PhotosynthesisResearch 42: 43-50, 1994. (~) 1994KluwerAcademicPublishers. Printedin the Netherlands. Regular paper

Deletion of eytochrome c oxidase genes from the cyanobaeterium Synechoeystis sp. PCC6803: Evidence for alternative respiratory pathways Georg Schmetterer,Daniel Alge & WolfgangGregor Membrane Protein Group, Institute of Physical Chemistry, Wdhringerstrafle 42, A-1090 Wien, Austria Received20 October 1993;acceptedin revisedform29 June 1994 Key words: bioenergetics, cox locus, cytochrome c oxidase mutant, respiratory chain

Abstract

An oligonucleotide directed against a highly conserved region of aa3-type cytochrome c oxidases was used to clone the cox genes from the cyanobacterium Synechocystis sp. PCC6803. Several overlapping clones were obtained that contained the coxB, coxA, and coxC genes, transcribed in the same direction in that order, coding for subunits II, I, and III, respectively. The deduced protein sequences of the three subunits showed high sequence similarity with the corresponding subunits of all known aa3-type cytochrome c oxidases. A 1.94-kb HindII fragment containing most of coxA and about half of coxC was deleted and replaced by a cassette coding for kanamycin resistance. Mutant cells that were homozygous for the deleted cox locus were obtained. They were viable under photoautotrophic and photoheterotrophic conditions, but contained no cytochrome c oxidase activity. Nevertheless, these mutant cells showed almost normal respiration, defined as cyanide-inhibitable 02 uptake by whole cells in the dark. It is concluded, therefore, that aa3-type cytochrome c oxidase is not the only terminal respiratory oxidase in Synechocystis sp. PCC6803. Abbreviations: CM - cytoplasmic membrane; DCMU - 3-(3,4-dichlorophenyl)-l,l-dimethylurea; HQNO 2-heptyl-4-hydroxyquinoline N-oxide; ICM - intracytoplasmic membranes; SU - subunit; TES - (Ntris(hydroxymethyl)methyl)-2-aminoethane sulfonic acid Introduction

Cyanobacteria are prokaryotes capable of oxygenic photosynthesis. Their main mode of life is therefore photosynthetic, but all cyanobacteria exhibit detectable levels of respiration in the dark with 02 as the terminal electron acceptor. Evidence for aaa-type cytochrome oxidase as a terminal respiratory oxidase in cyanobacteria came from spectroscopic data (Peschek et al. 1982), inhibitor studies (Peschek et al. 1988), EPR data (Fry et al. 1985), and cross reaction of the putative proteins with antibodies against the aa3type cytochrome oxidase of the bacterium Paracoccus denitrificans (Trnka and Peschek 1986). Immunological data also showed that purified cyanobacterial cytochrome oxidase possibly contains four subunits, corresponding to SU I, SU II, SU III, and SU IV in oth-

er aa3-type cytochrome oxidases (Niederhauser 1992). The complete subunit composition of cyanobacterial cytochrome oxidase, however, remains unknown at the present. The most important subunit of aa3-type cytochrome oxidases is undoubtedly SU I, since it could be shown in Paracoccus denitrificans that this subunit carries both hemes a and a3 and possesses significant enzymatic activity by itself (Mtiller et al. 1988). Therefore, an oligonucleotide homologous to a region of SU I of Paracoccus denitrificans that is highly conserved in all known aa3-type cytochrome oxidases was used to clone the corresponding gene from Synechocystis sp. PCC6803 (Alge et al. 1994; Alge and Peschek 1993). The genes for two other cytochrome oxidase subunits were identified by sequencing the flanking regions.

44 In contrast to several other cyanobacteria (such as Synechococcus sp. PCC6301), which apparently contain cytochrome oxidase activities in both their cytoplasmic membranes (CM) and their intracytoplasmic membranes (ICM) or 'thylakoids' (Molitor et al. 1987), Synechocystis sp. PCC6803 has negligible cytochrome oxidase activitiy in the CM (see Fig. 3). The original aim of this work was to verify the existence of an aa3-type cytochrome oxidase in this strain, and to address whether cyanobacteria are able to survive in the absence of this enzyme. Although the present study shows that a viable homozygous cytochrome oxidase mutant could be constructed, surprisingly this mutant had almost wild type respiratory activity, thus providing for the first time evidence that a cyanobacterium may contain branched respiratory electron transport pathways. Preliminary results about the cloning of the Synechocystis sp. PCC6803 cox genes were first reported at the 'EMBO Workshop on Comparative Structures and Function of Membranes in Chloroplasts and Cyanobacteria' in Korfu, Greece, 1989.

Materials and methods

Glucose-sensitive, wild type (see Flores and Schmetterer 1986) Synechocystis sp. PCC6803, obtained from C. P. Wolk, was grown in liquid in medium BG11 (Rippka et al. 1979) supplemented with 10 mM TES and on plates in the same medium supplemented with 10 g 1-1 agar and 3g 1-1 Na thiosulfate. For cells containing the kanamycin resistance gene from plasmid pRL446 (consisting of S.A1, L.EHE1 and C.K2, nomenclature of Elhai and Wolk 1988) 5 #g ml- l kanamycin was added. For the measurement of survival in the dark, cells were grown under photoautotrophic conditions as lawns on plates for 4 days, harvested, diluted to a concentration of 108 cells ml-1 and incubated at 30 °C in the dark on a rotary shaker. Total DNA from Synechocystis sp. PCC6803 cells was prepared by the method of Williams (1988). Six pooled gene banks in vector pUC 18 (Yanisch-Perron et al. 1985) prepared from DNA of the glucose-tolerant variant ofSynechocystis sp. PCC6803 (Williams 1988) were a gift of J.G.K. Williams and have been described previously (Schmetterer 1990). Restriction enzymes were from BoehringerMannheim and used according to the directions of the manufacturer. Southern blotting was performed with the Pharmacia VacuGene Blotter. Oligonucleotide 'C' (AACATRTGRTGNGCCCA) is complementary to the

non-coding strand of a region highly conserved in all known aa3-type cytochrome oxidases and was a gift of M. Saraste (EMBL, Heidelberg, Germany). Hybridization of Southern blots with oligonucleotide 'C' as a probe was performed as described (Raitio et al. 1987). DNA sequencing was performed on both strands by the dideoxy method (Sanger et al. 1977) using the Boehringer-Mannheim kit on overlapping subclones of the cox locus in pUCll8 or pUCll9 (Vieira and Messing 1987). Both the universal primers for pUC vectors and primers designed according to sequence already determined were employed. The search for sequence similarities between different cytochrome oxidases was performed by the method of Pearson and Lipman (Pearson and Lipman 1988) with the help of the GCG program package (Devereux et al. 1984). The mutant containing no cytochrome oxidase was made by insertional inactivation as described (Williams 1988). In order to test for homozygocity, cells from several single cell colonies were separately transferred to 50 ml of BG11 medium with kanamycin and grown to stationary phase. Total DNA prepared from these cultures was then used for Southern blots and hybridized with the HindII fragment lacking in pDAUV22 and pDAUV23 (see Fig. 1). DNA that gave no band corresponding to the wild type cox locus in the Southern blots was subjected to PCR (following the protocol of Sambrook et al. 1989) using oligonucleotides 5~-CTGGGCCCACCATATGq~FCAC-3 ' and 5'-GTCATTI~CGGGAACCAGTGG-3'. These correspond to nucleotides 2082-2102 and to the complement of 2403-2423, respectively, of the cox locus sequence and both lie within the coxA part of the HindlI fragment to be deleted. Membranes from Synechocystis sp. PCC6803 wild type and Cox- mutant Synechocystis UV22 were prepared essentially as described (Murata and Omata 1988). In short, the cells were broken in a French pressure cell and the debris removed by centrifugation (Beckman JA10 rotor, 6000 rpm, l0 rain, 4 °C). The supernatant was centrifuged in an ultracentrifuge (Beckman Ti60 rotor, 50000 rpm, 50 min, 4 °C) and the resulting pellet resuspended in 10 mM HEPES/ NaOH (pH 7.4). This suspension was called 'crude membranes'. It was used either directly for measuring cytochrome oxidase activity or applied onto a sucrose density gradient for further purification and the separation of CM and ICM (Murata and Omata 1988). Cytochrome oxidase activity of the isolated membranes was assayed in a Shimadzu double wave° length spectrophotometer by measuring oxidation of

45

E ~ z n ~ v v w

l/(f

I

H I L

/i

~

/Irfl

t

I

x

i

I

I{1!

I

(2) 1(3)

I

(4) (5) (8)

I I t~\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\~

L\\\\\\\\\\\\\\\\\\\\)

I

0

t000

~

coxC

f

I t

Ikk\\\\\\\k\\'O

coxA

cox B I

2000

~

t t

I

3000

t

I

1(8) I

4000

(7)

ORF4 I

4700 bp

Fig. 1. The cox locus ofSynechocystis sp. PCC6803 (1) Restriction map. (2) - (5) Clones obtained from gene banks of wild type Synechocystis

sp. PCC6803: (2) pDAUV1, (3) pDAUV2, (4) pDAUV26, (5) pDAUV27. (6) The binding site of oligonucleotide 'C'. (7) The open reading frames coding for the three subunits of cytochrome oxidase and the ORF4 of unknown significance; arrows indicate transcriptional direction. (8) The plasmids used for the construction of the Cox- mutant: the open box represents the 1.94-kb HindlI fragment missing; it is replaced by a kanamycin resistance cassette whose transcriptional direction is either parallel (pDAUV22) or antiparallel (pDAUV23) to that of the cox genes.

in vitro reduced beef heart cytochrome c (18 #M) in 10 mM Na phosphate pH 7.4, using A550-A540 at 20 °C as described (Molitor and Peschek 1986). Activity was tested for inhibition by 1.5 #M KCN. Protein content of the isolated membranes was determined by the method of Bradford using bovine serum albumin as the standard (Bradford 1976). Respiratory activity of whole cells was defined as KCN (2 mM) inhibitable O2-uptake by intact cells in the dark measured with a Clark-type electrode and a YSI Model 53 Oxygen Monitor.

Results

Cloning the cox locus from Synechocystis sp. PCC6803

Cloning part of the cox locus using oligonucleotide 'C' has been described (Alge et al. 1994). Two overlapping clones were obtained and named pDAUV1 and pDAUV2 (Fig. 1). Sequencing of the cloned inserts revealed an incomplete open reading frame at the 5' end of pDAUV1, so that the BamHI-HindlII fragment of this plasmid was used to isolate by colony hybridization two additional clones named pDAUV26 and pDAUV27 (Fig. 1), from other libraries. The sequence of the cox locus has been deposited at the EMBL Data Library under the accession number X53746. Four open reading frames were identified and named coxB, coxA, coxC, and ORF4 (Fig. 1): coxB codes for SU II (320 amino acids, calculated molecular mass 34816 Da), coxA for SU I (533 amino acids,

59388 Da) and coxC for SU III (218 amino acids, 24429 Da). The size of SU III assumes that the TTG at nucleotides 2929-2931, which is preceded by a ShineDalgarno type ribosomal binding site, is the start codon for coxC. Whether the small ORF4 (44 amino acids, 4878 Da) codes for a protein is not clear. Neither the DNA sequence nor the derived amino acid sequence show similarity with the sequences in the data bases. The Synechocystis sp. cytochrome oxidase

PCC6803 mutant lacking

Synechocystis sp. PCC6803 was originally chosen as the object of study because it is easy to introduce site specific mutations in this strain (Williams 1988). The Hindu fragment in pDAUV2 was removed and replaced by a kanamycin resistance cassette from pRL446 (Elhai and Wolk 1988) (see Fig. 1). Both orientations of the inserted cassette were obtained and called pDAUV22 and pDAUV23 (reading frame of the kanamycin resistance gene parallel and antiparallel to that of the cox genes, respectively). Synechocystis sp. PCC6803 was transformed with both plasmids with selection for kanamycin resistance. Transformants were obtained with both plasmids, but the strain constructed with pDAUV23 grew considerably more slowly than the one constructed with pDAUV22. Since cyanobacteria contain several copies of the chromosome (Herdman 1982), it was first checked whether the mutant cells were homozygous for kanamycin resistance - in other words, whether the cells had lost all copies of the original cox locus. It was observed that homozygocity was unusually difficult to

46

Fig. 2. Verification of homozygocity in Synechocystis UV22 by Southern hybridization with the Hindll fragment deleted in pDAUV22 (Figs. 2A, 2B) and by PCR (Fig. 2C). In the Southern experiments, the expected sizes for the wild type cox locus are 4.7kb (HindlII) and 6.4kb (XbaI). These bands appear prominently in wild type Synechocystis sp. PCC6803 (Fig. 2A, lanes 2 and 4, Fig. 2 B, lanes 1 and 3) and weaker in a heterozygous Cox mutant, where only a part of the chromosomes carry the wild type locus (Fig. 2A, lanes 3 and 5). The bands axe absent in the homozygous Cox- mutant, where all chromosomes carry the cox locus interrupted by the kanamycin resistance cassette (Fig. 2B, lanes 2 and 4). In contrast, the minor bands at 3.9kb (HindlII) and 4.2kb (XbaI) have the same intensity in all cases, showing that they are independent of the cox locus. Fig. 2A. DNA wild type Synechocystis sp. PCC6803 (lanes 2 and 4) and Synechocystis UV22 clone 3-4 (lanes 3 and 5) cut with HindlII (lanes 2 and 3) or XbaI (lanes 4 and 5). Lane 1 shows lambda DNA cut with HindlII and hybridized to itself. Fig. 2B. DNA from wild type Synechocystis sp. PCC6803 (lanes 1 and 3) and Synechocystis UV22 clone 3-1 (lanes 2 and 4) cut with HindlII (lanes 1 and 2) or XbaI (lanes 3 and 4). Fig. 2C. PCR products using the two oligonucleotides described in Materials and methods. Bands A correspond to the expected wild type size (342 bp). The significance of bands B and C is not yet known. In all lanes a total of 1/zg of DNA from different sources was loaded. Lane 1:1 #g (wild type(WT)). Lane 2:0.1 #g (WT) + 0.9/zg (clone 3-1 ). Lane 3:0.05 #g (WT) + 0.95 #g (clone 3-1). Lane 4:0.03 #g (WT) + 0.97 #g (clone 3-1). Lane 5:0.02 #g (WT) + 0.98/zg (clone 3-1). Lane 6:1 #g (clone 3-1). Lane 7:1 #g (clone 3-2). Lane 8:1 #g (clone 3-3 ).

a c h i e v e in this case. Starting f r o m the original colonies obtained after transformation with p D A U V 2 2 , only repeated restreaking of k a n a m y c i n resistant colonies finally led to c y t o c h r o m e oxidase minus strains that w e r e h o m o z y g o u s . A f t e r three restreakings the first

colonies w e r e found that appeared h o m o z y g o u s by Southern hybridization at the c o x locus (Fig. 2B), but a considerable n u m b e r o f colonies w e r e still heterozygous at the c o x locus (Fig. 2A). In order to ascertain further the h o m o z y g o c i t y o f the strain to be used for

47 phenotypic characterization, DNA from three independent colonies that gave no signal in Southern hybridization was subjected to PCR amplification using two oligonucleotides whose sequence is contained within the HindlI fragment to be removed in the mutants (see Materials and methods). Of the three colonies tested, only two (clones 3-1 and 3-2) gave no detectable signal corresponding to the expected wild type PCR product (342 bp, band A), while clone 3-3 clearly did (Fig. 2C). In order to estimate the detection limit for the presence of residual wild type cox locus in clone 3-1, mixtures of wild type D N A and mutant DNA were subjected to PCR (Fig. 2C, lanes 2 to 5). Since lane 5 contains only 2% wild type DNA but still gives a significant yield of wild type band A we could estimate that the concentration of wild type cox locus in Synechocystis UV22 clone 3-1 is at least 1000-fold less than in wild type cells. Since the chromosome copy number in Synechocystis sp. PCC6803 is about 10-20 (Herdman 1982), Fig. 2C implies that Synechocystis sp. PCC6803 can grow in the absence of cytochrome oxidase. Cells from clone 3-1 were picked for two further restreakings, after which no signal corresponding to the wild type bands were detected in either Southern hybridizations or PCR amplifications (data not shown) and the resulting strain was called Synechocystis UV22. However, not all kanamycin resistant clones were homozygous at the cox locus after five restreakings (data not shown). A homozygous strain Synechocystis UV23 was constructed in an analogous way using plasmid pDAUV23. Since Synechocystis UV22 (and similarly Synechocystis UV23) contains no intact coxA and coxC genes, these cells should contain no cytochrome oxidase activity and Fig. 3 shows this to be the case. Wild-type Synechocystis sp. PCC6803 displayed similar activities in its 'crude membranes' (see Materials and methods) and in the purified ICM, but no detectable activity in the CM. The specific activity of ICM from wild type Synechocystis sp. PCC6803 was 10 (nmol cytochrome c oxidized) min -1 (mg protein) -1, the specific activity of 'crude membranes' from the same strain varied from 6-12 (nmol cytochrome c oxidized) min -1 (mg protein) -x. In Synechocystis UV22 no activity was detected in either the 'crude membranes' or the ICM, even if a very high concentration of membranes was used (Fig. 3, trace 6). Table 1 shows measurements of respiratory activity in whole cells of wild type Synechocystis sp. PCC6803 and mutant Synechocystis UV22. It is clear that despite the absence of cytochrome oxidase, the mutant still

KCN

/.min [10 units

KCN

KIN /.

--5 ~ 6

Fig. 3. Cytochrome oxidase activities of isolated membranes from wild type Synechocystis sp. PCC6803 (traces 1. 2, 3, 4) and the Cox- mutant SynechocystisUV22 (traces 5, 6, 7). The membranes used were either 'crude membranes' (see Materials and methods) (traces 1, 2, 5, 6), ICM (traces 3, 7) or CM (trace 4). The amount of membranesused is given as #g protein per assay: 10 (traces 2, 3), 20 (traces 1, 4, 5, 7), 100 (trace 6). The right end of each trace indicates the start of the experiment by addition of the membrane suspension to the assay buffer (see Materials and methods) containing fully reduced horse heart cytochrome c. Proceeding from right to left the change in A550 - A540 is measured. Downward movement of the trace corresponds to oxidation of reduced cytochrome c. One unit correspondsto 51.3 nM cytochromec oxidised. At the time indicated by the arrows KCN was added to a final concentration of 1.5 #M.

respires. The difference from the wild type is only quantitative, but not qualitative: as in the wild type, the endogenous respiration can be significantly stimulated by exogenous glucose, and both H Q N O and KCN are inhibitors. The same experiments performed with Synechocystis UV23 gave qualitatively similar results (data not shown). The growth rate of Synechocystis UV22 under photoautotrophic conditions was a little higher than that of the wild type. Under the low-light regime used, the doubling time during exponential growth was about 3.4 days for the wild type and 2.7 days for the mutant. It is well known that Synechocystis sp. PCC6803 will not grow in complete darkness (Rippka et al. 1979; Anderson and Mclntosh 1991). However, when the cells were submitted to dark incubation, a significant phenotypic difference between wild type Synechocystis

48 Table 1. Oxygen consumption in (/zl O2).h-l.(mg Chl a) -1 by Synechocystis sp. PCC6803 wild type and C o x - mutant Synechocystis UV22 in BG11 medium in the dark at 30 o C. The numbers given represent the mean of two independent measurements, the variations axe shown in parentheses Additions to the medium

Wild type

None ('endogenous respiration') 20 mM glucose 20 mM glucose + 10 mM HQNO 20 mM glucose + 10 mM HQNO + 2 mM KCN

Table 2. Survival of Synechocystis sp. PCC6803 wild type and C o x - mutant Synechocystis UV22 suspended in BGll medium in the dark at 30 °C in the absence or presence of 5 mM glucose, n.d.: not determined Days

0

3 7 12 23

Viable counts per ml +Glucose Wild type Mutant

-Glucose Wild type

Mutant

1.108 8.107 4.107 4.105 0

1.108 n.d. n.d. n.d. 5.107

1.108 n.d. n.d. n.d. 1.107

1.108 3.106 1.104 0 n.d.

sp. PCC6803 and mutant Synechocystis UV22 became evident (Table 2). In the absence of added glucose, both the wild type and the mutant survived well for three weeks. In the presence of 5 mM glucose the Cox- mutant lost viability significantly faster than the wild type. Both the Cox- mutant and the wild type were able to grow photoheterotrophicaUy with 5 mM glucose in the presence of 10#M of the PS II inhibitor DCMU.

Discussion The genes for three subunits of the cytochrome oxidase of Synechocystis sp. PCC6803 have been cloned and sequenced (Alge and Peschek 1993; Alge et al. 1994). The results verify that the cytochrome oxidase in this cyanobacterium is of the aa3-type. In recognition of the fact that all prokaryotic aaa-type cytochrome oxidases whose sequences are available (Paracoccus denitrificans (Raitio et al. 1987, 1990), Bacillus sp. PS3 (Ishizuka et al. 1990), Bacillus subtilis (Saraste et

0.124 0.298 0.159 0.086

(0.020) (0.026) (0.008) (0.002)

Mutant 0.067 0.277 0.108 0.031

(0.013) (0.018) (0.006) (0.010)

al. 1991), Thermus thermophilus (Mather et al. 1991), Synechococcus vulcanus (Tano et al. 1991; Sone et al. 1992), Bradyrhizobium japonicum (Gabel and Maier 1990), Rhodobacter sphaeroides (Shapleigh and Gennis 1992), Bacillus firmus (Quirk et al. 1992)) show considerable sequence similarity with the homologous subunits of eukaryotic cytochrome oxidases (Saraste 1990; also see Alge et al. 1994) the genes for the three subunits were called coxB, coxA, and coxC. The cox genes from the two cyanobacteria Synechocystis sp. PCC6803 and Synechococcus vulcanus display high mutual sequence similarity: the deduced amino acid sequences for SU I, SU II, and SU III are 61.9%, 43.1%, and 45.4% identical, respectively. The proximity of the cox genes (Fig. 1) may make it likely that they constitute an operon, but attempts to identify the transcript(s) of the cox locus were not successful. The order of the four genes in the cox locus of Synechocystis sp. PCC6803 is the same as in Bacillus subtilis (Saraste et al. 1991), Bacillus sp. PS3 (Ishizuka et al. 1990) and Synechococcus vulcanus (Tano et al. 1991; Sone et al. 1992). However, this gene arrangement is not universal among prokaryotic cox loci. In Paracoccus denitrificans there are three nonlinked loci, one containing the genes for SU II and SU III, and two coding for two isozymes of SU I (Raitio et al. 1987, 1990). In B. subtilis, it could be shown that the gene encoding the small open reading frame at the 3' end of the cox locus probably codes for a fourth subunit of the cytochrome oxidase (Gai et al. 1990). Niederhauser (1992) reports values of 55 kDa, 32 kDa, 17.6 kDa, and 15.3 kDa for the sizes of the putative SU I, II, III, and IV in purified cytochrome oxidase from the cyanobacterium Anacystis nidulans (Synechococcus sp.). While the values for SU I, II, and III correspond fairly well to those calculated from the sequence data presented here,

49

NAD(P) H Quinones "

Cytochrome b6f

x

I

(?),,

I(+) \

Cytochrome c

\

,

Alternote Oxidose(s)

Cytochrome c Oxidose

O2

Fig. 4. Proposedschemefor the respiratoryelectrontransportchains of Synechocystis sp. PCC6803. ORF4 evidently does not code for a subunit similar to the putative A. nidulans SU IV. We report here for the first time the construction of a cytochrome oxidase mutant of a cyanobacterium. It proved rather diffcult to obtain a genuinely homozygous C o x - mutant. In the course of testing Synechocystis UV22 and Synechocystis UV23 for homozygocity (Fig. 2), it was observed that both under suitable Southern hybridization conditions and in the PCR experiments bands in addition to the expected ones were obtained. These bands were present both in the wild type and in the mutant (see Figs 2A and 2B; data not shown for PCR experiment). The possiblity that the one addtional band seen in Southern hybridization experiments with coxA contains (part of) the alternate terminal respiratory oxidase (see below) is currently under investigation. Despite the absence of cytochrome oxidase (Figs. 2 and 3), the respiratory activity of Synechocystis UV22 is similar to that of the wild type strain (Table 1), which must, therefore, contain at least one alternate terminal oxidase (Fig. 4). While branched respiratory chains are common in prokaryotes, so far no such proposal has been put forward for a cyanobacterium. Table 1 shows that there are some quantitative differences between respiratory activities of the wild type and the mutant. These are hard to interpret, however, since it is not known whether the alternate oxidase is expressed similarly in the wild type and in the mutant, and thus what the contribution - if any - of this alternate oxidase is in the wild type. Some properties of the alternate respiratory pathway can be inferred from the data: the alternate terminal oxidase must be cyanide sensitive, and the respiratory pathway to the alternate oxidase must include the quinone pool, since HQNO is an inhibitor of the oxidation of the quinone pool. The

existence of a cytochrome o, for instance, would be compatible with these data, since this enzyme is sensitive to both inhibitors (Kita et al. 1984). However, no spectroscopic evidence for such a cytochrome has so far been obtained in a cyanobacterium. Furthermore, the data currently do not distinguish whether the alternate respiratory pathway includes the cytochrome b6f complex, and thus both possibilities are presented in Fig. 4. In case the alternate oxidase should turn out to be of the cytochrome o type, the cytochrome b6f complex would not be expected to be part of the alternate respiratory pathway, since cytochrome o acts directly as a quinol-dioxygen oxidoreductase (Kita et al. 1984)

Acknowledgements M. Saraste kindly donated oligonucleotide 'C'. We thank J.G.K. Williams for the gene banks of Synechocystis sp. PCC6803, J. Elhai for plasmid pRL446, and A. Beyer for help with the GCG program. Thanks are also due to O. Kuntner for his dedicated and excellent technical assistance. Financial support from the Fonds zur F6rderung der wissenschaftlichen Forschung in Osterreich (Projekt No. P7015 and $6007) is gratefully acknowledged.

References Alge D and PeschekGA (1993) Identificationand characterization of the ctaD (coxB) gene as part of an operon encoding subunits I, II, and III of the cytochromec oxidase(eytochromeaa3)in the cyanobacteriumSynechocystisPCC6803. BiochemBiophysRes Commun 191:9-17 Alge D, SclunettererG and PeschekGA (1994) The gene encoding cytochrome c oxidase subunit I from Synechocystis PCC6803. Gene 138:127-132

50 Anderson SL and Mclntosh L (1991) Light-activated heterotrophic growth of the cyanobacterium Synechocystis sp. strain PCC6803: A blue-light-requiring process. J Bacteriol 173:2761-2767 Bradford M (1976) A rapid and sensitive method for the quantitation of microgram quantities of protein utilizing the principle of protein-dye binding. Anal Biochem 72:248-254 Devereux J, Haeberli P and Smithies O (1984) A comprehensive set of sequence analysis programs for the VAX. Nucl Acids Res 12: 387-395 Elhai J and Woik CP (1988) A versatile class of positive selection vectors based on the nonviability of palindrome containing plasraids that allows cloning into long polylinkers. Gene 68:119-138 Flores E and Schmetterer G (1986) Interaction of fructose with the glucose permease of the cyanobacterium Synechocystis sp. strain PCC6803. J Bacteriol 166:693-696 Fry IV, Peschek GA, Hufleijt M and Packer L (1985) EPR signals of redox active copper in EDTA washed membranes of the cyanobacterium Synechococcus 6311. Biochem Biophys Res Commun 129:109-116 Gabel C and Maier RJ (1990) Nncleotide sequence of the coxA gene encoding subunit I of cytochrome aa3 of Bradyrhizobium japonicum. Nucl Acids Res 18:6143 Gai WZ, Sun SM, Sone N and Chart SHP (1990) Cytochrome oxidase from thermophilic bacterium PS3 contains a fourth protein subunit. Biochem Biophys Res Commun 169:414--421 Grigorieva G and Shestakov S (1982) Transformation in the cyanobacterium Synechocystis sp. 6803. FEMS Microbiol Lett 13:367-370 Herdman M (1982) Evolution and genetic properties of cyanobacterial genomes. In: Carr NG and Whitton BA (eds) The Biology of Cyanobacteria, pp 263-305. Blackwell Scientific Publications, Oxford, UK Ishizuka M, Machida K, Shimada S, Mogi A, Tsuchiya T, Ohmori T, Souma Y, Gonda M and Sone N (1990) Nucleotide sequence of the gene coding for four subunits of cytochrome c oxidase from the thermophilic bacterium PS3. J Biochem 108:866-873 Kita K, Konishi K and Anraku Y (1984) Terminal oxidase of E. coli aerobic respiratory chain. I. Purification and properties of cytochrome b556-o complex from cells in the early exponential phase of aerobic growth. J Biol Chem 259:3368-3374 Mather MW, Springer P and Fee JA (1991) Cytochrome oxidase genes from Thermus thermophilus. Nucleotide sequence and analysis of the deduced primary structure of subunit IIc of cytochrome caa3. J Biol Chem 266:5025-5035 Molitor V and Peschek GA (1986) Respiratory electron transport in plasma and thylakoid membrane preparations from the cyanobacterium Anacystis nidulans. FEBS Lett 195:145-150 Molitor V, Tmka M and Peschek GA (1987) Isolated and purified plasma and thylakoid membranes of the cyanobacterium Anacystis nidulans contain immtmologically cross-reactive aa3-type cytochrome oxidase. Curr Microbiol 14:263-268 Mtiller M, Schliipfer M and Azzi A (1988) Cytochrome c oxidase from Paracoccus denitrificans: Both hemes are located in subunit I. Proc Nail Acad Sci USA 85:6647--6651 Murata N and Omata T (1988) Isolation of cyanobacterial plasma membranes. Meth Enzymol 167:245-251 Niederhauser H (1992) Isolierung und Charakterisierung der Cytochrom-c-Oxidase ans dem Cyanobakterium Anacystis nidulans. Dissertation, University of Vienna, Austria

Pearson WR and Lipman DJ (1988) Improved tools for biological sequence comparison. Proc Nat Acad Sci USA 85: 2A,~ 2448 Peschek GA, Molitor V, Trnka M, Wastyn M and Erber W (1988) Characterization of cytochrome-c oxidase in isolated and purified plasma and thylakoid membranes from cyanobacteria. Meth Enzymol 167:437-449 Peschek GA, Schmetterer G, Lauritsch G, Nitschmann WH, Kienzl PF and Muchl R (1982) Do cyanobacteria contain 'mammaliantype' cytochrome oxidase? Arch Microbiol 131:261-265 Quirk PG, Hicks DB and Krulwich TA (1992) Cloning of the cta operon from alkalophilic Bacillusfirmus OF4, and characterization of the pH regulated cytochrome caa3 oxidase it encodes. J Biol Chem 268:678--685 Raitio M, Jalli T and Saraste M (1987) Isolation and analysis of the genes for cytochrome c oxidase in Paracoccus denitrificans. EMBO J 6:2825-2833 Raitio M, Pispa JM, Metso T and Saraste M (1990) Are there isozymes of cytochrome c oxidase in Paracoccus denitrificans? FEBS Lett 261:431--435 Rippka R, Deruelles J, Waterbury JB, Herdman M and Stanier RY (1979) Generic assignments, strain histories and properties of pare cultures of cyanobacteria. J Gen Microbiol 111: 1-61 Sambrook J, Fritsch EF and Maniatis T (1989) Molecular cloning. Cold Spring Harbor Laboratory Press, Cold Spring Harbor, NY Sanger E Nicklen S and Coulson AR (1977) DNA sequencing with chain-terminating inhibitors. Proc Natl Acad Sci USA 74: 54635467 Saraste M (1990) Structural features of cytochrome oxidase. Quart Rev Biophys 23:331-366 Saraste M, Metso T, Nakari T, Jalli T, Lauraeus M and Van der Oost J (1991) The Bacillus subtilis cytochrome-c oxidase. Variations on a conserved protein theme. Ear J Biochem 195:517-525 Schmetterer G (1990) Sequence conservation among the glucose transporter from the cyanobacterium Synechocystis sp. PCC6803 and mammalian glucose transporters. Plant Mol Bol 14:697-706 Shapleigh JP and Gennis RB (1992) Cloning, sequencing and deletion from the chromosome of the gene encoding subunit I of the aaa-type cytochrome c oxidase ofRhodobacter sphaeroides. Mol Microbiol 6:635-642 Sone N, Tano H and Ishizuka M (1992) The genes in the thermophilic cyanobacterium Synechococcus vulcanus encoding cytochromec oxidase. Biochim Biophys Acta 1183:130-138 Tano H, Ishizuka M and Soue N (1991) The cytochrome c oxidase genes in blue-green algae and characteristics of the deduced protein sequence for subunit II of the thermophilic cyanobacterium Synechococcus vulcanus. Biochem Biophys Res Commnn 181: 437--442 Trnka M and Peschek GA (1986) Immunological identification of aa3-type cytochrome oxidase in membrane preparations of the cyanobacterium Anacystis nidulans. Biochem Biophys Res Commun 136:235-241 Vieira J and Messing J (1987) Production of single-stranded plasmid DNA. Meth Enzymol 153:3-11 Williams JGK (1988) Construction of specific mutations in photosystem II photosynthetic reaction center by genetic engineering methods in Synechocystis 6803. Meth Enzymol 167:766-778 Yanisch-Perron C, Vieira J and Messing J (1985) Improved M13 phage cloning vectors and host strains: nucleotide sequence of the M13mpl8 and pUC19 vectors. Geue 33:103-119