deletion of the 6-kda subunit affects the activity and yield of the bc1 complex from rhodovulum...

Eur. J. Biochem. 267, 3753±3761 (2000) q FEBS 2000

Deletion of the 6-kDa subunit affects the activity and yield of thebc1 complex from Rhodovulum sulfidophilum

Simon Rodgers1, Claudio Moser1,*, Marta Martinez-Julvez1 and Irmgard Sinning1,2

1European Molecular Biology Laboratory, Structural Biology Programme, Heidelberg, Germany; 2Biochemistry Centre Heidleberg (BZH),

Heidelberg, Germany

The cytochrome bc1 complex from Rhodovulum sulfidophilum purifies as a four-subunit complex: the

cytochrome b, cytochrome c1 and Rieske iron-sulphur proteins, which are encoded together in the fbc operon, as

well as a 6-kDa protein. The gene encoding the 6-kDa protein, named fbcS, has been identified. It is located

within the sox operon, which encodes the subunits of sarcosine oxidase. The encoded 6-kDa protein is very

hydrophobic and is predicted to form a single transmembrane helix. It shows no sequence homology to any

known protein. The gene has been knocked-out of the genome and a three-subunit complex can be purified.

This deletion leads to a large reduction in the yield of the isolated complex and in its activity compared to

wild-type. The high quinone content found in the wild-type complex is, however, maintained after removal of

the 6-kDa protein. Surprisingly, a fourth subunit of approximately 6 kDa is again found to copurify with the

Rhv. sulfidophilum bc1 complex when only the fbc operon is expressed heterologously in a near-relative,

Rhodobacter capsulatus, which lacks this small subunit in its own bc1 complex.

Keywords: cytochrome bc1 complex; transmembrane helix; fbc operon; quinone binding.

The cytochrome bc1 complex (EC 1.10.2.2.) is a membraneprotein complex that catalyses the electron transfer betweenquinol and a c-type cytochrome. The free energy of thisreaction is used to translocate protons across the membranethrough the complex in order to create a protonmotive forceto be used in later cellular energetic processes [1]. The cyto-chrome bc1 complex of purple bacteria has been extensivelystudied as a model enzyme because of its structural simplicityand its functional equivalence to the complexes from higherorganisms. It is composed of usually just the three essentialcatalytic subunits: cytochrome b, cytochrome c1 and the Rieskeiron-sulphur proteins, which are encoded and expressedtogether from the fbc operon [2,3]. In search of a bacterialcytochrome bc1 complex for structural studies we have startedto work with Rhodovulum sulfidophilum, a marine purple non-sulphur bacterium. The complex possesses a number ofinteresting new features: it contains a high content ofubiquinone (about six molecules per monomer), possesses ahigh activity [4] and is expressed to an unusually high level inthe biological membrane (the ratio to the reaction centre isabout 2 : 1; W. Nitschke, personal communication). It purifiesas a four subunit complex as it contains the three catalyticsubunits and an additional 6-kDa protein of so far unknownfunction [4].

The cytochrome bc1 complex from Rhodobacter sphaeroidesalso purifies with a fourth subunit, subunit IV, which has a mass

of 14 kDa [5,6]. This is the only other bc1 complex from apurple bacterium so far reported to contain a fourth subunit.The intrinsic nature of subunit IV to the complex was shown bythe fact that an anti-subunit IV antibody affinity columnadsorbs the whole bc1 complex [7]. Photoaffinity labellingtechniques using azido-Q derivatives showed that the subunit isinvolved in quinone binding, and that this occurs in the putativeC-terminal transmembrane helix region [8]. The generation andcharacterization of a Rb. sphaeroides mutant lacking sub-unit IV, showed that the subunit is not absolutely vital to thecomplex [9]. The strain can grow phototrophically and thecomposition, spectral properties and activity of the bc1 complexin the membrane are similar to wild-type. However, thecomplex has a Km for quinone that is four times higher and itis more labile to detergent treatment compared to wild-type.The results suggest that subunit IV has a quinone bindingrole as well as a structural role in the bc1 complex ofRb. sphaeroides.

The 6-kDa protein from Rhv. sulfidophilum might be animpurity (light harvesting complex subunits have a mass ofabout 6 kDa and are present in large amounts in thephotosynthetic bacterial membrane), a cleaved signal sequencefrom an unrelated protein, or a genuine subunit of the complex.Even if a genuine subunit, the 6-kDa protein might be acleavage product of a larger protein; subunit 9 of the mito-chondrial bc1 complex, for example, is cleaved from the Rieskeprecursor protein [10]. However, the fact that the 6-kDa proteinis only removed from the bc1 complex upon dissociation of thecomplex with a high detergent concentration [4] suggests that itis an intrinsic component.

In this paper, we analyze the role of the 6-kDa protein inmore detail. The gene encoding the protein, the fbcS gene, hasbeen cloned, sequenced and localized in the genome. To test therelevance of the 6-kDa protein to the stability and function ofthe cytochrome bc1 complex, the fbcS gene has been disruptedin the genome using double homologous recombination, and

Correspondence to I. Sinning, European Molecular Biology Laboratory,

Structural Biology Programme, Meyerhofstrasse 1, 69012 Heidelberg,

Germany. Fax: 1 49 6221 387306, Tel.: 1 49 6221 387274,

E-mail: [email protected]

Abbreviations: DIG, digoxigenin; Qi site, ubiquinol-reducing centre;

Qo site, ubiquinol-oxidizing centre.

Enzyme: cytochrome bc1 complex (EC 1.10.2.2.).

*Present address: Molecular Microbiology, GlaxoWellcome Research

Center, Via Fleming 4, I-37135 Verona, Italy.

(Received 24 January 2000, revised 4 April 2000, accepted 18 April 2000)

the cytochrome bc1 complex has been purified and charac-terized from this mutant strain.

E X P E R I M E N T A L P R O C E D U R E S

Cloning of the fbcS gene

Whenever possible standard molecular biological methodswere used [11] and will not be described. The N-terminalsequence of the 6-kDa protein was previously known to bePDNTSNDDVLVPAS [4] and a proteolytic cleavage product ofthe protein with an N-terminal sequence of AQVFPL was alsoobtained. Using this information degenerate primers weredesigned, taking into consideration the high G 1 C bias of theorganism, in order to synthesize a probe from the Rhv. sulfi-dophilum genome by polymerase chain reaction (PCR). PrimersP1 (5 0-P-CCSGAYAAYACNTCSAAYGA-3 0) and INT1 (5 0-P-AGSGGRAANACYTGNGC-3 0) were used in 30 cycles ofPCR performed on genomic DNA using Taq polymerase(PerkinElmer). The denaturation, annealing and elongationtimes were 95 8C, 62 8C and 72 8C, respectively, for 1 min,45 s and 30 s, respectively. The PCR product was cloned andsequenced to check its suitability. A product of 158 bp wasobtained, which from knowledge of the protein mass [4]corresponded to almost the complete encoding region. To makethe probe, digoxigenin (DIG)-dUTP was incorporated duringthe PCR using the manufacturer's protocol (Roche).

Rhv. sulfidophilum genomic DNA was digested with com-binations of different restriction enzymes and run on a 0.8%agarose gel. This was transferred by Southern blot onto a nylonfilter. Hybridization was performed overnight at 65 8C instandard hybridization buffer, followed by posthybridizationand immunological detection as per the DIG-System protocol(Roche). From this, a EcoRV±PvuII fragment of about 3300 bpwas selected and cloned into pBluescriptKS II1 (Stratagene).Colony hybridization was performed using the same protocolfollowed after the Southern hybridization above. One positiveclone, pEP9, was subsequently selected and a restriction mapwas constructed. A 1200-bp EcoRI±EcoRV fragment wassubcloned and sequenced using standard primers. Sequencingwas carried out using 7-deaza-dGTP in place of dGTP andfidelase (Fidelity Systems Inc.), due to the high G 1 C contentof the DNA.

Heterologous expression of Rhv. sulfidophilum cytochromebc1 complex

The plasmid pEB17, which is pBluescriptKS II1 with a5.7 kbp genomic fragment of Rhv. sulfidophilum, containingthe fbc operon, cloned into the EcoRI and BamHI sites(S. Rodgers, C. Moser & I. Sinning, unpublished results) wasused as template in two PCR reactions with primers oligo A (5 0-CATGGATCCCGACCCGGTCCTGGAAAG-3 0) and oligo B(5 0-CGCAGGCGAACGAATACCGG-3 0) and primers oligo 1(5 0-TAT-GCCGACGGCACCGAG-3 0) and oligo 2 (5 0-GTA-GAATTCAAGAAATACCCCGCCACG-3 0) using the Pfu poly-merase enzyme (Stratagene). The two PCR products were eachpurified and cut with AatII and NcoI, respectively. They werethen ligated with the 2378-bp AatII±NcoI restriction fragmentof pEB17. This 2865-bp band containing the complete fbcoperon was gel purified, cut with BamHI and EcoRI, andcloned into the expression plasmid pCHB500 [12] to createpCHB500FBC. Sequencing confirmed that no point mutationshad been incorporated into the PCR products.

The conjugation of the bc1 complex deficient Rb. capsulatusstrain, MT-RBC1 [13,14], with pCHB500FBC for heterologous

expression of the Rhv. sulfidophilum complex was performedusing the triparental mating system [15]. Positive colonies weredetected after 2 days of incubation at 30 8C on tetracycline,spectinomycin and streptomycin containing MPYE agar plates[16]. One positive clone was grown anaerobically in 15 LMPYE medium in 1 L screw-cap bottles under a light intensityof about 2000 lux (Osram, 120 W bulbs) at 30 8C for 3 days.The purification of the bc1 complex from this strain and fromthe wild-type Rb. capsulatus strain (DSM155) was by thestandard protocol [4].

Sequencing the sox operon of Rb. capsulatus

Primers for the known sequence to the 5 0-terminus of soxBfrom Rb. capsulatus [17], CAPPROBF (5 0-P-GAACACGA-GATGCGCGGCTACAAG-3 0), and a highly conserved regionin the 5 0-terminus of soxA seen by alignment of knownbacterial sequences, SARCOXAR (5 0-P-RTTNGCNAGNAG-NGCSGANGCNAGNGTRTCNCC-3 0), were used to PCR theintervening DNA from Rb. capsulatus genomic DNA. The PCRwas performed with 30 cycles of denaturation at 95 8C for1 min, annealing at 65 8C for 45 s, and elongation at 72 8C for2 min, using Taq polymerase (PerkinElmer). The 1400-bpproduct was cloned and sequenced as before.

Disrupting the fbcS gene

To make the suicide vector, pSUP2026KDAV, clone pEP9 wascut with AatII and purified. Linkers formed by oligonucleotidesAATLINHI (5 0-P-CCGTCTCCGTCGCACGCCG-3 0) andAATLINLO (5 0-CGGCGTGCGACGGAGACGGACGT-3 0)were ligated on to create blunt ends. Into this was ligated theSmaI fragment from pHP45V [18], containing the V inter-poson. The PstI fragment (< 3750 bp) was cut out and ligatedinto the suicide vector pSUP202 [19] cut with PstI. Theconjugation of pSUP2026KDAV into a rifampicin resistantstrain of Rhv. sulfidophilum was performed by the triparentalmating system (S. Rodgers, C. Moser & I. Sinning, unpublishedresults). Nine out of 200 colonies that were rifampicinR/spectinomycinR/streptomycinR were also rifampicinR/spectino-mycinR/streptomycinR/tetracyclineS. Clones were checked byPCR with primers 6KDPROBF (5 0-CCCGTGGCGTGG-TTTGCATAAGTTG-3 0), and 6KDPROBR (5 0-CCCTCGG-TCCGGTTCATGACATGAC-3 0) to check for the presenceof an intact vs. an V interposon interrupted fbcS gene(280-bp vs. 2400-bp bands, respectively); and primersOMEGFOR (5 0-CCGCATTAAAATCTAGCGAGGGCTTTAC-TAAGC-3 0) and 6KDAREV (5 0-CCGGCAGAACCGGCAA-CCGGCTGTTATCGCGCC-3 0) to check for the presence of theV cassette next to the soxA gene (1100-bp band), and thus fordouble homologous recombination. Both reactions wereperformed at 95 8C for 1 min, 63 8C for 45 s and 72 8C for75 s, using Taq polymerase (PerkinElmer). One of the positivecolonies, RHVSUL-DFBCS, was grown up in 35 L of M27Nmedium, and the cytochrome bc1 complex was purified usingthe standard protocol [4].

Biochemistry of the cytochrome bc1 complex

The purification, protein and ubiquinone quantification andactivity tests of the complex were performed as outlinedpreviously [4].

Computer analysis of nucleotide and protein sequences

The deduced 6-kDa protein sequence was analysed with blast2[20] to check for homologous protein matches. Using the

3754 S. Rodgers et al. (Eur. J. Biochem. 267) q FEBS 2000

phdhtm and phdtopology programmes [21], sequences wereanalysed to predict transmembrane helix location and topology.The sequence upstream of fbcS was examined using mactarg-search 2.0 [22] to analyse potential transcriptional promotersequences. Amino-acid sequence alignments were performedusing the clustal w 1.6 programme [23].

R E S U LT S

Localization of the fbcS gene

The sequence of the gene encoding the fourth subunit of theRhv. sulfidophilum bc1 complex, the 6-kDa protein, is shown inFig. 1, surrounded by its neighbouring genes. The gene hasbeen named fbcS. The fbcS gene is positioned between thesoxB and soxD genes in the middle of the sox operon, asidentified using a blast2 search [20]. This operon encodes thesubunits of the heterotetrameric sarcosine oxidase complex,which catalyses the oxidative demethylation of sarcosine toyield glycine, hydrogen peroxide, and 5,10-methylenetetra-hydrofolate [24].

The complete bacterial sequence of the sox operon has beendetermined for Corynebacterium sp. P-1 ([24] EMBL databankU23955 and a partial sequence exists for Sinorhizobium meliloti

(EMBL Data Bank AF055582). The operon contains four soxgenes arranged in the order soxBDAG and the genes are spacedvery closely together. Whereas the operons for the Coryne-bacterium sp. P-1 and S. meliloti have 11 bp and 25 bp betweentheir soxB and soxD genes, respectively, Rhv. sulfidophilum has472 bp (Figs 1 and 2A); comprising 238-bp upstream from thefbcS gene, 174 bp for the gene and 60 bp after it.

Upstream from fbcS a sequence region has been found withhomology to E. coli promoter regions (boxed regions in Fig. 1)using the mactargsearch program [22]. This transcriptionalpromoter would allow transcription of fbcS to be independentof sox operon expression. There is also a putative Shine±Dalgarno sequence, GGAGAG (residues 299±304), for the fbcSgene to enable translation of the gene. The soxD and soxAgenes overlap as is seen also in Corynebacterium sp. P-1,S. meliloti and the Rb. capsulatus sequences (Fig. 2A) and thisis associated with translational coupling.

The 6-kDa protein

The fbcS gene encodes a 57-amino-acid polypeptide (Figs 1and 3A). The expressed 6-kDa protein mass known fromMALDI-MS to be 5839 Da [4] corresponds to within 1 Da of

Fig. 1. The sequence of the fbcS gene, encoding the 6-kDa protein of Rhv. sulfidophilum. The 1197-bp EcoRI-EcoRV genomic fragment contains the

fbcS gene surrounded by the genes of the sox operon, soxB and soxD, which encode subunits of the heterotetrameric sarcosine oxidase. G 1 A rich stretches

before each gene, indicating Shine±Dalgarno sequences, are shown with a line above the sequence. A possible transcriptional promoter site before the fbcS

gene is indicated by the two boxed regions, showing the 235 site and the 210 site. The soxD and soxA genes overlap, which is a conserved feature of

bacterial sox operons and this is associated with translational coupling. Below the nucleotide sequence, the protein sequences of the 6-kDa protein and the sox

operon encoded proteins are given.

q FEBS 2000 The fbcS gene of Rhodovulum sulfidophilum (Eur. J. Biochem. 267) 3755

the theoretical mass of the expressed fbcS gene product, if theN-terminal methionine is excluded. The known N-terminal andinternal sequences are also seen from analysis of the sequencedgene. There is no protein sequence homologous to the 6-kDaprotein found from a blast2 search [20].

Overall, the sequence is hydrophobic, especially in thecentral part. The transmembrane protein prediction pro-grammes phdhtm and phdtopology predict a single trans-membrane helix, as shown in Fig. 3A. Following the positive-inside rule [25], the N-terminus would be inside the cytoplasm(negative side of the membrane), and the C-terminus would bein the periplasm (positive side of the membrane).

Expression of the Rhv. sulfidophilum bc1 complex inRb. capsulatus

The fbc operon of Rhv. sulfidophilum was heterologouslyexpressed in a Rb. capsulatus strain lacking the complex[13,14]. The bc1 complex was purified from this clone using thesame method as for Rhv. sulfidophilum [4], and gave a much

lower yield per mg chromatophore membrane protein, about15% compared to preparation of the complex from its nativesource. This purified complex had 70% of the activity of the bc1

complex purified from the native source.On a gel the purified complex looked identical to the

complex purified from the native source (Fig. 4, lanes B and C)and a protein of similar size to the 6-kDa protein is again found.N-terminal sequencing gave a protein sequence of DNQSNDD,which is very similar to that found for the 6-kDa protein fromRhv. sulfidophilum [4]. When the bc1 complex is purified fromwild-type Rb. capsulatus using the same protocol, this smallband is not seen (Fig. 4, lane A). A small subunit, similar to the6-kDa protein, is thus present in Rb. capsulatus membrane, butit has a much higher affinity for the Rhv. sulfidophilumcomplex.

Localization of the fbcS gene in Rb. capsulatus

A fourth subunit is not found to purify with the cytochrome bc1

complex of Rb. capsulatus (Fig. 4, lane A), a near relative of

Fig. 2. Organization of the sox operon in different bacteria. (A) The sox operon gene organization from (i) Corynebacterium sp. P1 (iii) S. meliloti and (iv)

Rb. capsulatus all have the same organization in the region shown, whereas (v) Rb. sphaeroides has a region of 491 bp inserted between the soxB and soxD

genes, containing the fbcQ gene, and (ii) Rhv. sulfidophilum has a region of 472 bp, containing the fbcS gene. The EcoRI±EcoRV genomic fragment

sequenced from Rhv. sulfidophilum and the AatII site used to insert the V cassette, for disruption of the fbcS gene from Rv. sulfidophilum is highlighted. For

reference, the organization of the Corynebacterium sp. P1 sox operon genes is shown at the top of the figure. The scale bar relates to parts i±v only. (B) Figure

showing that subunit IV of Rb. sphaeroides is encoded between the soxB and soxD genes. The sequences of the (i) C-terminus of the SoxB proteins and (ii)

N-terminus of the SoxD proteins of Rhv. sulfidophilum (SUL), Rb. sphaeroides (SPH) and Rb. capsulatus (CAP). For the C-terminus of the Rb. sphaeroides

SoxB protein only 16 bp of coding nucleotide sequence is available. Absolute conservation of a residue is indicated by *, whereas a dot shows similarity.


both Rhv. sulfidophilum and Rb. sphaeroides, although thesequences of their three catalytic subunits are highly homo-logous to each other ([26,27], S. Rodgers, C. Moser & I.Sinning, unpublished results).

When the equivalent region of the sox operon was isolatedfrom the Rb. capsulatus genome by PCR and sequenced, theregion between soxB and soxD was seen to be only 85-bp longand devoid of the fbcS gene (Fig. 2A). During this work thissequence also became available on the Capsulapedia web-site(http://rhodol.uchicago.edu/capsulapedia/capsulapedia/capsula-pedia.shtml). Although this is a longer stretch than seen forother bacteria, it does not contain a gene. Thus, a gene encodinga fourth subunit is not present in Rb. capsulatus in the samerelative position in the genome compared to Rhv. sulfidophilumand Rb. sphaeroides.

Disruption of the fbcS gene in Rhv. sulfidophilum

In order to assess the role of the 6-kDa protein in thecytochrome bc1 complex of Rhv. sulfidophilum, the fbcS genewas disrupted by insertion of the V cassette and introduced into

the genome by double homologous recombination. Afterconjugation, seven out of 200 colonies found to be spectino-mycin resistant but tetracycline sensitive were seen by PCR tobe disrupted in the targeted fbcS gene. The PCR check wasnecessary as two colonies were found to be spontaneouslyspectinomycin resistant.

The mutant strain, RHVSUL-DFBCS, was grown up in a35-L culture and the cytochrome bc1 complex was purified. Theyield of the complex is much less than when the wild-typeprotein is purified (7 mg of complex compared to . 30 mg,respectively). The activity of the purified complex is found tobe 4±5 mmol´s21 of cytochrome c reduced per mmol ofcytochrome c1 (for the wild-type protein it is 24 mmol´s21).Thus, disrupting the expression of the 6-kDa protein dramati-cally reduces both the yield and activity of the isolated bc1

complex from Rhv. sulfidophilum. In order to test whether theamount of bc1 complex is already decreased in membranes ofthe Rhv. sulfidophilum deletion strain monoclonal antibodieswhich recognize the Rieske protein and cytochrome b subunitof the bc1 complex from Paracoccus denitrificans (kindlyprovided by H. Michel, Frankfurt) and which recognize the

Fig. 3. The predicted transmembrane topology of the 6-kDa protein and comparison with subunit IV of Rb. sphaeroides. (A) Using the phdhtm and

phdtopology prediction programmes [21], the 6-kDa protein of Rhv. sulfidophilum was predicted to form a single transmembrane helix, with the more

positive N-terminus residing on the cytoplasmic (negative) side of the membrane; (B) the predicted topology of subunit IV of Rb. sphaeroides. The sidedness

of each molecule is predicted to be the same with the N-terminus on the cytoplasmic side of the membrane, on the opposite side to the cytochrome c1 and

Rieske protein extrinsic domains.


respective proteins from Rhv. sulfidophilum have been used inWestern blots. No significant difference in the amount of bothcomponents could be detected between membranes of wild typeand mutant strains of Rhv. sulfidophilum (data not shown). Inaddition, no significant difference in activity (measured asabove) could be found. This indicates, that the bc1 complex ofthe mutant strain is more sensitive towards solubilization and

purification which points towards a stabilizing role of the 6-kDaprotein.

Comparing the composition of the bc1 complex fromthe wild-type strain and the RHVSUL-DFBCS strain bySDS/PAGE (Fig. 5, lanes A and B, respectively), it is seen thatthe presence of a lower band in the gel is greatly reduced ondisrupting the fbcS gene, but it is not completely removed as avery faint band is still seen in lane B. The identity of theresidual protein has not been determined. This observationis similar to what was observed when the fbc operon ofRhv. sulfidophilum is expressed in Rb. capsulatus.

The quinone content was quantified for the complex lackingthe 6-kDa protein by RP-HPLC. In the wild-type complex,there are 5±6 quinones per monomeric complex [4]. Thecomplex lacking the 6-kDa protein contains five quinones percytochrome c1. Therefore, the deletion of the 6-kDa proteinfrom the complex had no significant effect on the binding ofquinone to the complex. This implies that the high quinonecontent of the Rhv. sulfidophilum complex is an intrinsicproperty of the conserved core subunits.

D I S C U S S I O N

We have cloned and sequenced the fbcS gene encoding the6-kDa protein which copurifies with the bc1 complex ofRhv. sulfidophilum [4]. From its protein sequence the 6-kDaprotein is predicted to form a single transmembrane helix.There are precedents for additional small subunits to beassociated with purified membrane proteins. Examples canbe seen directly in the mitochondrial bc1 complex structures[28±30]. Here five out of 11 subunits are less than 10 kDa(subunits 7±11). The subunits 7, 10 and 11 all contain atransmembrane helix. The latter two proteins form part of thecavities of the dimeric complex and have been implicated toplay a role in the proper assembly of the complex [31,32].Whereas subunit 11 is more peripherally located and can beeasily removed from the complex by delipidation without lossof activity [33], subunit 10 is tightly associated with thecomplex [34]. Similarly, the 6-kDa protein of Rhv. sulfi-dophilum is tightly associated with its complex; it can only beremoved by use of a high detergent concentration whichdissociates the complex [4]. The homologous cytochrome b6 fcomplex from higher plant chloroplasts purifies with threesmall hydrophobic polypeptides (PetG, PetL and PetM), eachwith a mass of about 4 kDa [35]. PetG and PetL have bothbeen shown by knock-out experiments in Chlamydomonasreinhardtii to be required for the stability and/or assembly ofthe complex [36,37]. Recently, PetN has been identified as anadditional small subunit of the cytochrome b6 f complex intobacco. This hydrophobic polypeptide of only 29 amino acidsplays a crucial role in complex assembly and/or stability [38].The 12-kDa protein component purifying with the cytochromec oxidase of Paracoccus denitrificans [39] has no homologywith any known protein and is known from its crystal structureto be a single transmembrane spanning helix [40]. The deletionof this component does not change the complex's spectral andenzymatic properties, and in contrast to the examples before, italso has no effect for the integrity of the complex. The role ofsubunit IV in the bc1 complex of Rb. sphaeroides for quinonebinding and complex stability will be discussed below.

The fbcS gene, encoding the 6-kDa protein, is located in acurious position. It is located in the only non-overlappingregion of the soxBDAG operon, between the soxB and soxDgenes (Fig. 2A). The sox operon encodes the four subunits ofthe heterotetrameric sarcosine oxidase complex. For expression

Fig. 4. Heterologous expression of the cytochrome bc1 complex. The

12.5% SDS-polyacrylamide protein gel is overloaded with Rb. capsulatus

(lane A) Rhv. sulfidophilum heterologously (lane B) and natively expressed

(lane C) bc1 complexes. Protein samples were treated for 10 min at 56 8C

in loading buffer and left overnight at room temperature prior to electro-

phoresis. The gel is stained with Coomassie brilliant blue R-250.

Molecular mass standards (in Da) are shown on the right-hand side.

Fig. 5. Comparison of the bc1 complex from wild-type and RHVSUL-

DFBCS strains of Rhv. sulfidophilum. A 15% SDS/polyacrylamide protein

gel overloaded with the wild-type protein complex, containing the 6-kDa

protein (lane A), and the complex isolated from the RHVSUL-DFBCS

(with the fbcS gene disrupted) strain, where this band is largely removed

(lane B). Samples are pretreated as in Fig. 4.


of the complex in Corynebacterium sp. P-1, sarcosine needs tobe present in the medium [24]. In the M27N minimal mediumused to grow Rhv. sulfidophilum there is no sarcosine present;however, the 6-kDa protein is known to be expressed [4]. Theregulation of sox operon expression might be different in thesetwo species, although they are both proteobacteria. The soxoperon in Corynebacterium sp. P-1 includes at the start the geneglyA (encoding a putative serine hydroxymethyltransferase),whereas the sox operon in Rhv. sulfidophilum has the cycHgene (involved in cytochrome biogenesis) upstream of it. Thisgene is not part of the operon, being transcribed in the oppositedirection. This latter sequence is known from the upstreamsequence of the cloned 3.3kbp EcoRV±PvuII Rhv. sulfi-dophilum genomic fragment (not shown) and this organizationis seen also in a sequence of the Rb. capsulatus genome [17].Due to this different organization at the start of the operon, it isnot inconceivable that the expression of the sox operon inRhv. sulfidophilum might not need sarcosine present in themedium. However, a search for a transcriptional promoter siteusing the mactargsearch programme [22] found a regionupstream of the fbcS gene that has homology to E. colipromoters (Fig. 1). The fbcS gene could thus have independenttranscription from the rest of the sox operon. This finding willhave to be checked with northern blot analysis.

The cytochrome bc1 complex of Rb. sphaeroides has a fourthlow molecular weight subunit, as Rhv. sulfidophilum does. Thebc1 complexes of other proteobacteria so far characterized donot copurify with a fourth subunit. Comparing the sequences ofthe 6-kDa protein of Rhv. sulfidophilum and subunit IV ofRb. sphaeroides, there is no sequence homology. They arehowever, both predicted to contain a single transmembranehelix, with the N-terminus on the cytoplasmic, negative side ofthe membrane (Fig. 3). The subunit IV has a long N-terminuscompared to the 6-kDa protein. It is predicted to be in thecytoplasm and thus would not interfere with the cytochrome c1

and Rieske extrinsic domains. A point mutation in the Trp79residue of subunit IV was seen to produce the same effects onthe four subunit complex as when the subunit IV is not present[41]. This suggests that it is a structurally important residue andinvolved in quinone binding. There is one tryptophan in the6 kDa sequence (Trp19) which is predicted to be in a similarposition, close to the negative side of the membrane.

The position of the fbcQ gene (encoding the subunit IV) andthe fbcS gene (encoding the 6-kDa protein) are seen in similarrelative positions to each other in their respective genomes(Fig. 2). Although not stated in the original sequence paper thebase position 508 of the sequenced 650-bp genomic fragmentof Rb. sphaeroides is the start of the soxD gene [8], as shown bythe comparison of the soxD genes of Rhv. sulfidophilum,Rb. sphaeroides and Rb. capsulatus (Fig. 2B). The first 16residues of the cloned fragment also encode identical aminoacids to the C-terminal end of soxB from Rhv. sulfidophilumand Rb. capsulatus (Gly-Ala-His-Opal stop codon) (Fig. 2B).From this the fbcQ gene can be positioned relative to thesox operon genes (Fig. 2A). It is seen that the distance betweenthe soxB and the soxD genes is very similar betweenRb. sphaeroides and Rhv. sulfidophilum, although the fbcQgene is 231-bp longer than the fbcS gene. The 5 0-end of thefbcQ gene encoding the N-terminal extension `replaces' thelarge intergenic region seen between soxB and fbcS. However,this intergenic region in the Rhv. sulfidophilum genome couldnot encode a protein sequence similar to the N-terminus of thesubunit IV.

In order to assess the importance of the 6-kDa protein to thebc1 complex in Rhv. sulfidophilum, the fbcS gene encoding this

protein was disrupted by V cassette mutagenesis. The straincreated grew anaerobically in a manner similar to the wild-typestrain and Western blot analysis showed that the amount of bc1

complex in the membrane is similar to the wild type strain.However, as the amount of purified complex is only about 25%of wild-type levels and the activity was only 20% of the wild-type complex the 6-kDa protein is structurally relevant to theRhv. sulfidophilum bc1 complex. It seems to play a stabilizingrole. A loss in activity is also seen with the Rb. sphaeroides bc1

complex on removal of its subunit IV which is also lesssignificant in the chromatophore membranes [9].

Based on the high homology between the mitochondrial andbacterial bc1 complex, one can predict the topology for a threesubunit bacterial complex, such as those from P. denitrificans,or Rb. capsulatus. The quinone cavities should be more open inthe bacterial complex which lacks additional subunits corres-ponding do subunits 10 and 11. This would lead to a less-tightbinding of the `pooled' quinone, as is the case for mostproteobacteria. However, the opposite has been observed forRhv. sulfidophilum [4]. The explanation could be that the 6-kDaprotein forms a transmembrane helix that narrows the cavity,thereby trapping the quinones. However, in this case thedeletion of the 6-kDa protein should have a significant effect onthe binding of quinone by the remaining bc1 complex. To oursurprise, the quinone content of the isolated bc1 complex fromRhv. sulfidophilum does not depend on the presence of the6-kDa protein. However, the 6-kDa protein, subunit IV andsubunit 10 are intrinsic to the stability of their respective bc1

complexes and their removal leads to a drop in enzymaticactivity. In Rhv. sulfidophilum the remaining bc1 complex stillcontains a small protein associated with it as it could beseen from SDS/PAGE, the identity of which is not clear. Thispoints towards the presence of additional, probably single-spanning proteins in the membrane which bind with low affinityto the complex but cannot provide the stabilizing effect of the6-kDa protein. This also relates to the copurification of a fourthsubunit when the fbc operon of Rhv. sulfidophilum is expressedheterologously in Rb. capsulatus. This subunit is not encodedbetween the soxB and soxD genes, as are the 6-kDa proteinand subunit IV. It is not known why the bc1 complexes ofRb. sphaeroides and Rhv. sulfidophilum require a fourthsubunit, whereas those of other closely related proteobacteria(such as Rb. capsulatus and P. denitrificans) appear not to. Itmight be that the affinity for the fourth subunit is moredetergent labile with these other complexes, just as a W79Lmutation in subunit IV makes it more labile to detergent [41].Therefore, a fourth subunit might be associated with thecomplex in the chromatophore membrane, but not copurifywith it. An example is seen with the reaction centre ofRubrivivax gelatinosus; in the chromatophore membrane thecomplex contains a bound tetrahaem cytochrome c, but whenpurified it is not attached [42]. With other proteobacterial bc1

complexes a fourth subunit might copurify, but the small bandmight not be detected by SDS/PAGE. This was the caseoriginally for the 12 kDa subunit found with the cytochrome coxidase of P. denitrificans [39]. It can be expected that moremembrane protein complexes, both novel and those alreadycharacterized, will be found to contain small, additionaltransmembrane proteins. These are easily overlooked, butthey might provide an important stabilizing function.

A C K N O W L E D G E M E N T S

We would like to thank Fevzi Daldal for pCHB500 and MT-RBC1, Joachim

Frey (University of Bern) for pHP45V, Dr W. Arnold (Bielefeld University)


for pSUP202 and Keith Ashman for N-terminal sequencing. We thank

H. Michel (Max-Planck-Institute for Biophysics, Frankfurt) for antibodies

against the Rieske protein and cyt.b from P. denitrificans bc1 complex. We

are also grateful to Guillermo Montoya and Kai te Kaat for stimulating

discussions. S. R. is supported by a DFG-grant Si-586/1 and M. M.-J. by a

Ramon Areces grant.

R E F E R E N C E S

1. Trumpower, B.L. & Gennis, R.B. (1994) Energy transduction by

cytochrome complexes in mitochondrial and bacteria respiration: the

enzymology of coupling electron transfer reactions to transmembrane

proton translocation. Annu. Rev. Biochem. 63, 675±716.

2. Gray, K.A. & Daldal, F. (1995) Mutational studies of the cytochrome bc1

complexes. In Anoxygenic Photosynthetic Bacteria (Blankenship, R.E.,

Madigan, M.T. & Bauer, C.E., eds), pp. 747±774. Kluwer Academic

Publishers, Dordrecht, the Netherlands.

3. Trumpower, B.L. (1990) Cytochrome bc1 complexes of micro-

organisms. Microbiol. Rev. 54, 101±129.

4. Montoya, G., te Kaat, K., Rodgers, S., Nitschke, W. & Sinning, I.

(1999) The cytochrome bc1 complex from Rhodovulum sulfidophilum

is a dimer with six quinones per monomer and an additional 6 kDa

component. Eur. J. Biochem. 259, 709±718.

5. Andrews, K.M., Crofts, A.R. & Gennis, R.B. (1990) Large-scale puri-

fication and characterisation of a highly active four subunit cyto-

chrome bc1 complex from Rhodobacter sphaeroides. Biochemistry

29, 2645±2651.

6. Yu, L., Mei, Q. & Yu, C. (1984) Characterisation of purified

cytochrome bc1 complex from Rhodopseudomonas sphaeroides

R-26. J. Biol. Chem. 259, 5752±5760.

7. Yu, L. & Yu, C.A. (1991) Essentiality of the molecular weight 15,000

protein (subunit IV) in the cytochrome b-c1 complex of Rhodobacter

sphaeroides. Biochemistry 30, 4934±4939.

8. Usui, S. & Yu, L. (1991) Subunit IV (Mr � 14,384) of the

cytochrome b-c1 complex from Rhodobacter sphaeroides: cloning,

DNA sequencing, and ubiquinone binding domain. J. Biol. Chem.

266, 15644±15649.

9. Chen, Y.-R., Usui, S., Yu, C.-A. & Yu, L. (1994) Role of subunit IV in

the cytochrome b-c1 complex from Rhodobacter sphaeroides.

Biochemistry 33, 10207±10214.

10. Brandt, U., Yu, L., Yu, C.A. & Trumpower, B.L. (1993) The

mitochondrial targeting presequence of the Rieske iron-sulfur protein

is processed in a single step after insertion into the cytochrome

bc1 complex in mammals and retained as a subunit in the

complex. J. Biol. Chem. 268, 8387±8390.

11. Sambrook, J., Fritsch, E.F. & Maniatis, T. (1989) Molecular Cloning: a

Laboratory Manual, 2nd edn. Cold Spring Harbor Laboratory Press,

Cold Spring Harbor, New York, USA.

12. Benning, C. & Sommerville, C.R. (1992) Isolation and genetic

complementation of a sulfolipid-deficient mutant of Rhodobacter

sphaeroides. J. Bacteriol. 174, 2352±2360.

13. Atta-Asafo-Adjei, E. & Daldal, F. (1991) Size of the amino acid side

chain at position 158 of cytochrome b is critical for an active

cytochrome bc1 complex and for photosynthetic growth of Rhodo-

bacter capsulatus. Proc. Natl Acad. Sci. USA 88, 492±496.

14. Tokito, M.K. & Daldal, F. (1992) petR, located upstream of the fbcFBC

operon encoding the cytochrome bc1 complex, is homologous to

bacterial response regulators and necessary for photosynthetic and

respiratory growth of Rhodobacter capsulatus. Mol. Microbiol. 6,

1645±1654.

15. Donohue, T.J. & Kaplan, S. (1991) Genetic techniques in Rhodo-

spirillaceae. Methods Enzymol. 204, 459±485.

16. Weaver, P.F., Wall, J.D. & Gest, H. (1975) Characterization of

Rhodopseudomonas capsulata. Arch. Microbiol. 105, 207±216.

17. Lang, S.E., Jenney, F.E. & Daldal, F. (1996) Rhodobacter capsulatus

CycH: a bipartite gene product with pleiotropic effects on the

biogenesis of structurally different c-type cytochromes. J. Bacteriol.

178, 5279±5290.

18. Prentki, P. & Krisch, H.M. (1984) In vitro insertional mutagenesis with

a selectable DNA fragment. Gene 29, 303±313.

19. Simon, R., Priefer, U. & PuÈhler, A. (1983) A broad host range

mobilisation system for in vivo genetic engineering: transposon

mutagenesis in Gram negative bacteria. Bio/Technol. 1, 784±791.

20. Altschul, S.F. & Gish, W. (1996) Local alignment statistics. Methods

Enzymol. 266, 460±480.

21. Rost, B., Casadio, R., Fariselli, P. & Sander, C. (1995) Prediction of

helical transmembrane segments at 95% accuracy. Protein Sci. 4,

521±533.

22. Mulligan, M.E., Hawley, D.K., Entriken, R. & McClure, W.R. (1984)

Escherichia coli promoter sequences predict in vitro RNA poly-

merase selectivity. Nucleic Acids Res. 12, 789±800.

23. Higgins, D.G., Thompson, J.D. & Gibson, T.J. (1996) Using

CLUSTAL for multiple sequence alignments. Methods Enzymol.

266, 383±402.

24. Chlumsky, L.J., Zhang, L. & Jorns, M.S. (1995) Sequence analysis of

sarcosine oxidase and nearby genes reveals homologies with key

enzymes of folate one-carbon metabolism. J. Biol. Chem. 270,

18252±18259.

25. von Heijne, G. (1986) The distribution of positively charged residues in

bacterial inner membrane proteins correlates with the transmembrane

topology. EMBO J. 5, 3021±3027.

26. Davidson, E. & Daldal, F. (1987) fbc operon, encoding the Rieske Fe-S

protein, cytochrome b, and cytochrome c1 apoproteins previously

described from Rhodopseudomonas sphaeroides, is from Rhodo-

pseudomonas capsulata. J. Mol. Biol. 195, 13±24.

27. Yun, C.-H., Beci, R., Crofts, A.R., Kaplan, S. & Gennis, R.B. (1990)

Cloning and DNA sequencing of the fbc operon encoding the

cytochrome bc1 complex from Rhodobacter sphaeroides. Eur. J.

Biochem. 194, 399±411.

28. Xia, D., Yu, C.-A., Xia, J.-Z., Kachurin, A.M., Zhang, L., Yu, L. &

Deisenhofer, J. (1997) Crystal structure of the cytochrome bc1

complex from bovine heart mitochondria. Science 277, 60±66.

29. Zhang, Z., Huang, L., Schulmeister, V.M., Chi, Y.-I., Kim, K.K., Hung,

L.-W., Crofts, A.C., Berry, E.A. & Kim, S.-H. (1998) Electron

transfer by domain movement in cytochrome bc1. Nature 392,

677±684.

30. Iwata, S., Lee, J., Okada, K.W., Lee, J.K., Iwata, M., Rasmussen, B.,

Link, T.A., Ramaswamy, S. & Jap, B.K. (1998) Complete structure of

the 11-subunit bovine mitochondrial cytochrome bc1 complex.

Science 281, 64±71.

31. Crofts, A.R. & Berry, E.A. (1998) Structure and function of the

cytochrome bc1 complex of mitochondria and photosynthetic

bacteria. Curr. Op. Str. Biol. 8, 501±509.

32. Phillips, J.D., Graham, L.A. & Trumpower, B.L. (1993) Subunit 9 of

Saccharomyces cerevisiae cytochrome bc1 complex is required for

insertion of EPR-detectable iron-sulfur cluster into the Rieske iron-

sulfur protein. J. Biol. Chem. 268, 11727±11736.

33. SchaÈgger, H., Hagen, T., Roth, B., Brandt, U., Link, T.A. & von Jagow,

G. (1990) Phospholipid specificity of bovine heart bc1 complex.

Eur. J. Biochem. 190, 123±130.

34. SchaÈgger, H., Link, T.A., Engel, W.D. & von Jagow, G. (1986)

Isolation of the eleven protein subunits of the bc1 complex from beef

heart. Methods Enzymol. 126, 224±237.

35. Pierre, Y., Breyton, C., Kramer, D. & Popot, J.L. (1995) Purification

and characterization of the cytochrome b6 f complex from Chlamy-

domonas reinhardtii. J. Biol. Chem. 270, 29342±29349.

36. Berthold, D.A., Schmidt, C.L. & Malkin, R. (1995) The deletion of

petG in Chlamydomonas reinhardtii disrupts the cytochrome b6 f

complex. J. Biol. Chem. 270, 29293±29298.

37. Takahashi, Y., Rahire, M., Breyton, C., Popot, J.L., Joliot, P. &

Rochaix, J.D. (1996) The chloroplast ycf 7 ( petL) open reading

frame of Chlamydomonas reinhardtii encodes a small functionally

important subunit of the cytochrome b6 f complex. EMBO J. 15,

3498±3506.

38. Hager, M., Biehler, K., Illerhaus, J., Ruf, S. & Bock, R. (1999) Targeted

inactivation of the smallest plastid genome-encoded open reading


frame reveals a novel and essential subunit of the cytochrome b6 f

complex. EMBO J. 18, 5834±5842.

39. Witt, H. & Ludwig, B. (1997) Isolation, analysis, and deletion of the

gene coding for subunit IV of cytochrome c oxidase in Paracoccus

denitrificans. J. Biol. Chem. 272, 5514±5517.

40. Iwata, S., Ostermeier, C., Ludwig, B. & Michel, H. (1995)

Structure at 2.8 AÊ resolution of cytochrome c oxidase from

Paracoccus denitrificans. Nature 376, 660±669.

41. Chen, Y.-R., Shenoy, S.K., Yu, C.-A. & Yu, L. (1995)

Identification of amino acid residues involved in structural and

ubiquinone-binding functions of subunit IV of the cytochrome

bc1 complex from Rhodobacter sphaeroides. J. Biol. Chem. 270,

11496±11501.

42. Agalidis, I. & Reiss-Husson, F. (1992) Purification and characterization

of Rhodocyclus gelatinosus photochemical reaction center. Biochim.

Biophys. Acta 1098, 201±208.


deletion of the 6-kda subunit affects the activity and yield of the bc1 complex from rhodovulum...

Documents