR

After cell disruption almost all SQR-activitywas found in the membrane fraction in allsystems that have been investigated until now.Except for the enzyme from R. capsulatus, allSQR enzymes are tightly bound to themembrane and were solubilized frommembranes only by detergent treatment[18,19,34].

In contrast, SQR from R. capsulatus wasdetached from the membrane by treatment with

Tab. 1: SQR-activity in bacterial membranes, substrate affinitiy and sensitivity towards

inhibitors. Km values for sulfide-quinone reductase (SQR), average specific activity of sulfide-

dependent quinone reduction at 20°C and the efficiency of inhibitors in membranes of Aquifex

aeolicus [34], Paracoccus denitrificans [32], Rhodobacter capsulatus [29], Allochromatium

vinosum [18], Oscillatoria limnetica [19,28] and Chlorobium limicola f. thiosulfatophilum [30].

For A. vinosum membranes, the quinone substrate was duroquinone [18]. I50-values give the

micromolar concentrations of the inhibitors for 50 % inhibition (unit, µmol quinone reduced min-1;

d-UQ, decyl-ubiquinone; PQ-1, plastoquinone-1; none, no inhibition; nd, not determined).

A. aeolicus P. denitrificans R. capsulatus A. vinosum O. limnetica C. limicola

Km (µM) H2S 11 26 5 nd 8 nd Quinone 5 (d-UQ) 3 (d-UQ) 2 (d-UQ) nd 32 (PQ-1) <20 (PQ-1)

Specific activity[unit (mg protein)-1 ]

3.515 (70°C)

0.3 0.1 0.14 0.02 nd

I50 (µM) Myxothiazole 43 22 43 4 none 6 Stigmatellin 20 20 2.5 nd 6.7 0.005 HQNO 12 nd nd 0.5 nd nd Antimycin A 10 15 50 ~10 none 0.096 NQNO 6 nd 2.4 nd 0.14 0.8 Aurachin C 0.0141 282 0.231 nd 0.0501 0.0121

KCN none none 120 500 12 10

1 2-Methyl-3-(3,7,11-trimethyl-2,6,10-dodekatrienyl)-1-hydroxy-chinolon2 2-Methyl-3-n-dodecyl-1-hydroxy-chinolon


2 M sodium bromide [20,39]. At highconcentrations, sodium bromide acts as achaotropic agent removing proteins frommembranes that are peripherically bound byhydrophic interactions. This raised thequestion, whether the enzyme faces thecytoplasmic or the periplasmic side of themembrane in R. capsulatus. Most of thesulfide-oxidizing bacteria deposit sulfur, theproduct of sulfide oxidation, outside the cells[4] or store the sulfur globules in theperiplasmic space [40]. Since R. capsulatusdeposits sulfur outside the cells as well [41], itseems reasonable that the SQR is attached tothe periplasmic surface of the cytoplasmicmembrane. Recently, a translational fusion ofSQR with the alkaline phosphatase of E. coliwas constructed and expressed in R.capsulatus. It was found that cells expressingthe fusion protein exhibited phosphataseactivity [42]. Alkaline phosphatase is onlyactive, if it is translocated to the periplasm [43].This result clearly indicated periplasmiclocalization of SQR in R. capsulatus andsuggests similar orientation of the sulfideoxidation site in other bacteria.

Generally, export of proteins across thecytoplasmic membrane of bacteria occurs viatwo distinct pathways, the sec-dependent andthe sec-independent pathway [44]. In bothpathways, the proteins that becometranslocated exhibit a characteristic N-terminalsignal peptide for translocation. Surprisingly,no signal peptide was found in the peptidesequence of SQR (see section 4.), although theenzyme becomes translocated. Deletion studieswith the alkaline phosphatase fusion proteinindicated that the C-terminal 38 amino acidresidues are essential for translocation [42]. Sofar, it is unknown, whether the C-terminusfunctions as signal peptide for translocation ordeletion of the C-terminus hinders translocationas a results from incorrect protein folding. Inthis context, it is worth mentioning that the

FAD-binding protein dihydroliponamidedehydrogenase from the cyanobacteriumSynechocystis is also localized at theperiplasmic surface of the membrane withoutexhibiting the characteristic sequences fortranslocation [45]. The fact that no significanthomology exists in the C-terminal regionbetween SQR and dihydroliponamidedehydrogenase argues against a function of theC-terminus as a signal peptide. Possibly, asurface-exposed pattern is necessary fortranslocation of membrane-boundflavoproteins by a so far unknown mechanism.

Cyanobacteria also deposit the sulfur fromoxidation of sulfide outside the cells as well[46,47]. Within cyanobacteria, however, SQR-activity was found in thylakoid membranes[28,48], the intracellular photosyntheticmembranes of cyanobacteria [49]. Arieli et al.[28] suggested that the sulfide oxidation site ofSQR is located at the cytoplasmic surface ofthylakoid membranes in O. limnetica. Thiswould require transport of sulfur through thecytoplasm and across the cytoplasmicmembrane. However, the thylakoid lumencorresponds phylogenetically to theperiplasmic space of bacteria, and, in analogyto the orientation in bacteria, localization of thesulfide oxidation site at the luminal side seemsmore plausible. If there is no continuitybetween the thylakoid lumen and theperiplasmic space - continuities between thecytoplasmic and the thylakoid membranes arestill a matter of debate [49] - , a more complexmachinery for export of sulfur would benecessary. For a variety of cyanobacteria,contact points between the plasma andthylakoid membrane have been noted.Concentration of SQR at these contact pointswould make transport of sulfur through thecytoplasm unnecessary. In order to clear up,whether cyanobacteria form an outgroup withinthe sulfide-oxidizing bacteria with respect tothe topography of SQR in membranes,

localization of the sulfide binding site incyanobacteria has to be investigated.

3.3. ISOLATED SQR ENZYMES

At present, only two SQR enzymes have beenpurified to homogeneity. The first enzymecatalyzing sulfide-dependent reduction ofquinone was isolated from O. limnetica [19]and named sulfide-quinone reductase (SQR;E.C.1.5.5.´.). Three years later SQR from R.capsulatus was isolated [20]. The SQRs wereidentified as single polypeptides with apparentmolecular masses of 55-57 kDa in SDS-PAGE.Fluorescence spectra of the enzymes revealedexcitation and emission maxima typical forflavoproteins (Tab. 2). This fluorescence isquenched by sulfide. Peptide sequencing of theN-termini revealed the characteristics of theFAD-binding βαβ-fold of many flavoproteins(see section 4.4.). The Km values of thepurified enzymes for sulfide and quinone weresimilar to that found for the membrane boundenzymes. The sensitivities of the isolatedenzymes towards quinone-analogues wereslightly increased, possibly due to increasedaccessibility of the quinone-binding site.Another reason might be that with purifiedenzyme samples less inhibitor is withdrawn bymembranes or other quinone-binding enzymespresent in membrane suspensions. As it wasfound for the membrane-bound enzymes, SQRpurified from O. limnetica was insensitivetowards myxothiazole and antimycin A [30], incontrast to the enzyme purified from R.capsulatus [50]. From these results it wassuggested that the quinone-binding site isformed by SQR alone and no additionalprotein is required for activity or for substratebinding.

SQR of R. capsulatus was heterologouslyexpressed in E. coli [20], and it was shown thatthe enzyme catalyzed the electron transfer fromsulfide to oxygen via the Q pool and quinoloxidase [51]. After cell disruption and

differential centrifugation, the main portion ofSQR-activity was found in the membranefraction. SQR isolated from E. coli and SQRisolated from R. capsulatus behaved identicalwith respect to their substrate affinity and theirfluorescence spectra (Tab. 2). Fluorescence ofthe heterologously expressed SQR was alsoquenched by sulfide. The extent of quenchingdepended on the sulfide concentration, andfluorescence was recovered by reoxidationwith quinone.

A substantial portion of SQR expressed in E.coli was found in inclusion bodies.Reconstitution of the denatured protein purifiedfrom inclusion bodies revealed that FAD is thesole prosthetic group of SQR and it is requiredfor enzyme activity (Griesbeck, unpublishedresults). Besides in membranes and in inclusionbodies, a decent part of SQR was also foundin the soluble fraction of E.coli after celldisruption and differential centrifugation [20].Although the portion of SQR in the solublefraction did not exceed 30 - 50 %, there weresignificantly higher amounts than found duringpurification of SQR from R. capsulatus.Similar results were recently reported byShibata et al. [51]. The reason for thisphenomenon is unknown. Possibly differencesin membrane lipid composition between bothspecies influence the interaction of SQR withthe membranes.

In summary, the following results wereestablished by studying the SQR expressionprotein:

− SQR is a single peptide;

− electron transfer from sulfide to membrane-intrinsic quinones and quinone-binding aremediated by SQR alone, and no additionalprotein is necessary;

− attachment of SQR to membranes is notmediated by an additional protein;

− FAD is the sole prosthetic group of SQR.


4. SQR GENES AND PROTEIN SEQUENCES

4.1. CLONING OF THE SQR GENE FROM R.CAPSULATUS

The first sqr gene to be cloned was the genefrom R. capsulatus [20,42]. By use of PCRamplified probes obtained witholigonucleotides deduced from tryptic peptidesof the purified enzyme, a 3.5-kb fragment ofgenomic DNA from R. capsulatus was isolatedand sequenced. This fragment contains the sqrgene, a second complete and one partial openreading frame in same orientation upstream thesqr gene, and the 3´-end of an open readingframe in opposite orientation downstream.

The sqr gene encodes a protein of 427 aminoacid residues with a theoretical molecularweight of 47 kDa and a net charge of +1. Asevident from N-terminal peptide sequencing,the first amino acid residue of the native proteinis alanine (ARc2 in Fig. 4). Five bp upstreamthe codon for MRc1 a nucleotide sequence wasfound that matches well with ribosome-bindingsites of R. capsulatus.

In order to test whether the isolated sqr gene isthe only gene encoding sulfide oxidizingactivity in R. capsulatus, mutants of R.capsulatus were constructed with insertionsand deletions in the sqr gene [42]. Thesestrains were unable to grow photoauto-

Tab. 2: Properties of purified SQR proteins. Comparison of the SQR proteins purified from

Oscillatoria limnetica, Rhodobacter capsulatus and Rhodobacter SQR expressed and purified

from E. coli. Data were taken from [19,20,39,48,50] and Griesbeck (unpublished data). I50-values

give the concentration of inhibitor for 50% inhibition of activity. (unit, µmol quinone reduced min-1;

nd, not determined; PQ-1, plastoquinone-1; d-UQ, decyl-ubiquinon).

O. limnetica R. capsulatus R. capsulatus

(expr. in E. coli)

Apparent size in SDS-PAGE (kDa) 57 55 55

Theoretical molecular mass (kDa) 48 47 47

Florescence maxima (nm) Excitation Emission

280 / 373 / 461527

280 / 375 / 458518

280 / 375 / 475525

Specific activity [unit (mg protein)-1 ] 1.9 3.5 10-15

Km (µM) H2S 8 2 2Quinone 32 (PQ-1) 2 (d-UQ) 2 (d-UQ)

pH optimum nd 6.3 6.5

I50 (µM)Stigmatellin 1.4 1.2Myxothiazole none 13Antimycin A none 32KCN 12 500

______________I____________ :: ::*.* .* *::::. : .R.caps. -------------------------------MAHIVVLGAGLGGAIMAYELREQVRKEDK 29S.putre. -------------------------------MKKIIVIGAGLGGVSVAFELKQRLPSDCE 29T.ferro. -------------------------------MAHVVILGAGTGGMPAAYEMKEALGSGHE 29A.aeol. ------------------------------MAKHVVVIGGGVGGIATAYNLRNLMP-DLK 29O.lim. -------------------------------MAHVAVIGAGLAGLPTAYELRHILPRQHR 29A.7120 -------------------------------MAHIVIVGAGLGGLPTAYELRHILPKQHQ 29A.halo. -------------------------------MAHIVIVGGGFGGLSAAYELKHLLHGKHK 29FCC MTLNRRDFIKTSGAAVAAVGILGFPHLAFGAGRKVVVVGGGTGGATAAKYIKLADP-SIE 59 :: ::*.* .* * :: .

: :*: : * *: * :... . : . : : * : . :.. :R.caps. VTVITKDPMYHFVPSNPWVAVGWRDRKEITVDLAPTMARKNIDFIPVAAKRLHPAENRVE 89S.putre. IGVISEGTDFQFVPSNPWVALGERTREDISLPIGTYLAKRKISYSGEGVTHIDPLNKKLT 89T.ferro. VTLISANDYFQFVPSNPWVGVGWKERDDIAFPIRHYVERKGIHFIAQSAEQIDAEAQNIT 89A.aeol. ITLISDRPYFGFTPAFPHLAMGWRKFEDISVPLAPLLPKFNIEFINEKAESIDPDANTVT 89O.lim. VTLISDKPNFTFTPSLPWVAFDLTPLERVQLDVGKLLKGRNIDWIHGKVNHIDPENKTSV 89A.7120 VTVISETPYFTFIPSLPWVAMGLTSLESIQVSLQQRLKQKGINWILGRVDYLNPQNQKIS 89A.halo. ITLISDETTFTFIPSLPWVAFNLRRLEDVQLPLAPLLARQGINWQHGRVTGLDPNQKRVS 89FCC VTLIEPNTDYYTCYLSNEVIGGDRKLESIKHGYDG-LRAHGIQVVHDSATGIDPDKKLVK 118 : :* : : . . : : * . :.. :

______II______ ------1∇---↓-- - . . ** ::: **..*. * *:* .** : : : *R.caps. LENGQSVSYDQIVIATGPELAFDEIEGFGP---EGHTQSICHIDHAEAAGAAFDRFCENP 146S.putre. LGDGTTNNYDYLVICTGPKLAFENVPGSGP---EGFTQSICTIDHAEKAYECFQQLLAKP 146T.ferro. LADGNTVHYDYLMIATGPKLAFENVPGSDPH--EGPVQSICTVDHAERAFAEYQALLREP 147A.aeol. TQSGKKIEYDYLVIATGPKLVFG-AEGQ-----EENSTSICTAEHALETQKKLQELYANP 143O.lim. AGE-QTLEYDYVVVATGPELATDAIAGLGPE--NGYTQSVCNPHHALMAKEAWQKFLQDP 146A.7120 LGE-QSISYDYLIIATGAELALDAVAGLGPD---GYTQSVCNPHHAIKAFQAWQNFLLAP 145A.halo. VGEDITFDYDYLVITTGASLAYHLMSGLGPE--EGYTQSVCNAHHAEMARDAWDEFLENP 147FCC TAGGAEFGYDRCVVAPGIELIYDKIEGYSEEAAAKLPHAWKAGEQTAILRKQLEDMADGG 178 ** :: .* .* * : .:: : :

---------2--∇------- -----3----↓- **:::** .**:*****:.: *::. * ::. :**::.*** **:*: *:. :R.caps. GPIIIGAAQGASCFGPAYEFTFILDTALRKRKIRDKVP-MTFVTSEPYVGHLGLDGVGDT 205S.putre. GPVVVGALQGASCFGPAYEYVLSLESLLRKHKVRQHIP-ITFVTSEPYVGHMGLGGVGDS 205T.ferro. GPIVIGAMAGASCFGPAYEYAMIVASDLKKRGMRDKIPSFTFITSEPYIGHLGIQGVGDS 207A.aeol. GPVVIGAIPGVSCFGPAYEFALMLHYELKKRGIRYKVP-MTFITSEPYLGHFGVGGIGAS 202O.lim. GPLVVGAVPGASCFGPAYEFALLADYVLRRKGMRDRVP-ITFVTPEPYVGHLGIGGMANS 205A.7120 GPLVVGALPKTSCLGPAYEFTLLADYVLRKQGLREQVS-ITFVTPEPYAGHLGIGGMANS 204A.halo. GPLLVGAVPGASCMGPAYEFALLADYALRQEGKRDQVP-ITFISPEPYLGHLGIGGMANS 206FCC TVVIAPPAAPFRCPPGPYERASQVAYYLKAHKPKSKVIILDSSQTFSKQSQFSKG----W 234 :: . * .** . *: . : :: : . . .::.

:: : :. : . .. . .* *::*.*:R.caps. KGLLEGNLRDKHIKWMTS--TRIKRVEPGKMVVEEVTEDGTVKPEKELPFGYAMMLPAFR 263S.putre. KSLMEHEFRERSIKWICN--AKTTHFESGIAYVDELDRKGEVEFQHTLPFSLAMFLPPFK 263T.ferro. KGILTKGLKEEGIEAYTN--CKVTKVEDNKMYVTQVDEKGETIKEMVLPVKFGMMIPAFK 265A.aeol. KRLVEDLFAERNIDWIAN--VAVKAIEPDKVIYEDLNGN-----THEVPAKFTMFMPSFQ 255O.lim. AELVTDLLENKGIRVLPN--TAVKEIHPEHMDLDSG---------EQLPFKYAMLLPPFR 254A.7120 AELVTKFMAERGVEVIEN--VAVTAIEANQIHLGNG---------RVLPFAYSMLLPPFR 253A.halo. GKLVTELMKQRNIDWVEN--AEIAEIKEDHVKLTDG---------REFPFNYSMFLPPFR 255FCC ERLYGFGTENAMIEWHPGPDSAVVKVDGGEMMVETAFG-------DEFKADVINLIPPQR 287 : : : . .. . ::*. :

____________III_____________ _III_ * : : : :: . : : .::..*: : . . . **R.caps. GIKALMGI-EGLVNPR-GFVIVDQHQQNPTFKNVFAVGVCVAIPPVGPTPVPCGVPKTG- 320S.putre. GSAAIAAV-PELCNEK-GFVLTDTFQRSPRYPEIYAAGVCVAVPPLENPQYRLAYPKQVL 321T.ferro. GVPAVAGV-EGLCNPG-GFVLVDEHQRSKKYANIFAAGIAIAIPPVETTPVPTGAPKTG- 322A.aeol. GPEVVASAGDKVANPANKMVIVNRCFQNPTYKNIFGVGVVTAIPPIEKTPIPTGVPKTG- 314O.lim. GPAFLREA-PELTNPK-GFVPVTNTYQHPKYESVYSAGVIVEINPPEKTPLPVGVPKTG- 311A.7120 GPRFVRQV-PGLSNQD-GFIPVLPTYRHPEYASIYAVGVVVEIKPSEVTPLPLGVPKTG- 310A.halo. GAQFLKEV-PGLTDEK-GFLPVLDTYQHPDYPSIYSAGVITQLAAPEETEVPLGAPKTG- 312FCC AGKIAQIA--GLTNDAGWCPVDIKTFESSIHKGIHVIGDACIANP---------MPKSG- 335 . : : . . :. *↑ . **

------4------ --∇-5----- * * * * ..** : .*:*:**:*: . *.* * *.R.caps. FMIESMVTATAHNIGRIVRG-LEPDEVGSWNAVCLADFGDKGIAFVAQPQIPP-----RN 374S.putre. FMIESMTTAIVDNILANLKG-EPPQTHPTLNAICLADMGDKGAAFVALPQNPP-----RN 375T.ferro. YMIESMVSAAVHNIKADLEGRKGEQTMGTWNAVCFADMGDRGAAFIALPQLKP-----RK 377A.aeol. MMIEQMAMAVAHNIVNDIRN-NPDKYAPRLSAICIADFGEDAGFFFADPVIPP-----RE 368O.lim. QMTEAMGMAAAHNIAIKLGVSKAKPVQPTLEAICIADFGDTGIVFVADPVLPDPKTGTRR 371A.7120 QMTEAMGMAVAHNIAIELGVFSAPPVTPTLDAICFADFGNSGILFLANPVLPDLATGKRR 370A.halo. QMTESMAMAVAHNIARELGEINARPVKPSLEAICMADFGDTGIIFIAAPVVPDPSVGHRR 372FCC YSANSQGKVAAAAVVALLKG--EEPGTPSYLNTCYSILAPAYGISVAAIYRPN----ADG 389 : . . : : * : :. .*

--------6--------- * *. * .** :*: *:: * . *: : :R.caps. VNWSSQGKWVHWAKEGFERYFMHKLRRGTSE----TFYEKAAMKLLGIDKLKAVKKG--- 427S.putre. ANWTSMGKWVHVAKIAFEKYFLYKVRNGTSE----PIYEKYVLGSLGIHRTHGPK----- 426T.ferro. VDVFAYGRWVHLAKVAFEKYFIRKMKMGVSEGCLRAFYEKVLFKMMGITRLK-------- 429A.aeol. RVITKMGKWAHYFKTAFEKYFLWKVRNGNIAP---SFEEKVLEIFLKVHPIELCKDCEGA 425O.lim. RAITKRGKWVSWSKTAFETFFLSKMRFGLAV----PWFERWGLRFMGLSLVEPLDTTRET 427A.7120 RAVALSGAWVTWAKAAFERYFLAKMRFGTAV----PWFEKLALKLLGLSLVAPLAVK-SS 425A.halo. HATALRGLWVNWAKNAFEWYFLAKMRWGTAV----PWFEKLGLYLLRLTLVTPISET-PT 427FCC SAIESVPDSGGVTPVDAPDWVLEREVQYAYS-----WYNNIVHDTFG------------- 431 :.: : :. :

R.caps. ---------- 427S.putre. ---------- 426T.ferro. ---------- 429A.aeol. PGSRC----- 430O.lim. GNQAFASKS- 436A.7120 RNISQENY-- 433A.halo. QQKDLTSIKG 437FCC ----------

Fig. 3: Alignment of SQR Sequences and FCC. SQR sequences are from Rhodobactercapsulatus (R. caps.) [20], Shewanella putrefaciens (S. putre.) [48], Thiobacillusferrooxidans (T. ferro) [34], Aquifex aeolicus (A. aeol.) [35], Oscillatoria limnetica (O.lim.)[48], Anabaena ATCC 7120 (A. 7120) (Kazusa DNA Research Institute,http://www.kazusa.or.jp) and Aphonathece halophytica (A. halo.) [48]. The sequence offlavocytochrome c (FCC) is from Allochromatium vinosum [17]. The FAD-binding domainsare indicated by continuous lines


trophically on sulfide, did not deposit anysulfur outside the cells under heterotrophicconditions and did not consume any sulfide inliquid medium. In addition, no SQR proteinwas detected in western blot analysis withmembranes. This defect was complemented bythe sqr gene on an autonomously replicatingplasmid. Overexpression of SQR in R.capsulatus significantly increased the sulfide-oxidation capability and increased SQR-activity in membranes 6-7fold. These resultsclearly indicate that SQR is the only sulfide-oxidizing enzyme in R. capsulatus and it isabsolutely required for sulfide-dependentgrowth.

The sqr gene of R. capsulatus is not part of anoperon. It is separated by 164 nucleotides fromthe upstream open reading frame, and asequence motif was found within this regionthat matches well with transcription startinitiation sites and regulatory sequences of R.capsulatus. In addition, insertional inactivationof the upstream ORF in R. capsulatus did notalter SQR expression, and SQR was expressedin Rhodobacter from an autonomouslyreplicating plasmid lacking the upstream ORF[42]. Because of significant peptide sequencesimilarity of the deduced amino acid sequencesto subunits of the ribose transporter inArchaeoglobus fulgidus it seems likely that thereading frames upstream sqr are part of anoperon encoding the ribose transporter in R.capsulatus. The peptide sequence deduced

from the 3´-end of the sequence downstreamsqr revealed significant similarity to the C-terminal EAL motif pattern of many regulatoryproteins [42]. Since SQR in R. capsulatus isan inducible protein (see section 6.),participation of this downstream sequence inthe regulation of SQR expression should betested.

4.2. CLONING OF CYANOBACTERIAL SQR GENES

Recently, the sqr genes from the cyanobacteriaO. limnetica and Aphonathece halophyticahave been isolated and sequenced [48]. Thetwo genes encode proteins of 436 and 437amino acid residues in size with a net charge of0 and –14, respectively. Interestingly,downstream both sqr genes, open readingframes encoding small proteins with significantsimilarity to transcription factors of bacteriaand cyanobacteria were found in the samedirection as sqr. In A. halophytica, the sqrgene and the downstream ORF are onlyseparated by 3 nucleotides suggestingcotranscription of both genes. In contrast, thetwo ORFs in O. limnetica are separated byapproximately 180 nucleotides. Since SQR isan inducible protein in both cyanobacterialsystems and regulatory proteins are oftenlocated in the vicinity of their targets ofregulation, these small proteins may play a rolein the regulation of SQR. The two smallcyanobacterial proteins do not share anysignificant sequence similarity with the

and roman numbers, the SQR-fingerprints are indicated by broken lines and arabic numbers. Thethree cysteines that are conserved among SQR-sequences are indicated by open triangles.Histidines that are possibly involved in quinone-binding are signed by arrows above thesequences. The aspartate that is conserved among all members of the glutathion reductase familyof flavoproteins but is missing in SQR is indicated by an arrow below the sequences. Positions ofidentical amino acid residues are marked by *, positions of similar amino acid residues withcolons (high similarity) and with dots (lower similarity). These positions for the SQR sequencesare given above the alignment and for SQRs and FCC below the alignment.

sequence deduced from the 3´-end of the ORFdownstream sqr in R. capsulatus. Therefore, ifall theses proteins are involved in regulation ofSQR expression, the regulatory mechanism incyanobacteria should be different from themechanism in R. capsulatus.

4.3 COMPARISON OF BACTERIAL SQRSEQUENCES

Besides the above mentioned sqr genes, openreading frames encoding proteins with highsimilarity to SQR sequences have beenidentified in sequence databases for A. aeolicus[34,35], Thiobacillus ferroxidans [34,48],Shewanella putrefaciens [48] and AnabaenaATCC7120 (Kazusa DNA Research Institute,http://www.kazusa.or.jp). The overall similarityamong the SQR sequences is in the range from53 to 79 % (identity 37-64 %, see Tab. 3). Thehighest similarity was found between the three

cyanobacterial sequences. For none of theSQR sequences membrane anchoring ormembrane spanning domains were found bysecondary structure prediction analysis. This isless surprising with respect to SQR of R.capsulatus, because this enzyme isperipherically bound to the membrane (seesection 3.2.), but it is remarkable with respectto the cyanobacterial enzymes and the SQR ofA. aeolicus that are all tightly membrane-boundproteins. However, secondary structureprediction analysis methods areapproximations, and the results in some casesdiffer from the real folding pattern of proteins.Therefore, additional structural data arenecessary to elucidate the way SQR interactswith membranes.

Tab. 3: Identity and similarity of SQR and FCC. Percentage values for identity (upper value)and similarity (lower value) are summarized. Values were obtained using BLASTP 2.0.9(http://www.ncbi.nlm.nih.gov/gorf/wblast2.cgi).

R. capsulatus100100

S. putrefaciens4664

100100

T. ferrooxidans4656

4965

100100

A. aeolicus3857

3953

4066

100100

O. limnetica4261

4256

3866

4461

100100

Anabaena 71204261

4157

3958

4057

6479

100100

A. halophytica4059

4259

3755

4056

5874

5975

100100

FCC (A. vinosum)2338

2436

2336

2440

2538

2236

2538

100100


4.4. FAD-BINDING

Although the overall similarity between SQRand FCC is significantly lower than similarityamong the different SQRs (22-25 % and 38-64% identity, respectively, see Tab. 3), all SQRsequences as well as FCC exhibit the threeFAD-binding domains which are characteristicamong proteins that belong to the glutathionreductase family of flavoproteins (indicated bycontinuous lines and roman numbers in Fig.3). These are the N-terminal βαβ-fold [52], asecond motif close to the N-terminus [20] anda motif close to the C-terminus [53].

Within the N-terminal βαβ-fold, all SQRs aswell as FCC contain the expected fingerprintresidues of the ADP-binding site characterizingmany NAD(P)/FAD containing proteins [52].The function of some of the residues in bindingFAD is known from the three-dimensionalstructure of several enzymes that belong to theglutathion reductase family of flavoproteins,e.g. glutathion reductase [54] and FCC [16,17].In the following section we summarizestructural details as evident from sequencecomparisons of SQR with glutathion reductaseand FCC. The addressed amino acid residuesare given with their position in the sequence ofthe SQR from R. capsulatus (Rc) and FCCfrom Allochromatium vinosum (FCC).

The three conserved glycine residues, GRc8,GRc10 and GRc13 in the first sequence motif ofR. capsulatus, are likely to provide space forthe phosphate groups of FAD and formhydrogen bonds to the phosphate O-atomsdirectly or via solvent molecules. Thecarboxyl-group of an acidic residue, DRc36, atthe end of the second β-sheet is in contact withthe hydroxyl-groups of the ribose.

Within the second motif (II), a conservedglycine (GRc106) is possibly located in a turn inthe Cα-chain at the end of a β-sheet. ARc104might be in contact with the phosphate bound

to the ribityl chain of FAD. Although the twoamino acid residues, YRc98 and DRc99, areconserved among most of the pyridinenucleotide oxidoreductases, indications abouttheir function in binding FAD were not found.

Within the third motif (IIIa,b), GRc299 ispossibly conserved because of a turn in the Cα

-chain to accomodate the phosphate groups ofFAD. The carbonyl O-atom of VRc298 couldbe in contact with the phosphate group near theribityl chain via a solvent molecule. Thehydrophobic and aromatic residues in position296 and 298 are possibly conserved to form ahydrophobic environment. Surprisingly, anaspartate that interacts with the ribityl chain ofthe flavin moiety of FAD and that is conservedamong all flavoproteins that belong to theglutathion reductase family (DFCC324 in FCC)is absent in all SQR proteins (position 300 inthe sequence from R. capsulatus, marked by ↑in Fig. 3). The reason for the absence of thisaspartate in SQRs is unknown. Possibly theconformation of FAD in SQR is somewhatdifferent. Near the benzenoid ring of FAD inFCC is a pentapeptide (PKSGYFCC332-336),which is also found in SQR (PKTGFRc317-321). The peptide nitrogen atoms of glycineand the aromatic amino acid residue formhydrogen bonds to N1 and O2 of FAD, andthe side chain of the lysine is close enough toform a hydrogen bond with the peptidyloxygen atom of DFCC324 (VRc300).

A cysteine residue, CFCC72, which is conservedamong FCCs, covalently binds FAD in FCCof A. vinosum [17]. This amino acid residue isabsent in SQR, indicating that FAD is notcovalently bound. This is in agreement with theobservation, that only a part of purified SQRcontains FAD (Griesbeck, unpublishedresults).

4.5. AMINO ACID RESIDUES THAT MIGHT BEINVOLVED IN QUINONE-BINDING

Besides the three FAD binding domains,several stretches were identified that areconserved among the SQR sequences(indicated by broken lines and arabic numbersin Fig. 3). Within two of these stretches thereare histidines (HRc131 and HRc196) that areconserved among all SQRs [48] and might beinvolved in quinone-binding (indicated by ↓ inFig. 3). Based on crystallographic data ofseveral quinone-binding proteins, Rich andFisher proposed a structural element forquinone binding sites [55]. It is composed of ahelical stretch that flanks one side of thequinone headgroup and contains a triad ofclose contact residues. The central residue ofthe triad is a histidine which forms a hydrogen-bond to one carbonyl of the quinone. Thefourth residue upstream of this histidine isaliphatic, usually leucine, and is the closest ofthe triad to the isoprenoid side chain. Anotherclose-contact residue three or four residuesdownstream the histidine makes up the contacttriad and is well conserved within homologoussequences of different Q-binding proteins. InSQR, HRc131 is located in a putative α-helixformed by the conserved stretch [S-I/V-C-(X)3-H-A-(X)2-A]. HRc196 is also located in aconserved stretch [E-P-Y-V/L-G-H-L/F-G-L/I]. Although the two regions are somewhatdifferent from the proposed quinone-bindingmotifs, they are good candidates for thequinone-binding site in SQR, especially thealiphatic and aromatic amino acid residues inclose vicinity to HRc131 might participate informing a hydrophic quinone-binding pocket.These histidines are absent in FCC. Thus bothof them, as well as the presence of valine(VRc300) instead of aspartate in FCC (s. Fig.3) are significant for SQR.

4.6. AMINO ACID RESIDUES THAT MIGHT BEINVOLVED IN SULFIDE OXIDATION

In the three-dimensional structure of FCC, twocysteines (CFCC191 and CFCC367) arepositioned close to the isoalloxazine moiety ofFAD [16]. These two cysteines are linked by adisulfide bridge and are located above thepyrimidine portion of the flavin. They possiblyplay a redox-active role in catalysis of FCC[15]. In the SQRs, three cysteine amino acidresidues (CRc127, CRc159 and CRc353) are wellconserved (indicated by open triangles in Fig.3). Two of them (CRc159 and CRc353) are in aposition identical to the position of the redox-active cysteines in the peptide sequence ofFCC. Preliminary results from replacements ofthese cysteines by site-directed mutagenesis inthe SQR of R. capsulatus indicate that all threecysteines are essential for SQR-activity(Griesbeck, unpublished data). Binding ofFAD by SQR was not impaired. These datasuggest, that three cysteine residues areinvolved in the catalytic reaction of SQR andtwo of them are possibly positioned close tothe pyrimidine portion of FAD.

In the sequence of SQR, glutamate ERc165 is ina position identical to the position of aglutamate in FCC. EFCC197 is near the N5position of FAD and was found to beresponsible for a pK-shift of sulfite binding toFAD [16]. In analogy, a similar function ofERc165 in SQR seems plausible.

Besides the above mentioned residues, severalstretches of conserved amino acid residues inSQR were identified (Fig. 3, broken lines).Although these stretches can serve as fingerprints for the identification of SQR sequences,their function needs to be investigated. Withinthe C-terminus, there is remarkable high degreeof similarity among SQR sequences. Fouraromatic amino acid residues, four positivelycharged and two negatively charged amino acidresidues are present in all sequences. This is ofspecial interest with respect to the function of


the C-terminus in glutathion reductase andlipoamide dehydrogenase. In thesehomodimeric enzymes, the C-terminal region ofone subunit is involved in the formation of thecatalytic site of the other subunit, and some ofthe C-terminal amino acid residues participatein stabilisation of the dimer, e.g. by formationof a salt bridge [54]. Size exclusionexperiments [20] and native gel electrophoresis([50], Schödl and Griesbeck, unpublishedresults) suggest that SQR of R. capsulatusmight be a homodimer as well. Possibly someof the C-terminal amino acid residues are alsoinvolved in dimer stabilisation and theformation of the catalytic site in SQR.

5. MECHANISM OF SULFIDE OXIDATION

At present, data about the mechanism ofsulfide oxidation and quinone reduction inSQR are scarce. However, in this section wetry to formulate a rough model for themechanism of sulfide-dependent reduction ofquinones by SQR on basis of available dataand in analogy to enzymatic systems that havebeen well investigated.

Most of the proteins that belong to theglutathion reductase family of flavoproteins, e.g. glutathion reductase, lipoamidedehydrogenase or FCC, catalyze reactions bywhich thiol groups or sulfide become eitherreduced or oxidized. In all these proteins twocysteines are involved in catalysis. Thereby,they change between the reduced thiol- and theoxidized disulfide-state. [15,54,56]. Except forFCC, in these enzymes electrons aretransferred from the two electron carrierNAD(P) via the two electron carrier FAD andvia two cysteine residues of the enzyme todisulfide, e.g. glutathion, that becomesreduced. Within SQR, the situation is similar,although the reaction occurs in oppositedirection. Electrons from sulfide reduce the

two electron carrier FAD, possibly via twocysteine residues, and the two electron carrierquinone becomes reduced.

There are some evidences in favour of thetransfer of two electrons from sulfide to thetwo electron carrier quinone by SQR. First, thesulfur-product from oxidation of sulfide (S2-)by R. capsulatus is zero-valent sulfur (S0) [41],(Griesbeck, Rethmeier and Fischer,unpublished). Neither sulfide-oxidation norformation of sulfur occured in a mutant strainof R. capsulatus, in which the sqr gene hadbeen deleted [42]. Second, membranes of anE. coli strain that expressed SQR from R.capsulatus catalyzed the electron transfer fromsulfide to oxygen (for reduction of 1 Mol O2 4Mol electrons are necessary). Thesulfide:oxygen ratio was 2:1, and the sulfurproduct was zero-valent sulfur [51]. Third, thesulfide:quinone ratio of the purified SQR wasfound to be 1:1 (Griesbeck, unpublished).Although a clear determination of the sulfurproduct of this reaction failed so far, transientformation of polysulfide was shown(Griesbeck, Rethmeier and Fischer,unpublished).

What is the sulfur-product of the reaction?Atomic sulfur is an unstable, highly reactivespecies, and it is very unlikely that it could bethe initial product. In nature, elemental sulfurmainly occurs in polymers, rings of mostly 8atoms (S8) [57]. Elemental sulfur is nearlyinsoluble in aqueous solutions. Therefore,organisms that do not store sulfur globules inthe periplasm would have the problem totransport it out of the cell prior to precipitation.This problem would not occur, if polysulfides(H-Sn-H) would be the product. Polysulfidesare soluble in aqueous solutions. They areunstable because of disproportionation to H2S+ Sn-1 [57]. This could explain the occurrenceof sulfur precipitates outside the cells, even ifpolysulfide is the immediate sulfur-product.Indeed, polysulfides deriving from biological

oxidation of sulfide have been detected inpurple and green sulfur bacteria [58].Formation of longer polysulfide chains as initialstates for formation of S8-sulfur would requirepolymerization of 8 sulfide molecules bysubsequent oxidation reactions with thegrowing polysulfide chain bound to theenzyme. Interestingly, Klimmek et al. [59] haverecently shown that the Sud protein ofWollinella succinogenes covalently bindspolysulfide. Sud catalyzes the formation ofthiocyanate from cyanide and polysulfide.Thereby, polysulfide with up to ten sulfuratoms is covalently bound to a cysteine residueof Sud.

Reaction model: From the above data wesuggest the following model for oxidation ofsulfide by SQR. In SQR there are twocysteines that are linked by a disulfide bridge inanalogy to FCC. This disulfide bridge isopened by addition of sulfide resulting in areduced cysteine residue and a thiolgroupbound to the second cysteine.

(1) E-S-S-E + H2S → E-S-SH + E-SH

In a second step, FAD becomes reduced and atrisulfide bridge is formed.

(2) E-S-SH + E-SH +FAD → E-S-S-S-E +FADH2

In a third step, the electrons from the flavin aretransferred to quinone and a second sulfidemolecule becomes bound to the enzyme.

(3) E-S-S-S-E + FADH2 + Q + H2S → E-S-S2H + E-SH +FAD + QH2

After several cycles through reaction (2) and(3), the polysulfide chain is released from theenzyme by formation of the initial disulfidebridge.

(4) E-S-S2H + E-SH → E-S-S-E + H-Sn-H

In view of the results from the site-directedmutagenesis the mechanism may actuallyinvolve three cysteines.

6. REGULATION OF EXPRESSION OF SQR

6.1. EXPRESSION OF SQR IN R. CAPSULATUS

The "non-sulfur" purple bacterium R.capsulatus holds a very versatile metabolism.It lives in pools and brooks that are oftenpolluted by waste water. There it has to adaptits metabolism to changing nutritionalconditions, e. g. hydrogen sulfide. Adaptationrequires de novo synthesis of the enzymaticsystems that are required for metabolizingnewly appearing substrates.

First indications that the sulfide oxidizingcapability in R. capsulatus is an inducibleprocess were given in the work of Wijbengaand van Gemerden [60]. They observed thatafter the first addition of sulfide to acetate-limited continuous cultures sulfide oxidationoccurred only after an initial lag phase. Afterapproximately 1 hour, a constant oxidation ratewas reached. No lag phase was observed aftera second addition of sulfide. In addition, lowergrowing cultures oxidized sulfide more rapidlythan fast growing cells and the initial lag phasewas elongated in fast growing culturessuggesting that the metabolic status of the cellsinfluences the sulfide oxidation capability.

In our laboratory, we observed that an increasein the sulfide oxidation capability of R.capsulatus cultures correlated with an increaseof membrane-bound SQR-activity (Meiler,Schaible, Schütz and Hauska, unpublished),and this increase was enhanced, if oxygen orCO2 were added as electron sinks to thecultures [42,61].

In order to simplify expression studies, weused reporter gene techniques. First, the genesencoding luciferase of Vibrio fischeri wereinserted into the genome of R. capsulatusunder control of the sqr-promotor [42,50].Addition of sulfide to heterotrophicallygrowing cultures caused a significant increaseof luciferase activity, and interestingly the levelof expression of luciferase was enhanced, if


Fig. 4: Model of the Regulation of SQR-expression in Rhodobacter capsulatus.Regulatory effects are shown by brokenarrows, electron transport steps by solid lines.(TCA cycle, tricarboxylic acid cycle; cyt,cytochrome).

oxygen was present. Since luciferase activitydepends on oxygen, this result was ambiguous,however. Therefore, we changed to the oxygeninsensitive phosphatase system.

An expression construct of sqr fused with thegene of the alkaline phosphatase of E. coli wasset under control of the sqr-promotor on an

autonomously replicating plasmid and wastransformed into R. capsulatus [42,61].Induction of heterotrophically growing cells bysulfide significantly increased phosphatase-activities of the cells, and the activities reacheda constant rate approximately 5 to 7 hours aftersulfide addition. After induction under oxicconditions, activities were several-fold higherthan under unoxic conditions. In order to testenhancement of expression by oxygen, parallelcultures were cultivated anoxically for 25hours. Then sterile air was added to half of thecultures. Addition of air increasedphosphatase-activities 2- to 3-fold within 7

hours. If the transcription inhibitor rifampicinwas added together with air, no increase ofphosphatase-activity was observed.

Based on the data from our studies and on theexperiments from Wijbenga and van Gemerden[60] two mechanisms are conceivable for themodulation of SQR expression:

− either there is a distinct regulator of SQRexpression for each stimulating or repressingagent

− or the activity and/or concentration of oneregulator protein is modulated by a signalthat reflects the metabolic status of the cell.

We favor the second model of regulation of theSQR expression in R. capsulatus as depictedin Fig. 4 for the following reason. Thequinone-pool plays a central role in the ETchain in R. capsulatus (Fig. 1). It is theconnection between phototrophic andchemotrophic metabolism [62,63]. Cyclic ETis only ensured in case of an appropriateQH2/Q ratio. An oversupply of reductants, e.g. acetate and sulfide, would result in a highQH2/Q ratio in absence of appropriateoxidants, e. g. oxygen. As a result, thegeneration of a proton gradient by the cyclicphotosynthetic ET in the light as well as by theET through the cytochrome bc complex in thedark would be inhibited. Consequently, ATPsynthesis and reduction of NAD+ would beattenuated. The negative control of theexpression of proteins that feed electrons intothe quinone-pool would prevent over-reductionof the pool. However, this model needs to betested by further experimentation.

6.2. EXPRESSION OF SQR IN CYANOBACTERIA

Cyanobacteria are exceptional in the world ofphototrophic prokaryotes. They are the onlybacteria that perform plant-typephotosynthesis, using two photosystems in

sqr

H S2

+

+

-O2

NAD+

CO2

bc complex

cyt oxidasec

UQH oxidase2

NADH:UQ oxidoreductase

Calvincycle

acetate

TCAcycle

UQHUQ

2

series and water as the electron donor.However, some species can facultatively shiftto bacteria-type, anoxygenic photosynthesis inwhich only photosystem I is involved (seereviews [46,47]). Efficient anoxygenicphotosynthesis depends on sulfide-induced denovo protein synthesis. After incubation in thepresence of sulfide and light, the cyanobacteriaO. limnetica and A. halophytica shift toanoxygenic photosynthesis within 2-3 hours[64]. The shift was found to be accompaniedby synthesis of SQR in membranes [19,48].Besides SQR, induction of several smallperiplasmic and membrane-bound proteins wasobserved in O. limnetica [65]. These proteinswere not found to be induced by dithionitesuggesting that sulfide is the sole stimulus. Upto now, neither additional information aboutmodulation of expression of proteins by sulfidenor information about the function of the smallsulfide-induced proteins is available.

7. OCCURENCE OF SQR

7.1. SQR IN PROKARYOTES

The occurence of SQR in either photo- orchemolithotrophic Eubacteria known so far isshown in Fig. 5. This occurence has beendocumented by SQR activity and/oridentification of genes encoding proteins withhigh similarity to known SQR sequences (seesection 4.). Both SQR-activity and theresponsible gene sequence are known for thecyanobacteria O. limnetica and A. halophytica[48], the purple non-sulfur bacterium R.capsulatus [20] (α-subgroup within theproteobacteria) and for A. aeolicus [34]. ThisSQR sequence exhibits high identity to theproteobacterial and cyanobacterial SQRs.Three additional gene sequences have beenfound in representatives of cyanobacteria(Anabaena ATCC7120) and the γ-subgroup ofproteobacteria (S. putrefaciens and T.ferrooxidans). For members of the latter group

SQR activity has been demonstrated withoutknowing the corresponding gene, i.e. for A.vinosum [18] and Thiobacillus sp. W5 [33], aswell as for one additional representative of theα-proteobacteria, Paracoccus denitrificans[32]. Beyond that, SQR activity has beenfound in green sulfur bacteria (Chlorobiumlimicola f. thiosulfatophilum [30] and C.tepidum [31]) and Chloroflexaceae (C.aurantiacus) [21]. This far distribution amongEubacteria and the occurence of SQR in A.aeolicus, a member of the Aquificaceae, themost deeply branching family within thebacterial domain [66], leads to the conclusionthat SQR could be an evolutionary old enzymethat was invented during the early periods oflife in a sulfidic atmosphere [67]. Therefore,occurrence of SQR also in Archaea [7] ispredictable, but investigations have not beencarried out so far.

Even in mitochondria of eukaryotes goodevidences for SQR activity have been found[21,68,69]. This is not surprising because ofthe prokaryotic origin of mitochondria [67].

7.2. THE HMT2 GENE FROMSCHIZOSACCHAROMYCES POMBE

Recently, sulfide-quinone oxidoreductaseactivity was reported for the mitochondrialenzyme HMT2 from fission yeast(Schizosaccharomyces pombe) [70]. Acadmium-hypersensitive mutant of S. pombewas found to accumulate abnormally highlevels of sulfide. The gene required for normalregulation of sulfide levels, hmt2+, was clonedby complementation of the cadmium-hypersensitive phenotype of the mutant. Cellfractionation and immunocytochemistryindicated that HMT2 protein is localized tomitochondria. HMT2 protein, expressed in andpurified from E. coli, was soluble, bound FADand catalyzed the reduction of quinone.However, the Km values of 2 mM for thebinding of sulfide and quinone by HMT2 are


almost a thousand times higher than the Km

values of all known SQR proteins (see Tab.1). A Km for sulfide of 2 mM seems rather highbecause of the enormous toxicity of sulfide.Inhibition by quinone-analogue inhibitors,which would be a good indication for theexistence of a specific quinone-binding site,

was not tested. In addition, sequencecomparison revealed a rather low similarity tothe known SQR sequences. Whereas the FADbinding domains and the two cyteines that forma disulfide bridge in FCC of A. vinosum arepresent (see Fig. 4 and section 4.), theaspartate (DAv324 in FCC) which is involved inFAD binding in all flavoproteins that belong tothe glutathion reductase familiy, and which isnot present in any of the known SQR

sequences, was found in HMT2. The domainsthat were found as fingerprint residues of SQR,were not conserved in HMT2 [34]. Overallidentity to SQR was 17-22 %, which is as lowas the identity between SQR and FCC. Incontrast the microbial SQR sequences exhibitan identity between 38 and 64 % (see Tab. 3).Therefore, the physiological role of HMT2 asSQR is questionable.

8. APPLICATIONS OF SULFIDE OXIDIZINGMICROORGANISMS

As a component of anthropogenic waste water,hydrogen sulfide often causes problemsbecause of its high toxicity even at lowconcentrations and its unpleasant odour. Inbiological sewage plants sulfide concentrations

Fig. 5: Occurence of SQR within the domain Bacteria. Phylogenetic tree of Bacteria based on16S rRNA (modified from [4]) and the occurence of SQR (SQR-activity and/or gene sequence) andFCC. SQR activity in Chloroflexaceae is not firmly established.

Thermotoga

deinococci& relativesPlanctomyces & relatives

chlamydiae

spirochetesbacteroides-flavobacteriaChlorobiaceae

Gram - positives(incl. Heliobacteriaeceae)

Cyanobacteria

Chloroflexaceae

Aquifex

δ groupγ group

β group

α group

Proteobacteria

SQR

SQR

SQR

FCC /

SQRFCC /

SQR

FCC /SQR

FCC /

FCC

(?)

can be found which are toxic not only to man,but also to the microorgranisms clarifiying thewater [71]. Therefore, operation of sewageplants can be inhibited by sulfide. In addition,sulfide can be oxidized to sulfuric acid bymicroorganisms which inhabit biofilms insidewaste water tubes. The resulting sulfuric acidleads to corrosion of metallic or concretematerial.

In order to remove sulfide from liquid andgaseous effluents, biotechnologicalapplications have been developed using photo-or chemotrophic bacteria. It can also beremoved by chemical methods, such aschemical oxidation or precipitation, but thesemethods are expensive and may cause newproblems because of the chemicals used.Bioreactors with anoxygenic phototrophicbacteria have been successfully tested forsulfide removal (for a review see [72]). Theadvantages of such systems are the almostcomplete removal of sulfide from the effluentsand no need for addition of chemical oxidants.The phototrophic bacteria acting in thesesystems are species of the genera Chlorobiumand Rhodobacter. The above mentionedspecies convert sulfide mainly to sulfate,therefore the sulfur is not removed completelyfrom the system and can be reduced to sulfideby sulfate-oxidizing bacteria again. In order tocircumvent this problem and to obtainelemental sulfur as the main product, it wouldbe useful to replace these organisms by otherphototrophic ones which only can oxidizesulfide to elemental sulfur, e.g. R. capsulatus.

From a sulfur-producing bioreactor thedominant sulfide-oxidizing bacterium has beenisolated and characterized. The isolate has beenshown to be a new Thiobacillus species andhas been given the working name Thiobacillussp. W5 [73]. From Thiobacillus sp. W5 anovel membrane-bound flavocytochrome csulfide dehydrogenase has been isolated, andSQR activity has been shown to be present inmembranes [33]. Studies of such bacteria may

help to explain under which conditions sulfideis oxidized only to sulfur instead of completelybeing oxidized to sulfate. The results can thenbe used for process control and optimizationof sulfide removal and sulfur production [73].

Elimination of unpleasant odor from animalfeeding has been tried by growing of sulfide-oxidizing phototrophic bacteria on swine feces.The level of substances responsible for theoffensive odor of swine sewage could bereduced by cultivation of R. capsulatus onwaste from animal feeding plants [74]. Anotherattempt is the genetic manipulation of intestinalbacteria. Hereby, expression of SQR from R.capsulatus in E. coli was studied, as well asexpression of genes for quinone biosynthesistogether with the sqr gene [51]. In both cases,sulfide oxidation to elemental sulfur by therecombinant bacteria as well as by isolatedmembranes was observed in laboratory scale,but only under aerobic conditions. Forapplication of this systems inside the strictlyanaerobic intestines, another terminal electronacceptor has to be found.

As a byproduct from sulfide-removing systemsusing phototrophic bacteria growing on waste,it is possible to produce foodstuff, medicalproducts or biomass that can be used asfertilizers or livestock feed (reviewed in [75]).

9. OPEN QUESTIONS AND PROSPECTS

Over the last ten years, investigations aboutSQR have led to new insight about sulfideoxidation in microorganisms. However, manyquestions remain open. SQR is regarded tohave evolved very early in bacterial evolutionbecause of its occurence in phylogeneticallyvery distant bacteria. Thus, occurence of SQRamongst sulfide-oxidizing Archaea would notbe surprising, but remains to be examined.Occurence of SQR within mitochondria is stilla matter of investigation. Although SQRactivity has been detected in mitochondria ofthe lugworm Arenicola marina, no SQR


protein has been purified up to now fromEukaryotes. HMT2 was suggested to catalyzesulfide-dependent reduction of quinones inmitochondria of S. pombe. However, aspointed out in section 4, it is not clear yet,whether this is the physiological function ofHMT2 in yeast and in other sulfide-oxidizingEukaryotes.

Whereas for SQR in R. capsulatus only anenergy gaining function has been documented[42], in A. halophytica the role of SQR seemsto be detoxification of sulfide [48]. A. vinosumexpresses two sulfide-oxidizing enzymes, SQRand FCC [18], and no effect on the sulfide-oxidizing capability of the cells was found afterinactivation of FCC. These results raise thequestions, whether the major role of SQR isenergy conversion or detoxification and whysome bacteria express both SQR and FCC.

Although some progress has been made,experimental data about the catalyticmechanism and about the structure of SQR arescarce. What is the immediate sulfur productof the SQR-reaction, what is the exact functionof the three cysteines in catalysis, which aminoacid residues are involved in binding ofquinone, and what is the function of the highlyconserved stretches? The SQR of R.capsulatus functionally expressed in E. colinow allows the isolation of larger quantities ofthe native and mutated enzyme for structuraland functional studies.

In the cyanobacteria O. limnetica and A.halophytica expression of SQR depends onsulfide, and in R. capsulatus the sulfide-dependent expression is modulated by varioussubstrates. However, up to now neitherinformation about the nature of the regulatoryproteins and the sulfide-sensing mechanism isavailable nor has the stimulus for themodulation of SQR-expression in R.capsulatus been identified. With thetranslational fusion of SQR with alkaline

phosphatase of E. coli in R. capsulatus, wenow have the tool in our hands to investigatethe modulation of expression in more detail andto test whether the downstream ORF isinvolved in regulation of SQR.

ACKNOWLEDGEMENT

This research is supported by the DeutscheForschungsgemeinschaft (DFG). We wouldlike to thank Yosepha Shahak, Etana Padanand Michal Bronstein for continuouscollaboration and inspiring discussion. UlrichFischer and Jörg Rethmeier (UniversitätBremen) carried out the sulfur analysis.Database searches for amino acid sequenceswere performed by BLASTP [76] with theamino acid sequence database at the NationalCenter for Biotechnology Information,Washington D.C. Preliminary sequence datawere obtained from The Institute for GenomicResearch website at http://www.tigr.org. andthe Kazusa DNA Research Institute athttp://www.kazusa.or.jp.

REFERENCES

1. Jannasch, H.W. 1989, AutotrophicBacteria, Schlegel, H.G. and Bowien, B.(Eds.) Springer-Verlag, Berlin Heidelberg,147-167.

2. Trudinger, P.A. 1992, Early OrganicEvolution, Schidlowski, M. e.a. (Eds.)Springer-Verlag, Berlin Heidelberg, 367-377.

3. National Research Council 1979, HydrogenSulfide, University Park Press, Baltimore

4. Brune, D.C. 1995, Anoxygenic photo-synthetic bacteria, Blankenship, R.E.,Madigan, M.T., and Bauer, C.E. (Eds.)Kluver, Dordrecht, Netherlands, 847-870.

5. Kelly, D.P., Shergill, J.K., Lu, W.P., andWood, A.P. 1997, Antonie VanLeeuwenhoek, 71, 95-107

6. Friedrich, C.G. 1998, Advances inmicrobial physiology, Poole, R.K. (Eds.)Academic Press, London, 236-289.

7. Stetter, K.O. 1996, FEMS Microbiol Rev,18, 149-158

8. Grieshaber, M.K. and Völkel, S. 1998,Annu Rev Physiol, 60, 33-53

9. Brune, D.C. 1989, Biochim Biophys Acta,975, 189-221

10. Blankenship, R.E. 1992, Photosyn Res, 33,91-111

11. Meyer, T.E. and Bartsch, R.G. 1976,Flavins and Flavoproteins, Singer, T.P.(Eds.) Elsevier, Amsterdam, 312-317.

12. Tollin, G., Meyer, T.E., and Cusanovich,M.A. 1982, Biochemistry, 21, 3849-3856

13. Cusanovich, M.A., Meyer, T.E., andTollin, G. 1985, Biochemistry, 24, 1281-1287

14. Guo, L.H., Hill, H.-A.O., Hopper, D.J.,Lawrence, G.A., and Sanghera, G.S. 1990,J Biol Chem, 265, 1958-1963

15. Meyer, T.E., Bartsch, R.G., andCusanovich, M.A. 1991, Biochemistry, 30,8840-8845

16. Chen, Z.W., Koh, M., Van Driessche, G.,Van Beeumen, J.J., Bartsch, R.G., Meyer,T.E., Cusanovich, M.A., and Mathews,F.S. 1994, Science, 266, 430-432

17. Van Driessche, G., Koh, M., Chen, Z.W.,Mathews, F.S., Meyer, T.E., Bartsch,R.G., Cusanovich, M.A., and VanBeeumen, J.J. 1996, Protein Science, 5,1753-1764

18. Reinartz, M., Tschaepe, J., Brueser, T.,Trueper, H.G., and Dahl, C. 1998, ArchMicrobiol, 170, 59-68

19. Arieli, B., Shahak, Y., Taglicht, D.,Hauska, G., and Padan, E. 1994, J BiolChem, 268, 5705-5711

20. Schütz, M., Shahak, Y., Padan, E., andHauska, G. 1997, J Biol Chem, 272, 9890-9894

21. Shahak, Y., Schütz, M., Bronstein, M.,Griesbeck, C., Hauska, G., and Padan, E.1999, Proceedings of the 9th InternationalSymposium Phototrophic Prokaryotes(Vienna 1997), Peschek, G.A., Löffelhardt,W., and Schmetterer, G. (Eds.) PlenumPress, New York, 217-228.

22. Knaff, D.B. and Buchanan, B.B. 1975,Biochim Biophys Acta, 376, 549-560

23. Von Jagow, G. and Link, T.A. 1986, MethEnzymol, 126, 253-271

24. Hauska, G., Schütz, M., and Büttner, M.1996, Advances in Photosynthesis, Vol. 4:Oxygenic Photosynthesis: The LightReactions, Ort, D.R. and Yokum, C.F.(Eds.) Kluwer Academic Publishers,Dordrecht, The Netherlands, 377-398.

25. Brune, D.C. and Trüper, H.G. 1986, ArchMicrobiol, 145, 295-301

26. Cohen, Y., Jørgensen, B.B., Padan, E.,and Shilo, M. 1975, Nature, 257, 489-492

27. Shahak, Y., Arieli, B., Binder, B., andPadan, E. 1987, Arch Biochem Biophys,259, 605-615

28. Arieli, B., Padan, E., and Shahak, Y. 1991,J Biol Chem, 266, 104-111

29. Shahak, Y., Klughammer, C., Padan, E.,Herrmann, I., and Hauska, G. 1994,Photosyn Res, 39, 175-181

30. Shahak, Y., Arieli, B., Padan, E., andHauska, G. 1992, FEBS Lett, 299, 127-130

31. Klughammer, C., Hager, C., Padan, E.,Schütz, M., Schreiber, U., Shahak, Y., andHauska, G. 1995, Photosyn Res, 43, 27-34

32. Schütz, M., Klughammer, C., Griesbeck,C., Quentmeier, A., Friedrich, C.G., andHauska, G. 1998, Arch Microbiol, 170,353-360

33. Visser, J.M., De Jong, G.A.H., Robertson,L.A., and Kuenen, J.G. 1997, ArchMicrobiol, 167, 295-301


34. Nübel, T., Klughammer, C., Huber, R.,Hauska, G., and Schütz, M. 2000, ArchMicrobiol, in press

35. Deckert, G., Warren, P.V., Gaasterland,T., Young, W.G., Lenox, A.L., Graham,D.E., Overbeek, R., Snead, M.A., Keller,M., Aujay, M., Huber, R., Feldman, R.A.,Short, J.M., Olsen, G.J., and Swanson,R.V. 1998, Nature, 392, 353-358

36. Schütz, M., Zirngibl, S., le Coutre, J.,Büttner, M., Xie, D.-L., Nelson, N.,Deutzmann, R., and Hauska, G. 1994,Photosyn Res, 39, 163-174

37. Massey, V. and Ghisla, S. 1983, BiologicalOxidations, Sund, H. and Ulrich, V. (Eds.)Springer, 114-139.

38. Yamanaka, T. and Kusai, K. 1976, Flavinsand Flavoproteins, Singer, T.P. (Eds.)Elsevier, Amsterdam, 292-301.

39. Shahak, Y., Hauska, G., Herrmann, I.,Arieli, B., Taglicht, D., and Padan, E.1992, Research in Photosynthesis, Murata,N. (Eds.) Kluwer Academic Publishers,Netherlands, 483-486.

40. Pattaragulwanit, K., Brune, D.C., Trüper,H.G., and Dahl, C. 1998, Arch Microbiol,169, 434-444

41. Hansen, T.A. and Van Gemerden, H. 1972,Arch Microbiol, 86, 49-56

42. Schütz, M., Maldener, I., Griesbeck, C.,and Hauska, G. 1999, J Bacteriol, 181,6516-6523

43. Manoil, C., Mekanalos, J.J., and Beckwith,J. 1990, J Bacteriol, 172, 515-518

44. Berks, B.C. 1996, Mol Microbiol, 22, 393-404

45. Engels, A., Kahmann, U., Ruppel, H.G.,and Pistorius, E.K. 1997, Biochim BiophysActa, 1340, 33-44

46. Cohen, Y., Jørgensen, B.B., Revsbech,N.P., and Poplawski, R. 1986, ApplEnviron Microbiol, 51, 398-407

47. Stal, L.J. 1995, New Phytol, 131, 1-32

48. Bronstein, M., Schütz, M., Hauska, G.,Padan, E., and Shahak, Y. 2000, JBacteriol, in press

49. Gantt, E. 1994, The Molecular Biology ofCyanobacteria, Bryant, D.A. (Eds.) KluwerAcademic Publishers, Netherlands, 119-138.

50. Schütz, M. (1997) Doctoral Thesis.Universität Regensburg, Germany

51. Shibata, H., Takahashi, M., Yamaguchi, I.,and Kobayashi, S. 1999, J Biosci Bioeng,88, 244-249

52. Wierenga, R.K., Terpstra, P., and Hol,W.G.I. 1986, J Mol Biol, 187, 101-107

53. Eggink, G., Engel, H., Vriend, G.,Terpstra, P., and Witholt, B. 1990, J MolBiol, 212, 135-142

54. Karplus, P.A. and Schulz, G.E. 1987, JMol Biol, 195, 701-729

55. Rich, P. and Fisher, N. 1999, BiochemicalSociety Transactions, 27, 561-565

56. Schierbeek, A.J., Swarte, M.B.A., Dijkstra,B.W., Vriend, G., Read, R.J., Hol, W.G.I.,Drenth, J., and Betzel, C. 1989, J Mol Biol,206, 365-379

57. Steudel, R. 1996, Ind Eng Chem Res, 35,1417-1423

58. Bloethe, M. and Fischer, U. (2000) Newinsights in sulfur metabolism of purple andgreen phototrophic sulfur bacteria and theirspheroblasts. Microbiology 2000, 12. -16.3. 2000, München

59. Klimmek, O., Kreis, V., Klein, C., Simon,J., Wittershagen, A., and Kröger, A. 1998,Eur J Biochem, 253, 263-269

60. Wijbenga, D.J. and Van Gemerden, H.1981, Arch Microbiol, 129, 115-118

61. Schütz, M., Griesbeck, C., Bronstein, M.,Maldener, I., Shahak, Y., Padan, E., andHauska, G. 1998, International Congresson Photosynthesis (Budapest, 1998),Garab, G. (Eds.)

62. Ferguson, S.J., Jackson, J.B., andMcEwan, A.G. 1987, FEMS MicrobiolRev, 46, 117-143

63. McEwan, A.G. 1994, Antonie VanLeeuwenhoek, 66, 151-164

64. Oren, A. and Padan, E. 1978, J Bacteriol,133, 558-563

65. Arieli, B., Binder, B., Shahak, Y., andPadan, E. 1989, J Bacteriol, 171, 699-702

66. Burggraf, S., Olsen, G.J., Stetter, K.O.,and Woese, C.R. 1992, System ApplMicrobiol, 15, 352-356

67. Brock, T.D., Madigan, M.T., Martinko,J.M., and Parker, J. (1997) Biology ofMicroorganisms. 8.,

68. Völkel, S. and Grieshaber, M.K. 1997, JExp Biol, 200, 83-92

69. Parrino, V., Kraus, D.W., and Doeller, J.E.2000, J Exp Biol, in press

70. Vande Weghe, J.G. and Ow, D.W. 1999, JBiol Chem, 274, 13250-13257

71. Zietz, U. 1995, Korrespondenz Abwasser,42, 257-262

72. Fischer, U. 1988, Special MicrobialProcesses (Volume 6b), Rehm, H.J. (Eds.)VCH Verlagsgesellschaft mbH, Weinheim,463-496.

73. Visser, J.M., Stefess, G.C., Robertson,L.A., and Kuenen, J.G. 1997, Antonie VanLeeuwenhoek, 72, 127-134

74. Lee, M.G. and Kobayashi, M. 1992, SoilSci Plant Nutr, 38, 767-770

75. Kobayashi, M. and Kobayashi, M. 1995,Anoxygenic Photosynthetic Bacteria,Blankenship, R.E., Madigan, M.T., andBauer, C.E. (Eds.) Kluver AcademicPublishers, The Netherlands, 1269-1282.

76. Altschul, S.F., Gish, W., Miller, W.,Myers, E.W., and Lipman, D.J. 1990, JMol Biol, 215, 403-410

Documents

Biological Sulfide Oxidation: Sulfide-Quinone Reductase ... · Biological Sulfide Oxidation: Sulfide-Quinone Reductase (SQR), the Primary Reaction Christoph Griesbeck, Günter Hauska