Different residues in periplasmic domains of the CcmC inner membrane protein of Pseudomonas fluorescens ATCC 17400 are critical for cytochrome c biogenesis and pyoverdine-mediated iron uptake


  • Ahmed Gaballa,

    1. Department of Immunology, Parasitology, and Ultrastructure, Flanders Interuniversity Institute for Biotechnology, Vrije Universiteit Brussels, Belgium.,
    2. Laboratory of Microbial Interactions, Vrije Universiteit Brussel, Paardenstraat 65, B-1640 Sint Genesius Rode, Belgium.,
    Search for more papers by this author
  • Christine Baysse,

    1. Department of Immunology, Parasitology, and Ultrastructure, Flanders Interuniversity Institute for Biotechnology, Vrije Universiteit Brussels, Belgium.,
    Search for more papers by this author
  • Nico Koedam,

    1. Laboratory of Microbial Interactions, Vrije Universiteit Brussel, Paardenstraat 65, B-1640 Sint Genesius Rode, Belgium.,
    Search for more papers by this author
  • Serge Muyldermans,

    1. Department of Immunology, Parasitology, and Ultrastructure, Flanders Interuniversity Institute for Biotechnology, Vrije Universiteit Brussels, Belgium.,
    Search for more papers by this author
  • Pierre Cornelis

    1. Department of Immunology, Parasitology, and Ultrastructure, Flanders Interuniversity Institute for Biotechnology, Vrije Universiteit Brussels, Belgium.,
    Search for more papers by this author

Pierre Cornelis. E-mail PCORNEL@vub.ac.be; Tel. (2) 359 02 21; Fax (2) 359 03 99.


The inner membrane protein CcmC (CytA) of Pseudomonas fluorescens ATCC17400, which has homologues in several bacteria and plant mitochondria, is needed for the biogenesis of cytochrome c. A CcmC-deficient mutant is also compromised in the production and utilization of pyoverdine, the high-affinity fluorescent siderophore. A topological model for CcmC, based on the analysis of alkaline phosphatase fusions, predicts six membrane-spanning regions with three periplasmic loops. Site-directed mutagenesis was used in order to assess the importance of some periplasm-exposed residues, conserved in all CcmC homologues, for cytochrome c biogenesis, and pyoverdine production/utilization. Despite the conservation of the residues His-61, Val-62 and Pro-63 in the first periplasmic loop, and Leu-184, His-185 and Gln-186 in the third periplasmic loop, their simultaneous replacement with Ala only partially affected cytochrome c biogenesis and pyoverdine production/utilization. Simultaneous replacements of residues Trp-115 and Gly-116 in the second periplasmic loop substantially affected pyoverdine production/utilization but not cytochrome c production. An Ala substitution of Asp-127, in the second periplasmic loop, resulted in decreased production of cytochrome c, slower growth in conditions of anaerobiosis and reduced pyoverdine production. On the other hand, a mutation in Trp-126, also in the second periplasmic loop, totally suppressed the production of cytochrome c, whereas it had no effect on the production and utilization of pyoverdine. These results show a differential involvement of amino acid residues in periplasmic domains of CcmC in cytochrome c biogenesis and pyoverdine production/utilization.


Under conditions of iron limitation, fluorescent pseudomonads (Pseudomonas aeruginosa, P. fluorescens, P. putida, P. syringae, P. chlororaphis) produce and excrete a fluorescent siderophore, named pyoverdine or pseudobactin, characterized by the presence of a chromophore part (a quinoline derivative) linked to a peptide arm (6–12 amino acids) (for reviews, Budzikiewicz, 1993; 1997). The chromophore is well conserved among different representatives of fluorescent pseudomonads, whereas the peptide arm is variable, resulting in a high specificity in the outer membrane receptor–siderophore interaction, which forms the first step in the pyoverdine-mediated high-affinity iron uptake (Hohnadel and Meyer, 1988). Several genes were reported to be involved in the biosynthesis of the fluorescent siderophore pyoverdine in P. aeruginosa, including pvdA which codes for an L-ornithine N5-oxygenase (Visca et al., 1994), pvdD encoding a peptide synthase (Merriman et al., 1995) and pvdE, the product of which is an ABC transporter (Mc Morran et al., 1996). A biosynthetic pathway for the pyoverdine chromophore has been proposed (Stintzi et al., 1996), and the corresponding genes have recently been sequenced (Stintzi, 1997).

P. fluorescens ATCC17400 produces a mixture of pyoverdines that differ not in their peptide arm but in the side chain-bound to the chromophore (Demange et al., 1990). Two main isoelectric forms of pyoverdine of P. fluorescens ATCC 17400 could be distinguished using isoelectric focusing (IEF) and their iron-binding property confirmed by chrome-azurol S overlay (CAS) (Koedam et al., 1994).

Recently, we have shown that a kanamycin cassette disruption of the ccmC gene (but not the downstream genes ccmD and ccmE ) from P. fluorescens ATCC 17400 severely compromised pyoverdine production (Gaballa et al., 1996). The isoelectric-focusing pattern of the mutant's pyoverdine was also altered because of the excretion of supplementary fluorescent as well as non-fluorescent iron-binding molecules with different isoelectric points (Gaballa et al., 1996). The ccmC gene is the first of eight open reading frames (ORFs), the deduced products of which showed sequence similarity to proteins involved in cytochrome c biogenesis in different bacteria (for recent reviews, Thöny-Meyer, 1997; Zumft, 1997; Page et al., 1998). Indeed, the same non-polar disruption of ccmC also abolished the production of cytochrome c that is essential for the growth under anaerobic conditions (Gaballa et al., 1996). CcmC is an inner membrane protein and homologues are found in different bacteria such as CcmC in E. coli (Thöny-Meyer et al., 1995), HelC in Rhodobacter capsulatus (Beckman et al., 1992), ORF263 of Bradyrhizobium japonicum (Ritz et al., 1993), CcmC of Paracoccus denitrificans (Page et al., 1997), CcmC of Haemophilus influenzae (Fleischmann et al., 1995) and another strain of P. fluorescens (Yang et al., 1996). ORFs coding for homologues of CcmC have also been described in mitochondrial genomes of wheat (Bonnard and Grienenberger, 1995) and Oenothera (Jekabsons and Schuster, 1995).

These inner membrane proteins form what could be a typical membrane component of an ABC-type transporter (Fath and Kolter, 1993; Goldman et al., 1997; Page et al., 1997; Thöny-Meyer, 1997). The tryptophan-rich sequence, W–G–X–X–W–X–W–D, present in all homologues of CcmC, is located in the periplasm (Gaballa et al., 1996) and is considered to form a haem binding site (Thöny-Meyer et al., 1994; Thöny-Meyer, 1997). As a first step in understanding the apparently divergent roles of ccmC in pyoverdine production and cytochrome c biogenesis, a topological model for CcmC was established through alkaline phosphatase fusions (Manoil, 1991). Further, the importance of different conserved amino acid residues, in periplasmic loops, in cytochrome c production or in pyoverdine production/utilization, was investigated through site-directed mutagenesis.


Topology of the CcmC protein

To determine the topology of CcmC we first used λ TnphoA to obtain random insertions in clone pPYOV35, containing the entire ccmC and part of ccmB. The combination of a primer in phoA and the reverse universal primer upstream of the 5′ end of ccmC allowed direct detection and localization by PCR of the Tn insertions in ccmC. More than 100 clones with a Tn5phoA insertion were further subjected to DNA sequencing. As only 12 insertions, at four different locations, corresponded to in-frame fusions (at Leu-32, Leu-89, Trp-180 and Val-229), other PhoA fusions were constructed using a PCR amplification approach: a PCR-amplified fragment corresponding to the 5′ region of the ccmC gene (from the non-coding upstream sequences down to the selected codon) was cloned in frame with the coding region of the mature PhoA protein. Three PhoA fusions in hydrophilic parts of CcmC were obtained using this approach, after Ser-54, Val-125 and Glu-155. The resulting ccmC-phoA fusions were mobilized to P. fluorescens ATCC 17400, and the transconjugants were plated on CAA medium containing X-phosphate for the determination of the phosphatase activity.

Clones Leu-32/PhoA, Ser-54/PhoA, Val-125/PhoA and Trp-180/PhoA (Fig. 1) conferred alkaline phosphatase activity to the corresponding transconjugants as indicated by the dark-blue colour of the colonies, whereas transconjugants of clones Leu-89/PhoA, Glu-155/PhoA and Val-229/PhoA remained white.

Figure 1.

. Topological model of the CcmC protein; +, indicates positively charged residues. The flags represent the positions of the different in-frame PhoA insertions (black for alkaline phosphatase-positive, white for negative). The circled residues are those conserved in all CcmC homologues (Gaballa et al., 1996).

Using the ‘positive inside’ rule (von Heijne, 1992) and the prediction of hydrophobic membrane helixes (Jones et al., 1994), the potential membrane-spanning segments were predicted. By combining the results of the alkaline phosphatase fusions and the predictions of the membrane-spanning segments, a topological model for CcmC was constructed (Fig. 1).

Site-directed mutagenesis of residues in periplasmic loops of CcmC

The residues that are conserved in CcmC and the homologous proteins from bacteria and plant mitochondria (Gaballa et al., 1996; Page et al., 1997) are clustered in the three periplasmic loops and in membrane helixes three and four (Fig. 1). To assess the importance of some of these conserved residues, especially in the tryptophan-rich motif, site-directed mutagenesis was carried out (Table 1). All ccmC mutant clones were fully sequenced, further subcloned in the vector pBBR1MCS and mobilized to a ccmC knock-out mutant (Gaballa et al., 1996). The resulting transconjugants were grown in iron-limited CAA for 36 h, their pyoverdines quantified, partially purified and separated by IEF in order to look for qualitative differences in pyoverdine isotypes (Koedam et al., 1994; Gaballa et al., 1996) (Fig. 2). The different c-type cytochromes in the total membrane, or in the soluble fraction, were detected by specific staining of the proteins covalently binding haem, after SDS–PAGE (Goodhew et al., 1986) (Fig. 3).

Table 1. . Codon changes of ccmC substitution mutants. The amino acid residues and the corresponding codon substitutions that were introduced in the mutagenic oligonucleotides are listed. Positions with degeneracy are shown in brackets.Thumbnail image of
Figure 2.

. IEF separation of C18-concentrated pyoverdines from 24 h cultures in CAA on PAG plates (Pharmacia) and detection by UV (top gel, for pyoverdine fluorescence) or by CAS overlay (for siderophore activity). Lane 1, wild-type pyoverdine; lane 2, ccmC knock-out mutant; lane 3, mutant D127 → A; lane 4, mutant W126 → I; lane 5, mutant H61 → A, P62 → A, V63 → A; lane 6, mutant L161 → A, H162 → A, Q163 → A; lane 7, mutant W115 → G, G116 → E.

Figure 3.

. SDS–PAGE (12%), followed by haem staining of total membrane proteins. Lane 1, protein markers; lane 2, beef hart cytochrome c (Sigma); lane 3, mutant W115 → G, G116 → E; lane 4, mutant L184 → A, H185A, Q186  → A; lane 5, mutant W126 → I; lane 6, mutant D127 → A; lane 7, knock-out mutant for ccmC ; lane 8, mutant H61 → A, P62 → A, V63 → A, lane 9, wild-type.

Growth in the absence of oxygen, and in the presence of nitrate as a terminal electron acceptor was also monitored (Fig. 4). Finally, the complementation of the ccmC knock-out for pyoverdine utilization was investigated by comparing the growth of different transconjugants (ccmC knock-out mutant with the different ccmC mutated genes in trans) in the presence of EDDHA, and wild-type pyoverdine (in order to compensate for variable defects in pyoverdine production) (Fig. 5).

Figure 4.

. Growth in condition of anaerobiosis with nitrate as a terminal electron acceptor. The bars from left to right represent OD measure- ments made after 26, 38, 135, 158 and 180 h of growth, respectively, for the wild-type (1), mutant ccmC (2), D127 → A (3), W-126 → I (4), H61 → AV62 → AP63 → A (5), L184 → AH185 → AQ186  → A (6) and W115 → GG116 → E (7). The culture was inoculated at an OD600 of 0.2. The result of only one experiment is shown (out of three).

Figure 5.

. Growth of wild-type (○), ccmC knock-out mutant (•), W126 → I (◊), D127 → A (▪), H61 → AP62 → AV63 → A (▵), L184 → AH185A → Q186A (□) and W115 → GG116 → E (+) in CAA medium containing 0.5 mg ml−1 EDDHA and 50 μM purified wild-type pyoverdine. The growth curves were realized in a Bio-Screen apparatus and each value represents the average of three experimental values.

The conserved residues His-61, Val-62, Pro-63 in the first periplasmic loop were replaced by a triad of Ala residues. These substitutions in the first periplasmic loop did not affect cytochrome c production as judged by haem staining after SDS–PAGE (Fig. 3, only the membrane-associated fraction is shown) but resulted in a delayed growth in the absence of oxygen and in the presence of nitrate, as compared with the wild type (Fig. 4). The same modified CcmC protein conferred production of pyoverdines to almost wild-type levels (results not shown), with a normal IEF pattern (Fig. 2) and also complemented the ccmC knock-out mutant for growth in the presence of EDDHA and pyoverdine (Fig. 5). The simultaneous substitutions, in the third periplasmic loop of residues Leu-184, His-185 and Gln-186 by alanines, resulted in halving the level of pyoverdine production (results not shown), which was correlated with a delayed growth in the presence of EDDHA and wild-type pyoverdine (Fig. 5), whereas the IEF pattern was unaltered (Fig. 2). The production of c-type cytochromes was also apparently unaffected in this mutant, except for a longer lag phase for the growth in anaerobic conditions (Fig. 4).

Simultaneous substitutions at residues Trp-115 and Gly-116 (Trp-115 → Gly, Gly-116 → Glu), in the second periplasmic domain, resulted in a severe reduction in pyoverdine production, accompanied by an alteration of the IEF pattern (Fig. 2), a phenotype almost comparable to the ccmC knock-out mutant (Gaballa et al., 1996). Likewise, these mutations also caused a strongly delayed growth in the presence of EDDHA and pyoverdine, comparable to what is observed for the ccmC mutant (Fig. 5). Substitution of Trp-126, which is a conserved residue of the tryptophan-rich motif, by isoleucine did not affect pyoverdine production, neither quantitatively nor qualitatively, and did not have any influence on the growth in the presence of EDDHA and pyoverdine (Fig. 5). However, the mutated gene was unable to restore normal cytochrome c production in the ccmC mutant (Fig. 3), resulting in a total absence of growth in conditions of anaerobiosis (Fig. 4). The same Trp-126 residue was replaced by different amino acids by introducing degeneracy in the primers used for the site-directed mutagenesis (Table 1). Mutants Trp-126 → Phe, Trp-126 → Tyr, Trp-126 → Thr, Trp-126 → Ser, Trp-126 → Ile, Trp-126 → Leu and Trp-126 → Lys were obtained. With the exception of Trp-126 → Phe, which conferred a wild-type phenotype, all other substitutions conferred the same phenotype as mutant Trp-126  → Ala (no cytochrome c, no growth in anaerobiosis, normal pyoverdine production, results not shown).

Finally, when Asp-127 was replaced by Ala, Thr, Ile, Asn or Lys, cytochrome c species were greatly reduced (Fig. 3), resulting in a very long lag phase (5 days) before growth could be observed in conditions of anaerobiosis (Fig. 4, results shown for the Ala substitution only). The same Asp-127 mutants also produced less pyoverdine than the wild-type, resulting in a delayed growth in the presence of EDDHA (Fig. 5). However, this mutation did not affect the IEF pattern of pyoverdines (Fig. 2).


In the biosynthesis of cytochrome c, the apocytochrome is translocated to the periplasmic space via the sec system, whereas it has been suggested that haem could be transported by an ABC transporter consisting of an ATP-binding protein, CcmA and two integral membrane proteins, CcmB and CcmC (Fath and Kolter, 1993; Goldman et al., 1997; Thöny-Meyer, 1997; Page et al., 1998). CcmC and its homologues contain a tryptophan-rich motif, which has been proposed to form a haem-binding site (Thöny-Meyer et al., 1994). In P. fluorescens ATCC 17400, CcmC (CytA), was found to be involved not only in cytochrome c biogenesis, but also to contribute to pyoverdine production and to its utilization (Gaballa et al., 1996). A topological model for CcmC has already been proposed, based on the prediction of six putative membrane-spanning helices (Page et al., 1997; Thöny-Meyer, 1997). Our model, confirmed experimentally by alkaline phosphatase fusions in the hydrophilic loops of the protein (Fig. 1), is in agreement with this prediction and corroborates our previous result indicating a periplasmic localization for the tryptophan-rich motif (Gaballa et al., 1996). Furthermore, a cytoplasmic localization for the C-terminal end is predicted, which is in agreement with the finding that the C-terminal part of the R. capsulatus HelC is located in the cytoplasm (Beckman and Kranz, cited in Bonnard and Grienenberger, 1995). Interestingly, most conserved residues in the different CcmC homologues are situated in the periplasmic domains of the protein, although some are found in the membrane helixes three and four (Fig. 1). In the study of Page et al. (1997) the CcmC proteins from Escherichia coli, P. denitrificans, P. fluorescens, H. influenzae, and HelC from R. capsulatus were predicted to be integral membrane components of bacterial periplasmic-binding protein-dependent transporters because they were found to contain some elements of a conserved motif, supposed to interact with the ATPase component of a typical ABC transporter (Dassa and Hofnung, 1985; Köster and Böhm, 1992; Page et al., 1997).

In cytochromes c, haem is covalently bound to the site C–X–X–C–H (Thöny-Meyer et al., 1994; Sambongi et al., 1996). Other haem-binding motifs remain speculative such as W–G–X–X–W–X–W–D in CcmC and its homologues, and R–C–P–V–C–Q in Ccl2 of R. capsulatus (Thöny-Meyer et al. 1994). In this study, the relative importance of conserved periplasmic residues of CcmC for cytochromes c and pyoverdine biogenesis was investigated by site-directed mutagenesis.

Mutant Trp-115 → Gly, Gly-116 → Glu was severely defective in pyoverdine production/utilization while being only marginally affected in cytochrome c biogenesis. Conversely, substitution of Trp-126 with Ile (or other residues, with the exception of Phe) totally abolished cytochromes c production, although it had a very limited effect on pyoverdine production and utilization. Interestingly, Trp-126 could only be replaced by Phe, suggesting that the presence of an aromatic amino acid side-chain at this position could be involved in the formation of a hydrophobic pocket where haem can fit (Thöny-Meyer, 1997). A ccmC mutant in which Asp-127 was replaced by Ala (or other residues, such as C, I, K, N or T, results not shown) showed both a decrease in pyoverdine and cytochrome c production. Interestingly, a mutant with a glutamic acid residue at position 127 had a wild-type phenotype, suggesting that the presence of a negatively charged amino acid at this position is important for the function of the protein (results not shown). Although our results confirm, experimentally, for the first time, the involvement of these two conserved residues (Trp-126, Asp-127) in cytochrome c biogenesis, they do not yet, as such, corroborate or contradict the suggestion that they are part of a putative haem-binding motif (Thöny-Meyer et al., 1994). It has been suggested recently (Page et al., 1997), and later experimentally demonstrated (Goldman et al., 1997), that CcmC interacts with CcmA and CcmB to form an ABC transporter that is proposed to be involved in the transport of haem or another, yet unidentified, component (Page et al., 1997; Page et al., 1998). Interestingly, in P. denitrificans, a disruption of ccmF, which is predicted to have 11 transmembrane helices and also to posses the tryptophan-rich motif, similarly resulted in decreased siderophore production (Pearce et al., 1998).

In a possible model, the CcmC protein could be a haem-binding transport protein, associated with CcmA and CcmB, and not (or not only) a haem exporter. In aerobic conditions, it would be involved in a step of pyoverdine maturation or in the reduction of ferripyoverdine, whereas, under conditions of anaerobiosis, it would participate in the biosynthesis of c-type cytochromes. We have indeed observed that c-type cytochromes are produced only under conditions of anaerobiosis while good aeration is needed for pyoverdine production (A. Gaballa et al., unpublished). The strongly reduced capacity of a ccmC mutant to use pyoverdine, although it has been demonstrated to take up the siderophore in uptake assays (Gaballa et al., 1996), suggests indeed that the defect in the utilization of pyoverdine resides in a step after the entry of the ferrisiderophore in the periplasm via its receptor. Our results would suggest a dual role for CcmC: a major one in the biogenesis of cytochromes c (as exemplified by the importance of residues W126 and D127), and one, maybe indirect, for the production/utilization of pyoverdine (as exemplified by mutation W115 → GG116 → E, which more particularly affects pyoverdine production and utilization). In this regard, it is interesting to mention that another protein, CcmG in E. coli, has recently been show to also have a dual function: one as a periplasmic disulphide reductase, and another as a polypeptide needed to stabilize other proteins involved in the biogenesis of cytochrome c (Fabianek et al., 1998).

In conclusion, this research indicates that all three conserved loops of CcmC are periplasmic and demonstrates, for the first time, the involvement of different amino acid residues of CcmC, especially in the tryptophan-rich motif, in haem and/or pyoverdine biogenesis/utilization.

Experimental procedures

Bacterial strains, vectors and growth conditions

P. fluorescens 17400 was maintained in casamino acids (CAA) medium (Cornelis et al., 1992). Unless otherwise indicated, 50 ml cultures were inoculated from an overnight preculture and incubated at 28°C at 200 r.p.m. (New-Brunswick Innova shaker). Growth in the presence of EDDHA (0.5 mg ml−1) with and without pyoverdine (50 μM) was carried out in a Bio-Screen apparatus (Life Technologies) using the following parameters: shaking for 30 s, every 3 min, reading every 10 min, temperature 28°C. Anaerobic growth was performed according to Pettigrew and Brown (1988) and as previously described (Gaballa et al., 1996), using an inoculum equivalent to a final OD600 = 0.2; the 10 ml cultures were performed in tube with a paraffin oil layer on top. Antibiotics were added to P. fluorescens strain 17400 derivatives at the following concentrations: kanamycin (Km) 600 μg ml−1 or chloramphenicol (Cm) 300 μg ml−1. E. coli strains were grown at 37°C in LB with the appropriate antibiotics, Km 100 μg ml−1, ampicillin (Ap) 100 μg ml−1 or Cm 25 μg ml−1. E. coli LE392 {supE hsdR galK trpR metB lacY tonA} was used for λTnphoA propagation (Gutierrez et al., 1987), whereas E. coli CC118 {araD139 Δ(ara, leu)7697 ΔlacX74 phoAΔ20 galE galK thi rpsE rpoB argE recA1} was used for the primary screening of alkaline phosphatase fusions (Gutierrez et al., 1987), E. coli DH5α (Gibco-BRL) {supE44 ΔlacU169 (φ80 lacZ(M15) hsdR17 recA1 endA1 gyrA96 thi-1 relA1} was used for regular DNA manipulations. The pBluescript SK(+) vector (Stratagene) was used for subcloning and sequencing while pBBR1MCS (Kovach et al., 1994) was used to express the ccmC or the phoA gene fusions in Pseudomonas.

DNA methodology

DNA manipulations were carried out according to standard protocols (Sambrook et al., 1989). Restriction endonucleases, T4 DNA ligase and Klenow fragment (Pharmacia) were used according to the manufacturer's instructions.

Construction of phoA fusions

A HindIII–NcoI DNA fragment from pPYOV4 (Gaballa et al., 1996) carrying the coding region of ccmC and the upstream sequence (830 bp) was cloned in HindIII–SpeI restricted pBBR1MCS or pBluescript SK(+) after treating the NcoI and the SpeI sites with Klenow fragment, yielding clones pPYOV35 and pPYOV303 respectively.

Clone pPYOV35 was used as a recipient for λTnphoA (Gutierrez et al., 1987) according to Manoil (1991). Clones were isolated from 10 independent transposon mutagenesis experiments in order to avoid the predominance of early transposition events and to obtain a wide variety of transposition events. The clones were initially analysed by PCR using primer phoa4 (5′-AGCACCGCCGGGTGCAGTAATAT-3′) in the phoA and a biotinylated reverse universal primer in the vector. The resulting PCR-amplified fragments were analysed by agarose gel electrophoresis and the junction between cytA and phoA was sequenced with fluorescent phoa4 primer using a solid-phase sequencing kit (Pharmacia) on an ALF-automated sequencer (Pharmacia). Clones with in-frame fusions were mobilized to P. fluorescens 17400 by triparental conjugation according to Cornelis et al. (1992). Transconjugants were grown in the presence of Xp and evaluated for the colony colour. ccmCphoA fusions at three different locations of the ccmC gene were carried out as previously described (Gaballa et al., 1996). Briefly, primers to amplify from S54, V125 or E155 (Fig. 3) were used in combination with the reverse universal sequencing primer to amplify the N-terminal coding region of ccmC together with the promoter region. The resulting PCR fragments were cloned in pGV4218 (T. Pattery, VUB, Belgium, personal communication) in-frame with the coding region corresponding to the mature PhoA protein. The constructs were confirmed by DNA sequencing and tested for alkaline phosphatase production in P. fluorescens 17400 as mentioned before.

Site-directed mutagenesis

Mutants in ccmC were obtained by polymerase chain reaction using the overlap extension technique described by Ho et al. (1989). For each mutation a set of two overlapping primers (Pharmacia) were used in a combination with the forward and the reverse universal sequencing primers in clone pPYOV303. Except for mutation in W115 and G116, where degeneracy in the codons was introduced in the primers, mutagenic oligonucleotides for other mutants contained the mismatches that corresponded to Ala substitutions (Table 1). The mutagenized fragments were subcloned in pBluescript SK(+) vector and confirmed by DNA dideoxy sequence analysis using a SequiTherm kit (Biozyme). Owing to the high GC content of Pseudomonas DNA, false stops of the polymerase were resolved by terminal transferase treatment as described by Gaballa et al. (1996). The mutagenized clones were further subcloned in the pBBR1MCS vector and introduced in a ccmC (cytA) mutant (Gaballa et al., 1996) by triparental conjugation and the transconjugants were analysed for their phenotypes. To ensure for the absence of any DNA polymerase misincorporation during the PCR reaction, the obtained ccmC mutants were fully sequenced.

Detection of siderophores

The concentration of pyoverdine was estimated spectrophotometrically at 403 nm using an extinction coefficient ε403 = 2 × 104 M−1 cm−1 (Höfte et al., 1993) and normalized by a biomass unit as OD600 of the culture. Measurements are averages of three independent experiments. Different forms of pyoverdine were detected with an IEF-CAS overlay technique according to Koedam et al. (1994).

SDS–PAGE and haem-staining

c-Type cytochromes were detected after SDS–PAGE separation of protein extracts and specific staining for haem (Goodhew et al., 1986). Soluble and membrane fractions were prepared according to the procedure of Cunningham and Williams (1995)).

Note added in proof

After this manuscript was submitted, a report from Goldman et al. (1998) described the topology of the HelC protein of Rhodobacter capsulatus, which corresponds perfectly to the topology of CcmC described in this work. The authors also found that the two histidinyl residues in the first and third periplasmic domains of HelC are essential for photosynthetic anaerobic growth, whereas in P. fluorescens the replacement of each of these residues by an alanine only results in a delayed growth in conditions of anaerobiosis. This discrepancy could maybe be explained by the difference between the two anaerobiosis growth assays (photosynthetic versus growth in rich medium with nitrate and high inoculum). The same authors also describe the replacement of the D residue in the W–W–D motif by an E, which does not result in a defect in cytochromes c biogenesis, in agreement with our results showing that only E, a negatively charged residue, can substitute for D in this position.


The Bio-Screen equipment was obtained thanks to a fund (‘krediet aan navorsers’) from the Fonds voor Wetenschappelijk Onderzoek (FWO). Ahmed Gaballa wishes to thank the Koedam family for their financial support during his PhD.