The respiratory chain of cyanobacteria appears to be branched rather than linear; furthermore, respiratory and photosynthetic electron-transfer chains co-exist in the thylakoid membrane and even share components. This review will focus on the three types of terminal respiratory oxidases identified so far on a genetic level in cyanobacteria: aa3-type cytochrome c oxidase, cytochrome bd-quinol oxidase and the alternative respiratory terminal oxidase. We summarize here their genetic, biochemical and biophysical characterization to date and discuss their interactions with electron donors as well as their physiological roles.

Cyanobacterial respiration

Cyanobacterial respiration is much less well understood than photosynthesis. This is not surprising, as most cyanobacteria are obligate photoautotrophs, although, in the dark, they depend completely on respiration to maintain energy levels. The presence of a respiratory electron-transfer chain at the cytoplasmic membrane is essential to provide energy for various transport processes (e.g. sugar import and ion pumps) and its importance increases under stress conditions such as high salt (the so-called ‘salt respiration’) and nitrogen fixation [1,2]. However, understanding the respiratory chain has been hampered by the high complexity of its components and their interactions.

Respiration in cyanobacteria is especially intriguing because it is not physically separated from the photosynthetic electron-transfer chain. Whereas the cytoplasmic membrane can only perform respiration, photosynthetic and respiratory electron-transfer chains co-exist in cyanobacterial thylakoids. They even share common components, such as the cytochrome bf complex and the redox carriers plastoquinol, Cyt c6 (cytochrome c6; previously known as cytochrome c-553) and Pc (plastocyanin) [38], so that electron transfer in the thylakoid resembles more a branched network than distinct chains. How these processes are regulated to avoid futile cycles has been reviewed by Scherer [6]. The emphasis of this review will be placed on the terminal oxidases found in cyanobacteria. We will introduce the different types of cyanobacterial terminal oxidases characterized so far and discuss their functional importance.

Terminal oxidases

Most respiratory terminal oxidases belong to the haem-copper superfamily, characterized by a binuclear centre (ligated by subunit I) comprising a CuB centre electronically coupled with a high-spin haem a. They reduce molecular oxygen to water and couple electron transfer with the pumping of protons, leading to ATP production. Depending on their substrate specificity, they can be further subdivided into COXs (cytochrome c oxidases; such as aa3-type COXs) or quinol oxidases (such as bo-type quinol oxidases).

In contrast, the bd-quinol oxidases (cytochrome bd-quinol oxidases) do not belong to the haem-copper superfamily: they lack the binuclear centre and instead contain two haem proteins. Although the bd-quinol oxidases are not capable of pumping protons, they still generate a membrane potential due to the exclusive use of protons from the negative side of the membrane in the reaction to form water; however, this results in a reduced coupling efficiency.

The aa3-type COX

The main type of oxidase, which has been found in all cyanobacteria characterized so far, is the aa3-type COX. Its genes are always found in a single operon: ctaCDE (or coxBAC), encoding subunit II (ligates CuA and is the main docking site for soluble donors), subunit I (ligates haem a and the haem a3-CuB binuclear centre) and subunit III respectively. Some cyanobacteria, e.g. Anabaena sp. PCC 7120, have two copies of the ctaCDE operon (coxBAC1 and coxBAC2) [9].

This enzyme has been characterized using spectroscopic, immunological and kinetic assays [8,10]. Evidence for its localization in both the cytoplasmic and thylakoid membranes has been reviewed by Peschek [11]. Its expression and distribution varies with growth conditions, especially salt concentrations [10,12].

Its kinetic activity has been investigated using whole cells, membranes and more recently the purified enzyme, with a range of donor proteins (for a recent review, see [8]). Most of the initial studies were carried out with horse heart cytochrome c; cyanobacterial COX generally shows high reactivity towards this protein [14]. In terms of physiological donors, Pc and Cyt c6, the electron donors to Photosystem I in cyanobacterial photosynthesis, are capable of interacting with COX [3,1517]. This has been reviewed by Peschek et al. [8,11]. Recent work has used a soluble CtaC protein, which, within limits, has enabled estimation of the rate constants for the interaction between the donor and the CuA centre, and will facilitate further probing of the role of specific residues through mutagenesis studies [1820].

There have been suggestions that a small c-type cytochrome, Cyt cM (cytochrome cM), is capable of interacting with COX (and indeed Photosystem I) as an alternative to Pc and Cyt c6, perhaps under stress conditions such as high light and low temperatures [2123]. However, evidence from several studies indicates that Cyt cM does not mediate electron transfer to Photosystem I and that no third mediator can compensate fully for the absence of Cyt c6 and Pc in an interaction with COX [24,25].

Based on sequence similarity to the cytochrome c-binding region of COX from Bacillus sp. and Thermus thermophilus, Cyt cM was proposed to be an additional subunit of COX [26]. This is consistent with the notion that the hydrophobic N-terminus of Cyt cM may function as a transmembrane region rather than a targeting sequence. Although studies with soluble CtaC indicate that Cyt cM is not absolutely required for electron transfer [1820], it may promote binding or act as a redox mediator between soluble electron carriers and COX [26]. Although the measured redox potential (Em) of soluble Cyt cM was only 150 mV [27] (which would raise a problem for its reduction by Cyt f (cytochrome f), Em≈330 mV), membrane localization and/or the presence of other COX subunits could potentially increase this to allow Cyt cM to act as a redox mediator.

Although COX appears to be the main terminal oxidase, sequencing projects and genetic approaches [28,29] have revealed the presence of other potential terminal oxidases: the ARTO (alternative respiratory terminal oxidase) and a bd-quinol oxidase.

The bd-quinol oxidase

Genes homologous with the cydAB operon encoding a bd-quinol oxidase in Escherichia coli were identified in Synechocystis sp. PCC 6803 [29]. Introduction of these cyanobacterial cydAB genes into an E. coli mutant lacking both the bo- and bd-quinol oxidases could restore respiration, indicating that they encode a functional bd-quinol oxidase [30].

Transcription of cydAB was detected by reverse transcriptase–PCR in Synechocystis sp. PCC 6803 [30] and Anabaena variabilis [31]. Whilst native CydAB has never been isolated, several studies would suggest the presence of a bd-quinol oxidase. Measurement of the flash-induced turnover of Cyt f suggested that there was competition with Cyt f for electrons downstream of Photosystem II, consistent with a functional bd-quinol oxidase [32]. In addition, in vivo studies using dyes responsive to proton concentrations demonstrated that a quinol oxidase is able to energize both the thylakoid and plasma membranes in the presence of DBMIB (2,5-dibromo-3-methyl-6-isopropylbenzoquinone), an inhibitor of the cytochrome bf complex [33]. However, a study investigating the haem content in both cytoplasmic and thylakoid membranes found no evidence of a functional bd-quinol oxidase [34,35].

The major quinone in cyanobacteria is plastoquinone-9; however, as discussed by Peschek et al. [8], no plastoquinone-supported bd-quinol oxidase has yet been identified.

The ARTO

Analysis of the genome sequence of Synechocystis sp. PCC 6803 [29] revealed a set of genes very similar in sequence to the ctaCDE genes, encoding an ARTO. These genes have subsequently been found in most cyanobacteria, with multiple ARTO operons in certain N2-fixing species (e.g. Nostoc punctiforme [36]). However, no ARTO operons have been identified in the thermophilic cyanobacteria Thermosynechococcus elongatus BP-1 {[37] and Prochlorococcus sp. MED4 and MIT9313 (genomes completed; Joint Genome Institute, Walnut Creek, CA, U.S.A.)}.

In general, the genes show the same organization as the COX operon. However, the usual organization of the ARTO operon (ctaCDEII) is fragmented in Synechocystis sp. PCC 6803: the ctaCII gene is in a different region of the genome. The only other example of a different genetic organization found so far is Gloeobacter violaceus PCC 7421 (which lacks thylakoids and has the photosynthetic electron transfer chain in the cytoplasmic membrane), where only the ctaCDII genes are found and no ctaEII gene can be identified in the genome sequence [38].

Two small genes, encoding hypothetical transmembrane proteins, were found only 24 bp upstream of the ctaCII in Anabaena sp. PCC 7120 and were shown to be co-transcribed with the ARTO genes [36]. Their presence just upstream of ctaCII appears to be conserved among cyanobacteria.

The ARTO complex also appears to be a member of the haem-copper superfamily, as the residues responsible for ligation of the binuclear centre are conserved. However, on the basis of subunit composition, it is more similar to the bo-quinol oxidases (cytochrome bo-quinol oxidases), as the ARTO subunit II homologue lacks the CuA site (and is highly conserved where this site would be) [30].

Transcription of the ARTO genes has been detected in Anabaena sp. PCC 7120 [36] and at lower levels in A. variabilis [31] and Synechocystis sp. PCC 6803 [30]. In Synechocystis sp. PCC 6803, ctaEII transcripts could only be detected in cells grown under low light (hence avoiding light stress that is associated with loss of functional COX) in strains with deletions in both ctaDE and cydAB [30].

Currently, there is no consensus in the literature on the physiological substrates or role of the ARTO complex. Reactivity towards bovine cytochrome c could not be detected in vitro [28]. However, this is a basic protein (pI∼10), whereas the soluble redox proteins in unicellular species of cyanobacteria are generally acidic [39]. Support for Cyt c6 as a donor to ARTO came from the presence of an HQNO (2-heptyl-4-hydroxyquinoline N-oxide; an inhibitor of bd-quinol oxidases)-insensitive pathway in a ctaCDE strain of Synechocystis sp. PCC 6803, which could be blocked by deletion of the gene for Cyt c6 [40]. In addition, a mutant of Synechocystis where both ctaCDE and cydAB were deleted, leaving only the ARTO operon intact, remained capable of energizing the cytoplasmic membrane unless the gene for Cyt c6 was also deleted [41]. These results indicate a Cyt c6-supported COX activity for the ARTO; however, it has been suggested that the result with HQNO may have been due to incomplete inhibition of the bd-quinol oxidase, and that ARTO operates as a bo-quinol oxidase [30]. If so, it is likely to interact with the same redox carriers as the bd-quinol oxidase.

Physiology

Energy metabolism

COX

While COX is not necessary for photoautotrophic growth [30], it does appear to be needed for chemoheterotrophic growth. Furthermore, chemoheterotrophic growth of Nostoc MAC led to increased levels of Cyt c6 in the periplasm and increased activity of COX [42]. Indeed, the cox genes appear to be essential to energy metabolism, since they cannot be deleted from a Photosystem I-less (psaAB) strain [43].

The bd-quinol oxidase

The existence of an HQNO- or pentachlorophenol-sensitive respiratory pathway in a COX-less strain of Synechocystis sp. PCC 6803 led to the conclusion that the bd-quinol oxidase could contribute to energy metabolism, at least when COX was not active [41]. The bd-quinol oxidases are associated with a lower coupling efficiency. Thus they may have a physiological role to remove reducing equivalents e.g. under high-light conditions, when COX alone cannot prevent overreduction of the plastoquinone pool [33].

On the other hand, Km estimates for the oxygen affinity of COX and bd-quinol oxidase of Synechocystis sp. PCC 6803 were 1.0 and 0.35 μM respectively [41]. Therefore the bd-quinol oxidase may operate under conditions of low oxygen tension.

ARTO

Two previous studies using deletion mutants of Synechocystis sp. PCC 6803 and inhibitors proposed that ARTO was active and could contribute to energy metabolism [40,41]. Low levels of residual oxygen consumption in a cydABctaCDE strain of Synechocystis sp. PCC 6803 could be inhibited by KCN, suggesting a functional ARTO [41]; however, even low levels of respiration could not be observed in an earlier study using the same mutant strain [30].

However, the deletion of the ARTO genes had no further impact on the fluorescence quenching of a COX-less strain of Synechocystis sp. PCC 6803 [33]. The presence of a putative FNR binding site (a transcriptional regulator of the switch from aerobic to anaerobic growth in E. coli) upstream of the ARTO operon may indicate that the ARTO complex functions under anaerobic or micro-aerobic conditions [30].

Nitrogen fixation

Nitrogen fixation is associated with high respiratory rates, reflecting the high energetic costs of the process and also the need to minimize oxygen levels within the heterocysts to prevent damage to nitrogenase (reviewed in [44]).

COX

Two COX operons have been found in the cyanobacteria capable of diazotrophic growth. One (COX1) is expressed constitutively, whereas the other (COX2) is specifically up-regulated during heterocyst differentiation in Anabaena sp. PCC 7120 [9,36] and A. variabilis [31]. A potential binding site for NtcA (a global regulator of the genes involved in nitrogen fixation) could be identified upstream of the COX2 operon in A. variabilis [31]. Indications are that COX2 dominates under conditions of nitrogen fixation [31]. However, a COX2 deletion mutant of Anabaena sp. PCC 7120 was still able to fix nitrogen [9], suggesting that other terminal respiratory oxidases are able to compensate for its loss.

ARTO

ARTO (also named COX3) transcripts were exclusively detected in the heterocysts in Anabaena sp. PCC 7120; expression increased in parallel with COX2 [36]. This up-regulation of ARTO was lost in ntcA and hetR mutants (genes involved in regulating the development of nitrogen fixation). Although only low level transcription and no up-regulation of ARTO was seen in A. variabilis, there was some evidence that activity of the ARTO increases in A. variabilis grown on dinitrogen compared with the cyanobacteria grown with nitrate [31]. Mutants of Anabaena sp. PCC 7120 in which either COX2 or ARTO was deleted were able to grow diazotrophically; however, deletion of both resulted in almost zero growth under N2-fixing conditions [36].

However, the ARTOs present in non-N2-fixing bacteria, e.g. Synechocystis sp. PCC 6803 [29], if functional, must have a different physiological role.

Energization of the cytoplasmic membrane and salt stress

Although there are components of the photosystems in the cytoplasmic membrane, it is not photosynthetically active [45,46]. Therefore the respiratory electron-transfer chain is chiefly responsible for the energization of the cytoplasmic membrane (ATPases located in the cytoplasmic membrane are also implicated [13]), and energization is KCN-sensitive [41].

Growth of Synechococcus sp. PCC 6311 at increased salt concentrations led to an increase in the rate of respiration (‘salt respiration’) and the amount of COX at the cytoplasmic (but not thylakoid) membrane [2]. The increased respiration is proposed to generate an H+ gradient required to drive H+/Na+ exchange. H+/Na+ antiporters have been identified in Synechocystis sp. PCC 6803 [47]. COX was shown to be necessary and sufficient for ‘salt respiration’ [41].

Conclusions

Whereas the localization and activity of the COX are fairly well understood, the activities and roles of the bd-quinol oxidase and ARTO complex remain in dispute. This is due in part to ambiguities in the interpretation of inhibitor studies. For example, pentachlorophenol has been interpreted as an inhibitor of bd-quinol oxidase and COX, as well as an uncoupler [31,41,48]. Similarly, HQNO inhibits both bd- and bo-quinol oxidases, but has been used in studies to determine the activity of the ARTO complex [31], requiring the assumption that ARTO operates as a COX rather than a bo-quinol oxidase. Ensuring adequate levels of an inhibitor can also be problematic (e.g. [40] versus [30]).

Interpretation of mutant studies can be difficult, as it is often unclear whether effects are direct or indirect, perhaps as a result of stress (e.g. [22,23]). Furthermore, results of genetic experiments can produce quantitatively very different results from inhibitor experiments. For example, the apparent respiratory activity remaining in a COX-less strain of Synechocystis inhibited with HQNO was nearly 30%, whereas residual activity in COX Cyd double deletion was only 2% [41]. To explain this, the possibility of an uncharacterized gene regulatory system is proposed; however, this discrepancy could be due to incomplete inhibition.

Certainly, one major obstacle to accepting ARTO and CydAB as functional terminal oxidases is the absence of their characterization at the protein level. The greatest need is for a direct method to assay the relative contributions of the three oxidases in vivo [31]; however, the low transcript levels (e.g. [30,31]) suggest that any characterization at the protein level may remain problematic.

Overall, branched electron-transfer chains are common in prokaryotes [49]. Thus the idea that multiple oxidases should operate downstream of plastoquinol in cyanobacteria is not surprising. Branched electron-transfer chains would allow cyanobacteria to meet their energetic demands under a variety of environmental conditions and differentiation states, e.g. nitrogen fixation. While sequencing projects have proved to be invaluable tools for revealing the presence of these genes, characterization at the protein level is now crucial to understand fully the roles of the cyanobacterial terminal oxidases.

Mechanisms of Bioenergetic Membrane Proteins: Structures and Beyond: First Joint German/British Bioenergetics Meeting, held at Wilhelm-Kempf Haus, Wiesbaden, Germany, 20–24 March 2005. Organized by F. MacMillan (Frankfurt), B. Ludwig (Frankfurt), U. Brandt (Frankfurt), H. Michel (Frankfurt), P. Rich (London), M. Jones (Bristol) and J. Walker (Cambridge). Edited by F. MacMillan, P. Rich and M. Jones.

Abbreviations

     
  • ARTO

    alternative respiratory terminal oxidase

  •  
  • bd-quinol oxidase

    cytochrome bd-quinol oxidase

  •  
  • bo-quinol oxidase

    cytochrome bo-quinol oxidase

  •  
  • COX

    cytochrome c oxidase

  •  
  • Cyt

    c6, cytochrome c6

  •  
  • Cyt

    cM, cytochrome cM

  •  
  • Cyt

    f, cytochrome f

  •  
  • HQNO

    2-heptyl-4-hydroxyquinoline N-oxide

  •  
  • Pc

    plastocyanin

This work was supported by the Biotechnology and Biological Sciences Research Council.

References

References
1
Wolk
C.P.
Ernst
A.
Elhai
J.
Bryant
D.A.
The Molecular Biology of Cyanobacteria
1994
Dordrecht, The Netherlands
Kluwer Academic Publishers
(pg. 
769
-
823
)
2
Fry
I.V.
Huflejt
M.
Erber
W.W.
Peschek
G.A.
Packer
L.
Arch. Biochem. Biophys.
1986
, vol. 
244
 (pg. 
686
-
691
)
3
Lockau
W.
Arch. Microbiol.
1981
, vol. 
128
 (pg. 
336
-
340
)
4
Scherer
S.
Stürzl
E.
Böger
P.
Arch. Microbiol.
1982
, vol. 
132
 (pg. 
333
-
337
)
5
Scherer
S.
Almon
H.
Böger
P.
Photosynth. Res.
1988
, vol. 
15
 (pg. 
95
-
114
)
6
Scherer
S.
Trends Biochem. Sci.
1990
, vol. 
15
 (pg. 
458
-
462
)
7
Peschek
G.A.
Peschek
G.A.
Löffelhardt
W.
Schmetterer
G.
The Phototrophic Prokaryotes
1999
New York
Kluwer/Plenum Publishers
(pg. 
201
-
209
)
8
Peschek
G.A.
Obinger
C.
Paumann
M.
Physiol. Plant.
2004
, vol. 
120
 (pg. 
358
-
369
)
9
Jones
K.M.
Haselkorn
R.
J. Bacteriol.
2002
, vol. 
184
 (pg. 
2491
-
2499
)
10
Peschek
G.A.
Wastyn
M.
Trnka
M.
Molitor
V.
Fry
I.V.
Packer
L.
Biochemistry
1989
, vol. 
28
 (pg. 
3057
-
3063
)
11
Peschek
G.A.
Biochim. Biophys. Acta
1996
, vol. 
1275
 (pg. 
27
-
32
)
12
Peschek
G.A.
Molitor
V.
Trnka
M.
Wastyn
M.
Erber
W.W.
Methods Enzymol.
1988
, vol. 
167
 (pg. 
437
-
449
)
13
Peschek
G.A.
Obinger
C.
Fromwald
S.
Bergman
B.
FEMS Microbiol. Lett.
1994
, vol. 
124
 (pg. 
431
-
438
)
14
Kienzl
P.F.
Peschek
G.A.
Plant Physiol.
1982
, vol. 
69
 (pg. 
580
-
584
)
15
Stürzl
E.
Scherer
S.
Böger
P.
Photosynth. Res.
1982
, vol. 
3
 (pg. 
191
-
201
)
16
Binder
A.
Hauser
R.
Krogmann
D.W.
Biochim. Biophys. Acta
1984
, vol. 
765
 (pg. 
241
-
246
)
17
Alpes
I.
Stürzl
E.
Scherer
S.
Böger
P.
Z. Naturforsch.
1984
, vol. 
36c
 (pg. 
623
-
627
)
18
Paumann
M.
Bernroitner
M.
Lubura
B.
Peer
M.
Jakopitsch
C.
Furtmuller
P.G.
Peschek
G.A.
Obinger
C.
FEMS Microbiol. Lett.
2004
, vol. 
239
 (pg. 
301
-
307
)
19
Paumann
M.
Feichtinger
M.
Bernroitner
M.
Goldfuhs
J.
Jakopitsch
C.
Furtmuller
P.G.
Regelsberger
G.
Peschek
G.A.
Obinger
C.
FEBS Lett.
2004
, vol. 
576
 (pg. 
101
-
106
)
20
Paumann
M.
Lubura
B.
Regelsberger
G.
Feichtinger
M.
Kollensberger
G.
Jakopitsch
C.
Furtmuller
P.G.
Peschek
G.A.
Obinger
C.
J. Biol. Chem.
2004
, vol. 
279
 (pg. 
10293
-
10303
)
21
Zhang
L.
Pakrasi
H.B.
Whitmarsh
J.
J. Biol. Chem.
1994
, vol. 
269
 (pg. 
5036
-
5042
)
22
Malakhov
M.P.
Wada
H.
Los
D.A.
Semenenko
V.E.
Murata
N.
J. Plant Physiol.
1994
, vol. 
144
 (pg. 
259
-
264
)
23
Malakhov
M.P.
Malakhova
O.A.
Murata
N.
FEBS Lett.
1999
, vol. 
444
 (pg. 
281
-
284
)
24
Molina-Heredia
F.P.
Balme
A.
Hervas
M.
Navarro
J.A.
De la Rosa
M.A.
FEBS Lett.
2002
, vol. 
517
 (pg. 
50
-
54
)
25
Duran
R.V.
Hervas
M.
De La Rosa
M.A.
Navarro
J.A.
J. Biol. Chem.
2004
, vol. 
279
 (pg. 
7229
-
7233
)
26
Manna
P.
Vermaas
W.
Plant Mol. Biol.
1997
, vol. 
35
 (pg. 
407
-
416
)
27
Cho
Y.S.
Pakrasi
H.B.
Whitmarsh
J.
Eur. J. Biochem.
2000
, vol. 
267
 (pg. 
1068
-
1074
)
28
Schmetterer
G.
Alge
D.
Gregor
W.
Photosynth. Res.
1994
, vol. 
42
 (pg. 
43
-
50
)
29
Kaneko
T.
Sato
S.
Kotani
H.
Tanaka
A.
Asamizu
E.
Nakamura
Y.
Miyajima
N.
Hirosawa
M.
Sugiura
M.
Sasamoto
S.
, et al. 
DNA Res.
1996
, vol. 
3
 (pg. 
109
-
136
)
30
Howitt
C.A.
Vermaas
W.F.J.
Biochemistry
1998
, vol. 
37
 (pg. 
17944
-
17951
)
31
Pils
D.
Wilken
C.
Valladares
A.
Flores
E.
Schmetterer
G.
Biochim. Biophys. Acta
2004
, vol. 
1659
 (pg. 
32
-
45
)
32
Schneider
D.
Berry
S.
Rich
P.
Seidler
A.
Rogner
M.
J. Biol. Chem.
2001
, vol. 
276
 (pg. 
16780
-
16785
)
33
Berry
S.
Schneider
D.
Vermaas
W.F.
Rogner
M.
Biochemistry
2002
, vol. 
41
 (pg. 
3422
-
3429
)
34
Peschek
G.A.
Wastyn
M.
Fromwald
S.
Mayer
B.
FEBS Lett.
1995
, vol. 
371
 (pg. 
89
-
93
)
35
Fromwald
S.
Zoder
R.
Wastyn
M.
Lubben
M.
Peschek
G.A.
Arch. Biochem. Biophys.
1999
, vol. 
367
 (pg. 
122
-
128
)
36
Valladares
A.
Herrero
A.
Pils
D.
Schmetterer
G.
Flores
E.
Mol. Microbiol.
2003
, vol. 
47
 (pg. 
1239
-
1249
)
37
Nakamura
Y.
Kaneko
T.
Sato
S.
Ikeuchi
M.
Katoh
H.
Sasamoto
S.
Watanabe
A.
Iriguchi
M.
Kawashima
K.
Kimura
T.
, et al. 
DNA Res.
2002
, vol. 
9
 (pg. 
123
-
130
)
38
Nakamura
Y.
Kaneko
T.
Sato
S.
Mimuro
M.
Miyashita
H.
Tsuchiya
T.
Sasamoto
S.
Watanabe
A.
Kawashima
K.
Kishida
Y.
, et al. 
DNA Res.
2003
, vol. 
10
 (pg. 
137
-
145
)
39
Kerfeld
C.A.
Krogmann
D.W.
Annu. Rev. Plant Physiol. Plant Mol. Biol.
1998
, vol. 
49
 (pg. 
397
-
425
)
40
Pils
D.
Gregor
W.
Schmetterer
G.
FEMS Microbiol. Lett.
1997
, vol. 
152
 (pg. 
83
-
88
)
41
Pils
D.
Schmetterer
G.
FEMS Microbiol. Lett.
2001
, vol. 
203
 (pg. 
217
-
222
)
42
Obinger
C.
Knepper
J.-C.
Zimmermann
U.
Peschek
G.A.
Biochem. Biophys. Res. Commun.
1990
, vol. 
169
 (pg. 
492
-
501
)
43
Vermaas
W.F.J.
Shen
G.
Styring
S.
FEBS Lett.
1994
, vol. 
337
 (pg. 
103
-
108
)
44
Fay
P.
Microbiol. Rev.
1992
, vol. 
56
 (pg. 
340
-
373
)
45
Smith
D.
Bendall
D.S.
Howe
C.J.
Mol. Microbiol.
1992
, vol. 
6
 (pg. 
1821
-
1827
)
46
Smith
D.
Howe
C.J.
FEMS Microbiol. Lett.
1993
, vol. 
110
 (pg. 
341
-
347
)
47
Inaba
M.
Sakamoto
A.
Murata
N.
J. Bacteriol.
2001
, vol. 
183
 (pg. 
1376
-
1384
)
48
Barstad
A.W.
Peyton
D.H.
Smejtek
P.
Biochim. Biophys. Acta
1993
, vol. 
1140
 (pg. 
262
-
270
)
49
Poole
R.K.
Cook
G.M.
Adv. Microb. Physiol.
2000
, vol. 
43
 (pg. 
165
-
224
)