Membrane topology and mutational analysis of Escherichia coli CydDC, an ABC-type cysteine exporter required for cytochrome assembly

Hugo Cruz-Ramos{dagger}, Gregory M. Cook{ddagger}, Guanghui Wu§, Michael W. Cleeter|| and Robert K. Poole

Department of Molecular Biology and Biotechnology, The University of Sheffield, Firth Court, Western Bank, Sheffield S10 2TN, UK

Correspondence
Robert K. Poole
r.poole{at}sheffield.ac.uk


   ABSTRACT
TOP
ABSTRACT
INTRODUCTION
METHODS
RESULTS AND DISCUSSION
REFERENCES
 
Cytochrome bd is a respiratory quinol oxidase in Escherichia coli. Besides the structural genes (cydA and cydB) encoding the oxidase complex, the cydD and cydC genes, encoding an ABC-type transporter, are required for assembly of this oxidase. Recently, cysteine has been identified as a substrate (allocrite) that is transported from the cytoplasm by CydDC, but the mechanism of cysteine export to the periplasm and its role there remain unknown. To initiate an understanding of structure–function relationships in CydDC, its membrane topography was analysed by generating protein fusions between random and selected residues in the two polypeptides with both alkaline phosphatase and {beta}-galactosidase. CydD and CydC are experimentally shown each to have six transmembrane segments, two major cytoplasmic loops and three minor periplasmic loops; both termini of each protein face the cytoplasm. The cydD1 allele is shown to have two point mutations (G319D, G429E) within the ATP-binding domain of CydD; either mutation alone is sufficient to cause loss or severe reduction of cytochrome bd assembly. A comparative sequence analysis prompted the targeting of residues in CydD for site-directed mutational analysis, which identified (i) the ‘start’ methionine residue, (ii) essential residues in the ATP-binding site (Walker sequence A) and (iii) a duplicated positively charged heptameric motif, R-G/T-L/M-X-T/V-L-R, in CydD cytoplasmic loop II. The replacement of arginines in these motifs with glycines resulted in Cyd phenotypes; however, activity could be restored at these positions by replacing the glycine with lysine or histidine and hence returning the positive charge. The conservation of these charges in CydD-like proteins indicates functional importance. Evolutionary aspects of bacterial cyd genes are discussed.


Abbreviations: ABC, ATP-binding cassette; AP, alkaline phosphatase; BG, {beta}-galactosidase; CL, cytoplasmic loop; PL, periplasmic loop

{dagger}Present address: Equipe AVENIR-INSERM d'Immunite Anti-Microbienne des Muqueuses, E0364, Institut de Biologie de Lille, 1 rue du Professeur Calmette – BP 447 – 59021 Lille, France.

{ddagger}Present address: Department of Microbiology, Otago School of Medical Sciences, University of Otago, PO Box 56, Dunedin, New Zealand.

§Present address: Veterinary Laboratories Agency – Weybridge, New Haw, Addlestone, Surrey KT15 3ND, UK.

||Present address: Department of Clinical Neurosciences, Royal Free Hospital, Rowland Hill Street, London NW3 2QG, UK.


   INTRODUCTION
TOP
ABSTRACT
INTRODUCTION
METHODS
RESULTS AND DISCUSSION
REFERENCES
 
Escherichia coli uses two membrane-bound terminal oxidases for aerobic respiration, cytochromes bo' and bd (Gennis & Stewart, 1996). The genes encoding the cytochrome bd quinol oxidase, cydAB, are expressed both aerobically and anaerobically, but expression is maximal during aerobic stationary phase or in low oxygen environments (Cotter et al., 1990; Poole & Cook, 2000). Consistent with this expression profile is the extraordinarily high affinity of cytochrome bd for O2 (Km=3–8 nM; D'Mello et al., 1996). Cytochrome bd comprises two integral membrane polypeptides, subunit I (CydA, 58 kDa) and subunit II (CydB, 43 kDa). Subunit I contains haem b558 directly involved in ubiquinol oxidation whilst two further haems, b595 and d, in a catalytic binuclear centre are shared by both subunits (Jünemann, 1997; Borisov et al., 2001). All three haems are probably near the periplasmic side of the membrane (Osborne & Gennis, 1999). Loss of cytochrome bd causes a complex phenotype that extends beyond the lack of a functional oxidase. The Cyd phenotype in E. coli includes (i) an inability to exit aerobically from stationary phase at 37 °C (Siegele et al., 1996); (ii) sensitivity to high temperature (Delaney et al., 1992), (iii) increased sensitivity to H2O2 (Delaney et al., 1992); (iv) sensitivity to iron chelators (Cook et al., 1998); and loss of motility and increased benzylpenicillin sensitivity (Pittman et al., 2002). The need to understand cytochrome bd assembly is underlined by the importance of this oxidase in many bacteria. Thus, in Azotobacter vinelandii and Klebsiella pneumoniae, cytochrome bd deficiency severely impairs nitrogen-fixing ability under aerobic or microaerobic conditions (Kelly et al., 1990; Juty et al., 1997; Edwards et al., 2000). Furthermore, in Shigella flexneri and Brucella abortis, the Cyd phenotype includes attenuation of intracellular survival and virulence (Way et al., 1999; Endley et al., 2001).

In E. coli, two other genes, cydD and cydC, constituting an operon unlinked to cydAB, are necessary for cytochrome bd assembly (Georgiou et al., 1987; Poole et al., 1989, 1993; Bebbington & Williams, 1993; Delaney et al., 1993). These genes encode the strikingly similar, but distinct, components of an ATP-binding cassette (ABC)-type transporter (Poole et al., 1993), thought to be the first found in bacteria composed of only two non-identical subunits. ABC transporters are central to many physiological processes, from bacteria to man, being responsible for solute uptake and for secretion of toxins, drugs and substrates that range from small ions to large proteins (for a review, see Schmitt & Tampe, 2002). The genome of E. coli encodes some 80 ABC transporters (Linton & Higgins, 1998), many of which, like CydDC, are poorly characterized. It was hypothesized that the substrate (‘allocrite’, i.e. the translocated but not transformed molecule; Holland & Blight, 1999) of CydDC might be haem (Poole et al., 1993, 1994). However, the assembly of haem into heterologous apoproteins (e.g. Ascaris haemoglobin) exported to the periplasm of E. coli does not require cydC (Goldman et al., 1996a) and, critically, transport studies using inside-out membrane vesicles revealed no discernible differences between wild-type and cydD mutant strains (Cook & Poole, 2000). Intriguingly though, overexpression of CydDC results in synthesis of a novel haem pigment in anaerobically grown cells (Cook et al., 2002), suggesting a link with haem metabolism.

An important clue to the function of CydDC was the finding (Goldman et al., 1996a) that the periplasm of a cydC mutant is more oxidizing than that of a wild-type strain. This, and the observation that Cyd defects can be corrected by extracellular provision of reductants (Goldman et al., 1996b; Pittman et al., 2002), suggests that CydDC exports a reducing molecule to the periplasm for redox maintenance. Recently, we have demonstrated, using everted membrane vesicles, that CydDC exports cysteine in an ATP-dependent, uncoupler-independent manner (Pittman et al., 2002). The sensitivity of a cydD mutant to benzylpenicillin and dithiothreitol, and the loss of motility, were all reversed by exogenous cysteine. The amino acid also increases cytochrome c levels in the periplasm of a cydD mutant, but does not restore cytochrome d. Consistent with CydDC being a cysteine exporter, a cydD mutant is hypersensitive to high cysteine concentrations in the medium and accumulates higher cytoplasmic cysteine levels. CydDC overexpression confers resistance to high extracellular cysteine concentrations.

The structure of this novel cysteine transporter is unknown. In an effort to understand its organization within the cytoplasmic membrane, we have determined the topography of each of the subunits. We report here agreement between topological models for these proteins, derived largely from hydrophobicity considerations, with experimental results from protein fusion studies. We also report two mutations in the ATP-binding domain found in the cydD1 chromosomal allele, the first to be isolated (Poole et al., 1989). Site-directed mutagenesis reveals other essential residues in CydD including, in a cytoplasmic loop (CL), a duplicated positively charged heptameric domain in which the positive charge has a key functional role.


   METHODS
TOP
ABSTRACT
INTRODUCTION
METHODS
RESULTS AND DISCUSSION
REFERENCES
 
Bacterial strains, plasmids and phages.
The bacterial strains used were derivatives of E. coli K-12, namely AN2342 (F, referred to as wild-type), AN2343 (F cydD1) (Poole et al., 1989), XL-1 Blue (Stratagene), CC118 [araD139 {Delta}(ara, leu)7697 {Delta}lacX74 {Delta}phoA20 galE galK thi rpsE rpoB argEam recA1] (Manoil & Beckwith, 1985; Manoil, 1990) and TG1 [{Delta}(lac-pro), supE, thi, hsdD5/F' traD36, proAB+, lacIq, lacZDM15] (Amersham Biosciences). Plasmid pRP33 is a pBR328 derivative carrying the trxB and cydDC genes (Poole et al., 1993). Plasmid pRKP1602 originated from a 2·53 kb Csp45I deletion (ablating the trxB and cat genes) of plasmid pRP33 (Cook et al., 2002). pRKP1604 harbours most of the cydD gene and is the religated product of a 3·46 kb Eco47III deletion of pRKP1602. The protein fusion vector pRS414 contains the lacZ' gene (Simons et al., 1987). The phagemid pBluescript (Stratagene) was used for cloning and single-stranded DNA preparation. The phages {lambda}TnphoA/in and {lambda}TnlacZ/in (Manoil & Bailey, 1997) were provided by C. Manoil (University of Washington).

Construction of fusion proteins.
Insertions of TnphoA/in and TnlacZ/in were made as previously described by Manoil & Bailey (1997). Strain CC118 (Manoil, 1990), harbouring plasmid pRP33, was grown for 2–3 h (OD600=1·0) and aliquots (0·2 ml) were infected with {lambda}TnphoA/in or {lambda}TnlacZ/in at a m.o.i. of 0·1–0·3. The mixture was incubated at 37 °C for 10 min without aeration followed by the addition of 0·8 ml LB medium (Miller, 1972). Cells were aerobically grown at 37 °C for 1–2 h followed by plating of dilutions on LB supplemented with tetracycline (10 µg ml–1) and kanamycin (50 µg ml–1). The resulting colonies were then replica-plated onto LB agar containing the chromogenic substrates 5-bromo-4-chloro-3-indolyl phosphate, toluidine salt (X-P; 40 µg ml–1) or 5-bromo-4-chloro-3-indolyl {beta}-D-galactoside (X-G; 40 µg ml–1), tetracycline, and an increased concentration of kanamycin (300 µg ml–1) to select for transposition of TnphoA/in or TnlacZ/in into plasmids. Plasmid DNA was isolated by alkaline-SDS extraction of pooled lysogenic colonies that had grown after 1–2 days of incubation at 37 °C. The pooled plasmid DNA was used to transform strain CC118 and selection was on LB agar containing tetracycline, kanamycin (300 µg ml–1) and X-P or X-G. Individual blue colonies were purified and plasmids were isolated from these clones. Transpositions were mapped with restriction digests and were localized precisely by DNA sequencing using the primers ‘TnphoA-I’ or ‘TnlacZ-II’ (Table 1). To construct specific cydDphoA fusions, the phoA gene was amplified by PCR using {lambda}TnphoA/in as template (GenBank accession no. U25548) and primers RP253 and RP254, both including an AgeI site, or primers RP161 and RP162, both including an MfeI site (Table 1). Plasmid pRKP1602 was cut by AgeI restriction enzyme and ligated to the PCR-generated phoA gene to construct a protein fusion at residue G284. The 1·06 kb MfeI fragment of plasmids pRKP1602 and derivatives (R210G/R216G, R238G/R244G, R210K/R216K and R238H/R244H) was exchanged with a 1·38 kb PCR-generated MfeI fragment containing the phoA gene. The cydD–phoA constructs have an in-frame fusion after the I261 residue of CydD and the mature alkaline phosphatase (AP). The lacZ gene was amplified by PCR using pRS414 as template. Specific cydDlacZ gene fusions were constructed in pRKP1602 as follows. The lacZ gene amplified with primers RP249 and RP250 (with NsiI sites) or RP251 and RP252 (with Bst98I sites; Table 1), was inserted either in NsiI or Bst98I unique sites of the plasmid to construct protein fusions at residues H134 and S202, respectively. The orientations of the fusions were determined by restriction analysis. Sequences and reading frames were checked by DNA sequencing, using primers ‘TnphoA-I’ (Manoil & Bailey, 1997) or RP140 (Table 1, for cydDlacZ gene fusions). The constructs were transformed into E. coli CC118 and transformants were screened on nutrient agar containing ampicillin (100 µg ml–1) and X-P or X-G as before.


View this table:
[in this window]
[in a new window]
 
Table 1. Oligonucleotide sequences

 
Assays of enzymic reporter activity.
AP activity was determined by measuring the hydrolysis at 37 °C of p-nitrophenylphosphate in CHCl3- and SDS-permeabilized cells as described elsewhere (Michaelis et al., 1983; Maloy et al., 1996). Cell pellets were washed once in 1 M Tris/HCl, pH 8·0, resuspended in 5–10 ml of the same buffer and stored on ice. {beta}-Galactosidase (BG) activity was measured in CHCl3- and SDS-permeabilized cells by monitoring the hydrolysis of o-nitrophenyl-{beta}-D-galactopyranoside (Giacomini et al., 1992). Cell pellets were washed once with PBS and resuspended in 5–10 ml of Z buffer (Miller, 1972). Each culture was assayed in triplicate. Absorbance (A) at 420, 550 and 660 nm was measured to express both reporter activities according to the following equation: 1000{bullet}[A420–(1·75{bullet}A550)/time(min){bullet}volume(ml){bullet}A600] (Miller, 1972).

Analysis of the cydD1 allele.
The cydD1 mutation was found by screening survivors of nitrosoguanidine mutagenesis for loss of spectroscopically detectable cytochrome d (Poole et al., 1989). The mutation in strain AN801 was transduced to other strains as described by Poole et al. (1989). A mutant allele in strain AN2343 (cydD1) was obtained and amplified from chromosomal DNA by PCR using primers flanking the entire gene sequence. The double-stranded product was used as template for another series of reactions using each primer separately to generate single-stranded DNA. Single-stranded template was purified by agarose gel electrophoresis, phenol/chloroform and Centricon 100 (Amicon) microconcentration, prior to sequencing using the ABI protocol for PCR products.

Construction of mutant CydD proteins.
Site-directed mutagenesis of the cydD gene was performed using the QuickChange method (Stratagene) or the M13-based Sculptor (Amersham) method. Table 1 shows the details for QuickChange mutagenesis of pRKP1604, namely the residue substitutions, mutagenic oligonucleotide sequences, changed nucleotides, altered codons, restriction sequences and newly introduced or ablated restriction sites. pRKP1604 derivatives carrying the mutated sequences were amplified in E. coli XL-1 Blue. Mutations from these plasmids were transferred to pRKP1602 for complementation analysis by replacing restriction fragments containing the altered sequences.

For the M13-based Eckstein method (Olsen et al., 1993), a 1·8 kb DdeI fragment was excised from pRP39 and cloned into the SmaI site of pBluescript SK. Single-stranded DNA was prepared by superinfection with helper phage M13K07 (Vieira & Messing, 1987) and annealed with the mutagenic primers shown and described in Table 1. A mutant heteroduplex was generated using Klenow polymerase and T4 DNA ligase and the non-mutant strand removed by virtue of incorporation of a thionucleotide in the mutant strand during in vitro synthesis. Homoduplex mutant plasmids were reconstructed in vitro according to the Amersham protocol and used to transform E. coli TG1. Mutated plasmids were detected by endonuclease restriction analysis or sequencing and all changes were confirmed by sequencing. Complementation of AN2343 (cydD) was performed using the constructs in pBluescript, or after cloning the mutated cydD gene into pBR328 by replacement of an EcoRV–BamHI fragment.

Immunoblot analysis.
Detection of the CydD protein in cell membrane fractions was performed by immunoblot analysis with an antibody raised to a peptide comprising amino acids Q302 to V318 in the cytoplasmic carboxy-terminus (Cook et al., 2002). Microaerobic cultures were grown with shaking in Erlenmeyer flasks containing 40 % of their volume of Luria medium supplemented with D-glucose (0·5 %) and ampicillin (100 µg ml–1) with shaking. Membrane fractions prepared from shear-disrupted cells (Poole & Haddock, 1974) were washed and resuspended in buffer (50 mM Tris/HCl pH 7·4, 2 mM MgCl2, 1 mM EGTA). Protein was estimated by the method of Markwell et al. (1978) using bovine serum albumen as standard. After SDS-PAGE, gels were electroblotted onto Hybond-C nitrocellulose (Amersham Biosciences). Blots were developed by using anti-CydD serum (diluted 1 : 200), monoclonal anti-rabbit IgG peroxidase conjugate (Sigma, diluted 1 : 2000) and the ECL Western blotting detection system (Amersham Biosciences).

Assay of the Cyd phenotype.
Strains were challenged on zinc-azide medium (ZnSO4-NaN3, 0·15 mM each) supplemented with Casamino acids (ZABC plates; Poole et al., 1989) and on nutrient agar containing 0·1 mM of the metal-chelating agent ethylenediamine-di(o-hydroxyphenyl acetic acid) (EDDHA) (Cook et al., 1998).

Spectrophotometric analysis and protein assays for determining the cytochrome bd levels of the mutants were performed on cells from microaerobic cultures, grown as described above. Cells were washed and resuspended in Tris/HCl buffer (pH 7·5). A few grains of sodium dithionite were added to reduce the sample, and a few grains of ammonium persulfate and potassium ferricyanide were used with shaking to oxidize the sample. The reduced sample was bubbled with a steady stream of CO gas for 2 min. The reduced minus oxidized and CO-reduced minus reduced spectra were recorded on a Johnson Research Foundation SDB4 dual-wavelength scanning spectrophotometer (Kalnenieks et al., 1998).

The H2O2 sensitivity of strains was examined on solid and in liquid media. Bacterial growth inhibition was measured in a disk diffusion assay as follows. A sample (500 µl) of liquid culture was plated in top agar on NA plates and a sterile mini-disc containing 20 µl of a 3 % (w/v) H2O2 solution was placed on top. Plates were incubated overnight at 37 °C and the diameter of growth inhibition was measured. Second, the viability method of Gort et al. (1999) was used. LB cultures with ampicillin (100 µg ml–1) were inoculated at a starting OD600~0·02. After cultures reached mid-exponential phase (OD600~0·6), aliquots were removed and diluted 1 : 10 000 in PBS, pH 7·4, before adding 2 mM H2O2. Surviving cells were enumerated by plating in top agar on NA plates with appropriate antibiotic, after overnight incubation at 37 °C.

All complementation tests were done by transforming strain AN2343 (cydD1 recA+) with plasmids bearing wild-type or mutant copies of cydDC. We have not formally ruled out the possibility of intragenic complementation, specifically the energization by ATPase activity in one polypeptide, of solute transport via the transmembrane segments of another polypeptide. However, we are not aware of such activity and, in the case of CydDC, mutations in one ABC region are sufficient to inactivate all cysteine transport by the heterodimer (Pittman et al., 2002).

In silico analysis of CydDC sequences.
Sequence searches were performed with the BLAST 2.0 and WU-BLAST 2.0 programs accessible at the National Center for Biotechnology Information (http://www.ncbi.nlm.nih.gov), and the Genome Sequencing Center (http://genome.wustl.edu/blast/client.pl), respectively. Searches were also made using the Bordetella BLAST Server from the Sanger Centre (http://www.sanger.ac.uk/Projects/B_pertussis/blast_server.shtml). Protein sequences were obtained from GenBank, the Comprehensive Microbial Resource of the Institute for Genomic Research (http://www.tigr.org), the Colibri and SubtiList databases (http://genolist.pasteur.fr/Colibri/ and http://genolist.pasteur.fr/SubtiList/, respectively). The program MEME was employed to detect conserved (repeated) motifs in protein sequences.

To predict the topography of CydD and CydC, TMHMM 2.0 (Sonnhammer et al., 1998; Krogh et al., 2001) was used. Multiple sequence alignments and additional analyses of transmembrane segments in the CydDC sequences were performed using the program CLUSTALW and both the von Heijne Helix (transmembrane) and Kyte–Doolittle hydrophilicity algorithms in MacVector 7.0 software. Gap creation and extension penalties used were 3·0 and 1·0, respectively. In certain regions, the alignments had to be adjusted manually.

The sequence of CydDC was compared with the sequences of all known protein structures (the pdb sequence database at http://www.ncbi.nlm.nih.gov) using BLAST (Altschul et al., 1997). Very significant similarities were found to the MsbA proteins from Vibrio cholerae and from E. coli. The former was the more significant and includes the whole sequence of CydDC except for the short N-terminal helix and the first transmembrane helix (25 % sequence identity over amino acids 66–577), whilst the latter (26 % identity, but only over residues 154–580) omits the first three transmembrane sequences. Threading using 3D-PSSM (Kelley et al., 2000) also yields the V. cholerae MsbA as having the most significant similarity over the whole sequence range of CydD. The structure of V. cholerae MsbA (Chang, 2003) therefore provides a good structural model for CydDC. The two point mutations were modelled and their immediate environments in the structure examined using TurboFrodo (Roussel & Cambillau, 1991).


   RESULTS AND DISCUSSION
TOP
ABSTRACT
INTRODUCTION
METHODS
RESULTS AND DISCUSSION
REFERENCES
 
Determination of the start methionine of CydD
In earlier hydrophobicity plots and in the analyses shown in Figs 1 and 2, the start methionine residue of CydD suggested by Poole et al. (1993) was used (labelled M1 in Fig. 2a, see later). However, the amino acid sequence shows another methionine, M45, according to this assumption, which the hydrophobicity and topological analyses place within the membrane. Protein synthesis in vivo shows CydD to have an apparent molecular mass of 61 kDa (Poole et al., 1993) and comparison with the value (64 kDa) from the assumed protein sequence does not allow the start methionine to be identified unambiguously. However, confirmation of M1 was provided by construction of an M1L mutation, which, when present on a plasmid, failed to complement the cydD1 allele, while mutation of M45 was fully complementing when transformants were tested for the Cyd phenotype (see Methods; results not shown).



View larger version (35K):
[in this window]
[in a new window]
 
Fig. 1. Prediction of the topographies of the CydD (top) and CydC (bottom) proteins using TMHMM. For details see text. For each protein, the probability of regions being transmembrane (red), inside (i.e. cytoplasmic, blue) or outside (i.e. periplasmic, magenta) are plotted as a function of amino acid residue number.

 


View larger version (40K):
[in this window]
[in a new window]
 
Fig. 2. Topological model of the CydD (top) and CydC (bottom) polypeptides. In each, the six transmembrane segments are numbered I–VI, the periplasmic loops PL I–III and cytoplasmic loops CL I–II. Amino acid residues (one letter code) at the fusion junctions are indicated with white arrows for LacZ fusions and with black arrows for PhoA fusions. Table 2 gives the measured enzyme activity for each fusion protein. Residues altered by site-directed mutagenesis are indicated by bold black arrowheads. The two point mutations in the cydD1 allele (G319, G429) are indicated by bold type. The white arrowhead points to the site of in-frame cydD–phoA fusions used for topology studies of CL II (see text). The number of positively and negatively charged residues within the periplasmic and CLs is indicated in boxes. The underlined sequences in the ATP-binding domain (C terminus) of both polypeptides show, respectively, the Walker sequence A, the linker peptide, the Walker sequence B and the conserved His residue common to ABC transporters (Linton & Higgins, 1998).

 
Predictions and experimental determination of membrane topography of the CydDC proteins
The CydDC proteins each comprise an N-terminal hydrophobic domain and a C-terminal hydrophilic domain (the ATP-binding or ABC domain) (Poole et al., 1993). Early hydrophobicity predictions of the CydD and CydC sequences revealed stretches of hydrophobic amino acids corresponding to six putative transmembrane helices (Poole et al., 1993) each of sufficient length to span the membrane. A variety of newer tools to predict the topography of transmembrane proteins is now available, of which TMHMM is considered the best-performing (Moller et al., 2001). This implements circular hidden Markov models and determines the most probable topography for the whole protein. The topographies for CydD and CydC are strikingly similar (Fig. 1); each protein is predicted to comprise six transmembrane regions, with both the N and C termini in the cytoplasm. However, in CydC, the probability of a sixth, C-terminus-proximal transmembrane region is lower than for other transmembrane regions. According to Boyd & Beckwith (1989), the distribution of charged residues may function as a topogenic determinant, since they play an important role during protein translocation and membrane protein assembly (Andersson & von Heijne, 1993). The overall amino acid composition and net charge of the N-terminal regions (residues 1–25 and 1–15 for CydD and CydC, respectively) is more similar to that of a typical cytoplasmic than a periplasmic region (Nakashima & Nishikawa, 1992). All proposed CLs possess a net positive charge, whereas the periplasmic loops (PLs) display a net negative (excluding CydD loop I) or no net charge at all, consistent with the ‘positive-inside’ rule (von Heijne, 1992). Finally, the cytoplasmic location of the ABC-bearing C termini is consistent with their putative function, namely ATP binding and hydrolysis.

However, experimental evidence in support of topological models has been lacking. To assess which of these regions actually span the membrane, fusions between random amino acid residues in CydD and CydC with both AP and BG proteins were generated using TnphoA/in and TnlacZ/in transposition into pRP33, resulting in plasmids bearing in-frame cydD–, cydC–phoA and cydD–, cydC–lacZ fusions. A total of 26 fusions were well distributed, covering both protein sequences (Fig. 2a, b; Table 2). To obtain each hydrophobic segment flanked by either phoA or lacZ fusions, we also constructed fusion proteins with specific residues in CydD (i.e. H134, S202 and G284) as described in Methods.


View this table:
[in this window]
[in a new window]
 
Table 2. Locations and measured enzyme activities of AP and BG fusions to CydD and CydC

 
To determine the subcellular locations of the PhoA/LacZ fusions, AP and BG activities of fusion proteins were measured (Table 2). AP has low activity in the cytoplasm (Manoil, 1990), whereas BG remains inactive in the periplasm (Lee et al., 1989). The LacZ fusion proteins K7, Q15, H134, S202, A317, E353 and S465 for CydD, as well as L135, E242, Q248, T329, T429, L483 and L559 for CydC exhibited high levels of BG activity (>450 Miller units). These data show that the N and C termini of each protein are in the cytoplasm and also allowed tentative identification of two CLs for each protein, as shown in Figs 2(a) and (b). Furthermore, the AP fusion proteins A46, L49, M52, L276 and G284 for CydD, as well as L35, L38, L43, L162, L166 and L282 for CydC showed high AP activity (>290 units), consistent with periplasmic exposure of these residues in each protein. The AP fusions place two loops of each polypeptide (labelled PLI and PLIII in each of Figs 2a, b) in the periplasm. Despite the lack of LacZ fusions within the CydC N terminus, the predicted location of the first transmembrane section (I) (Fig. 2b) and the active PhoA fusions L35, L38 and L43 (residues labelled 17 to 19 in Fig. 2b) allowed us to place the N terminus in the cytoplasm. In CydD, AP fusions L75 and L167 (residues labelled 6 and 8, respectively in Fig. 2a) gave low activity; both are predicted to lie within transmembrane regions. The BG fusions A188 and R192 displayed activity that, although corresponding to a cytoplasmic location, are not readily explained. In the CydD sequence, beyond residue 300, further possible transmembrane regions are predicted but at very low probability (see Fig. 1) but the high BG activities of the LacZ fusions E353 and S465 clearly place this region in the cytoplasm.

Problems can arise from the use of these fusions through instability of the fusion proteins. For example, when a LacZ fusion protein is cleaved during synthesis, it is released into the cytoplasm giving a false positive. However, PhoA fusion proteins must be translocated to the periplasm to be measurable, so that stability is of less concern. The results from the LacZ and PhoA fusions are complementary and internally consistent and have been used to construct the transmembrane topological models of CydD and CydC shown in Figs 2(a) and (b), respectively. Both polypeptides span the membrane six times (TM I–VI) and each protein has two large CLs (labelled CL I–II in Fig. 2a, b) and three small PLs (PL I–III in Fig. 2). The positions of the helices predicted from TMHMM lie within the regions defined by protein fusion activity. von Heijne transmembrane and Kyte–Doolittle hydrophilicity plots also reveal hydrophobic clusters in CydD and CydC sequences similar to those shown in Fig. 2 (results not shown). Although the cytoplasmic location of the ATPase domains cannot be deduced from the positive LacZ fusions obtained alone (see above), there is ample evidence for a cytoplasmic location for the nucleotide-binding domain in related proteins (e.g. MsbA; Chang & Roth, 2001) and this location is consistent with the exclusively intracellular availability of ATP. Our experimental findings are also consistent with the topography of other prokaryotic and eukaryotic ABC transporters. For example, the Drosophila melanogaster pigment precursor transporter is also a heterodimer, in which both the White and Brown subunits possess six transmembrane domains with both N and C termini in the cytoplasm (Ewart et al., 1994). Also, in the crystal structure of E. coli MsbA – the lipid A flippase – the transmembrane region is composed of six tilted {alpha}-helices per monomer, with contacts between helices from different monomers (Chang & Roth, 2001).

Mutations in the cydD1 allele and the nucleotide-binding domain
The first mutant allele of cydD was isolated after random mutagenesis and screening spectroscopically for loss of the distinctive absorbance properties of cytochrome bd (Poole et al., 1989). Despite use of this strain over many years, during which time the stability of this allele has been noted, the mutated residue(s) have not been determined. Here we report two point mutations, G319D and G429E in the cydD1 mutant allele. Both are in the C-terminal, cytoplasmic portion of the CydD protein, flanking the Walker A-motif (Fig. 2a). To test the importance of each mutation in conferring a Cyd phenotype, we individually engineered these changes into the cydD gene cloned into pBluescript and tested the ability of the mutant constructs to complement the cydD1 allele in strain AN2343. The G319D mutation alone was sufficient to eliminate complementation as judged by the absence of spectroscopically detectable cytochrome bd. The G429E mutation in pBluescript reduced cytochrome d content to less than one-third the level of that afforded by the wild-type cydD+ gene cloned into either pBluescript or pBR328. Thus, either mutation affects cytochrome bd assembly, perhaps explaining the failure to find spontaneous revertants at this locus. Fig. 3 shows a model of the CydD protein based on the solved structure of MsbA in Vibrio cholera (Chang, 2003). The cytoplasmic domain comprises the N terminus, two large CLs, and the C terminus bearing the ATP-binding cassette, including Walker motifs A and B, as demonstrated in the fusion analysis. The upper half of the structure in Fig. 3 comprises the six membrane-spanning helices and small PLs, again as demonstrated in this work. G429 (in E. coli; Ser423 in Vibrio) is close to a conserved Asp residue. Substitution with Glu, a large negatively charged residue, next to a conserved Asp (510 in E. coli; 505 in Vibrio) may be responsible for disruption of structure near the Walker B region. G319 is buried in a pocket and may make van der Waals contact with Leu104 at the extreme end of a transmembrane helix, although the reason for the Cyd phenotype is unclear at this time. Mutation of G319 might slow down conformational changes that are required during the catalytic cycle. This region of the ABC domain may serve an intramolecular signalling function coupling the hydrolysis of ATP with the membrane-spanning components (Ames & Lecar, 1992). Glycines are expected to effect transmission of conformational changes within proteins, since they lack a lateral chain and appear crucial in dynamic functionality. Glycine residues have been shown to interact with the phosphoryl groups of ATP. A G313A mutation within the G box of HPK was observed to diminish greatly the formation of a covalent adduct with ATP upon UV irradiation (Ninfa et al., 1993).



View larger version (22K):
[in this window]
[in a new window]
 
Fig. 3. Model of the solved structure of one ‘half-molecule’ of MsbA in V. cholerae and the positions of residues of significance in CydD. The lower half of the structure is the cytoplasmic domain (N terminus, two large cytoplasmic loops and C terminus bearing the ATP-binding cassette, including Walker motifs A and B. The upper half comprises six membrane-spanning helices and small periplasmic loops. The mutated residues in the cydD1 mutant allele (Cyd phenotype) are G429E and G319D. In green are the Walker motifs. For clarity, the positions of residues 429 and 319 are shown in a space-filling representation (magenta and red, respectively) and the nearby residues Asp 510/505 and Leu 104 are shown as ball-and-stick. Most of the heptameric repeat region (207–246) is disordered and not observed in either the V. cholerae or the E. coli MsbA structure. Its approximate location is indicated by a blue ellipse. Residues 238–246, which are observed, are coloured in blue. For other details, see text.

 
The invariant amino acid doublet Gly-Lys of the Walker motif A has been proposed to be directly involved in binding of ATP (Parsonage et al., 1987) and mutation of these residues in various members of the traffic ATPase family eliminates or severely compromises function (Ewart et al., 1994 and references therein). To establish the significance of the proposed nucleotide-binding fold in CydD, we introduced the mutation G388L/K389Q in the Walker sequence A (Fig. 3). This change in CydD was sufficient to disrupt function, as assessed by failure to restore growth of a cydD mutant under selective conditions and total lack of cytochrome d spectral signals (not shown). Several studies of P-glycoprotein (the multidrug-resistance protein) indicate that the two ATP-binding sites cannot function independently and a strong cooperative interaction occurs between the two (Urbatsch et al., 1995). The requirement for two ABC domains in CydDC is demonstrated by the fact that the directed G388K/K389Q mutation of CydD in the nucleotide-binding domain disrupts function. It has been previously reported that mutation of G135 and K136 to LQ within the nucleotide-binding fold motif of the white gene of D. melanogaster is unable to complement eye colour in recipient transgenic flies with a defective white gene (Ewart et al., 1994).

Sequence-informed targeting of residues for mutagenesis in the CydD protein
A search for similarities between two transmembrane regions of both CydDC proteins (TM III and TM IV) with other proteins in GenBank has shown numerous hits with mitochondrial b-type cytochrome sequences from both invertebrates and vertebrates (Cook et al., 2002). Moreover, TM IV of CydD has a pair of glycines (G171 and G184; Fig. 2a) separated by 12 residues, which, with a similar pair in TM I of CydC (G21 and G34; Fig. 2b) resemble the quartet of conserved glycines in cytochromes b involved in configuration of the haem pocket (Esposti et al., 1993). Inspection of putative CydD and CydC sequences in other bacteria is confounded by a lack of information on features that distinguish CydDC transporters (i.e. that export Cys and GSH) from other members of the ABC transporter family. However, CLUSTALW alignments show that the Salmonella typhimurium LT2 CydD and CydC proteins exhibit precisely the same quartet of glycines as in E. coli, and similar groupings of glycines occur in the presumed CydD and CydC proteins of other Gram-negative bacteria. Thus, in both Vibrio vulnificus YJ016 (accession NP_934247) and Photobacterium profundum (accession CAG19569) a pair of glycines is found in CydD, separated by 17 residues, the second of the pair aligning with E. coli G184. In Brucella melitensis 16M (accession NP_459933), the alignment and spacing are as in E. coli and Salmonella. In all these bacteria, glycine pairs also occur in CydC, the first of each pair aligning with E. coli G21: in V. vulnificus, there are 12 residues between the glycines, whereas in Photobacterium profundum and Brucella melitensis, 10 residues separate the glycines. Interestingly, no such glycine pairs occur in the putative CydD and CydC proteins of the Gram-positive bacteria Mycobacterium tuberculosis, Bacillus cereus or Listeria monocytogenes. Another glycine pair (G274 and G284) is seen in CydD of E. coli, Salmonella and Photorhabdus luminescens (accession NP_928889). In this preliminary examination of the possible roles of these paired glycines, we targeted glycine residues G171, G274 and G284 in CydD TM IV, PL III and TM VI, respectively (Fig. 2a), for mutagenesis.

Statistical analysis of transmembrane segments of ion channels (Brandl & Deber, 1986) reveals a frequent distribution of prolines. Furthermore, in light of the recent demonstration that CydDC transports cysteine (Pittman et al., 2002), it is notable that most prolines in the TMs of prokaryote amino acid transporters are highly conserved (Pi et al., 1998). Therefore, we decided to change the proline residues in CydD TM III (P142 and P151, Fig. 2a) for leucines whose relatively large size may change the local structure. These two prolines are strongly conserved in bacterial CydD-like proteins (data not shown).

When CL II of the CydD protein was analysed using the MEME program, we noticed two positively charged heptameric motifs separated by 21 residues. These are shown in Fig. 4; the first starts at residue 210 in E. coli CydD and the second at residue 238. Each of these is well conserved in CydD proteins of other species. In the first heptamer, the second position is G in 17 of the 19 cases examined, whereas the second position in the second heptamer is T in 17 of 19 cases. We therefore describe the ‘consensus' at this position as G/T. The third residue in the first heptamer is usually L (12L/19) and the third residue in the second heptamer is usually M (17M/19). We therefore describe the ‘consensus' at this position as L/M. Overall, we suggest that the consensus is R-G/T-L/M-X-T/V-L-R, where ‘/’ separates the most probable occupants of certain positions (Fig. 4). The R210, R216 and R244 residues are frequently preceded by another positively charged residue, i.e. lysine. Thus, we decided to introduce the R210G/R216G and R238G/R244G changes into the CydD protein to annul the positive charges of these regions. Fig. 3 shows the heptameric repeat region. Residues 207–246 are disordered in the structures of both the V. cholerae and the E. coli MsbA; the approximate location of this region is indicated by a blue ellipse. Residues 238–246, which are observed, are coloured in blue.



View larger version (54K):
[in this window]
[in a new window]
 
Fig. 4. Positively charged heptameric regions of CydD-like sequences in a multiple alignment. Residue numbers defining the limits of these regions in each protein are shown. Identical and similar residues are in dark grey or light grey boxes, respectively. Positive charges define the start and end points of each proposed heptameric region and are indicated above the aligned sequences. Gene accession, or sequence identifier, or translated base-pair numbers (in parentheses) for each sequence follow: E. coli K-12-MG1655 (2506099), Salmonella enterica serovar Typhimurium (g1421_saltyLT2), Providencia stuartii (939711), Pasteurella multocida PM70 (12720839), V. cholerae El TorN16961 (9655658), Haemophilus influenzae Rd KW20 (1574714), Shewanella violacea (1536829), Bordetella pertussis (translated sequences from base pair 270836 to 272530), Streptomyces coelicolor (*N-terminus of ‘CydCD’: 3928723), Caulobacter crescentus CB15 (13421998), Bacillus subtilis 168 (2829798), Enterococcus faecalis V583 (EF2059), Mycobacterium tuberculosis H37Rv (2113905), Mycobacterium smegmatis (6289096), Lactococcus lactis subsp. lactis IL1403 (12723617), Staphylococcus aureus subsp. aureus N315 (13700575), Bacillus halodurans C-125 (10176596), Streptococcus pyogenes M1 GAS strain SF370 (13622838), Proteus mirabilis (4097161).

 
Phenotypic analysis of site-directed cydD mutations
Cyd mutants have a complex phenotype, including sensitivity to: (i) zinc and azide, inhibitors of the remaining cytochrome o-mediated respiratory chain (Poole et al., 1989); (ii) iron chelators (Cook et al., 1998) and (iii) H2O2 (Delaney et al., 1992; Lindqvist et al., 2000). We exploited these sensitivities to test strain AN2343 (cydD1) that had been transformed with pRKP1602 (bearing the wild-type cydDC genes) or site-directed cydD mutant derivatives, by plating on ZABC medium (containing Zn2+ and azide ions) or plates supplemented with EDDHA or H2O2. The non-conservative substitutions of the targeted glycine or proline residues (G171V, G274F, G284A, P142L and P151L) did not prevent complementation of strain AN2343 by the pRKP1602 derivatives. Furthermore, spectral analysis revealed that all transformants synthesized cytochrome bd (data not shown). Therefore, these residues do not play key roles in CydD function and do not lend support to the notion that cytochrome b-like sequences in CydD and CydC play essential roles in haem binding. The ability of these mutant constructs to form the enigmatic P-574 pigment (Cook et al., 2002) has not been tested, but the identification of cysteine as an allocrite for CydDC (Pittman et al., 2002) suggests that changes in haemoproteins in CydDC-overexpressing strains probably result from alterations in periplasmic redox status, rather than a direct involvement of CydDC in haem transport.

To assess the functional importance of the duplicated heptameric motif in CydD CL II, we screened arginine-to-glycine substitutions. pRKP1602 (carrying wild-type cydD+) complemented strain AN2343 (cydD1) as shown by the high level of cytochrome d, lack of inhibition by EDDHA, zinc or azide, and small zones of inhibition around disks impregnated with H2O2 (Fig. 5). However, plasmids encoding the R210G/R216G and R238G/R244G mutant forms of CydD exhibited significantly lower levels of cytochrome d than the wild-type (2·3- and 1·4-fold less, respectively; Fig. 5) and slower growth when challenged with the above agents.



View larger version (15K):
[in this window]
[in a new window]
 
Fig. 5. Cytochrome d levels in cydD1 strain carrying different plasmids. pRKP1602 carries the wild-type cydDC operon and pBR328 is the vector control. The remaining four plasmids carry the site-directed mutations shown. Growth inhibition under stress conditions (EDDHA, zinc/azide), percentage survival, and diameter of inhibition halo in the presence of H2O2, are shown below. For details, see text. Cytochrome d levels were measured in at least two different cultures for each strain; variations in duplicate measurements were <2–5 %. Survival counts and growth inhibition on plates were repeated three times with similar results.

 
Studies of membrane insertion and topology of CydD mutants
We investigated the possibility that unsuccessful insertion of mutant proteins into the membrane was responsible for lack of complementation. Membrane fractions were probed with the anti-CydD serum as described in Methods. As can be seen in Fig. 6, the G319D/G429E CydD protein was detected at similar levels to that of the wild-type protein. Thus, the AN2343 strain phenotype is not due to a defective insertion of the CydD subunit into the membrane. Additionally, we detected a stronger signal when the CydD overproducer plasmid pRKP1602 and its derivatives were present. This suggests that the mutations described here do not disrupt CydD protein localization.



View larger version (8K):
[in this window]
[in a new window]
 
Fig. 6. Immunoblot analysis of the CydD protein. Identical amounts of protein membrane fractions (20 µg) were loaded for 12 % SDS-PAGE, and the CydD proteins were detected with anti-CydD serum. Lanes 1 and 2 correspond to cydD1 AN2343 and wild-type AN2342 strains, respectively, transformed with pBR328. Lanes 3 and 4 are wild-type and cydD1 strains, respectively, transformed with pRKP1602. Construct derivatives of pRKP1602, R210G/R216G (lane 5), R238G/R244G (lane 6), R210K/R216K (lane 7) and R238H/R244H (lane 8), were put into the cydD1 strain. Molecular mass markers (in kDa) are indicated to the left of the blot.

 
The importance of the positively charged domains located in the CLs of CydD was investigated by site-directed mutagenesis. The charges in the highly conserved R-X-G-R-R domains of the glucose transporter Glut1 form local cytoplasmic anchor points, ensuring the correct membrane topography (Sato & Mueckler, 1999). The overall charge of the CydD CL II is also rather positive with 12 positively and 10 negatively charged residues (Fig. 2a). We asked whether the removal of two positive charges in the R210G/R216G and R238G/R244G substitutions might provoke the ‘looping-out’ of CL II into the periplasm. A series of in-frame cydD–phoA fusions after residue I261 (see Fig. 2a) for mutations R210G/R216G, R238G/R244G, R210K/R216K and R238H/R244H was constructed and transformed into strain CC118. Choice of residue I261 allowed use of a unique MfeI site for construction of such fusions. Aberrant translocation of CL II into the periplasm would yield high AP activity, but neither the wild-type cydD–phoA fusion nor the other mutated fusions gave a strong AP activity (9–18 Miller units). This demonstrates that CL II remains located in the cytoplasm and that the targeted positive charges are not essential determinants of membrane topography.

Arg to Gly mutations are overcome by replacement of positive charges
We constructed plasmids including the R210K/R216K and R238H/R244H changes in CydD, and introduced them into the cydD1 background. These substitutions prevented the growth inhibition in challenging conditions, and the cytochrome d levels were increased close to those from the wild-type strain (Fig. 5). Thus, regardless of the different residue side-chains, lysine and histidine are functional substitutions for arginines in the duplicated heptameric motif of CydD CL II, and are able to restore the wild-type phenotype by the sole supply of a positive charge. The function of this region is unclear but it may be involved in cysteine binding.

Evidence from sequence analysis for conservation, but shuffling, of cydDC genes
In E. coli, the genes encoding the ATP transporter are denoted cydDC because the second gene of the operon was discovered first (Georgiou et al., 1987; Poole et al., 1989). Gene nomenclature in Bacillus subtilis follows an alphabetical order and, in contrast with E. coli, the genes are transcribed as a cydABCD polycistronic message (Winstedt et al., 1998). The E. coli nomenclature has been used in mycobacteria as well as in Gram-negative bacteria, whereas the Bacillus subtilis nomenclature has been reserved for Gram-positive bacteria. A further 17 and 18 complete orthologue sequences for cydD and cydC, respectively, were analysed. Interestingly, the cyd locus of Streptomyces coelicolor A3(2) encodes three proteins, namely CydA, CydB and a protein (GI:3928723; 1172 amino acids) that has two ABC transporter family signatures, one at the middle (‘ABC1’) and the second at its C terminus (‘ABC2’). This apparent fusion protein (‘CydDC’) resembles, therefore, some eukaryotic ABC transport systems where a tandemly duplicated molecule yields a single protein with two ABC domains. In this case, we considered separately the N-terminal ‘ABC1’ (575 residues, here named CydD) and the C-terminal ‘ABC2’ (remaining 597 residues, named CydC here) segments, and used these truncated proteins as supplementary sequences for CydD and CydC analyses. An alignment of the CydDC sequences was performed. Alignment and analysis of these sequences using the von Heijne Helix (transmembrane) and Kyte–Doolittle hydrophilicity algorithms revealed similar topological features, namely six TMs, two major CLs and three minor PLs with both N and C termini located in the cytoplasm. Almost all the CLs from CydDC sequences showed a net positive charge. We also investigated the organization in bacterial genomes of the cyd genes and found four different types of genetic organization. The cydAB and cydDC genes can constitute (i) independent clusters, (ii) neighbouring divergent clusters (<-cydCD cydAB->), or show a tandem cydABDC arrangement encoding (iii) three or (iv) four proteins. In the first case, we traced the presence of the trxB gene that is upstream of the E. coli cydD gene. It appears that across short phylogenetic distances this gene organization is conserved but vanishes in organisms such as Streptomyces violacea and Gram-positive bacteria. A similar analysis of the aat gene downstream of E. coli cydC in these organisms showed no conservation. In this context, it is interesting to mention that the Aeromonas jandaei locus (GI:1903398) displays the trxB and aat genes separated not by the cydDC cluster but by a totally unrelated gene encoding a response regulator protein, RrpX. A search of the upstream genes of the cydAB clusters from a few genomes showed weak conservation of gene ordering. Intriguingly, Lactococcus lactis exhibits the tra983B gene encoding the transposase of IS983B. Immediately upstream of the cydAB operon of Brucella melitensis biovar Abortus, two genes encoding ABC-type transporters of 331 and 322 residues were found (Endley et al., 2001). Tentatively, these genes were named cydC and cydD by the authors although no functional evidence was provided.

Conservation of the cydD and cydC genes in an operon possibly facilitates interactions of the encoded proteins (Dandekar et al., 1998), may favour lateral gene transfer (Lawrence & Roth, 1996) or allow co-localization of the mRNAs in the same region of the cell (Danchin et al., 2000). The conservation of gene order is obvious between closely related species, but rapidly becomes less conspicuous among more distantly related organisms (Tamames, 2001). Nevertheless, cydDC gene ordering in Staphylococcus aureus, Bacillus halodurans or Streptococcus pyogenes resembles more the E. coli organization (except for the upstream trxB gene) than the corresponding one found in Gram-positive bacteria. On the basis of the presence of a small number of insertion sequences among 11 complete genome sequences, Bacillus subtilis was found to have the most stable genome retaining the ancestral genome structure (Itoh et al., 1999). Gene translocation events may be responsible for loss of gene order in some bacteria, as Itoh et al. (1999) have proposed for the elimination of operons.

Conclusions
We propose a heterodimeric structure for the ABC-type transporter CydDC that is absolutely required for cytochrome bd assembly in E. coli. The two constituent polypeptides are strikingly similar and each is shown here to comprise six N-terminal transmembrane segments and a C-terminal, cytoplasmic ATP-binding cassette. The function of this transporter appears to be the export of reductant, specifically cysteine (Pittman et al., 2002), to the periplasm. Although cydD- and cydC-like genes have been reported in most bacteria that synthesize cytochrome bd, it is possible that the heterodimeric structure proposed may not in all cases have the same transport function. Thus, in D. melanogaster, the White and Scarlet proteins that constitute a heterodimeric ABC transporter form a tryptophan transporter, whereas the White and Brown proteins form a guanine transporter (Tearle et al., 1989; Mackenzie et al., 1999). Different pairings of CydD- and/or CydC-like proteins might export diverse substrate molecules.


   ACKNOWLEDGEMENTS
 
This work was supported by grants P12980 and C02603 from the BBSRC. We are indebted to Frank Gibson and Graeme Cox for their interest in this work over many years, Lyndall Hatch for contributions to cloning and sequencing the cydD1 allele and Gary Ewart for advice on mutagenesis. We are also grateful to Peter Artymiuk for producing Fig. 3 and Colin Manoil for providing plasmids and strains.


   REFERENCES
TOP
ABSTRACT
INTRODUCTION
METHODS
RESULTS AND DISCUSSION
REFERENCES
 
Altschul, S. F., Madden, T. L., Schäffer A. A., Zhang, J., Zhang, Z., Miller, W. & Lipman, D. J. (1997). Gapped BLAST and PSI-BLAST: a new generation of protein database search programs. Nucleic Acids Res 25, 3389–3402.[Abstract/Free Full Text]

Ames, G. F.-L. & Lecar, H. (1992). ATP-dependent bacterial transporters and cystic fibrosis: analogy between channels and transporters. FASEB J 6, 2660–2666.[Abstract/Free Full Text]

Andersson, H. & von Heijne, G. (1993). Position-specific Asp-Lys pairing can affect signal sequence function and membrane protein topology. J Biol Chem 268, 21389–21393.[Abstract/Free Full Text]

Bebbington, K. J. & Williams, H. D. (1993). Investigation of the role of the cydD gene product in production of a functional cytochrome d oxidase in Escherichia coli. FEMS Microbiol Lett 112, 19–24.[CrossRef][Medline]

Borisov, V. B., Sedelnikova, S. E., Poole, R. K. & Konstantinov, A. A. J. (2001). Interaction of cytochrome bd with carbon monoxide at low and room temperatures: evidence that only a small fraction of heme b595 reacts with CO. J Biol Chem 276, 22095–22099.[Abstract/Free Full Text]

Boyd, D. & Beckwith, J. (1989). Positively charged amino acid residues can act as topogenic determinants in membrane proteins. Proc Natl Acad Sci U S A 86, 9446–9450.[Abstract]

Brandl, C. J. & Deber, C. M. (1986). Hypothesis about the function of membrane-buried proline residues in transport proteins. Proc Natl Acad Sci U S A 83, 917–921.[Abstract]

Chang, G. (2003). Structure of MsbA from Vibrio cholera: a multidrug resistance ABC transporter homolog in a closed conformation. J Mol Biol 330, 419–430.[CrossRef][Medline]

Chang, G. & Roth, C. B. (2001). Structure of MsbA from E. coli: a homolog of the multidrug resistance ATP binding cassette (ABC) transporters. Science 293, 1793–1800.[Abstract/Free Full Text]

Cook, G. M. & Poole, R. K. (2000). Oxidase and periplasmic cytochrome assembly in Escherichia coli K-12: CydDC and CcmAB are not required for haem-membrane association. Microbiology 146, 527–536.[Medline]

Cook, G. M., Loder, C., Søballe, B., Stafford, G. P., Membrillo-Hernández, J. & Poole, R. K. (1998). A factor produced by Escherichia coli K-12 inhibits the growth of E. coli mutants defective in the cytochrome bd quinol oxidase complex: enterochelin rediscovered. Microbiology 144, 3297–3308.[Medline]

Cook, G. M., Cruz-Ramos, H., Moir, A. J. & Poole, R. K. (2002). A novel haem compound accumulated in Escherichia coli overexpressing the cydDC operon, encoding an ABC-type transporter required for cytochrome assembly. Arch Microbiol 178, 358–369.[CrossRef][Medline]

Cotter, P. A., Chepuri, V., Gennis, R. B. & Gunsalus, R. P. (1990). Cytochrome o (cyoABCDE) and d (cydAB) oxidase gene expression in Escherichia coli is regulated by oxygen, pH, and the fnr gene product. J Bacteriol 172, 6333–6338.[Medline]

Danchin, A., Guerdoux-Jamet, P., Moszer, I. & Nitschke, P. (2000). Mapping the bacterial cell architecture into the chromosome. Philos Trans R Soc Lond B Biol Sci 355, 179–190.[CrossRef][Medline]

Dandekar, T., Snel, B., Huynen, M. & Bork, P. (1998). Conservation of gene order: a fingerprint of proteins that physically interact. Trends Biochem Sci 23, 324–328.[CrossRef][Medline]

Delaney, J. M., Ang, D. & Georgopoulos, C. (1992). Isolation and characterization of the Escherichia coli htrD gene, whose product is required for growth at high temperatures. J Bacteriol 174, 1240–1247.[Abstract]

Delaney, J. M., Wall, D. & Georgopoulos, C. (1993). Molecular characterization of the Escherichia coli htrD gene: cloning, sequence, regulation, and involvement with cytochrome d oxidase. J Bacteriol 175, 166–175.[Abstract]

D'Mello, R., Hill, S. & Poole, R. K. (1996). The cytochrome bd quinol oxidase in Escherichia coli has an extremely high oxygen affinity and two oxygen-binding haems: implications for regulation of activity in vivo by oxygen inhibition. Microbiology 142, 755–763.[Medline]

Edwards, S. E., Loder, C. S., Wu, G., Corker, H., Bainbridge, B. W., Hill, S. & Poole, R. K. (2000). Mutation of cytochrome bd quinol oxidase results in reduced stationary phase survival, iron deprivation, metal toxicity and oxidative stress in Azotobacter vinelandii. FEMS Microbiol Lett 185, 71–77.[CrossRef][Medline]

Endley, S., McMurray, D. & Ficht, T. A. (2001). Interruption of the cydB locus in Brucella abortus attenuates intracellular survival and virulence in the mouse model of infection. J Bacteriol 183, 2454–2462.[Abstract/Free Full Text]

Esposti, D. E., De Vries, S., Crimi, M., Ghelli, A., Patarnello, T. & Meyer, A. (1993). Mitochondrial cytochrome b: evolution and structure of the protein. Biochim Biophys Acta 1143, 243–271.[Medline]

Ewart, G. D., Cannell, D., Cox, G. B. & Howells, A. J. (1994). Mutational analysis of the traffic ATPase (ABC) transporters involved in uptake of eye pigment precursors in Drosophila melanogaster. Implications for structure-function relationships. J Biol Chem 269, 10370–10377.[Abstract/Free Full Text]

Gennis, R. B. & Stewart, V. (1996). Respiration. In Escherichia coli and Salmonella. Cellular and Molecular Biology, pp. 217–261. Washington, DC: American Society for Microbiology.

Georgiou, C. D., Fang, H. & Gennis, R. B. (1987). Identification of the cydC locus required for expression of the functional form of the cytochrome d terminal oxidase complex in Escherichia coli. J Bacteriol 169, 2107–2112.[Medline]

Giacomini, A., Corich, V., Ollero, F. J., Squartini, A. & Nuti, M. P. (1992). Experimental conditions may affect reproducibility of the beta-galactosidase assay. FEMS Microbiol Lett 79, 87–90.[Medline]

Goldman, B. S., Gabbert, K. K. & Kranz, R. G. (1996a). Use of heme reporters for studies of cytochrome biosynthesis and heme transport. J Bacteriol 178, 6338–6347.[Abstract]

Goldman, B. S., Gabbert, K. K. & Kranz, R. G. (1996b). The temperature-sensitive growth and survival phenotypes of Escherichia coli cydDC and cydAB strains are due to deficiencies in cytochrome bd and are corrected by exogenous catalase and reducing agents. J Bacteriol 178, 6348–6351.[Abstract]

Gort, A. S., Ferber, D. M. & Imlay, J. A. (1999). The regulation and role of the periplasmic copper, zinc superoxide dismutase of Escherichia coli. Mol Microbiol 32, 179–191.[CrossRef][Medline]

Holland, I. B. & Blight, M. A. (1999). ABC-ATPases, adaptable energy generators fuelling transmembrane movement of a variety of molecules in organisms from bacteria to humans. J Mol Biol 293, 381–399.[CrossRef][Medline]

Itoh, T., Takemoto, K., Mori, H. & Gojobori, T. (1999). Evolutionary instability of operon structures disclosed by sequence comparisons of complete microbial genomes. Mol Biol Evol 16, 332–346.[Abstract]

Jünemann, S. (1997). Cytochrome bd terminal oxidase. Biochim Biophys Acta 1321, 107–127.[Medline]

Juty, N. S., Moshiri, F., Merrick, M., Anthony, C. & Hill, S. (1997). The Klebsiella pneumoniae cytochrome bd' terminal oxidase complex and its role in microaerobic nitrogen fixation. Microbiology 143, 2673–2683.[Medline]

Kalnenieks, U., Galinina, N., Bringer-Meyer, S. & Poole, R. K. (1998). Membrane D-lactate oxidase in Zymomonas mobilis: evidence for a branched respiratory chain. FEMS Microbiol Lett 168, 91–97.[CrossRef][Medline]

Kelley, L. A., MacCallum, R. M. & Sternberg, M. J. E. (2000). Enhanced genome annotation using structural profiles in the program 3D-PSSM. J Mol Biol 299, 499–520.[Medline]

Kelly, M. J., Poole, R. K., Yates, M. G. & Kennedy, C. (1990). Cloning and mutagenesis of genes encoding the cytochrome bd terminal oxidase complex in Azotobacter vinelandii: mutants deficient in the cytochrome d complex are unable to fix nitrogen in air. J Bacteriol 172, 6010–6019.[Medline]

Krogh, A., Larsson, B., von Heijne, G. & Sonnhammer, E. L. L. (2001). Predicting transmembrane protein topology with a hidden Markov Model: application to complete genomes. J Mol Biol 305, 567–580.[CrossRef][Medline]

Lawrence, J. G. & Roth, J. R. (1996). Selfish operons: horizontal transfer may drive the evolution of gene clusters. Genetics 143, 1843–1860.[Abstract/Free Full Text]

Lee, C., Li, P., Inouye, H., Brickman, E. R. & Beckwith, J. (1989). Genetic studies on the inability of {beta}-galactosidase to be translocated across the Escherichia coli cytoplasmic membrane. J Bacteriol 171, 4609–4616.[Medline]

Lindqvist, A., Membrillo-Hernández, J., Poole, R. K. & Cook, G. M. (2000). Role of respiratory oxidases in protecting Escherichia coli K12 from oxidative stress. Antonie van Leeuwenhoek 78, 23–31.[CrossRef][Medline]

Linton, K. J. & Higgins, C. F. (1998). The Escherichia coli ATP-binding cassette (ABC) proteins. Mol Microbiol 28, 5–13.[CrossRef][Medline]

Mackenzie, S. M., Brooker, M. R., Gill, T. R., Cox, G. B., Howells, A. J. & Ewart, G. D. (1999). Mutations in the white gene of Drosophila melanogaster affecting ABC transporters that determine eye colouration. Biochim Biophys Acta 1419, 173–185.[Medline]

Maloy, S. R., Stewart, V. J. & Taylor, R. K. (1996). Genetic Analysis of Pathogenic Bacteria. Cold Spring Harbor, NY: Cold Spring Harbor Laboratory.

Manoil, C. (1990). Analysis of protein localization by use of gene fusions with complementary properties. J Bacteriol 172, 1035–1042.[Medline]

Manoil, C. & Bailey, J. (1997). A simple screen for permissive sites in proteins: analysis of Escherichia coli lac permease. J Mol Biol 267, 250–263.[CrossRef][Medline]

Manoil, C. & Beckwith, J. (1985). TnphoA: a transposon probe for protein export signals. Proc Natl Acad Sci U S A 82, 8129–8133.[Abstract]

Markwell, M. A., Hass, S. M., Bieber, L. L. & Tolbert, N. E. (1978). A modification of the Lowry procedure to simplify protein determination in membrane and lipoprotein samples. Anal Biochem 87, 206–210.[Medline]

Michaelis, S., Inouye, H., Oliver, D. & Beckwith, J. (1983). Mutations that alter the signal sequence of alkaline phosphatase in Escherichia coli. J Bacteriol 154, 366–374.[Medline]

Miller, J. H. (1972). Experiments in Molecular Genetics. Cold Spring Harbor, NY: Cold Spring Harbor Laboratory.

Moller, S., Croning, M. D. R. & Apweiler, R. (2001). Evaluation of methods for the prediction of membrane spanning regions. Bioinformatics 17, 646–653.[Abstract/Free Full Text]

Nakashima, H. & Nishikawa, K. (1992). The amino acid composition is different between the cytoplasmic and extracellular sides in membrane proteins. FEBS Lett 303, 141–146.[CrossRef][Medline]

Ninfa, E. G., Atkinson, M. R., Kamberov, E. S. & Ninfa, A. J. (1993). Mechanisms of autophosphorylation of Escherichia coli nitrogen regulator II (NRII or NtrB): trans-phosphorylation between subunits. J Bacteriol 175, 7024–7032.[Abstract]

Olsen, D. B., Sayers, J. R. & Eckstein, F. (1993). Site-directed mutagenesis of single-stranded and double-stranded DNA by phosphorothionate approach. Methods Enzymol 217, 189–217.[CrossRef][Medline]

Osborne, J. P. & Gennis, R. B. (1999). Sequence analysis of cytochrome bd oxidase suggests a revised topology for subunit I. Biochim Biophys Acta 1410, 32–50.[Medline]

Parsonage, D., Wilke-Mounts, S. & Senior, A. E. (1987). Directed mutagenesis of the {beta}-subunit of F1-ATPase from Escherichia coli. J Biol Chem 262, 8022–8026.[Abstract/Free Full Text]

Pi, J., Dogovski, C. & Pittard, A. J. (1998). Functional consequences of changing proline residues in the phenylalanine-specific permease of Escherichia coli. J Bacteriol 180, 5515–5519.[Abstract/Free Full Text]

Pittman, M. S., Corker, H., Wu, G., Binet, M. B., Moir, A. J. G. & Poole, R. K. (2002). Cysteine is exported from the Escherichia coli cytoplasm by CydDC, an ABC-type transporter required for cytochrome assembly. J Biol Chem 277, 49841–49849.[Abstract/Free Full Text]

Poole, R. K. & Cook, G. M. (2000). Redundancy of aerobic respiratory chains in bacteria? Routes, reasons and regulation. Adv Microb Physiol 43, 165–224.[Medline]

Poole, R. K. & Haddock, B. A. (1974). Energy-linked reduction of nicotinamide-adenine dinucleotide in membranes derived from normal and various respiratory-deficient mutant strains of Escherichia coli K12. Biochem J 144, 77–85.[Medline]

Poole, R. K., Williams, H. D., Downie, J. A. & Gibson, F. (1989). Mutations affecting the cytochrome d-containing oxidase complex of Escherichia coli K12: identification and mapping of a fourth locus, cydD. J Gen Microbiol 135, 1865–1874.[Medline]

Poole, R. K., Hatch, L., Cleeter, M. W., Gibson, F., Cox, G. B. & Wu, G. (1993). Cytochrome bd biosynthesis in Escherichia coli: the sequences of the cydC and cydD genes suggest that they encode the components of an ABC membrane transporter. Mol Microbiol 10, 421–430.[Medline]

Poole, R. K., Gibson, F. & Wu, G. (1994). The cydD gene product, component of a heterodimeric ABC transporter, is required for assembly of periplasmic cytochrome c and of cytochrome bd in Escherichia coli. FEMS Microbiol Lett 117, 217–223.[CrossRef][Medline]

Roussel, A. & Cambillau, C. (1991). ‘TurboFrodo’, Silicon Graphics Geometry Partners Directory, p. 86. Mountain View, CA: Silicon Graphics.

Sato, M. & Mueckler, M. (1999). A conserved amino acid motif (R-X-G-R-R) in the Glut1 glucose transporter is an important determinant of membrane topology. J Biol Chem 274, 24721–24725.[Abstract/Free Full Text]

Schmitt, L. & Tampe, R. (2002). Structure and mechanism of ABC transporters. Curr Opin Struct Biol 12, 754–760.[CrossRef][Medline]

Siegele, D. A., Imlay, K. R. & Imlay, J. A. (1996). The stationary-phase-exit defect of cydC (surB) mutants is due to the lack of a functional terminal cytochrome oxidase. J Bacteriol 178, 6091–6096.[Abstract]

Simons, R. W., Houman, F. & Kleckner, N. (1987). Improved single and multicopy lac-based cloning vectors for protein and operon fusions. Gene 53, 85–96.[CrossRef][Medline]

Sonnhammer, E. L. L., von Heijne, G. & Krogh, A. (1998). A hidden Markov model for predicting transmembrane helices in protein sequences. In Proceedings of the Sixth International Conference on Intelligent Systems for Molecular Biology, vol. 5, pp. 175–182. Edited by J. Glasgow, T. Littlejohn, F. Major, R. Lathrop, D. Sankoff & C. Sensen. Menlo Park, CA: AAAI/MIT Press.

Tamames, J. (2001). Evolution of gene order conservation in prokaryotes. Genome Biol 2, 1–11.

Tearle, R. G., Belote, J. M., McKeown, M., Baker, B. S. & Howells, A. J. (1989). Cloning and characterisation of the scarlet gene of Drosophila melanogaster. Genetics 122, 595–606.[Abstract/Free Full Text]

Urbatsch, I. L., Sankaran, B., Weber, J. & Senior, A. E. (1995). P-glycoprotein is stably inhibited by vanadate-induced trapping of nucleotide at a single catalytic site. J Biol Chem 270, 19383–19390.[Abstract/Free Full Text]

Vieira, J. & Messing, J. (1987). Production of single-stranded plasmid DNA. Methods Enzymol 153, 3–11.[Medline]

von Heijne, G. (1992). Membrane protein structure prediction. Hydrophobicity analysis and the positive-inside rule. J Mol Biol 225, 487–494.[Medline]

Way, S. S., Sallustio, S., Magliozzo, R. S. & Goldberg, M. B. (1999). Impact of either elevated or decreased levels of cytochrome bd expression on Shigella flexneri virulence. J Bacteriol 181, 1229–1237.[Abstract/Free Full Text]

Winstedt, L., Yoshida, K., Fujita, Y. & von Wachenfeldt, C. (1998). Cytochrome bd biosynthesis in Bacillus subtilis: characterization of the cydABCD operon. J Bacteriol 180, 6571–6580.[Abstract/Free Full Text]

Received 25 March 2004; revised 6 July 2004; accepted 7 July 2004.



This Article
Abstract
Full Text (PDF)
Alert me when this article is cited
Alert me if a correction is posted
Citation Map
Services
Email this article to a friend
Similar articles in this journal
Similar articles in PubMed
Alert me to new issues of the journal
Download to citation manager
Google Scholar
Articles by Cruz-Ramos, H.
Articles by Poole, R. K.
Articles citing this Article
PubMed
PubMed Citation
Articles by Cruz-Ramos, H.
Articles by Poole, R. K.
Agricola
Articles by Cruz-Ramos, H.
Articles by Poole, R. K.


HOME HELP FEEDBACK SUBSCRIPTIONS ARCHIVE SEARCH TABLE OF CONTENTS
INT J SYST EVOL MICROBIOL MICROBIOLOGY J GEN VIROL
J MED MICROBIOL ALL SGM JOURNALS
Copyright © 2004 Society for General Microbiology.